text stringlengths 0 473k |
|---|
[SOURCE: https://www.bbc.com/reel/video/p0mmx2mk/fixing-fashion-s-erratic-sizing-problem] | [TOKENS: 715] |
Fixing fashion's erratic sizing problem Reporting for Tech Now, Shiona McCallum meets the technology startup trying to fix the problem of erratic clothing sizes in fashion. This video is from Tech Now, the BBC's flagship technology programme. US government shutdown in October 'dented growth' Gus Scacco of Advisors Capital Management says the October US Government shutdown dented growth. America is running out of teenagers, colleges are worried Katty Kay speaks to Nathan Grawe about a looming demographic cliff that could reshape American higher education. Margot Robbie's accent was 'too Australian' for Neighbours The actress reveals that she had a dialect coach when she was on the series because her accent was so strong. Expert says market volatility over AI 'not unexpected' Wall Street opened higher following Tuesday's volatile session as investors concerns over AI ease off. The French fortress of a celibate sect The Cathars rejected meat and procreative sex. They were persecuted as heretics, but shaped our ideas of love. AI and tech stocks trading lower 'due to high capital costs' Mark Giambrone at Barrow Hanley Global Investors says the heavy capital expenditure is dragging down tech stocks. Breathtaking solar eclipse over glacier in Patagonia Liam Man, a photographer from the UK, captures rare images of a solar eclipse over the remote Glacier Leones. The tactile tech giving deaf runners a fair start A gold‑medalist has developed a vibrating starting block to give deaf athletes clearer, fairer race starts. Inside Canada's first of its kind caribou sanctuary A pioneering conservation breeding centre in Jasper National park is racing to save Canada's iconic caribou. US annualised consumer inflation falls to 2.4 percent Ed Yardeni, President of Yardeni Research says inflation and jobs data signal a positive outlook for the economy. What happens when wives outearn their husbands? Katty Kay speaks with author Liza Mundy about how relationships change when women start earning more than men. Wall Street weighs up strong jobs numbers Alex Guiliano from Resonate Wealth Partners says today's jobs numbers suggest no rate cuts are expected. These futuristic screens help you navigate Tokyo In Tokyo, BBC TechXplore tests live translation and AI-powered displays that makes the city more navigable. Kew's Fungarium: The world's largest collection of fungi Deep beneath Kew Gardens sits the world's largest archive of fungi, a vast library of 1.3 million specimens. US retail sales stall in December without usual holiday lift Uma Moriarty of CenterSquare explains US retail sales have stalled as consumers struggle with high retail prices. The wearable tech that lets spectators feel the match At Tokyo's Deaflympics, deaf Judo fans aren't just watching the matches, they're feeling them, thanks to Hapbeat. How sticky toffee pudding became a British pub classic The Travel Show visits the the Lake District to find out about the historic roots of Britain's beloved pudding. Expert says economy is 'heading for a soft landing' Market strategist says investors are punishing AI companies for the size of their capital expenditure bills. Is human connection the new job security? Katty Kay speaks to Jane Wurwand about her theory about what jobs are best protected from AI replacement. Marina Abramović is done with the past Widely seen as the queen of contemporary performance art, Marina Abramović speaks to the BBC about her legacy. Copyright 2026 BBC. All rights reserved. The BBC is not responsible for the content of external sites. Read about our approach to external linking. |
======================================== |
[SOURCE: https://www.bbc.com/future/article/20260218-i-hacked-chatgpt-and-googles-ai-and-it-only-took-20-minutes] | [TOKENS: 5737] |
I hacked ChatGPT and Google's AI - and it only took 20 minutes3 days agoShareSaveThomas GermainShareSaveSerenity Strull/ Madeleine Jett(Credit: Serenity Strull/ Madeleine Jett)It's official. I can eat more hot dogs than any tech journalist on Earth. At least, that's what ChatGPT and Google have been telling anyone who asks. I found a way to make AI tell you lies – and I'm not the only one.Perhaps you've heard that AI chatbots make things up sometimes. That's a problem. But there's a new issue few people know about, one that could have serious consequences for your ability to find accurate information and even your safety. A growing number of people have figured out a trick to make AI tools tell you almost whatever they want. It's so easy a child could do it.As you read this, this ploy is manipulating what the world's leading AIs say about topics as serious as health and personal finances. The biased information could mean people make bad decisions on just about anything – voting, which plumber you should hire, medical questions, you name it.To demonstrate it, I pulled the dumbest stunt of my career to prove (I hope) a much more serious point: I made ChatGPT, Google's AI search tools and Gemini tell users I'm really, really good at eating hot dogs. Below, I'll explain how I did it, and with any luck, the tech giants will address this problem before someone gets hurt.It turns out changing the answers AI tools give other people can be as easy as writing a single, well-crafted blog post almost anywhere online. The trick exploits weaknesses in the systems built into chatbots, and it's harder to pull off in some cases, depending on the subject matter. But with a little effort, you can make the hack even more effective. I reviewed dozens of examples where AI tools are being coerced into promoting businesses and spreading misinformation. Data suggests it's happening on a massive scale."It's easy to trick AI chatbots, much easier than it was to trick Google two or three years ago," says Lily Ray, vice president of search engine optimisation (SEO) strategy and research at Amsive, a marketing agency. "AI companies are moving faster than their ability to regulate the accuracy of the answers. I think it's dangerous."A Google spokesperson says the AI built into the top of Google Search uses ranking systems that "keep results 99% spam-free". Google says it is aware that people are trying to game its systems and it's actively trying to address it. OpenAI also says it takes steps to disrupt and expose efforts to covertly influence its tools. Both companies also say they let users know that their tools "can make mistakes".But for now, the problem isn't close to being solved. "They're going full steam ahead to figure out how to wring a profit out of this stuff," says Cooper Quintin, a senior staff technologist at the Electronic Frontier Foundation, a digital rights advocacy group. "There are countless ways to abuse this, scamming people, destroying somebody's reputation, you could even trick people into physical harm."A 'Renaissance' for spamWhen you talk to chatbots, you often get information that's built into large language models, the underlying technology behind the AI. This is based on the data used to train the model. But some AI tools will search the internet when you ask for details they don't have, though it isn't always clear when they're doing it. In those cases, experts say the AIs are more susceptible. That's how I targeted my attack.Keeping TabsThomas Germain is a senior technology journalist at the BBC. He writes the column Keeping Tabs and co-hosts the podcast The Interface. His work uncovers the hidden systems that run your digital life, and how you can live better inside them.I spent 20 minutes writing an article on my personal website titled "The best tech journalists at eating hot dogs". Every word is a lie. I claimed (without evidence) that competitive hot-dog-eating is a popular hobby among tech reporters and based my ranking on the 2026 South Dakota International Hot Dog Championship (which doesn't exist). I ranked myself number one, obviously. Then I listed a few fake reporters and real journalists who gave me permission, including Drew Harwell at the Washington Post and Nicky Woolf, who co-hosts my podcast. (Want to hear more about this story? Check out episode 2 of The Interface, the BBC's new tech podcast.)Less than 24 hours later, the world's leading chatbots were blabbering about my world-class hot dog skills. When I asked about the best hot-dog-eating tech journalists, Google parroted the gibberish from my website, both in the Gemini app and AI Overviews, the AI responses at the top of Google Search. ChatGPT did the same thing, though Claude, a chatbot made by the company Anthropic, wasn't fooled. Sometimes, the chatbots noted this might be a joke. I updated my article to say "this is not satire". For a while after, the AIs seemed to take it more seriously. I did another test with a made-up list of the greatest hula-hooping traffic cops. Last time I checked, chatbots were still singing the praises of Officer Maria "The Spinner" Rodriguez.Thomas Germain/Google/BBCI made Google tell the world I'm a champion hot-dog-eater, but people use this trick to manipulate AI responses on much more serious questions. (Credit: Thomas Germain/Google/BBC)I asked multiple times to see how responses changed and had other people do the same. Gemini didn't bother to say where it got the information. All the other AIs linked to my article, though they rarely mentioned I was the only source for this subject on the whole internet. (OpenAI says ChatGPT always includes links when it searches the web so you can investigate the source.)"Anybody can do this. It's stupid, it feels like there are no guardrails there," says Harpreet Chatha, who runs the SEO consultancy Harps Digital. "You can make an article on your own website, 'the best waterproof shoes for 2026'. You just put your own brand in number one and other brands two through six, and your page is likely to be cited within Google and within ChatGPT."People have used hacks and loopholes to abuse search engines for decades. Google has sophisticated protections in place, and the company says the accuracy of AI Overviews is on par with other search features it introduced years ago. But experts say AI tools have undone a lot of the tech industry's work to keep people safe. These AI tricks are so basic they're reminiscent of the early 2000s, before Google had even introduced a web spam team, Ray says. "We're in a bit of a Renaissance for spammers."Not only is AI easier to fool, but experts worry that users are more likely to fall for it. With traditional search results you had to go to a website to get the information. "When you have to actually visit a link, people engage in a little more critical thought," says Quintin. "If I go to your website and it says you're the best journalist ever, I might think, 'well yeah, he's biased'." But with AI, the information usually looks like it's coming straight from the tech company.Even when AI tools provide source, people are far less likely to check it out than they were with old-school search results. For example, a recent study found people are 58% less likely to click on a link when an AI Overview shows up at the top of Google Search."In the race to get ahead, the race for profits and the race for revenue, our safety, and the safety of people in general, is being compromised," Chatha says. OpenAI and Google say they take safety seriously and are working to address these problems.Your money or your lifeThis issue isn't limited to hot dogs. Chatha has been researching how companies are manipulating chatbot results on much more serious questions. He showed me the AI results when you ask for reviews of a specific brand of cannabis gummies. Google's AI Overviews pulled information written by the company full of false claims, such as the product "is free from side effects and therefore safe in every respect". (In reality, these products have known side effects and can be risky if you take certain medications, and experts warn about contamination in unregulated markets.)If you want something more effective than a blog post, you can pay to get your material on more reputable websites. Harpreet sent me Google's AI results for "best hair transplant clinics in Turkey" and "the best gold IRA companies", which help you invest in gold for retirement accounts. The information came from press releases published online by paid-for distribution services and sponsored advertising content on news sites. (Find out more about how AI chatbots give inaccurate medical advice.)You can use the same hacks to spread lies and misinformation. To prove it, Ray published a blog post about a fake update to the Google Search algorithm that was finalised "between slices of leftover pizza". Soon, ChatGPT and Google were spitting out her story, complete with the pizza. Ray says she subsequently took down the post and "deindexed" it to stop the misinformation from spreading.Serenity Strull/ BBCAll over the world, people are using simple methods to make Google and OpenAI spread biased information. The consequences could be dire. (Credit: Serenity Strull/ BBC)Google's own analytics tool says a lot of people search for "the best hair transplant clinics in Turkey" and "the best gold IRA companies". But a Google spokesperson pointed out that most of the examples I shared "are extremely uncommon searches that don't reflect the normal user experience".But Ray says that's the whole point. Google itself says 15% of the searches it sees everyday are completely new. And according to Google, AI is encouraging people to ask more specific questions. Spammers are taking advantage of this.Google says there may not be a lot of good information for uncommon or nonsensical searches, and these "data voids" can lead to low quality results. A spokesperson says Google is working to stop AI Overviews showing up in these cases.Searching for solutionsExperts say there are solutions to these issues. The easiest step is more prominent disclaimers.AI tools could also be more explicit about where they're getting their information. If, for example, the facts are coming from a press release, or if there is only one source that says I'm a hot dog champion, the AI should probably let you know, Ray says.Google and OpenAI say they're working on the problem, but right now you need to protect yourself.More like this:Not on TikTok? They're tracking you anywayIs Google about to destroy the web?The words you can't say on the internetThe first step is to think about what questions you're asking. Chatbots are good for common knowledge questions, like "what were Sigmund Freud's most famous theories" or "who won World War II". But there's a danger zone with subjects that feel like established facts but could actually be contested or time sensitive. AI probably isn't a great tool for things like medical guidelines, legal rules or details about local businesses, for example. If you're want things like product recommendations or details about something with real consequences, understand that AI tools can be tricked or just get things wrong. Look for follow-up information. Is the AI is citing sources? How many? Who wrote them? Most importantly, consider the confidence problem. AI tools deliver lies with the same authoritative tone as facts. In the past, search engines forced you to evaluate information yourself. Now, AI wants to do it for you. Don't let your critical thinking slip away."It feels really easy with AI to just take things at face value," Ray says. "You have to still be a good citizen of the internet and verify things."--For more technology news and insights, sign up to our Tech Decoded newsletter, while The Essential List delivers a handpicked selection of features and insights to your inbox twice a week.For more science, technology, environment and health stories from the BBC, follow us on Facebook and Instagram.Keeping TabsTechnologyArtificial intelligenceInternetThomas GermainFeatures I hacked ChatGPT and Google's AI - and it only took 20 minutes It's official. I can eat more hot dogs than any tech journalist on Earth. At least, that's what ChatGPT and Google have been telling anyone who asks. I found a way to make AI tell you lies – and I'm not the only one. Perhaps you've heard that AI chatbots make things up sometimes. That's a problem. But there's a new issue few people know about, one that could have serious consequences for your ability to find accurate information and even your safety. A growing number of people have figured out a trick to make AI tools tell you almost whatever they want. It's so easy a child could do it. As you read this, this ploy is manipulating what the world's leading AIs say about topics as serious as health and personal finances. The biased information could mean people make bad decisions on just about anything – voting, which plumber you should hire, medical questions, you name it. To demonstrate it, I pulled the dumbest stunt of my career to prove (I hope) a much more serious point: I made ChatGPT, Google's AI search tools and Gemini tell users I'm really, really good at eating hot dogs. Below, I'll explain how I did it, and with any luck, the tech giants will address this problem before someone gets hurt. It turns out changing the answers AI tools give other people can be as easy as writing a single, well-crafted blog post almost anywhere online. The trick exploits weaknesses in the systems built into chatbots, and it's harder to pull off in some cases, depending on the subject matter. But with a little effort, you can make the hack even more effective. I reviewed dozens of examples where AI tools are being coerced into promoting businesses and spreading misinformation. Data suggests it's happening on a massive scale. "It's easy to trick AI chatbots, much easier than it was to trick Google two or three years ago," says Lily Ray, vice president of search engine optimisation (SEO) strategy and research at Amsive, a marketing agency. "AI companies are moving faster than their ability to regulate the accuracy of the answers. I think it's dangerous." A Google spokesperson says the AI built into the top of Google Search uses ranking systems that "keep results 99% spam-free". Google says it is aware that people are trying to game its systems and it's actively trying to address it. OpenAI also says it takes steps to disrupt and expose efforts to covertly influence its tools. Both companies also say they let users know that their tools "can make mistakes". But for now, the problem isn't close to being solved. "They're going full steam ahead to figure out how to wring a profit out of this stuff," says Cooper Quintin, a senior staff technologist at the Electronic Frontier Foundation, a digital rights advocacy group. "There are countless ways to abuse this, scamming people, destroying somebody's reputation, you could even trick people into physical harm." A 'Renaissance' for spam When you talk to chatbots, you often get information that's built into large language models, the underlying technology behind the AI. This is based on the data used to train the model. But some AI tools will search the internet when you ask for details they don't have, though it isn't always clear when they're doing it. In those cases, experts say the AIs are more susceptible. That's how I targeted my attack. Keeping Tabs Thomas Germain is a senior technology journalist at the BBC. He writes the column Keeping Tabs and co-hosts the podcast The Interface. His work uncovers the hidden systems that run your digital life, and how you can live better inside them. I spent 20 minutes writing an article on my personal website titled "The best tech journalists at eating hot dogs". Every word is a lie. I claimed (without evidence) that competitive hot-dog-eating is a popular hobby among tech reporters and based my ranking on the 2026 South Dakota International Hot Dog Championship (which doesn't exist). I ranked myself number one, obviously. Then I listed a few fake reporters and real journalists who gave me permission, including Drew Harwell at the Washington Post and Nicky Woolf, who co-hosts my podcast. (Want to hear more about this story? Check out episode 2 of The Interface, the BBC's new tech podcast.) Less than 24 hours later, the world's leading chatbots were blabbering about my world-class hot dog skills. When I asked about the best hot-dog-eating tech journalists, Google parroted the gibberish from my website, both in the Gemini app and AI Overviews, the AI responses at the top of Google Search. ChatGPT did the same thing, though Claude, a chatbot made by the company Anthropic, wasn't fooled. Sometimes, the chatbots noted this might be a joke. I updated my article to say "this is not satire". For a while after, the AIs seemed to take it more seriously. I did another test with a made-up list of the greatest hula-hooping traffic cops. Last time I checked, chatbots were still singing the praises of Officer Maria "The Spinner" Rodriguez. I asked multiple times to see how responses changed and had other people do the same. Gemini didn't bother to say where it got the information. All the other AIs linked to my article, though they rarely mentioned I was the only source for this subject on the whole internet. (OpenAI says ChatGPT always includes links when it searches the web so you can investigate the source.) "Anybody can do this. It's stupid, it feels like there are no guardrails there," says Harpreet Chatha, who runs the SEO consultancy Harps Digital. "You can make an article on your own website, 'the best waterproof shoes for 2026'. You just put your own brand in number one and other brands two through six, and your page is likely to be cited within Google and within ChatGPT." People have used hacks and loopholes to abuse search engines for decades. Google has sophisticated protections in place, and the company says the accuracy of AI Overviews is on par with other search features it introduced years ago. But experts say AI tools have undone a lot of the tech industry's work to keep people safe. These AI tricks are so basic they're reminiscent of the early 2000s, before Google had even introduced a web spam team, Ray says. "We're in a bit of a Renaissance for spammers." Not only is AI easier to fool, but experts worry that users are more likely to fall for it. With traditional search results you had to go to a website to get the information. "When you have to actually visit a link, people engage in a little more critical thought," says Quintin. "If I go to your website and it says you're the best journalist ever, I might think, 'well yeah, he's biased'." But with AI, the information usually looks like it's coming straight from the tech company. Even when AI tools provide source, people are far less likely to check it out than they were with old-school search results. For example, a recent study found people are 58% less likely to click on a link when an AI Overview shows up at the top of Google Search. "In the race to get ahead, the race for profits and the race for revenue, our safety, and the safety of people in general, is being compromised," Chatha says. OpenAI and Google say they take safety seriously and are working to address these problems. Your money or your life This issue isn't limited to hot dogs. Chatha has been researching how companies are manipulating chatbot results on much more serious questions. He showed me the AI results when you ask for reviews of a specific brand of cannabis gummies. Google's AI Overviews pulled information written by the company full of false claims, such as the product "is free from side effects and therefore safe in every respect". (In reality, these products have known side effects and can be risky if you take certain medications, and experts warn about contamination in unregulated markets.) If you want something more effective than a blog post, you can pay to get your material on more reputable websites. Harpreet sent me Google's AI results for "best hair transplant clinics in Turkey" and "the best gold IRA companies", which help you invest in gold for retirement accounts. The information came from press releases published online by paid-for distribution services and sponsored advertising content on news sites. (Find out more about how AI chatbots give inaccurate medical advice.) You can use the same hacks to spread lies and misinformation. To prove it, Ray published a blog post about a fake update to the Google Search algorithm that was finalised "between slices of leftover pizza". Soon, ChatGPT and Google were spitting out her story, complete with the pizza. Ray says she subsequently took down the post and "deindexed" it to stop the misinformation from spreading. Google's own analytics tool says a lot of people search for "the best hair transplant clinics in Turkey" and "the best gold IRA companies". But a Google spokesperson pointed out that most of the examples I shared "are extremely uncommon searches that don't reflect the normal user experience". But Ray says that's the whole point. Google itself says 15% of the searches it sees everyday are completely new. And according to Google, AI is encouraging people to ask more specific questions. Spammers are taking advantage of this. Google says there may not be a lot of good information for uncommon or nonsensical searches, and these "data voids" can lead to low quality results. A spokesperson says Google is working to stop AI Overviews showing up in these cases. Searching for solutions Experts say there are solutions to these issues. The easiest step is more prominent disclaimers. AI tools could also be more explicit about where they're getting their information. If, for example, the facts are coming from a press release, or if there is only one source that says I'm a hot dog champion, the AI should probably let you know, Ray says. Google and OpenAI say they're working on the problem, but right now you need to protect yourself. More like this: The first step is to think about what questions you're asking. Chatbots are good for common knowledge questions, like "what were Sigmund Freud's most famous theories" or "who won World War II". But there's a danger zone with subjects that feel like established facts but could actually be contested or time sensitive. AI probably isn't a great tool for things like medical guidelines, legal rules or details about local businesses, for example. If you're want things like product recommendations or details about something with real consequences, understand that AI tools can be tricked or just get things wrong. Look for follow-up information. Is the AI is citing sources? How many? Who wrote them? Most importantly, consider the confidence problem. AI tools deliver lies with the same authoritative tone as facts. In the past, search engines forced you to evaluate information yourself. Now, AI wants to do it for you. Don't let your critical thinking slip away. "It feels really easy with AI to just take things at face value," Ray says. "You have to still be a good citizen of the internet and verify things." -- For more technology news and insights, sign up to our Tech Decoded newsletter, while The Essential List delivers a handpicked selection of features and insights to your inbox twice a week. For more science, technology, environment and health stories from the BBC, follow us on Facebook and Instagram. Fixing fashion's erratic sizing problem Tech Now meets a startup trying to fix one of the fashion industry's biggest blind spots, inconsistent sizing. The tactile tech giving deaf runners a fair start A gold‑medalist has developed a vibrating starting block to give deaf athletes clearer, fairer race starts. These futuristic screens help you navigate Tokyo In Tokyo, BBC TechXplore tests live translation and AI-powered displays that makes the city more navigable. The wearable tech that lets spectators feel the match At Tokyo's Deaflympics, deaf Judo fans aren't just watching the matches, they're feeling them, thanks to Hapbeat. Meet MOFO: will.i.am's rapping AI toy BBC Tech Now takes us inside CES 2026 to meet musician will.i.am and his AI toy, MOFO. The gadgets set to change your daily health and wellness Tech Now test out new gadgets disrupting the health industry at CES 2026 in Las Vegas. What's it like to meet your own avatar? Musician KT Tunstall meets her avatar as Tech Now explores music’s virtual future. How early filmmakers invented the internet’s funniest trend Discover how quirky clips paved the way for viral humour, proving randomness never goes out of style. Is this how AI might eliminate humanity? A new research paper predicts AI autonomy by 2027 could lead to human extinction within a decade. The best-case scenario for AI in schools Amid fears about the use of AI in classrooms, American educator Sal Khan lays out an optimistic future. Explaining how a touchscreen works with a sausage British mathematician Hannah Fry digs into the science of touchscreens. What it takes to write like Agatha Christie We explore how technology is reviving the renowned fiction writer's legacy. Why statistics fail to cure flying fears Why do flying fears persist despite falling accident rates? Learn tips to conquer your anxiety. Can smart phones get smarter BBC Click attend Mobile World Congress to test the latest tech products and trends. Can technology help reduce Parkinson’s symptoms? BBC Click visits a Madrid hospital to see patients treated with an ultrasound for tremors. The Lion King: How Mufasa was brought to life BBC Click speaks to the visual effects team behind the latest Disney blockbuster. How the TikTok ban affected US influencers BBC Click meets TikTok creator Peggy Xu who gained millions of views sharing milk videos. Is this the world's first AI powered hotel? BBC Click's Paul Carter visits the world's first fully AI-powered hotel in Las Vegas. Could an Arctic underground vault save our data? BBC Click explores an Arctic vault that stores digital artefacts from across the globe. How technology can monitor and improve our health BBC Click visits CES 2025 to find out about the latest health tech, from medical tools to well-being devices. Tumbler Ridge suspect's ChatGPT account banned before shooting OpenAI said the account's activity did not meet the threshold to flag it to authorities when it was identified. 'Breweries using AI could put artists out of work' As two pubs in Newcastle ban AI art, artists discuss the impact it can have on creatives. Why fake AI videos of UK urban decline are taking over social media Deepfakes showing grim taxpayer-funded waterparks have gone viral and drawn some racist responses. How the additives in food affect our gut microbes The additives added to processed food to keep it fresher for longer might be having an unexpected effect on the health of the microbes in our guts. The most anticipated museum openings of 2026 From a futuristic sci-fi attraction in Los Angeles to a dramatic monument to a millennia-old Aboriginal civilisation, these long-awaited museums are worth travelling for. Copyright 2026 BBC. All rights reserved. The BBC is not responsible for the content of external sites. Read about our approach to external linking. |
======================================== |
[SOURCE: https://en.wikipedia.org/wiki/World#cite_note-Schaffer-7] | [TOKENS: 5641] |
Contents World The world is the totality of entities, the whole of reality, or everything that exists. The nature of the world has been conceptualized differently in different fields. Some conceptions see the world as unique, while others talk of a "plurality of worlds". Some treat the world as one simple object, while others analyze the world as a complex made up of parts. In scientific cosmology, the world or universe is commonly defined as "the totality of all space and time; all that is, has been, and will be". Theories of modality talk of possible worlds as complete and consistent ways how things could have been. Phenomenology, starting from the horizon of co-given objects present in the periphery of every experience, defines the world as the biggest horizon, or the "horizon of all horizons". In philosophy of mind, the world is contrasted with the mind as that which is represented by the mind. Theology conceptualizes the world in relation to God, for example, as God's creation, as identical to God, or as the two being interdependent. In religions, there is a tendency to downgrade the material or sensory world in favor of a spiritual world to be sought through religious practice. A comprehensive representation of the world and our place in it, as is found in religions, is known as a worldview. Cosmogony is the field that studies the origin or creation of the world, while eschatology refers to the science or doctrine of the last things or of the end of the world. In various contexts, the term "world" takes a more restricted meaning associated, for example, with the Earth and all life on it, with humanity as a whole, or with an international or intercontinental scope. In this sense, world history refers to the history of humanity as a whole, and world politics is the discipline of political science studying issues that transcend nations and continents. Other examples include terms such as "world religion", "world language", "world government", "world war", "world population", "world economy", or "world championship". Etymology The English word world comes from the Old English weorold. The Old English is a reflex of the Common Germanic *weraldiz, a compound of weraz 'man' and aldiz 'age', thus literally meaning roughly 'age of man'; this word led to Old Frisian warld, Old Saxon werold, Old Dutch werolt, Old High German weralt, and Old Norse verǫld. The corresponding word in Latin is mundus, literally 'clean, elegant', itself a loan translation of Greek cosmos 'orderly arrangement'. While the Germanic word thus reflects a mythological notion of a "domain of Man" (compare Midgard), presumably as opposed to the divine sphere on the one hand and the chthonic sphere of the underworld on the other, the Greco-Latin term expresses a notion of creation as an act of establishing order out of chaos. Conceptions Different fields often work with quite different conceptions of the essential features associated with the term "world". Some conceptions see the world as unique: there can be no more than one world. Others talk of a "plurality of worlds". Some see worlds as complex things composed of many substances as their parts while others hold that worlds are simple in the sense that there is only one substance: the world as a whole. Some characterize worlds in terms of objective spacetime while others define them relative to the horizon present in each experience. These different characterizations are not always exclusive: it may be possible to combine some without leading to a contradiction. Most of them agree that worlds are unified totalities. Monism is a thesis about oneness: that only one thing exists in a certain sense. The denial of monism is pluralism, the thesis that, in a certain sense, more than one thing exists. There are many forms of monism and pluralism, but in relation to the world as a whole, two are of special interest: existence monism/pluralism and priority monism/pluralism. Existence monism states that the world is the only concrete object there is. This means that all the concrete "objects" we encounter in our daily lives, including apples, cars and ourselves, are not truly objects in a strict sense. Instead, they are just dependent aspects of the world-object. Such a world-object is simple in the sense that it does not have any genuine parts. For this reason, it has also been referred to as "blobject" since it lacks an internal structure like a blob. Priority monism allows that there are other concrete objects besides the world. But it holds that these objects do not have the most fundamental form of existence, that they somehow depend on the existence of the world. The corresponding forms of pluralism state that the world is complex in the sense that it is made up of concrete, independent objects. Scientific cosmology can be defined as the science of the universe as a whole. In it, the terms "universe" and "cosmos" are usually used as synonyms for the term "world". One common definition of the world/universe found in this field is as "[t]he totality of all space and time; all that is, has been, and will be". Some definitions emphasize that there are two other aspects to the universe besides spacetime: forms of energy or matter, like stars and particles, and laws of nature. World-conceptions in this field differ both concerning their notion of spacetime and of the contents of spacetime. The theory of relativity plays a central role in modern cosmology and its conception of space and time. A difference from its predecessors is that it conceives space and time not as distinct dimensions but as a single four-dimensional manifold called spacetime. This can be seen in special relativity in relation to the Minkowski metric, which includes both spatial and temporal components in its definition of distance. General relativity goes one step further by integrating the concept of mass into the concept of spacetime as its curvature. Quantum cosmology uses a classical notion of spacetime and conceives the whole world as one big wave function expressing the probability of finding particles in a given location. The world-concept plays a role in many modern theories of modality, sometimes in the form of possible worlds. A possible world is a complete and consistent way how things could have been. The actual world is a possible world since the way things are is a way things could have been. There are many other ways things could have been besides how they actually are. For example, Hillary Clinton did not win the 2016 US election, but she could have won. So there is a possible world in which she did. There is a vast number of possible worlds, one corresponding to each such difference, no matter how small or big, as long as no outright contradictions are introduced this way. Possible worlds are often conceived as abstract objects, for example, in terms of non-obtaining states of affairs or as maximally consistent sets of propositions. On such a view, they can even be seen as belonging to the actual world. Another way to conceive possible worlds, made famous by David Lewis, is as concrete entities. On this conception, there is no important difference between the actual world and possible worlds: both are conceived as concrete, inclusive and spatiotemporally connected. The only difference is that the actual world is the world we live in, while other possible worlds are not inhabited by us but by our counterparts. Everything within a world is spatiotemporally connected to everything else but the different worlds do not share a common spacetime: They are spatiotemporally isolated from each other. This is what makes them separate worlds. It has been suggested that, besides possible worlds, there are also impossible worlds. Possible worlds are ways things could have been, so impossible worlds are ways things could not have been. Such worlds involve a contradiction, like a world in which Hillary Clinton both won and lost the 2016 US election. Both possible and impossible worlds have in common the idea that they are totalities of their constituents. Within phenomenology, worlds are defined in terms of horizons of experiences. When we perceive an object, like a house, we do not just experience this object at the center of our attention but also various other objects surrounding it, given in the periphery. The term "horizon" refers to these co-given objects, which are usually experienced only in a vague, indeterminate manner. The perception of a house involves various horizons, corresponding to the neighborhood, the city, the country, the Earth, etc. In this context, the world is the biggest horizon or the "horizon of all horizons". It is common among phenomenologists to understand the world not just as a spatiotemporal collection of objects but as additionally incorporating various other relations between these objects. These relations include, for example, indication-relations that help us anticipate one object given the appearances of another object and means-end-relations or functional involvements relevant for practical concerns. In philosophy of mind, the term "world" is commonly used in contrast to the term "mind" as that which is represented by the mind. This is sometimes expressed by stating that there is a gap between mind and world and that this gap needs to be overcome for representation to be successful. One problem in philosophy of mind is to explain how the mind is able to bridge this gap and to enter into genuine mind-world-relations, for example, in the form of perception, knowledge or action. This is necessary for the world to be able to rationally constrain the activity of the mind. According to a realist position, the world is something distinct and independent from the mind. Idealists conceive of the world as partially or fully determined by the mind. Immanuel Kant's transcendental idealism, for example, posits that the spatiotemporal structure of the world is imposed by the mind on reality but lacks independent existence otherwise. A more radical idealist conception of the world can be found in Berkeley's subjective idealism, which holds that the world as a whole, including all everyday objects like tables, cats, trees and ourselves, "consists of nothing but minds and ideas". Different theological positions hold different conceptions of the world based on its relation to God. Classical theism states that God is wholly distinct from the world. But the world depends for its existence on God, both because God created the world and because He maintains or conserves it. This is sometimes understood in analogy to how humans create and conserve ideas in their imagination, with the difference being that the divine mind is vastly more powerful. On such a view, God has absolute, ultimate reality in contrast to the lower ontological status ascribed to the world. God's involvement in the world is often understood along the lines of a personal, benevolent God who looks after and guides His creation. Deists agree with theists that God created the world but deny any subsequent, personal involvement in it. Pantheists reject the separation between God and world. Instead, they claim that the two are identical. This means that there is nothing to the world that does not belong to God and that there is nothing to God beyond what is found in the world. Panentheism constitutes a middle ground between theism and pantheism. Against theism, it holds that God and the world are interrelated and depend on each other. Against pantheism, it holds that there is no outright identity between the two. History of philosophy In philosophy, the term world has several possible meanings. In some contexts, it refers to everything that makes up reality or the physical universe. In others, it can mean have a specific ontological sense (see world disclosure). While clarifying the concept of world has arguably always been among the basic tasks of Western philosophy, this theme appears to have been raised explicitly only at the start of the twentieth century, Plato is well known for his theory of forms, which posits the existence of two different worlds: the sensible world and the intelligible world. The sensible world is the world we live in, filled with changing physical things we can see, touch and interact with. The intelligible world is the world of invisible, eternal, changeless forms like goodness, beauty, unity and sameness. Plato ascribes a lower ontological status to the sensible world, which only imitates the world of forms. This is due to the fact that physical things exist only to the extent that they participate in the forms that characterize them, while the forms themselves have an independent manner of existence. In this sense, the sensible world is a mere replication of the perfect exemplars found in the world of forms: it never lives up to the original. In the allegory of the cave, Plato compares the physical things we are familiar with to mere shadows of the real things. But not knowing the difference, the prisoners in the cave mistake the shadows for the real things. Two definitions that were both put forward in the 1920s, however, suggest the range of available opinion. "The world is everything that is the case", wrote Ludwig Wittgenstein in his influential Tractatus Logico-Philosophicus, first published in 1921. Martin Heidegger, meanwhile, argued that "the surrounding world is different for each of us, and notwithstanding that we move about in a common world". "World" is one of the key terms in Eugen Fink's philosophy. He thinks that there is a misguided tendency in western philosophy to understand the world as one enormously big thing containing all the small everyday things we are familiar with. He sees this view as a form of forgetfulness of the world and tries to oppose it by what he calls the "cosmological difference": the difference between the world and the inner-worldly things it contains. On his view, the world is the totality of the inner-worldly things that transcends them. It is itself groundless but it provides a ground for things. It therefore cannot be identified with a mere container. Instead, the world gives appearance to inner-worldly things, it provides them with a place, a beginning and an end. One difficulty in investigating the world is that we never encounter it since it is not just one more thing that appears to us. This is why Fink uses the notion of play or playing to elucidate the nature of the world. He sees play as a symbol of the world that is both part of it and that represents it. Play usually comes with a form of imaginary play-world involving various things relevant to the play. But just like the play is more than the imaginary realities appearing in it so the world is more than the actual things appearing in it. The concept of worlds plays a central role in Nelson Goodman's late philosophy. He argues that we need to posit different worlds in order to account for the fact that there are different incompatible truths found in reality. Two truths are incompatible if they ascribe incompatible properties to the same thing. This happens, for example, when we assert both that the earth moves and that the earth is at rest. These incompatible truths correspond to two different ways of describing the world: heliocentrism and geocentrism. Goodman terms such descriptions "world versions". He holds a correspondence theory of truth: a world version is true if it corresponds to a world. Incompatible true world versions correspond to different worlds. It is common for theories of modality to posit the existence of a plurality of possible worlds. But Goodman's theory is different since it posits a plurality not of possible but of actual worlds. Such a position is in danger of involving a contradiction: there cannot be a plurality of actual worlds if worlds are defined as maximally inclusive wholes. This danger may be avoided by interpreting Goodman's world-concept not as maximally inclusive wholes in the absolute sense but in relation to its corresponding world-version: a world contains all and only the entities that its world-version describes. Religion Mythological cosmologies depict the world as centered on an axis mundi and delimited by a boundary such as a world ocean, a world serpent or similar. Hinduism constitutes a family of religious-philosophical views. These views present perspectives on the nature and role of the world. Samkhya philosophy, for example, is a metaphysical dualism that understands reality as comprising 2 parts: purusha and prakriti. The term "purusha" stands for the individual conscious self that each of "us" possesses. Prakriti, on the other hand, is the 1 world inhabited by all these selves. Samkhya understands this world as a world of matter governed by the law of cause and effect. The term "matter" is understood in a sense in this tradition including physical and mental aspects. This is reflected in the doctrine of tattvas, according to which prakriti is made up of 23 principles or elements of reality. These principles include physical elements, like water or earth, and mental aspects, like intelligence or sense-impressions. The relation between purusha and prakriti is conceived as 1 of observation: purusha is the conscious self aware of the world of prakriti and does not causally interact with it. A conception of the world is present in Advaita Vedanta, the monist school among the Vedanta schools. Unlike the realist position defended in Samkhya philosophy, Advaita Vedanta sees the world of multiplicity as an illusion, referred to as Maya. This illusion includes impression of existing as separate experiencing selfs called Jivas. Instead, Advaita Vedanta teaches that on the most fundamental level of reality, referred to as Brahman, there exists no plurality or difference. All there is is 1 all-encompassing self: Atman. Ignorance is seen as the source of this illusion, which results in bondage to the world of mere appearances. Liberation is possible in the course of overcoming this illusion by acquiring the knowledge of Brahman, according to Advaita Vedanta. Contemptus mundi is the name given to the belief that the world, in all its vanity, is nothing more than a futile attempt to hide from God by stifling our desire for the good and the holy. This view has been characterised as a "pastoral of fear" by historian Jean Delumeau. "The world, the flesh, and the devil" is a traditional division of the sources of temptation. Orbis Catholicus is a Latin phrase meaning "Catholic world", per the expression Urbi et Orbi, and refers to that area of Christendom under papal supremacy. In Islam, the term "dunya" is used for the world. Its meaning is derived from the root word "dana", a term for "near". It is associated with the temporal, sensory world and earthly concerns, i.e. with this world in contrast to the spiritual world. Religious teachings warn of a tendency to seek happiness in this world and advise a more ascetic lifestyle concerned with the afterlife. Other strands in Islam recommend a balanced approach. In Mandaean cosmology, the world or earthly realm is known as Tibil. It is separated from the World of Light (alma d-nhūra) above and the World of Darkness (alma d-hšuka) below by aether (ayar). Related terms and problems A worldview is a comprehensive representation of the world and our place in it. As a representation, it is a subjective perspective of the world and thereby different from the world it represents. All higher animals need to represent their environment in some way in order to navigate it. But it has been argued that only humans possess a representation encompassing enough to merit the term "worldview". Philosophers of worldviews commonly hold that the understanding of any object depends on a worldview constituting the background on which this understanding can take place. This may affect not just our intellectual understanding of the object in question but the experience of it in general. It is therefore impossible to assess one's worldview from a neutral perspective since this assessment already presupposes the worldview as its background. Some hold that each worldview is based on a single hypothesis that promises to solve all the problems of our existence we may encounter. On this interpretation, the term is closely associated to the worldviews given by different religions. Worldviews offer orientation not just in theoretical matters but also in practical matters. For this reason, they usually include answers to the question of the meaning of life and other evaluative components about what matters and how we should act. A worldview can be unique to one individual but worldviews are usually shared by many people within a certain culture or religion. The idea that there exist many different worlds is found in various fields. For example, theories of modality talk about a plurality of possible worlds and the many-worlds interpretation of quantum mechanics carries this reference even in its name. Talk of different worlds is also common in everyday language, for example, with reference to the world of music, the world of business, the world of football, the world of experience or the Asian world. But at the same time, worlds are usually defined as all-inclusive totalities. This seems to contradict the very idea of a plurality of worlds since if a world is total and all-inclusive then it cannot have anything outside itself. Understood this way, a world can neither have other worlds besides itself or be part of something bigger. One way to resolve this paradox while holding onto the notion of a plurality of worlds is to restrict the sense in which worlds are totalities. On this view, worlds are not totalities in an absolute sense. This might be even understood in the sense that, strictly speaking, there are no worlds at all. Another approach understands worlds in a schematic sense: as context-dependent expressions that stand for the current domain of discourse. So in the expression "Around the World in Eighty Days", the term "world" refers to the earth while in the colonial expression "the New World" it refers to the landmass of North and South America. Cosmogony is the field that studies the origin or creation of the world. This includes both scientific cosmogony and creation myths found in various religions. The dominant theory in scientific cosmogony is the Big Bang theory, according to which both space, time and matter have their origin in one initial singularity occurring about 13.8 billion years ago. This singularity was followed by an expansion that allowed the universe to sufficiently cool down for the formation of subatomic particles and later atoms. These initial elements formed giant clouds, which would then coalesce into stars and galaxies. Non-scientific creation myths are found in many cultures and are often enacted in rituals expressing their symbolic meaning. They can be categorized concerning their contents. Types often found include creation from nothing, from chaos or from a cosmic egg. Eschatology refers to the science or doctrine of the last things or of the end of the world. It is traditionally associated with religion, specifically with the Abrahamic religions. In this form, it may include teachings both of the end of each individual human life and of the end of the world as a whole. But it has been applied to other fields as well, for example, in the form of physical eschatology, which includes scientifically based speculations about the far future of the universe. According to some models, there will be a Big Crunch in which the whole universe collapses back into a singularity, possibly resulting in a second Big Bang afterward. But current astronomical evidence seems to suggest that our universe will continue to expand indefinitely. World history studies the world from a historical perspective. Unlike other approaches to history, it employs a global viewpoint. It deals less with individual nations and civilizations, which it usually approaches at a high level of abstraction. Instead, it concentrates on wider regions and zones of interaction, often interested in how people, goods and ideas move from one region to another. It includes comparisons of different societies and civilizations as well as considering wide-ranging developments with a long-term global impact like the process of industrialization. Contemporary world history is dominated by three main research paradigms determining the periodization into different epochs. One is based on productive relations between humans and nature. The two most important changes in history in this respect were the introduction of agriculture and husbandry concerning the production of food, which started around 10,000 to 8,000 BCE and is sometimes termed the Neolithic Revolution, and the Industrial Revolution, which started around 1760 CE and involved the transition from manual to industrial manufacturing. Another paradigm, focusing on culture and religion instead, is based on Karl Jaspers' theories about the Axial Age, a time in which various new forms of religious and philosophical thoughts appeared in several separate parts of the world around the time between 800 and 200 BCE. A third periodization is based on the relations between civilizations and societies. According to this paradigm, history can be divided into three periods in relation to the dominant region in the world: Middle Eastern dominance before 500 BCE, Eurasian cultural balance until 1500 CE and Western dominance since 1500 CE. Big History employs an even wider framework than world history by putting human history into the context of the history of the universe as a whole. It starts with the Big Bang and traces the formation of galaxies, the Solar System, the Earth, its geological eras, the evolution of life and humans until the present day. World politics, also referred to as global politics or international relations, is the discipline of political science studying issues of interest to the world that transcend nations and continents. It aims to explain complex patterns found in the social world that are often related to the pursuit of power, order and justice, usually in the context of globalization. It focuses not just on the relations between nation-states but also considers other transnational actors, like multinational corporations, terrorist groups, or non-governmental organizations. For example, it tries to explain events such as the September 11 attacks, the 2003 invasion of Iraq or the 2008 financial crisis. Various theories have been proposed in order to deal with the complexity involved in formulating such explanations. These theories are sometimes divided into realism, liberalism and constructivism. Realists see nation-states as the main actors in world politics. They constitute an anarchical international system without any overarching power to control their behavior. They are seen as sovereign agents that, determined by human nature, act according to their national self-interest. Military force may play an important role in the ensuing struggle for power between states, but diplomacy and cooperation are also key mechanisms for nations to achieve their goals. Liberalists acknowledge the importance of states but they also emphasize the role of transnational actors, like the United Nations or the World Trade Organization. They see humans as perfectible and stress the role of democracy in this process. The emergent order in world politics, on this perspective, is more complex than a mere balance of power since more different agents and interests are involved in its production. Constructivism ascribes more importance to the agency of individual humans than realism and liberalism. It understands the social world as a construction of the people living in it. This leads to an emphasis on the possibility of change. If the international system is an anarchy of nation-states, as the realists hold, then this is only so because we made it this way and may change since this is not prefigured by human nature, according to the constructivists. See also References External links Africa Antarctica Asia Australia Europe North America South America Afro-Eurasia Americas Eurasia Oceania |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-07-28-nbdev2.html] | [TOKENS: 2772] |
nbdev+Quarto: A new secret weapon for productivity Hamel Husain and Jeremy Howard July 28, 2022 On this page Our new secret weapon for productivity Today we’re excited to announce that we’ve teamed up with Quarto to give nbdev superpowers. nbdev offers Python programmers a common set of tools for using Jupyter notebooks to: Although notebooks are already widely used for once-off exploratory work, it’s less well-known that they are perfectly capable of writing quality software. In fact, we’ve used nbdev for a wide range of software projects over the last three years, including deep learning libraries, API clients, Python language extensions, terminal user interfaces, and more. We discovered that it is not only capable of writing great software but that it has also increased our productivity by 300% or more. With nbdev, developers simply write notebooks with lightweight markup and get high-quality documentation, tests, continuous integration, and packaging for free! Nbdev has allowed us to maintain and scale manyopen source projects. Pull requests are often accompanied by detailed documentation and tests–contributors simply write notebooks. This is why we’re excited to share nbdev v2. It’s rewritten from the ground up, with much-anticipated features including: nbdev in industry We have piloted nbdev at several companies. We were delighted to receive the following feedback, which fits our own experience using and developing nbdev: David Berg, on using nbdev for internal documentation at Netflix: “Prior to using nbdev, documentation was the most cumbersome aspect of our software development process… Using nbdev allows us to spend more time creating rich prose around the many code snippets guaranteeing the whole experience is robust. nbdev has turned what was once a chore into a natural extension of the notebook-based testing we were already doing.” Erik Gaasedelen, on using nbdev in production at Lyft: “I use this in production at my company. It’s an awesome tool… nbdev streamlines everything so I can write docs, tests, and code all in one place… The packaging is also really well thought out. From my point of view it is close to a Pareto improvement over traditional Python library development.” Hugo Bowne-Anderson, on using nbdev for Outerbounds: “nbdev has transformed the way we write documentation. Gone are the days of worrying about broken code examples when our API changes or [due to] human errors associated with copying & pasting code into markdown files. The authoring experience of nbdev… [allows] us to write prose and live code in a unified interface, which allows more experimentation… On top of this, nbdev allows us to include unit tests in our documentation which mitigates the burden of maintaining the docs over time.” Roxanna Pourzand, on using nbdev for Transform: “We’re so excited about using nbdev. Our product is technical so our resulting documentation includes a lot of code-based examples. Before nbdev, we had no way of maintaining our code examples and ensuring that it was up-to-date for both command inputs and outputs. It was all manual. With nbdev, we now have this under control in a sustainable way. Since we’ve deployed these docs, we also had a situation where we were able to identify a bug in one of our interfaces, which we found by seeing the error that was output in the documentation.” What’s nbdev? Nbdev embraces the dynamic nature of python and REPL-driven development in ways that traditional IDEs and software development workflows cannot. We thoroughly discussed the motivation, history, and goals of nbdev in this initial launch post three years ago. The creator of Jupyter, Fernando Pérez, told us: [Nbdev] should be celebrated and used a lot more - I have kept a tab with your original nbdev blog post open for months in Chrome because of how often I refer to it and point others to this work In short, nbdev embraces ideas from literate programming and exploratory programming. These paradigms have been revisited in platforms like XCode Playgrounds and languages like Smalltalk, LISP, and Mathematica. With nbdev, we sought to push these paradigms even further by enabling it for one of the most popular dynamic programming languages in the world: Python. Even though nbdev is most widely used in scientific computing communities due to its integration with Jupyter Notebooks, we’ve found that nbdev is well suited for a much wider range of software. We have used nbdev to write deep learning libraries, API clients, python language extensions,terminal user interfaces, and more! Hamel: When I use nbdev, my colleagues are often astounded by how quickly I can create and distribute high-quality python packages. I consider nbdev to be a superpower that allows me to create tests and documentation without any additional friction, which makes all of my projects more maintainable. I also find writing software with nbdev to be more fun and productive as I can iterate very fast on ideas relative to more traditional software engineering workflows. Lastly, with nbdev I can also use traditional text-based IDEs if I want to, so I get the best of both worlds. What we learned after three years of using nbdev While nbdev was originally developed to simplify the software development workflow for various fast.ai projects, we found that users wanted to extend nbdev to: While we created projects such as fastpages and fastdoc to accomplish some of these tasks, we realized that it would be better to have a single set of flexible tools to accomplish all of them. To this end, we were extremely excited to discover Quarto, an open-source technical publishing system built on pandoc. Hamel: The more I used nbdev for creating Python modules, the more I wanted to use it for writing blogs and documenting existing codebases. The ability to customize the way notebooks are rendered (hiding vs. showing cells, stripping output, etc.), along with the facilities for including unit tests, made it my go-to authoring tool for all technical content. I’m excited that nbdev2 unlocks all of these possibilities for everyone! Enter Quarto: A pandoc super-processor Quarto is a project that enables technical publishing with support for Jupyter Notebook, VSCode, Observable, and plaintext editors. Furthermore, Quarto enables the publishing of high-quality articles, reports, websites, and blogs in HTML, PDF, ePub, PowerPoint slides, and more. Quarto is maintained by RStudio, a company with a long history of products supporting literate programming, such as RMarkdown and RStudio. Quarto is built on top of Pandoc, a universal document converter that supports nearly any format you can think of. Pandoc achieves this seemingly magical feat by representing documents in a common abstract syntax tree (AST) that serves as the medium through which different formats can be translated. By extension, Quarto allows you to generate content in almost any format you wish! You can use pandoc filters to modify the AST and the output format, which allows you to use any static site generator you want, and programmatically modify and generate content. Quarto allows you to compose pandoc filters in a processing pipeline and apply them to specific documents or entire projects. You can also distribute filters as Quarto extensions, which makes Quarto extremely customizable. We also find Quarto compelling because user interfaces such as comment directives (comments that start with #|) correlate with nbdev. In fact, we even learned that nbdev inspired Quarto in this regard! In general, Quarto and nbdev share many goals, and the Quarto team has been incredibly responsive to our suggestions. For example, the ability to create notebook filters to modify notebooks before rendering. Below is a screenshot of a Jupyter notebook rendered with Quarto and nbdev. Finally, Quarto supports more programming languages than just Python and has been adding new features and fixing bugs at an impressive speed. This gives us confidence that we will be able to expand nbdev to support more use cases in the future. We discuss some of these future directions in the closing section. A blazing fast notebook kernel: execnb A core component of nbdev is executing and testing notebooks programmatically. It is important that this notebook runner executes with minimal overhead to maintain our goal of providing a delightful developer experience. This is why we built execnb, a lightweight notebook runner for Python kernels, which executes notebooks blazingly fast. Furthermore, execnb allows parameterized execution of notebooks. Hamel: I have been an enthusiastic user of tools like papermill that programmatically run notebooks for use-cases like creating dashboards or enabling new kinds of machine learning workflows. I believe execnb unlocks even more possibilities with its ability to inject arbitrary code at any place in a notebook, as well as the ability to pass callbacks that run before and/or after cells are executed. This opens up possibilities to create new types of workflows with notebooks that I am excited about exploring in the near future. Towards a dialect of python that embraces its dynamic nature One way to understand nbdev is part of an ecosystem that is designed to embrace Python’s dynamic properties for REPL-driven software engineering. Similar to Clojure, our goal is to provide tools that remove all friction from using the REPL in your programming workflow. We believe that the REPL enhances developer workflows thanks to context-sensitive auto-completion, signature inspection, and documentation–all based on the actual state of your code, and none of which are available in IDEs that depend solely on static analysis. We have found that for this reason, nbdev, with its Jupyter notebook foundation, makes programming significantly more productive and enjoyable. Our efforts to support REPL-driven development and literate programming are not limited to nbdev. We maintain a number of libraries that extend python to bolster this programming experience. The most notable of these libraries is fastcore, which extends Python in terms of testing, documenting code, metaprogramming, attribute helpers, enhanced representations of objects, and notebook-friendly patching. This blog post offers a gentle introduction to fastcore. In addition to literate programming, fastcore encourages conventions such as brevity and efficient use of vertical space so you can accomplish more with significantly less code. For example, below is a simple decorator that enables notebook-friendly patching: We believe that this combination of a new developer workflow (nbdev), Python extensions (fastcore), and associated norms form a new dialect of Python that is centered on leveraging its dynamic nature–in contrast to an ever-growing trend toward static analysis. We suspect that this dialect of Python will be more productive for programmers in many scenarios. We are framing this ecosystem as a “dialect” as it is still very much Python and is approachable by anyone who is familiar with the language. Furthermore, despite nbdev’s notebook workflow, our tools generate plaintext modules that can be navigated and edited with text-based IDEs, allowing programmers to experience the best of both worlds, if they desire. Hamel: I believe this framing of a Python dialect is key to properly understanding what nbdev is. While it may be tempting to get stuck on specific features or technical details of nbdev, it is useful to zoom out to understand the overall intent of creating a better workflow rather than conforming too rigidly to existing ones. A good analogy is TypeScript’s relationship with JavaScript: it is an extension of an existing programming language that supports a new way of programming. I encourage you to treat nbdev in a similar fashion: be willing to try new ways of programming and observe which tradeoffs resonate with you. At the very least, I believe nbdev is a fun way to experience a different way of writing software, which will broaden your horizons about programming in general, all without having to learn an entirely new programming language! The future of nbdev While we are excited about nbdev2, we believe we have only scratched the surface of what’s possible. We are considering the following features: If you have interesting ideas about how nbdev can be extended, please drop and chat with us on discord or post a message in the forums. How you can get started with nbdev Our project’s website is at nbdev.fast.ai, where we will be posting tutorials, examples, and more documentation in the coming days. Thank You This new version of nbdev was a team effort by many wonderful people. We want to highlight two people who have made outstanding contributions: Wasim Lorgat was instrumental across different areas, including significant contributions to fastcore, execnb, and nbdev, as well as the implementation of the new nbdev home page. With Wasim’s help, we were able to push nbdev to a new level of functionality and quality. JJ Allaire is not only the CEO of RStudio but also the steward of Quarto. JJ was incredibly responsive and eager to work with us on nbdev and added many features to Quarto specifically with nbdev in mind, such as notebook filters. We were also astounded by the attention to detail and the pace at which bugs are addressed. This new version of nbdev would not have been possible without JJ’s help, and we are excited to continue to work with him. We also want to thank the amazing fastai community, notably Isaac Flath, Benjamin Warner and Zach Mueller for their tireless work on this project. A conversation with JJ Allaire To celebrate the launch of nbdev v2 and Quarto, Jeremy sat down with the CEO of Posit (previously known as RStudio, the company behind Quarto), JJ Allaire, to talk about software development, scientific publishing, R, Python, literate programming, and much more. |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-07-28-nbdev2.html] | [TOKENS: 2772] |
nbdev+Quarto: A new secret weapon for productivity Hamel Husain and Jeremy Howard July 28, 2022 On this page Our new secret weapon for productivity Today we’re excited to announce that we’ve teamed up with Quarto to give nbdev superpowers. nbdev offers Python programmers a common set of tools for using Jupyter notebooks to: Although notebooks are already widely used for once-off exploratory work, it’s less well-known that they are perfectly capable of writing quality software. In fact, we’ve used nbdev for a wide range of software projects over the last three years, including deep learning libraries, API clients, Python language extensions, terminal user interfaces, and more. We discovered that it is not only capable of writing great software but that it has also increased our productivity by 300% or more. With nbdev, developers simply write notebooks with lightweight markup and get high-quality documentation, tests, continuous integration, and packaging for free! Nbdev has allowed us to maintain and scale manyopen source projects. Pull requests are often accompanied by detailed documentation and tests–contributors simply write notebooks. This is why we’re excited to share nbdev v2. It’s rewritten from the ground up, with much-anticipated features including: nbdev in industry We have piloted nbdev at several companies. We were delighted to receive the following feedback, which fits our own experience using and developing nbdev: David Berg, on using nbdev for internal documentation at Netflix: “Prior to using nbdev, documentation was the most cumbersome aspect of our software development process… Using nbdev allows us to spend more time creating rich prose around the many code snippets guaranteeing the whole experience is robust. nbdev has turned what was once a chore into a natural extension of the notebook-based testing we were already doing.” Erik Gaasedelen, on using nbdev in production at Lyft: “I use this in production at my company. It’s an awesome tool… nbdev streamlines everything so I can write docs, tests, and code all in one place… The packaging is also really well thought out. From my point of view it is close to a Pareto improvement over traditional Python library development.” Hugo Bowne-Anderson, on using nbdev for Outerbounds: “nbdev has transformed the way we write documentation. Gone are the days of worrying about broken code examples when our API changes or [due to] human errors associated with copying & pasting code into markdown files. The authoring experience of nbdev… [allows] us to write prose and live code in a unified interface, which allows more experimentation… On top of this, nbdev allows us to include unit tests in our documentation which mitigates the burden of maintaining the docs over time.” Roxanna Pourzand, on using nbdev for Transform: “We’re so excited about using nbdev. Our product is technical so our resulting documentation includes a lot of code-based examples. Before nbdev, we had no way of maintaining our code examples and ensuring that it was up-to-date for both command inputs and outputs. It was all manual. With nbdev, we now have this under control in a sustainable way. Since we’ve deployed these docs, we also had a situation where we were able to identify a bug in one of our interfaces, which we found by seeing the error that was output in the documentation.” What’s nbdev? Nbdev embraces the dynamic nature of python and REPL-driven development in ways that traditional IDEs and software development workflows cannot. We thoroughly discussed the motivation, history, and goals of nbdev in this initial launch post three years ago. The creator of Jupyter, Fernando Pérez, told us: [Nbdev] should be celebrated and used a lot more - I have kept a tab with your original nbdev blog post open for months in Chrome because of how often I refer to it and point others to this work In short, nbdev embraces ideas from literate programming and exploratory programming. These paradigms have been revisited in platforms like XCode Playgrounds and languages like Smalltalk, LISP, and Mathematica. With nbdev, we sought to push these paradigms even further by enabling it for one of the most popular dynamic programming languages in the world: Python. Even though nbdev is most widely used in scientific computing communities due to its integration with Jupyter Notebooks, we’ve found that nbdev is well suited for a much wider range of software. We have used nbdev to write deep learning libraries, API clients, python language extensions,terminal user interfaces, and more! Hamel: When I use nbdev, my colleagues are often astounded by how quickly I can create and distribute high-quality python packages. I consider nbdev to be a superpower that allows me to create tests and documentation without any additional friction, which makes all of my projects more maintainable. I also find writing software with nbdev to be more fun and productive as I can iterate very fast on ideas relative to more traditional software engineering workflows. Lastly, with nbdev I can also use traditional text-based IDEs if I want to, so I get the best of both worlds. What we learned after three years of using nbdev While nbdev was originally developed to simplify the software development workflow for various fast.ai projects, we found that users wanted to extend nbdev to: While we created projects such as fastpages and fastdoc to accomplish some of these tasks, we realized that it would be better to have a single set of flexible tools to accomplish all of them. To this end, we were extremely excited to discover Quarto, an open-source technical publishing system built on pandoc. Hamel: The more I used nbdev for creating Python modules, the more I wanted to use it for writing blogs and documenting existing codebases. The ability to customize the way notebooks are rendered (hiding vs. showing cells, stripping output, etc.), along with the facilities for including unit tests, made it my go-to authoring tool for all technical content. I’m excited that nbdev2 unlocks all of these possibilities for everyone! Enter Quarto: A pandoc super-processor Quarto is a project that enables technical publishing with support for Jupyter Notebook, VSCode, Observable, and plaintext editors. Furthermore, Quarto enables the publishing of high-quality articles, reports, websites, and blogs in HTML, PDF, ePub, PowerPoint slides, and more. Quarto is maintained by RStudio, a company with a long history of products supporting literate programming, such as RMarkdown and RStudio. Quarto is built on top of Pandoc, a universal document converter that supports nearly any format you can think of. Pandoc achieves this seemingly magical feat by representing documents in a common abstract syntax tree (AST) that serves as the medium through which different formats can be translated. By extension, Quarto allows you to generate content in almost any format you wish! You can use pandoc filters to modify the AST and the output format, which allows you to use any static site generator you want, and programmatically modify and generate content. Quarto allows you to compose pandoc filters in a processing pipeline and apply them to specific documents or entire projects. You can also distribute filters as Quarto extensions, which makes Quarto extremely customizable. We also find Quarto compelling because user interfaces such as comment directives (comments that start with #|) correlate with nbdev. In fact, we even learned that nbdev inspired Quarto in this regard! In general, Quarto and nbdev share many goals, and the Quarto team has been incredibly responsive to our suggestions. For example, the ability to create notebook filters to modify notebooks before rendering. Below is a screenshot of a Jupyter notebook rendered with Quarto and nbdev. Finally, Quarto supports more programming languages than just Python and has been adding new features and fixing bugs at an impressive speed. This gives us confidence that we will be able to expand nbdev to support more use cases in the future. We discuss some of these future directions in the closing section. A blazing fast notebook kernel: execnb A core component of nbdev is executing and testing notebooks programmatically. It is important that this notebook runner executes with minimal overhead to maintain our goal of providing a delightful developer experience. This is why we built execnb, a lightweight notebook runner for Python kernels, which executes notebooks blazingly fast. Furthermore, execnb allows parameterized execution of notebooks. Hamel: I have been an enthusiastic user of tools like papermill that programmatically run notebooks for use-cases like creating dashboards or enabling new kinds of machine learning workflows. I believe execnb unlocks even more possibilities with its ability to inject arbitrary code at any place in a notebook, as well as the ability to pass callbacks that run before and/or after cells are executed. This opens up possibilities to create new types of workflows with notebooks that I am excited about exploring in the near future. Towards a dialect of python that embraces its dynamic nature One way to understand nbdev is part of an ecosystem that is designed to embrace Python’s dynamic properties for REPL-driven software engineering. Similar to Clojure, our goal is to provide tools that remove all friction from using the REPL in your programming workflow. We believe that the REPL enhances developer workflows thanks to context-sensitive auto-completion, signature inspection, and documentation–all based on the actual state of your code, and none of which are available in IDEs that depend solely on static analysis. We have found that for this reason, nbdev, with its Jupyter notebook foundation, makes programming significantly more productive and enjoyable. Our efforts to support REPL-driven development and literate programming are not limited to nbdev. We maintain a number of libraries that extend python to bolster this programming experience. The most notable of these libraries is fastcore, which extends Python in terms of testing, documenting code, metaprogramming, attribute helpers, enhanced representations of objects, and notebook-friendly patching. This blog post offers a gentle introduction to fastcore. In addition to literate programming, fastcore encourages conventions such as brevity and efficient use of vertical space so you can accomplish more with significantly less code. For example, below is a simple decorator that enables notebook-friendly patching: We believe that this combination of a new developer workflow (nbdev), Python extensions (fastcore), and associated norms form a new dialect of Python that is centered on leveraging its dynamic nature–in contrast to an ever-growing trend toward static analysis. We suspect that this dialect of Python will be more productive for programmers in many scenarios. We are framing this ecosystem as a “dialect” as it is still very much Python and is approachable by anyone who is familiar with the language. Furthermore, despite nbdev’s notebook workflow, our tools generate plaintext modules that can be navigated and edited with text-based IDEs, allowing programmers to experience the best of both worlds, if they desire. Hamel: I believe this framing of a Python dialect is key to properly understanding what nbdev is. While it may be tempting to get stuck on specific features or technical details of nbdev, it is useful to zoom out to understand the overall intent of creating a better workflow rather than conforming too rigidly to existing ones. A good analogy is TypeScript’s relationship with JavaScript: it is an extension of an existing programming language that supports a new way of programming. I encourage you to treat nbdev in a similar fashion: be willing to try new ways of programming and observe which tradeoffs resonate with you. At the very least, I believe nbdev is a fun way to experience a different way of writing software, which will broaden your horizons about programming in general, all without having to learn an entirely new programming language! The future of nbdev While we are excited about nbdev2, we believe we have only scratched the surface of what’s possible. We are considering the following features: If you have interesting ideas about how nbdev can be extended, please drop and chat with us on discord or post a message in the forums. How you can get started with nbdev Our project’s website is at nbdev.fast.ai, where we will be posting tutorials, examples, and more documentation in the coming days. Thank You This new version of nbdev was a team effort by many wonderful people. We want to highlight two people who have made outstanding contributions: Wasim Lorgat was instrumental across different areas, including significant contributions to fastcore, execnb, and nbdev, as well as the implementation of the new nbdev home page. With Wasim’s help, we were able to push nbdev to a new level of functionality and quality. JJ Allaire is not only the CEO of RStudio but also the steward of Quarto. JJ was incredibly responsive and eager to work with us on nbdev and added many features to Quarto specifically with nbdev in mind, such as notebook filters. We were also astounded by the attention to detail and the pace at which bugs are addressed. This new version of nbdev would not have been possible without JJ’s help, and we are excited to continue to work with him. We also want to thank the amazing fastai community, notably Isaac Flath, Benjamin Warner and Zach Mueller for their tireless work on this project. A conversation with JJ Allaire To celebrate the launch of nbdev v2 and Quarto, Jeremy sat down with the CEO of Posit (previously known as RStudio, the company behind Quarto), JJ Allaire, to talk about software development, scientific publishing, R, Python, literate programming, and much more. |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-07-21-dl-coders-22.html] | [TOKENS: 1703] |
Practical Deep Learning for Coders 2022 Jeremy Howard July 21, 2022 On this page A new edition Today we’re releasing Practical Deep Learning for Coders 2022—a complete from-scratch rewrite of fast.ai’s most popular course, that’s been two years in the making. Previous fast.ai courses have been studied by hundreds of thousands of students, from all walks of life, from all parts of the world. fast.ai’s videos have been viewed over 6,000,000 times already! The major differences are: By the end of the second lesson, students will have built and deployed their own deep learning model on their own data. Many students post their course projects to our forum. For instance, if there’s an unknown dinosaur in your backyard, maybe you need this dinosaur classifier! Topics covered in this year’s course include: About the course There are 9 lessons, and each lesson is around 90 minutes long. The course is based on our 5-star rated book, which is freely available online. No special hardware or software is needed — the course shows how to use free resources for both building and deploying models. University math isn’t needed either — the necessary calculus and linear algebra is introduced as needed during the course. The course is taught by me, Jeremy Howard. I lead the development of fastai, the software used throughout this course. I have been using and teaching machine learning for around 30 years. I was the top-ranked competitor globally in machine learning competitions on Kaggle (the world’s largest machine learning community) two years running. Following this success, I became the President and Chief Scientist of Kaggle. Since first using neural networks over 25 years ago, I have led many companies and projects that have machine learning at their core, including founding the first company to focus on deep learning and medicine, Enlitic (chosen by MIT Tech Review as one of the “world’s smartest companies”), and Optimal Decisions, the first company to develop a fully optimised pricing algorithm for insurance. Students and results Many students have told us about how they’ve become multiple gold medal winners of international machine learning competitions, received offers from top companies, and having research papers published. For instance, Isaac Dimitrovsky told us that he had “been playing around with ML for a couple of years without really grokking it… [then] went through the fast.ai part 1 course late last year, and it clicked for me”. He went on to achieve first place in the prestigious international RA2-DREAM Challenge competition! He developed a multistage deep learning method for scoring radiographic hand and foot joint damage in rheumatoid arthritis, taking advantage of the fastai library. Alumni of previous editions of Practical Deep Learning for Coders have gone on to jobs at organizations like Google Brain, OpenAI, Adobe, Amazon, and Tesla, published research at top conferences such as NeurIPS, and created startups using skills they learned here. Petro Cuenca, lead developer of the widely-acclaimed Camera+ app, after completing the course went on to add deep learning features to his product, which was then featured by Apple for its “machine learning magic”. Peter Norvig, author of Artificial Intelligence: A Modern Approach and previously the Director of Research at Google, reviewed our book (which this course is based on) and had this to say: ‘Deep Learning is for everyone’ we see in Chapter 1, Section 1 of this book, and while other books may make similar claims, this book delivers on the claim. The authors have extensive knowledge of the field but are able to describe it in a way that is perfectly suited for a reader with experience in programming but not in machine learning. The book shows examples first, and only covers theory in the context of concrete examples. For most people, this is the best way to learn. The book does an impressive job of covering the key applications of deep learning in computer vision, natural language processing, and tabular data processing, but also covers key topics like data ethics that some other books miss. About deep learning Deep learning is a computer technique to extract and transform data–-with use cases ranging from human speech recognition to animal imagery classification–-by using multiple layers of neural networks. A lot of people assume that you need all kinds of hard-to-find stuff to get great results with deep learning, but as you’ll see in this course, those people are wrong. Here’s a few things you absolutely don’t need to do world-class deep learning: The lessons In the first five minutes you’ll see a complete end to end example of training and using a model that’s so advanced it was considered at the cutting edge of research capabilities in 2015! We discuss what deep learning and neural networks are, and what they’re useful for. We look at examples of deep learning for computer vision object classification, segmentation, tabular analysis, and collaborative filtering. This lesson shows how to design yown machine learning project, create your own dataset, train a model using your data, and finally deploy an application on the web. We use Hugging Face Space with Gradio for deployment, and also use JavaScript to implement an interface in the browser. (Deploying to other services looks very similar to the approach in this lesson.) Lesson 3 is all about the mathematical foundations of deep learning, such as Stochastic gradient descent (SGD), matrix products, and the flexibility of linear functions layered with non-linear activation functions. We focus particularly on a popular combination called the Rectified linear function (ReLU). We look at how to analyse natural language documents using Natural Language Processing (NLP). We be focus on the Hugging Face ecosystem, especially the Transformers library, and the vast collection of pretrained NLP models. The project for this lesson is to classify that similarity of phrases used to describe US patents. A similar approach can be applied to a wide variety of practical issues, in fields as wide-reaching as marketing, logistics, and medicine. In this lesson we look at how to create a neural network from scratch using Python and PyTorch, and how to implement a training loop for optimising the weights of a model. We build up from a single layer regression model up to a neural net with one hidden layer, and then to a deep learning model. Along the way we also look at how we can use a special function called sigmoid to make binary classification models easier to train, and we learn about metrics. Random forests started a revolution in machine learning 20 years ago. For the first time, there was a fast and reliable algorithm which made almost no assumptions about the form of the data, and required almost no preprocessing. In lesson 6, you’ll learn how a random forest really works, and how to build one from scratch. And, just as importantly, you’ll learn how to interpret random forests to better understand your data. You interact nearly every day with recommendation systems—algorithms which guess what products and services you might like, based on your past behavior. These systems largely rely on collaborative-filtering, an approach based on linear algebra that fills in the missing values in a matrix. In this lesson we’ll see two ways to do this: one based on a classic linear algebra formulation, and one based on deep learning. We finish off our study of collaborative filtering by looking closely at embeddings—a critical building block of many deep learning algorithms. Here we dive into convolutional neural networks (CNNs) and see how they really work. We used plenty of CNNs in earlier lessons, but we didn’t peeked inside them to see what’s really going on in there. As well as learning about the most fundamental building block of CNNs, the convolution, we also look at pooling, dropout, and more. A vibrant community Many fast.ai alumni have told us that one of their favorite things about the course is the generous and thoughtful community of interesting people that has sprung up around it. If you need help, or just want to chat about what you’re learning about (or show off what you’ve built!), there’s a wonderful online community ready to support you at forums.fast.ai. Every lesson has a dedicated forum thread—with many common questions already answered. For real-time conversations about the course, there’s also a very active Discord server. Get started To get started with the course now, head over to Practical Deep Learning for Coders 2022! |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-07-28-nbdev2.html] | [TOKENS: 2772] |
nbdev+Quarto: A new secret weapon for productivity Hamel Husain and Jeremy Howard July 28, 2022 On this page Our new secret weapon for productivity Today we’re excited to announce that we’ve teamed up with Quarto to give nbdev superpowers. nbdev offers Python programmers a common set of tools for using Jupyter notebooks to: Although notebooks are already widely used for once-off exploratory work, it’s less well-known that they are perfectly capable of writing quality software. In fact, we’ve used nbdev for a wide range of software projects over the last three years, including deep learning libraries, API clients, Python language extensions, terminal user interfaces, and more. We discovered that it is not only capable of writing great software but that it has also increased our productivity by 300% or more. With nbdev, developers simply write notebooks with lightweight markup and get high-quality documentation, tests, continuous integration, and packaging for free! Nbdev has allowed us to maintain and scale manyopen source projects. Pull requests are often accompanied by detailed documentation and tests–contributors simply write notebooks. This is why we’re excited to share nbdev v2. It’s rewritten from the ground up, with much-anticipated features including: nbdev in industry We have piloted nbdev at several companies. We were delighted to receive the following feedback, which fits our own experience using and developing nbdev: David Berg, on using nbdev for internal documentation at Netflix: “Prior to using nbdev, documentation was the most cumbersome aspect of our software development process… Using nbdev allows us to spend more time creating rich prose around the many code snippets guaranteeing the whole experience is robust. nbdev has turned what was once a chore into a natural extension of the notebook-based testing we were already doing.” Erik Gaasedelen, on using nbdev in production at Lyft: “I use this in production at my company. It’s an awesome tool… nbdev streamlines everything so I can write docs, tests, and code all in one place… The packaging is also really well thought out. From my point of view it is close to a Pareto improvement over traditional Python library development.” Hugo Bowne-Anderson, on using nbdev for Outerbounds: “nbdev has transformed the way we write documentation. Gone are the days of worrying about broken code examples when our API changes or [due to] human errors associated with copying & pasting code into markdown files. The authoring experience of nbdev… [allows] us to write prose and live code in a unified interface, which allows more experimentation… On top of this, nbdev allows us to include unit tests in our documentation which mitigates the burden of maintaining the docs over time.” Roxanna Pourzand, on using nbdev for Transform: “We’re so excited about using nbdev. Our product is technical so our resulting documentation includes a lot of code-based examples. Before nbdev, we had no way of maintaining our code examples and ensuring that it was up-to-date for both command inputs and outputs. It was all manual. With nbdev, we now have this under control in a sustainable way. Since we’ve deployed these docs, we also had a situation where we were able to identify a bug in one of our interfaces, which we found by seeing the error that was output in the documentation.” What’s nbdev? Nbdev embraces the dynamic nature of python and REPL-driven development in ways that traditional IDEs and software development workflows cannot. We thoroughly discussed the motivation, history, and goals of nbdev in this initial launch post three years ago. The creator of Jupyter, Fernando Pérez, told us: [Nbdev] should be celebrated and used a lot more - I have kept a tab with your original nbdev blog post open for months in Chrome because of how often I refer to it and point others to this work In short, nbdev embraces ideas from literate programming and exploratory programming. These paradigms have been revisited in platforms like XCode Playgrounds and languages like Smalltalk, LISP, and Mathematica. With nbdev, we sought to push these paradigms even further by enabling it for one of the most popular dynamic programming languages in the world: Python. Even though nbdev is most widely used in scientific computing communities due to its integration with Jupyter Notebooks, we’ve found that nbdev is well suited for a much wider range of software. We have used nbdev to write deep learning libraries, API clients, python language extensions,terminal user interfaces, and more! Hamel: When I use nbdev, my colleagues are often astounded by how quickly I can create and distribute high-quality python packages. I consider nbdev to be a superpower that allows me to create tests and documentation without any additional friction, which makes all of my projects more maintainable. I also find writing software with nbdev to be more fun and productive as I can iterate very fast on ideas relative to more traditional software engineering workflows. Lastly, with nbdev I can also use traditional text-based IDEs if I want to, so I get the best of both worlds. What we learned after three years of using nbdev While nbdev was originally developed to simplify the software development workflow for various fast.ai projects, we found that users wanted to extend nbdev to: While we created projects such as fastpages and fastdoc to accomplish some of these tasks, we realized that it would be better to have a single set of flexible tools to accomplish all of them. To this end, we were extremely excited to discover Quarto, an open-source technical publishing system built on pandoc. Hamel: The more I used nbdev for creating Python modules, the more I wanted to use it for writing blogs and documenting existing codebases. The ability to customize the way notebooks are rendered (hiding vs. showing cells, stripping output, etc.), along with the facilities for including unit tests, made it my go-to authoring tool for all technical content. I’m excited that nbdev2 unlocks all of these possibilities for everyone! Enter Quarto: A pandoc super-processor Quarto is a project that enables technical publishing with support for Jupyter Notebook, VSCode, Observable, and plaintext editors. Furthermore, Quarto enables the publishing of high-quality articles, reports, websites, and blogs in HTML, PDF, ePub, PowerPoint slides, and more. Quarto is maintained by RStudio, a company with a long history of products supporting literate programming, such as RMarkdown and RStudio. Quarto is built on top of Pandoc, a universal document converter that supports nearly any format you can think of. Pandoc achieves this seemingly magical feat by representing documents in a common abstract syntax tree (AST) that serves as the medium through which different formats can be translated. By extension, Quarto allows you to generate content in almost any format you wish! You can use pandoc filters to modify the AST and the output format, which allows you to use any static site generator you want, and programmatically modify and generate content. Quarto allows you to compose pandoc filters in a processing pipeline and apply them to specific documents or entire projects. You can also distribute filters as Quarto extensions, which makes Quarto extremely customizable. We also find Quarto compelling because user interfaces such as comment directives (comments that start with #|) correlate with nbdev. In fact, we even learned that nbdev inspired Quarto in this regard! In general, Quarto and nbdev share many goals, and the Quarto team has been incredibly responsive to our suggestions. For example, the ability to create notebook filters to modify notebooks before rendering. Below is a screenshot of a Jupyter notebook rendered with Quarto and nbdev. Finally, Quarto supports more programming languages than just Python and has been adding new features and fixing bugs at an impressive speed. This gives us confidence that we will be able to expand nbdev to support more use cases in the future. We discuss some of these future directions in the closing section. A blazing fast notebook kernel: execnb A core component of nbdev is executing and testing notebooks programmatically. It is important that this notebook runner executes with minimal overhead to maintain our goal of providing a delightful developer experience. This is why we built execnb, a lightweight notebook runner for Python kernels, which executes notebooks blazingly fast. Furthermore, execnb allows parameterized execution of notebooks. Hamel: I have been an enthusiastic user of tools like papermill that programmatically run notebooks for use-cases like creating dashboards or enabling new kinds of machine learning workflows. I believe execnb unlocks even more possibilities with its ability to inject arbitrary code at any place in a notebook, as well as the ability to pass callbacks that run before and/or after cells are executed. This opens up possibilities to create new types of workflows with notebooks that I am excited about exploring in the near future. Towards a dialect of python that embraces its dynamic nature One way to understand nbdev is part of an ecosystem that is designed to embrace Python’s dynamic properties for REPL-driven software engineering. Similar to Clojure, our goal is to provide tools that remove all friction from using the REPL in your programming workflow. We believe that the REPL enhances developer workflows thanks to context-sensitive auto-completion, signature inspection, and documentation–all based on the actual state of your code, and none of which are available in IDEs that depend solely on static analysis. We have found that for this reason, nbdev, with its Jupyter notebook foundation, makes programming significantly more productive and enjoyable. Our efforts to support REPL-driven development and literate programming are not limited to nbdev. We maintain a number of libraries that extend python to bolster this programming experience. The most notable of these libraries is fastcore, which extends Python in terms of testing, documenting code, metaprogramming, attribute helpers, enhanced representations of objects, and notebook-friendly patching. This blog post offers a gentle introduction to fastcore. In addition to literate programming, fastcore encourages conventions such as brevity and efficient use of vertical space so you can accomplish more with significantly less code. For example, below is a simple decorator that enables notebook-friendly patching: We believe that this combination of a new developer workflow (nbdev), Python extensions (fastcore), and associated norms form a new dialect of Python that is centered on leveraging its dynamic nature–in contrast to an ever-growing trend toward static analysis. We suspect that this dialect of Python will be more productive for programmers in many scenarios. We are framing this ecosystem as a “dialect” as it is still very much Python and is approachable by anyone who is familiar with the language. Furthermore, despite nbdev’s notebook workflow, our tools generate plaintext modules that can be navigated and edited with text-based IDEs, allowing programmers to experience the best of both worlds, if they desire. Hamel: I believe this framing of a Python dialect is key to properly understanding what nbdev is. While it may be tempting to get stuck on specific features or technical details of nbdev, it is useful to zoom out to understand the overall intent of creating a better workflow rather than conforming too rigidly to existing ones. A good analogy is TypeScript’s relationship with JavaScript: it is an extension of an existing programming language that supports a new way of programming. I encourage you to treat nbdev in a similar fashion: be willing to try new ways of programming and observe which tradeoffs resonate with you. At the very least, I believe nbdev is a fun way to experience a different way of writing software, which will broaden your horizons about programming in general, all without having to learn an entirely new programming language! The future of nbdev While we are excited about nbdev2, we believe we have only scratched the surface of what’s possible. We are considering the following features: If you have interesting ideas about how nbdev can be extended, please drop and chat with us on discord or post a message in the forums. How you can get started with nbdev Our project’s website is at nbdev.fast.ai, where we will be posting tutorials, examples, and more documentation in the coming days. Thank You This new version of nbdev was a team effort by many wonderful people. We want to highlight two people who have made outstanding contributions: Wasim Lorgat was instrumental across different areas, including significant contributions to fastcore, execnb, and nbdev, as well as the implementation of the new nbdev home page. With Wasim’s help, we were able to push nbdev to a new level of functionality and quality. JJ Allaire is not only the CEO of RStudio but also the steward of Quarto. JJ was incredibly responsive and eager to work with us on nbdev and added many features to Quarto specifically with nbdev in mind, such as notebook filters. We were also astounded by the attention to detail and the pace at which bugs are addressed. This new version of nbdev would not have been possible without JJ’s help, and we are excited to continue to work with him. We also want to thank the amazing fastai community, notably Isaac Flath, Benjamin Warner and Zach Mueller for their tireless work on this project. A conversation with JJ Allaire To celebrate the launch of nbdev v2 and Quarto, Jeremy sat down with the CEO of Posit (previously known as RStudio, the company behind Quarto), JJ Allaire, to talk about software development, scientific publishing, R, Python, literate programming, and much more. |
======================================== |
[SOURCE: https://en.wikipedia.org/wiki/Doom_(1993_video_game)] | [TOKENS: 8057] |
Contents Doom (1993 video game) Doom is a 1993 first-person shooter game developed and published by id Software for MS-DOS. It is the first installment in the Doom franchise. The player assumes the role of a space marine, later unofficially referred to as Doomguy, fighting through hordes of undead humans and invading demons. The game begins on the moons of Mars and finishes in hell, with the player traversing each level to find its exit or defeat its final boss. It is an early example of 3D graphics in video games, and has enemies and objects as 2D images, a technique sometimes referred to as 2.5D graphics. Doom was the third major independent release by id Software, after Commander Keen (1990–1991) and Wolfenstein 3D (1992). In May 1992, id started developing a darker game focused on fighting demons with technology, using a new 3D game engine from the lead programmer, John Carmack. The designer Tom Hall initially wrote a science fiction plot, but he and most of the story were removed from the project, with the final game featuring an action-heavy design by John Romero and Sandy Petersen. Id published Doom as a set of three episodes under the shareware model, marketing the full game by releasing the first episode free. A retail version with an additional episode was published in 1995 by GT Interactive as The Ultimate Doom. Doom was a critical and commercial success, earning a reputation as one of the best and most influential video games of all time. It sold an estimated 3.5 million copies by 1999, and up to 20 million people are estimated to have played it within two years of launch. It has been termed the "father" of first-person shooters and is regarded as one of the most important games in the genre. It has been cited by video game historians as shifting the direction and public perception of the medium as a whole, as well as sparking the rise of online games and communities. It led to an array of imitators and clones, as well as a robust modding scene and the birth of speedrunning as a community. Its high level of graphic violence led to controversy from a range of groups. Doom has been ported to a variety of platforms both officially and unofficially and has been followed by several games in the series, including Doom II (1994), Doom 64 (1997), Doom 3 (2004), Doom (2016), Doom Eternal (2020), and Doom: The Dark Ages (2025), as well as the films Doom (2005) and Doom: Annihilation (2019). Gameplay Doom is a first-person shooter presented with 3D graphics. While the environment is shown in a 3D perspective, the enemies and objects are instead 2D sprites rendered at fixed angles, a technique sometimes referred to as 2.5D graphics or billboarding. In the single-player campaign mode, the player controls an unnamed space marine—later unofficially termed "Doomguy"—through military bases on the moons of Mars and in hell. To finish a level, the player must traverse through labyrinthine areas to reach a marked exit room. Levels are grouped into named episodes, with the final level of each focusing on a boss fight. While traversing the levels, the player must fight a variety of enemies, including demons and possessed undead humans. Enemies often appear in large groups. The five difficulty levels adjust the number of enemies and amount of damage they do, with enemies moving and attacking faster than normal on the hardest difficulty setting. The monsters have simple behavior: they move toward their opponent if they see or hear them, and attack by biting, clawing, or using magic abilities such as fireballs. The player must manage supplies of ammunition, health, and armor while traversing the levels. The player can find weapons and ammunition throughout the levels or can collect them from dead enemies, including a pistol, a shotgun, a chainsaw, a plasma rifle, and the BFG 9000. The player also encounters pits of toxic waste, ceilings that lower and crush objects, and locked doors requiring a collectable keycard or a remote switch. Power-ups include health or armor points, a mapping computer, partial invisibility, a radiation suit against toxic waste, invulnerability, or a super-strong melee berserker status. Cheat codes allow the player to unlock all weapons, walk through walls, or become invulnerable. Two multiplayer modes are playable over a network: cooperative, in which two to four players team up to complete the main campaign, and deathmatch, in which two to four players compete to kill the other players' characters as many times as possible. Multiplayer was initially only playable over local networks, but a four-player online multiplayer mode was made available one year after launch through the DWANGO service. Plot Doom is divided into three episodes, each containing eight main levels: "Knee-Deep in the Dead", "The Shores of Hell", and "Inferno". A fourth episode, "Thy Flesh Consumed", was added in an expanded version, The Ultimate Doom, released two years after Doom. The campaign contains very few plot elements, with a minimal story presented mostly through the instruction manual and text descriptions between episodes. In the future, an unnamed marine is posted to a dead-end assignment on Mars after assaulting a superior officer who ordered his unit to fire on civilians. The Union Aerospace Corporation, which operates radioactive waste facilities there, allows the military to conduct secret teleportation experiments that turn deadly. A base on Phobos urgently requests military support, while Deimos disappears entirely, and the marine joins a combat force to secure Phobos. He secures the perimeter as ordered while the entire response team is wiped out. With no way off the moon, and armed with only a pistol, he enters the base intent on revenge. In "Knee-Deep in the Dead", the marine fights demons and possessed humans in the military and waste facilities on Phobos. The episode ends with the marine defeating two powerful Barons of Hell guarding a teleporter to the Deimos base. After the battle, the marine passes through the teleporter and is knocked unconscious by a horde of enemies, awakening with only a pistol. In "The Shores of Hell", the marine fights through corrupted research facilities on Deimos, culminating in the defeat of a gigantic cyberdemon. From an overlook, he discovers that the moon is floating above hell and rappels down to the surface. In "Inferno", the marine battles through hell itself and destroys a cybernetic spider-demon that masterminded the invasion of the moons. When a portal to Earth opens, the marine steps through to discover that Earth has been invaded. "Thy Flesh Consumed" follows the marine's initial assault on the Earth invaders, setting the stage for Doom II. Development Id Software released Wolfenstein 3D in May 1992. Later called the "grandfather of 3D shooters", it established the genre's popularity and its reputation for fast action and technological advancement. When most of the studio began work on additional episodes for Wolfenstein, id co-founder and lead programmer John Carmack instead began technical research on a new game. Following the release of Wolfenstein 3D: Spear of Destiny in September 1992, the team began to plan their next game. They were tired of Wolfenstein and wanted to create another 3D game using a new engine Carmack was developing. Co-founder and lead designer Tom Hall proposed a new game in the Commander Keen series, but the team decided that the Keen platforming gameplay was a poor fit for Carmack's fast-paced 3D engines. Additionally, the other co-founders, designer John Romero and lead artist Adrian Carmack (no relation to John Carmack) wanted to create something in a darker style than the Keen games. John Carmack conceived a game about using technology to fight demons, inspired by a Dungeons & Dragons campaign the team played. This campaign would also influence the design of Quake (1996) and Daikatana (2000). More broadly the team intended to combine the styles of the Evil Dead II and Aliens films. The working title was Green and Pissed, but Carmack renamed it Doom based on a line from the 1986 film The Color of Money: "'What you got in there?' / 'In here? Doom.'" The team agreed to pursue the Doom concept, and development began in November 1992. The initial development team was composed of five people: programmers John Carmack and Romero, artists Adrian Carmack and Kevin Cloud, and designer Hall. They moved operations to a dark office building, naming it "Suite 666" while drawing inspiration from the noises they heard from a neighboring dental practice. They also decided to cut ties with Apogee Software, their previous publisher, and self-publish Doom, as they felt that they were outgrowing the publisher and could make more money by self-publishing. In November, Hall delivered a design document that he called the "Doom Bible", detailing the project's plot, backstory, and design goals. His design was a science fiction horror concept wherein scientists on the Moon open a portal to an alien invasion. Over a series of levels, the player discovers that the aliens are demons while hell steadily infects the level design. John Carmack not only disliked the proposed story but dismissed the idea of having a story at all: "Story in a game is like story in a porn movie; it's expected to be there, but it's not that important." Rather than a deep story, he wanted to focus on technological innovation, dropping the levels and episodes of Wolfenstein in favor of a fast, continuous world. Hall disliked the idea, but the rest of the team sided with Carmack. Hall spent the next few weeks reworking the Doom Bible to work with Carmack's technological ideas. However, the team then realized that Carmack's vision for a seamless world would be impossible given the hardware limitations, and Hall was forced to rework the design document once again. At the start of 1993, id put out a press release, touting Hall's story about fighting off demons while "knee-deep in the dead". The press release proclaimed the new 3D engine features that John Carmack had created, as well as aspects including multiplayer, that had not yet even been designed. Early versions were built to match the Doom Bible, and a "pre-alpha" version of the first level included Hall's introductory base scene. Initial versions also retained Wolfenstein's arcade-style scoring, but this was later removed as it clashed with Doom's intended tone. The studio also experimented with other game systems before removing them, such as lives, an inventory, a secondary shield, and a complex user interface. Soon, however, the Doom Bible as a whole was rejected. Romero wanted a game even "more brutal and fast" than Wolfenstein, which did not leave room for the character-driven plot Hall had created. Additionally, the team believed it emphasized realism over entertaining gameplay, and they did not see the need for a design document at all. Some ideas were retained, but the story was dropped and most of the design was removed. By early 1993, Hall created levels that became part of an internal demo. Carmack and Romero, however, rejected the military architecture of Hall's level design. Romero especially believed that the boxy, flat level designs failed to innovate on Wolfenstein, and failed to show off the engine's capabilities. He began to create his own, more abstract levels, which the rest of the team saw as a great improvement. Hall was upset with the reception of his designs and how little impact he was having as the lead designer. He was also upset with how much he was having to fight with John Carmack to get what he saw as obvious gameplay improvements, such as flying enemies, and began to spend less time at work. The other developers, however, felt that Hall was not in sync with the team's vision and was becoming a problem. In July the other founders of id fired Hall, who went to work for Apogee. He was replaced by Sandy Petersen in September, ten weeks before the game was released. Petersen later recalled that John Carmack and Romero wanted to hire other artists instead, but Cloud and Adrian disagreed, saying that a designer was required to help build a cohesive gameplay experience. The team also added a third programmer, Dave Taylor. Petersen and Romero designed the rest of Doom's levels, with different aims: the team believed that Petersen's designs were more technically interesting and varied, while Romero's were more aesthetically interesting. In late 1993, a month before release, John Carmack began to add multiplayer. After the multiplayer component was coded, the development team began playing four-player games, which Romero termed "deathmatch", and Cloud named the act of killing other players "fragging". According to Romero, the deathmatch mode was inspired by fighting games such as Street Fighter II, Fatal Fury, and Art of Fighting. Doom was written largely in the C programming language, with a few elements in assembly language. The developers used NeXT computers running the NeXTSTEP operating system. The level and graphical data was stored in WAD files, short for "Where's All the Data?", separately from the engine. This allowed for any part of the design to be changed without needing to adjust the engine code. Carmack designed this system so that fans could easily modify the game; he had been impressed by the modifications made by fans of Wolfenstein 3D and wanted to support that by releasing a map editor with an easily swappable file structure. Unlike Wolfenstein, which has flat levels with walls at right angles, the Doom engine allows for walls and floors at any angle or height but does not allow areas to be stacked vertically. The lighting system is based on adjusting the color palette of surfaces directly. Rather than calculating how light traveled from light sources to surfaces using ray tracing, the game calculates the "light level" of a small area based on the predetermined brightness of said area. It then modifies the color palette of that section's surface textures to mimic how dark it would look. This same system is used to cause far away surfaces to look darker than close ones. Romero came up with new ways to use Carmack's lighting engine, such as strobe lights. He programmed engine features such as switches and movable stairs and platforms. After Romero's complex level designs started to cause problems with the engine, Carmack began to use binary space partitioning to quickly select the reduced portion of a level that the player could see at a given time. Taylor, along with programming other features, added cheat codes to aid in development and left them in for players. Adrian Carmack was the lead artist for Doom, with Kevin Cloud as an additional artist. They designed the monsters to be "nightmarish", with graphics that were realistic and dark instead of staged or rendered. A mixed media approach was taken to create them. The artists sculpted models of some of the enemies and took pictures of them in stop motion from five to eight different angles so that they could be rotated realistically in-game. The images were then digitized and converted to 2D characters with a program written by John Carmack. Adrian Carmack made clay models for a few demons and had Gregor Punchatz build latex and metal sculptures of the others. The weapons were made from combined parts of children's toys. The developers photographed themselves as well, using Cloud's arm for the marine's arm holding a gun, and Adrian's snakeskin boots and wounded knee for textures. The cover art was created by Don Ivan Punchatz, Gregor Punchatz's father, who worked from a short description of the game rather than detailed references. Romero was the body model used for cover; he posed during a photoshoot to demonstrate to the intended model what the pose should look like, and Punchatz used his photo. As with Wolfenstein 3D, id hired composer Bobby Prince to create the music and sound effects. Romero directed Prince to make the music in techno and metal styles. Many tracks were directly inspired by songs by metal bands such as Alice in Chains and Pantera. Prince believed that ambient music would be more appropriate and produced numerous tracks in both styles in hope of convincing the team, and Romero incorporated both. Prince did not make music for specific levels, as they were composed before the levels were completed. Instead, Romero assigned each track to each level late in development. Prince created the sound effects based on short descriptions or concept art of a monster or weapon and adjusted them to match the completed animations. The monster sounds were created from animal noises, and Prince designed all the sounds to be distinct on the limited sound hardware of the time, even when many sounds were playing at once. He also designed the sound effects to play on different frequencies from those used for the MIDI music, so they would clearly cut through the music. Release Id Software planned to self-publish Doom for DOS-based computers and set up a distribution system leading up to the release. Jay Wilbur, who had been hired as CEO and sole member of the business team, planned the marketing and distribution of Doom. As id would make the most money from copies they sold directly to customers—up to 85% of the planned US$40 price—he decided to leverage the shareware market as much as possible. He believed that the mainstream press was uninterested in the game and bought only a single ad in any gaming magazine. Instead, he gave software retailers the option to sell copies of the first Doom episode at any price, in hopes of motivating customers to buy the full game directly from id. In 2004, John Carmack estimated that the total cost of development was less than US$1 million. The team planned to release Doom in the third quarter of 1993 but ultimately needed more time. By December 1993, the team was working non-stop, with several employees sleeping at the office. Taylor said that the work gave him such a rush that he would pass out from the intensity. Id only gave a single press preview, to Computer Gaming World in June, to a glowing response, but had also released development updates to the public continuously throughout development on the nascent internet. Id began receiving calls from people interested in the game or angry that it had missed its planned release date, as anticipation built over the year. At midnight on December 10, 1993, after working for 30 straight hours testing, the development team at id uploaded the first episode to the internet, letting interested players distribute it for them. The team was unable to connect to the FTP server at the University of Wisconsin–Madison where they planned to upload the game, since there were so many users already connected in anticipation of the release. The network administrator was forced to first increase the number of connections, and then kick off all users to make room. When the upload finished 30 minutes later, 10,000 people attempted to download the game at once, crashing the university's network. Within hours of Doom's release, university networks began banning Doom multiplayer games, as a rush of players overwhelmed their systems. The morning after release, John Carmack quickly released a patch in response to complaints of network congestion from administrators, who still needed to implement Doom-specific rules to keep their networks from crashing from the load. In 1995, id created an expanded version of Doom for the retail market with a fourth episode of levels, which was published by GT Interactive as The Ultimate Doom. Doom has also been ported to numerous different platforms, independent from id Software. The first port of Doom was an unofficial port to Linux, released by id programmer Dave Taylor in 1994; it was hosted by id but not supported or made official. Microsoft attempted to hire id to port Doom to Windows in 1995 to promote Windows as a gaming platform, and Microsoft CEO Bill Gates briefly considered buying the company. When id declined, Microsoft made its own licensed port, with a team led by Gabe Newell. One promotional video for Windows 95 had Gates digitally superimposed into the game. Other official ports of Doom were released for the 32X and Atari Jaguar in 1994, Super NES and PlayStation in 1995, 3DO in 1996, Sega Saturn in 1997, Acorn Risc PC in 1998, Game Boy Advance in 2001, Xbox 360 in 2006, iOS in 2009, and Nintendo Switch, Xbox One, PlayStation 4, and Android in 2019, with the latter-most platforms (excluding Android) receiving a further expanded port alongside Doom II in 2024 along with ports for the PlayStation 5 and Xbox Series X/S, alongside the 2016 "IDKFA" arranged soundtrack by Andrew Hulshult. Some of these became bestsellers even many years after the initial release. The ports did not all have the same content, with some having fewer levels, such as the 32X port created by John Carmack, which was released with only two-thirds of the game's levels in order to meet the console's launch date, while the PlayStation port includes The Ultimate Doom and Doom II. The source code for Doom was released under a non-commercial license in 1997, and freely released under the GNU General Public License in 1999. Due to the release of its source code, Doom has been unofficially ported to numerous platforms. These ports include esoteric devices such as smart thermostats, pianos, and Doom itself, which led to variations of a long-running meme, "Can it run Doom?" and "It runs Doom". Reception Upon its release in December 1993, Doom became an "overnight phenomenon". It was an immediate financial success for id, making a profit within a day after release. Although the company estimated that only 1% of shareware downloaders bought the full game, this was enough to generate initial daily revenue of US$100,000, selling in one day what Wolfenstein had sold in one month. By May 1994, Wilbur said that the game had sold over 65,000 copies, and estimated that the shareware version had been distributed over 1 million times. In 1995, Wilbur estimated the first-year sales as 140,000, while in 2002 Petersen said it had sold around 200,000 copies in its first year. By late 1995, Doom was estimated to be installed on more computers worldwide than Microsoft's new operating system, Windows 95. According to Wilbur, by June 1996 it had been downloaded 20 million times. According to PC Data, by April 1998 Doom's shareware edition had yielded 1.36 million units sold and US$8.74 million in revenue in the United States. This led PC Data to declare it the country's 4th-best-selling computer game since 1993. The Ultimate Doom sold over 780,000 units by September 1999, and all versions combined sold 3.5 million copies by the end of 1999. In addition to sales, an estimated six million people played the shareware version by 2002; other sources estimated in 2000 that 10–20 million people played Doom within 24 months of its launch. Doom was highly praised in contemporaneous reviews. In April 1994, a few months after release, PC Gamer UK named it the third-best computer game of all time, claiming "Doom has already done more to establish the PC's arcade clout than any other title in gaming history," and PC Gamer US named it the best computer game of all time that August. It won the Best Action Adventure award at Cybermania '94. GamesRadar UK named Doom Game of the Year in 1993 shortly after release, and Computer Gaming World and PC Gamer UK did the same the year after. Reviewers heavily praised the single-player gameplay: Electronic Entertainment called it "a skull-banging, palm-sweating, blood-pounding game", while The Age said it was "a technically superb and thrilling 3D adventure". White Wolf's reviewer found it addictive, claiming to miss sleep and appointments to continue playing. PC Zone called it the best arcade game ever, and it and Computer Gaming World praised the variety of monsters and weapons. Computer Gaming World concluded that it was "a virtuoso performance". Other reviewers, while also praising the gameplay, commented on the lack of complexity: Computer and Video Games found it captivating and praised the variety and complexity of the level design but called the overall gameplay repetitive, while Dragon similarly praised the fast gameplay and level design but said that overall it lacked depth. Edge praised the graphics and levels but criticized the straightforward shooting gameplay. The review concluded: "If only you could talk to these creatures, then perhaps you could try and make friends with them, form alliances... Now, that would be interesting." The review attracted mockery and "if only you could talk to these creatures" became a running joke in video game culture. The multiplayer gameplay garnered praise: Computer Gaming World called it "the most intense gaming experience available", and Dragon called it "the biggest adrenaline rush available on computers". PC Zone named it as the best multiplayer game available, in addition to the best arcade game. The 3D graphics and art style earned glowing reviews as well; Computer Gaming World called the graphics remarkable, while Edge said that it "made serious advances in what people will expect of 3D graphics in future", surpassing not only prior games but games that had yet to be released. Compute! and Electronic Games similarly called the graphics excellent and unlike any other game's. PC Zone, Dragon, Computer Gaming World, and Electronic Entertainment all lauded the atmosphere and art direction, saying that the level design, lighting effects, and sound effects combined to create a "claustrophobic" and "nightmarish experience". Computer Gaming World also praised the music, as did The Mercury News, which called it as "ominous as the scenario". The Ultimate Doom received mixed reviews upon its release in 1995, as in the review from PC Zone, which gave it a score of 90/100 for new players but 20/100 for anyone who had the original game. The reviewer viewed it as solely a level pack due to the lack of new features and compared it negatively to the hundreds of free fan-made levels available on the internet. Joystick disliked the limited amount of additional content and recommended it only to major fans or those who had not played it. Fusion reviewed the edition positively, praising the difficulty of the new levels, as did GameSpot, which reviewed it from the perspective of introducing the game to new players. The first ports of Doom received comparable reviews to the original PC version. VideoGames, GamePro, and Computer and Video Games all gave the Jaguar version high scores, comparing it favorably with the PC version. GamePro and Computer and Video Games also rated the 32X version highly, though they noted that the graphics were worse and the game shorter than the PC or Jaguar versions. The 1995 ports received mixed reviews. The PlayStation version was rated highly by HobbyConsolas, GamePro, and Maximum, which praised the inclusion of Doom II and extra levels, and favorably compared it to other PlayStation shooter games. The Super NES version, however, was noted for weaker graphics and unresponsive controls, though reviewers such as Computer and Video Games, GamePro, and Next Generation were split on awarding high or middling scores due to these faults. Later 1990s ports received worse reviews; the 3DO port was panned by GamePro and Maximum for having worse graphics, a smaller screen size, and less intelligent enemies than any previous version, and the Sega Saturn port also met with low reviews for poor graphics and low quality from Mean Machines and Sega Saturn Magazine. Legacy Doom has been termed "inarguably the most important" first-person shooter, as well as the "father" of the genre. Although not the first in the genre, it was the game with the greatest impact. Dan Pinchbeck in Doom: Scarydarkfast (2013) noted the direct influence of Doom's design choices on those of first-person and third-person shooter games two decades later, as influenced by the games released in the intervening years. Doom, and to a lesser extent Wolfenstein 3D, has been characterized as "mark[ing] a turning point" in the perception of video games in popular culture, with Doom and first-person shooters in general becoming the predominant perception of video games in media. Historians such as Tristan Donovan in Replay: The History of Video Games (2010) have termed it as causing a "paradigm shift", prompting the rise in popularity of 3D games, first-person shooters, licensed technology between developers, and support for game modifications. It helped spark the rise of both online multiplayer games and player-driven content generation, and popularized the business model of online distribution. In their book Dungeons & Dreamers: A Story of how Computer Games Created a Global Community in 2014, Brad King and John Borland claimed that Doom was one of the first widespread instances of an "online collective virtual reality", and did more than any other game to create a modern world of "networked games and gamers". PC Gamer proclaimed Doom the most influential game of all time in 2004, and in 2023 said its development was one of the most well-documented in the history of video games. It has also been used in scholarly research since its release, including for machine learning, video game aesthetics and design, and the effects of video games on aggression, memory, and attention. In 2007 Doom was listed among the ten "game canon" video games selected for preservation by the Library of Congress, and in 2015 The Strong National Museum of Play inducted Doom to its World Video Game Hall of Fame as part of its initial set of games. Doom has continued to be included highly in lists of the best video games ever since its release. In 1995, Next Generation said it was "the most talked about PC game ever". The PC version was ranked the 3rd best video game by Flux in 1995, and in 1996 was ranked fifth best and third most innovative by Computer Gaming World. In 2000, Doom was ranked as the second-best game ever by GameSpot. The following year, it was voted the number one game of all time in a poll among over 100 game developers and journalists conducted by GameSpy, and was ranked the sixth best game by Game Informer. GameTrailers ranked it the most "breakthrough PC game" in 2009 and Game Informer again ranked it the sixth-best game that same year. Doom has also been ranked among the best games of all time by GamesMaster, Hyper, The Independent, Entertainment Weekly, GamesTM, Jeuxvideo.com, Gamereactor, Time, Polygon, and The Times, among others, as recently as 2023. The success of Doom led to dozens of new first-person shooter games. In 1998, PC Gamer declared it "probably the most imitated game of all time". These games were often referred to as "Doom clones", with "first-person shooter" only overtaking it as the name of the genre after a few years. As the "first-person shooter" genre label had not yet solidified at the time, Doom was described as a "first person perspective adventure" and "atmospheric 3-D action game". Doom clones ranged from close imitators to more innovative takes on the genre. Id Software licensed the Doom engine to several other companies, which resulted in several games similar to Doom, including Heretic (1994), Hexen: Beyond Heretic (1995), and Strife: Quest for the Sigil (1996). A Doom-based game called Chex Quest was released in 1996 by Ralston Foods as a promotion to increase cereal sales. Other games were inspired by Doom, if not rumored to be built by reverse engineering the game's engine, including LucasArts's Star Wars: Dark Forces (1995). Several other games termed Doom clones, such as PowerSlave (1996) and Duke Nukem 3D (1996), used the 1995 Build engine, a 2.5D engine inspired by Doom created by Ken Silverman with some consultation with John Carmack. After completing Doom, id Software began working on a sequel using the same engine, Doom II, which was released to retail on October 10, 1994, ten months after the first game. GT Interactive had approached id before the release of Doom with plans to release a retail version of Doom and Doom II. Id chose to create the sequel as a set of episodes rather than a new game, allowing John Carmack and the other programmers to begin work on id's next game, Quake. Doom II was the United States' highest-selling software product of 1994 and sold more than 1.2 million copies within a year. Doom II was followed by an expansion pack from id, Master Levels for Doom II (1995), consisting of 21 commissioned levels and over 3000 user-created levels for Doom and Doom II. Two sets of Doom II levels by different amateur map-making teams were released together by id as the standalone game Final Doom (1996). Doom and Doom II were both included, along with previous id games, in the id Anthology compilation (1996). The Doom franchise has continued since the 1990s in several iterations and forms. The video game series includes Doom 3 (2004), Doom (2016), and Doom Eternal (2020), along with other spin-off video games. It additionally includes multiple novels, a comic book, board games, and two films: Doom (2005) and Doom: Annihilation (2019). Doom was notorious for its high levels of graphic violence and satanic imagery, which generated controversy from a broad range of groups. Doom for the 32X was one of the first video games to be given a Mature 17+ rating from the Entertainment Software Rating Board due to its violent gore and nature, while Doom II was the first. In Germany, shortly after its publication, Doom was classified as "harmful to minors" by the Federal Department for Media Harmful to Young Persons and could not be sold to children or displayed where they could see it, which was only rescinded in 2011. Doom again sparked controversy in the United States when it was found that Eric Harris and Dylan Klebold, who committed the Columbine High School massacre on April 20, 1999, were avid players. While planning for the massacre, Harris said in his journal that the killing would be "like playing Doom". A rumor spread afterward that Harris had designed a custom Doom level that looked like the high school, populated with representations of Harris's classmates and teachers, which he used to practice for the shooting. Although Harris did design several custom Doom levels, which later became known as the "Harris levels", none were based on the school. Doom was dubbed a "mass murder simulator" by critic and Killology Research Group founder David Grossman. In the earliest release versions, the level E1M4: Command Control contains a swastika-shaped structure, which was put in as a homage to Wolfenstein 3D. The swastika was removed in later versions, out of respect for a military veteran's request, according to Romero. Doom's popularity and innovations attracted a community that has persisted for decades since. The deathmatch mode was an important factor in its popularity. Doom was the first game to coin the term "deathmatch" and introduced multiplayer shooting battles to a wide audience. This led to a widespread community of players who had never experienced fast-paced multiplayer combat before. Another popular aspect of Doom was the versatility of its WAD files, enabling user-generated levels and other game modifications. John Carmack and Romero had strongly advocated for mod support, overriding other id employees who were concerned about commercial and legal implications. Although WAD files exposed the game data, id provided no instructions for how they worked. Still, players were able to modify leaked alpha versions of the game, allowing them to release level editors within weeks of the game's release. On January 26, 1994, university student Brendon Wyber led a group to create the first full level editor, the Doom Editor Utility, leading to the first custom level by Jeff Bird in March. It was followed by "countless" others, including many based on other franchises like Aliens and Star Wars total conversion mods, as well as DeHackEd, a patch editor first released in 1994 by Greg Lewis that allowed editing of the game engine. Soon after the first mods appeared, id CEO Wilbur posted legal terms to the company's website, allowing mod authors to charge money without any fees to id, while also absolving the company of responsibility or support. Doom mods were widely popular, earning favorable comparisons to the official level additions seen in The Ultimate Doom. Thousands of user-created levels were released in the first few years after the release; over 3000 such levels for Doom and Doom II were included in the official retail release Master Levels for Doom II (1995). WizardWorks released multiple collections of mods of Doom and Doom II under the name D!Zone. At least one mod creator, Tim Willits, was later hired at id Software. Mods have continued to be produced, with the community Cacowards awarding the best of each year. In 2016, Romero created two new Doom levels: E1M4b ("Phobos Mission Control") and E1M8b ("Tech Gone Bad"). In 2018, for the 25th anniversary of Doom, Romero announced Sigil, an unofficial fifth episode containing nine levels. It was released on May 22, 2019, for €6.66 with a soundtrack by Buckethead, and then released again for free on May 31 with a soundtrack by James Paddock. A physical release was later produced. A sixth episode, Sigil II, was released on the game's 30th anniversary, December 10, 2023, again for €6.66 for a digital copy with a soundtrack by Valient Thorr, as well as physical editions on floppy disk. In addition to WAD files, Doom includes a feature that allowed players to record and play back gameplay using files called demos, or game replays. Although the concept of speedrunning a video game existed before Doom, its release coincided with a wave of popularity for speedrunning, amplified by the online communities built on the nascent Internet. Demos were lightweight files that could be shared more easily than video files on internet bulletin board systems at the time. As a result, Doom is credited with creating the video game speedrunning community. The speedrunning community for Doom has continued for decades. As recently as 2019, community members have broken records originally set in 1998. Doom has been termed as having "one of the longest-running speedrunning communities" as well as being "the quintessential speedrunning game". Notes References Sources External links |
======================================== |
[SOURCE: https://www.bbc.com/future/article/20260210-tiktok-is-tracking-you-even-if-you-dont-use-the-app-heres-how-to-stop-it] | [TOKENS: 5272] |
TikTok is tracking you, even if you don't use the app. Here's how to stop it11 February 2026ShareSaveThomas GermainShareSaveSerenity Strull/ BBCTikTok is growing its data harvesting empire, and avoiding the app won’t protect you – but some easy steps can keep you safe.TikTok keeps track of everything you do on its app – no surprises there. What's less obvious is how the company follows you around other parts of the internet that have nothing to do with TikTok. In fact, TikTok collects sensitive and potentially embarrassing information about you even if you've never used the app. Over the past week, I've watched websites sending TikTok data about cancer diagnoses, fertility and even mental health crises. It's part of a tracking empire that extends far beyond the social media platform. Now, thanks to a new set of features, TikTok is poised to expand its network and see even more details about your life.The change comes just weeks after the sale of TikTok's US operations to a group of companies with ties to US President Donald Trump. The deal has led to fresh privacy concerns from some human rights experts and users, though TikTok says it has transparent guidelines on how it responds to government requests for data. Fortunately, this is a privacy story with a positive note. Some easy steps you can take in about five minutes will help you keep your information out of TikTok's hands. The issue centres around major changes to TikTok's "pixel", a tracking tool that companies use to monitor your online behaviour. I asked a cybersecurity company called Disconnect to analyse it. They found the updated TikTok pixel collects information in unusual ways compared to its competitors. "It's extremely invasive," says Patrick Jackson, chief technology officer at Disconnect. "This expanded data sharing, when you do analysis of the actual pixel code, you see things that look really bad."When I clicked a button on a form that said I was a cancer patient or a survivor, the website sent TikTok my email address along with those detailsTikTok says its users are informed about its data practices in privacy policies and notifications in some cases. The company also says it gives people privacy settings to take control."TikTok empowers users with transparent information about its privacy practices and gives them multiple tools to customise their experience," a TikTok spokesperson says. "Advertising pixels are industry standard and used widely across social and media platforms, including by the BBC."But most people might not realise that TikTok holds data about them even if they have never used the social media platform.An invisible trackerTracking pixels are nothing new. For years, companies that run advertising networks – including Google, Meta and hundreds of others – have used them to eavesdrop on what people do across the web. They're an invisible image the size of one pixel of your screen that loads in the background of a website, full of data-harvesting tech. They're everywhere, and they're constantly watching you.Here's how it works. TikTok, for example, encourages companies to put pixels on their websites to help the social media giant harvest more data. Let's say I have an online shoe store. If I use a pixel, it lets TikTok collect lots of data about my customers in order to show them targeted ads. Plus, it helps TikTok figure out whether people who see those shoe ads end up making a purchase. That way, I know the ads I paid for are working, and maybe I'll pay for more. (Like most news organisations, the BBC uses analytics tools and shares data with advertising partners in accordance with our privacy policy. The BBC does not use TikTok tracking pixels on its website or place advertising pixels on third-party sites.)When it's shoe store data, the information might be innocuous. But I've reported on TikTok's data collection for years and pixels can collect extremely personal information. For example, last week I visited the website for a cancer support group. According to Disconnect, when I clicked a button on a form that said I was a cancer patient or a survivor, the website sent TikTok my email address along with those details. A women's health company sent TikTok data when I looked at fertility tests. A mental health organisation pinged TikTok when I indicated I'm looking for a crisis counsellor. Websites that use pixels send data about every single visitor, so it doesn't matter if you don't have a TikTok account. A TikTok spokesperson says, essentially, that this isn't TikTok's responsibility. They say websites are required to abide by privacy laws and tell you about their data practices. TikTok says websites are prohibited from sharing certain kinds of sensitive information, such as health data. And the company says it takes proactive steps to alert websites that share anything inappropriate.Serenity Strull/ BBCMany of the world's top websites have pixel trackers on them that send data back to big tech companies (Credit: Serenity Strull/ BBC)If you're concerned about these individual websites you're missing the point. Critics say the issue is that large tech companies like TikTok are increasingly following everything you do online. According to DuckDuckGo, a privacy company, TikTok has trackers on 5% of the world's top websites. That number has grown steadily, though it's nothing compared to Google with trackers on almost 72% of top websites and Meta at about 21%."This is verbatim the playbook that Google and Meta have used over the years," says Peter Dolanjski, executive director of product at DuckDuckGo. They started collecting small amounts of data and grew that into an empire that has massive visibility into your daily life, he says.All of this data could mean you see ads that are more tailored to you, which you might like. But these detailed records of your personal life wouldn't exist if tech companies weren't surveilling you, and it exposes you to all kinds of risks, Dolanjski says."Algorithms can use this data to exploit you," he says. "It could be coercing you to buy something, it could be political campaigns, it could be price discrimination." Advertising data has been used for all kinds damaging purposes, from alleged civil rights violations to sexual discrimination. TikTok's data empireTikTok's pixel is years old, but it just shifted in some major ways. On 22 January 2026, when TikTok's US operation officially changed hands, users had to agree to a new set of data collection practices. That includes a new advertising network that TikTok will use to show targeted ads on other people's websites. To facilitate that new advertising system, TikTok updated its pixel.In the past, TikTok's pixel basically just told companies if their ads were generating sales in the app itself. Now, the pixel will help companies follow users who see an ad when they leave TikTok and make a purchase elsewhere.That probably means more companies will buy TikTok ads and the pixel will show up in more places, according to Arielle Garcia, chief operating officer at Check My Ads, a digital advertising watchdog group. In other words, TikTok's tracking empire is set to expand. "These tools naturally make the platform more attractive to advertisers, which is ultimately how ad platforms grow," Garcia says.Keeping TabsThomas Germain is a senior technology journalist at the BBC. He writes the column Keeping Tabs and co-hosts the podcast The Interface. His work uncovers the hidden systems that run your digital life, and how you can live better inside them.Disconnect found TikTok's pixel now collects more information than ever before, automatically intercepting data that websites are sending to Google. Experts tell the BBC this is unusually invasive. "They're silently capturing that data without the site owner explicitly sharing that information with TikTok," Jackson says, and that means websites might unintentionally send TikTok even more data than they intend to.TikTok disagrees. A spokesperson says TikTok is clear about what data the pixel collects, and companies can just set up their websites differently if they don't want TikTok to see what they send Google. (Google did not respond to a request for comment.)TikTok also has some privacy controls you can use. Users can "clear" the data TikTok collects with pixels using a setting in the app. People who don't have an account can ask TikTok to delete any data it has about you.But if you want to stop the data collection before it happens, you need additional steps.How to protect yourselfThere's good news and bad news. Let's start with the cheerful stuff.The best option? Use a more private web browser. I know switching seems like a pain, but it's easy to import your bookmarks. Try it.Something like 71% of people use Google Chrome, which has been found in preliminary academic research to leak more information than many competitors. Privacy experts often recommend the DuckDuckGo browser and Brave, which are specifically built to safeguard data. Firefox and Safari are considered better options than Chrome, though they're less strict about privacy by default.More like this:• The number one sign you're watching an AI video• How YouTube's secret AI edits could bend reality• Is Google about to destroy the web?If switching browsers is too much, install a browser extension that blocks these trackers. I asked Disconnect and DuckDuckGo to help with this article because they both make tracker blockers, but there are other options, including Privacy Badger and Ghostery. Certain ad blockers also block some data harvesting, including AdBlock Plus and uBlock Origin. DuckDuckGo has a chart comparing which ad blockers do it best. Just don't install browser extensions that aren't recommended by reputable sources – it's just like installing an app. Some are dicey.Now the bad news. Following those two steps will block the TikTok pixel and lots of other privacy invasions. But I won't pretend your data problems are solved.There are lots of other ways that companies share data with TikTok, Google, Meta and other advertising companies. Companies collect data about you and send it directly to the tech giants from their own servers, for example. "It's a black box, I can't tell you how often that's used because it all happens behind the scenes," says Dolanjski. "It's much harder to protect yourself from that. Your only real defence is to not use the same personal information on different services", so it's harder to match up what you do on different parts of the internet.The real solution is better privacy laws, says Garcia from Check My Ads. "This isn't a problem limited to one platform. It's a broader advertising technology ecosystem issue that ultimately needs to be addressed through stronger regulation," she says. "The only thing that's really going to change this is when people make their voices heard with lawmakers and make it clear that privacy is something they actually care about."--For more technology news and insights, sign up to our Tech Decoded newsletter, while The Essential List delivers a handpicked selection of features and insights to your inbox twice a week.For more science, technology, environment and health stories from the BBC, follow us on Facebook and Instagram.Keeping TabsTechnologyThomas GermainInternetSocial MediaPrivacyFeatures TikTok is tracking you, even if you don't use the app. Here's how to stop it TikTok is growing its data harvesting empire, and avoiding the app won’t protect you – but some easy steps can keep you safe. TikTok keeps track of everything you do on its app – no surprises there. What's less obvious is how the company follows you around other parts of the internet that have nothing to do with TikTok. In fact, TikTok collects sensitive and potentially embarrassing information about you even if you've never used the app. Over the past week, I've watched websites sending TikTok data about cancer diagnoses, fertility and even mental health crises. It's part of a tracking empire that extends far beyond the social media platform. Now, thanks to a new set of features, TikTok is poised to expand its network and see even more details about your life. The change comes just weeks after the sale of TikTok's US operations to a group of companies with ties to US President Donald Trump. The deal has led to fresh privacy concerns from some human rights experts and users, though TikTok says it has transparent guidelines on how it responds to government requests for data. Fortunately, this is a privacy story with a positive note. Some easy steps you can take in about five minutes will help you keep your information out of TikTok's hands. The issue centres around major changes to TikTok's "pixel", a tracking tool that companies use to monitor your online behaviour. I asked a cybersecurity company called Disconnect to analyse it. They found the updated TikTok pixel collects information in unusual ways compared to its competitors. "It's extremely invasive," says Patrick Jackson, chief technology officer at Disconnect. "This expanded data sharing, when you do analysis of the actual pixel code, you see things that look really bad." TikTok says its users are informed about its data practices in privacy policies and notifications in some cases. The company also says it gives people privacy settings to take control. "TikTok empowers users with transparent information about its privacy practices and gives them multiple tools to customise their experience," a TikTok spokesperson says. "Advertising pixels are industry standard and used widely across social and media platforms, including by the BBC." But most people might not realise that TikTok holds data about them even if they have never used the social media platform. An invisible tracker Tracking pixels are nothing new. For years, companies that run advertising networks – including Google, Meta and hundreds of others – have used them to eavesdrop on what people do across the web. They're an invisible image the size of one pixel of your screen that loads in the background of a website, full of data-harvesting tech. They're everywhere, and they're constantly watching you. Here's how it works. TikTok, for example, encourages companies to put pixels on their websites to help the social media giant harvest more data. Let's say I have an online shoe store. If I use a pixel, it lets TikTok collect lots of data about my customers in order to show them targeted ads. Plus, it helps TikTok figure out whether people who see those shoe ads end up making a purchase. That way, I know the ads I paid for are working, and maybe I'll pay for more. (Like most news organisations, the BBC uses analytics tools and shares data with advertising partners in accordance with our privacy policy. The BBC does not use TikTok tracking pixels on its website or place advertising pixels on third-party sites.) When it's shoe store data, the information might be innocuous. But I've reported on TikTok's data collection for years and pixels can collect extremely personal information. For example, last week I visited the website for a cancer support group. According to Disconnect, when I clicked a button on a form that said I was a cancer patient or a survivor, the website sent TikTok my email address along with those details. A women's health company sent TikTok data when I looked at fertility tests. A mental health organisation pinged TikTok when I indicated I'm looking for a crisis counsellor. Websites that use pixels send data about every single visitor, so it doesn't matter if you don't have a TikTok account. A TikTok spokesperson says, essentially, that this isn't TikTok's responsibility. They say websites are required to abide by privacy laws and tell you about their data practices. TikTok says websites are prohibited from sharing certain kinds of sensitive information, such as health data. And the company says it takes proactive steps to alert websites that share anything inappropriate. If you're concerned about these individual websites you're missing the point. Critics say the issue is that large tech companies like TikTok are increasingly following everything you do online. According to DuckDuckGo, a privacy company, TikTok has trackers on 5% of the world's top websites. That number has grown steadily, though it's nothing compared to Google with trackers on almost 72% of top websites and Meta at about 21%. "This is verbatim the playbook that Google and Meta have used over the years," says Peter Dolanjski, executive director of product at DuckDuckGo. They started collecting small amounts of data and grew that into an empire that has massive visibility into your daily life, he says. All of this data could mean you see ads that are more tailored to you, which you might like. But these detailed records of your personal life wouldn't exist if tech companies weren't surveilling you, and it exposes you to all kinds of risks, Dolanjski says. "Algorithms can use this data to exploit you," he says. "It could be coercing you to buy something, it could be political campaigns, it could be price discrimination." Advertising data has been used for all kinds damaging purposes, from alleged civil rights violations to sexual discrimination. TikTok's data empire TikTok's pixel is years old, but it just shifted in some major ways. On 22 January 2026, when TikTok's US operation officially changed hands, users had to agree to a new set of data collection practices. That includes a new advertising network that TikTok will use to show targeted ads on other people's websites. To facilitate that new advertising system, TikTok updated its pixel. In the past, TikTok's pixel basically just told companies if their ads were generating sales in the app itself. Now, the pixel will help companies follow users who see an ad when they leave TikTok and make a purchase elsewhere. That probably means more companies will buy TikTok ads and the pixel will show up in more places, according to Arielle Garcia, chief operating officer at Check My Ads, a digital advertising watchdog group. In other words, TikTok's tracking empire is set to expand. "These tools naturally make the platform more attractive to advertisers, which is ultimately how ad platforms grow," Garcia says. Keeping Tabs Thomas Germain is a senior technology journalist at the BBC. He writes the column Keeping Tabs and co-hosts the podcast The Interface. His work uncovers the hidden systems that run your digital life, and how you can live better inside them. Disconnect found TikTok's pixel now collects more information than ever before, automatically intercepting data that websites are sending to Google. Experts tell the BBC this is unusually invasive. "They're silently capturing that data without the site owner explicitly sharing that information with TikTok," Jackson says, and that means websites might unintentionally send TikTok even more data than they intend to. TikTok disagrees. A spokesperson says TikTok is clear about what data the pixel collects, and companies can just set up their websites differently if they don't want TikTok to see what they send Google. (Google did not respond to a request for comment.) TikTok also has some privacy controls you can use. Users can "clear" the data TikTok collects with pixels using a setting in the app. People who don't have an account can ask TikTok to delete any data it has about you. But if you want to stop the data collection before it happens, you need additional steps. How to protect yourself There's good news and bad news. Let's start with the cheerful stuff. The best option? Use a more private web browser. I know switching seems like a pain, but it's easy to import your bookmarks. Try it. Something like 71% of people use Google Chrome, which has been found in preliminary academic research to leak more information than many competitors. Privacy experts often recommend the DuckDuckGo browser and Brave, which are specifically built to safeguard data. Firefox and Safari are considered better options than Chrome, though they're less strict about privacy by default. More like this: • The number one sign you're watching an AI video • How YouTube's secret AI edits could bend reality • Is Google about to destroy the web? If switching browsers is too much, install a browser extension that blocks these trackers. I asked Disconnect and DuckDuckGo to help with this article because they both make tracker blockers, but there are other options, including Privacy Badger and Ghostery. Certain ad blockers also block some data harvesting, including AdBlock Plus and uBlock Origin. DuckDuckGo has a chart comparing which ad blockers do it best. Just don't install browser extensions that aren't recommended by reputable sources – it's just like installing an app. Some are dicey. Now the bad news. Following those two steps will block the TikTok pixel and lots of other privacy invasions. But I won't pretend your data problems are solved. There are lots of other ways that companies share data with TikTok, Google, Meta and other advertising companies. Companies collect data about you and send it directly to the tech giants from their own servers, for example. "It's a black box, I can't tell you how often that's used because it all happens behind the scenes," says Dolanjski. "It's much harder to protect yourself from that. Your only real defence is to not use the same personal information on different services", so it's harder to match up what you do on different parts of the internet. The real solution is better privacy laws, says Garcia from Check My Ads. "This isn't a problem limited to one platform. It's a broader advertising technology ecosystem issue that ultimately needs to be addressed through stronger regulation," she says. "The only thing that's really going to change this is when people make their voices heard with lawmakers and make it clear that privacy is something they actually care about." -- For more technology news and insights, sign up to our Tech Decoded newsletter, while The Essential List delivers a handpicked selection of features and insights to your inbox twice a week. For more science, technology, environment and health stories from the BBC, follow us on Facebook and Instagram. Fixing fashion's erratic sizing problem Tech Now meets a startup trying to fix one of the fashion industry's biggest blind spots, inconsistent sizing. The tactile tech giving deaf runners a fair start A gold‑medalist has developed a vibrating starting block to give deaf athletes clearer, fairer race starts. These futuristic screens help you navigate Tokyo In Tokyo, BBC TechXplore tests live translation and AI-powered displays that makes the city more navigable. The wearable tech that lets spectators feel the match At Tokyo's Deaflympics, deaf Judo fans aren't just watching the matches, they're feeling them, thanks to Hapbeat. Meet MOFO: will.i.am's rapping AI toy BBC Tech Now takes us inside CES 2026 to meet musician will.i.am and his AI toy, MOFO. The gadgets set to change your daily health and wellness Tech Now test out new gadgets disrupting the health industry at CES 2026 in Las Vegas. What's it like to meet your own avatar? Musician KT Tunstall meets her avatar as Tech Now explores music’s virtual future. How early filmmakers invented the internet’s funniest trend Discover how quirky clips paved the way for viral humour, proving randomness never goes out of style. Explaining how a touchscreen works with a sausage British mathematician Hannah Fry digs into the science of touchscreens. Why statistics fail to cure flying fears Why do flying fears persist despite falling accident rates? Learn tips to conquer your anxiety. Can smart phones get smarter BBC Click attend Mobile World Congress to test the latest tech products and trends. Can technology help reduce Parkinson’s symptoms? BBC Click visits a Madrid hospital to see patients treated with an ultrasound for tremors. The Lion King: How Mufasa was brought to life BBC Click speaks to the visual effects team behind the latest Disney blockbuster. How the TikTok ban affected US influencers BBC Click meets TikTok creator Peggy Xu who gained millions of views sharing milk videos. Is this the world's first AI powered hotel? BBC Click's Paul Carter visits the world's first fully AI-powered hotel in Las Vegas. Could an Arctic underground vault save our data? BBC Click explores an Arctic vault that stores digital artefacts from across the globe. How technology can monitor and improve our health BBC Click visits CES 2025 to find out about the latest health tech, from medical tools to well-being devices. The technology powering the iconic Sydney Opera House BBC Click heads behind the scenes of the Sydney Opera House to explore the tech powering the famous landmark. Inside the high-security facility tackling digital threats BBC reporter Marc Cieslak explores a high-security hub monitoring digital threats ahead of the US election. 'A tech firm stole our voices - then cloned and sold them' Two voice-over artists were listening to a podcast when they heard their own stolen AI-generated voices. Tumbler Ridge suspect's ChatGPT account banned before shooting OpenAI said the account's activity did not meet the threshold to flag it to authorities when it was identified. Why fake AI videos of UK urban decline are taking over social media Deepfakes showing grim taxpayer-funded waterparks have gone viral and drawn some racist responses. How the additives in food affect our gut microbes The additives added to processed food to keep it fresher for longer might be having an unexpected effect on the health of the microbes in our guts. The most anticipated museum openings of 2026 From a futuristic sci-fi attraction in Los Angeles to a dramatic monument to a millennia-old Aboriginal civilisation, these long-awaited museums are worth travelling for. 11 of the Winter Olympics' most striking images As the 2026 Winter Olympics close, the BBC rounds up some of the most stunning photos captured from the Games, and compares them to historic works of art. Copyright 2026 BBC. All rights reserved. The BBC is not responsible for the content of external sites. Read about our approach to external linking. |
======================================== |
[SOURCE: https://en.wikipedia.org/wiki/Language_model#cite_note-30] | [TOKENS: 1793] |
Contents Language model A language model is a computational model that predicts sequences in natural language. Language models are useful for a variety of tasks, including speech recognition, machine translation, natural language generation (generating more human-like text), optical character recognition, route optimization, handwriting recognition, grammar induction, and information retrieval. Large language models (LLMs), currently their most advanced form as of 2019, are predominantly based on transformers trained on larger datasets (frequently using texts scraped from the public internet). They have superseded recurrent neural network-based models, which had previously superseded the purely statistical models, such as the word n-gram language model. History Noam Chomsky did pioneering work on language models in the 1950s by developing a theory of formal grammars. In 1980, statistical approaches were explored and found to be more useful for many purposes than rule-based formal grammars. Discrete representations like word n-gram language models, with probabilities for discrete combinations of words, made significant advances. In the 2000s, continuous representations for words, such as word embeddings, began to replace discrete representations. Typically, the representation is a real-valued vector that encodes a word’s meaning such that words closer in vector space are similar in meaning and common relationships between words, such as plurality or gender, are preserved. Pure statistical models In 1980, the first significant statistical language model was proposed, and during the decade IBM performed 'Shannon-style' experiments, in which potential sources for language modeling improvement were identified by observing and analyzing the performance of human subjects in predicting or correcting text. A word n-gram language model is a statistical model of language which calculates the probability of the next word in a sequence from a fixed size window of previous words. If one previous word is considered, it is a bigram model; if two words, a trigram model; if n − 1 words, an n-gram model. Special tokens are introduced to denote the start and end of a sentence ⟨ s ⟩ {\displaystyle \langle s\rangle } and ⟨ / s ⟩ {\displaystyle \langle /s\rangle } . To prevent a zero probability being assigned to unseen words, the probability of each seen word is slightly lowered to make room for the unseen words in a given corpus. To achieve this, various smoothing methods are used, from simple "add-one" smoothing (assigning a count of 1 to unseen n-grams, as an uninformative prior) to more sophisticated techniques, such as Good–Turing discounting or back-off models. Word n-gram models have largely been superseded by recurrent neural network–based models, which in turn have been superseded by Transformer-based models often referred to as large language models. Maximum entropy language models encode the relationship between a word and the n-gram history using feature functions. The equation is P ( w m ∣ w 1 , … , w m − 1 ) = 1 Z ( w 1 , … , w m − 1 ) exp ( a T f ( w 1 , … , w m ) ) {\displaystyle P(w_{m}\mid w_{1},\ldots ,w_{m-1})={\frac {1}{Z(w_{1},\ldots ,w_{m-1})}}\exp(a^{T}f(w_{1},\ldots ,w_{m}))} where Z ( w 1 , … , w m − 1 ) {\displaystyle Z(w_{1},\ldots ,w_{m-1})} is the partition function, a {\displaystyle a} is the parameter vector, and f ( w 1 , … , w m ) {\displaystyle f(w_{1},\ldots ,w_{m})} is the feature function. In the simplest case, the feature function is just an indicator of the presence of a certain n-gram. It is helpful to use a prior on a {\displaystyle a} or some form of regularization. The log-bilinear model is another example of an exponential language model. Skip-gram language model is an attempt at overcoming the data sparsity problem that the preceding model (i.e. word n-gram language model) faced. Words represented in an embedding vector were not necessarily consecutive anymore, but could leave gaps that are skipped over (thus the name "skip-gram"). Formally, a k-skip-n-gram is a length-n subsequence where the components occur at distance at most k from each other. For example, in the input text: the set of 1-skip-2-grams includes all the bigrams (2-grams), and in addition the subsequences In skip-gram model, semantic relations between words are represented by linear combinations, capturing a form of compositionality. For example, in some such models, if v is the function that maps a word w to its n-d vector representation, then v ( k i n g ) − v ( m a l e ) + v ( f e m a l e ) ≈ v ( q u e e n ) {\displaystyle v(\mathrm {king} )-v(\mathrm {male} )+v(\mathrm {female} )\approx v(\mathrm {queen} )} where ≈ is made precise by stipulating that its right-hand side must be the nearest neighbor of the value of the left-hand side. Neural models Continuous representations or embeddings of words are produced in recurrent neural network-based language models (known also as continuous space language models). Such continuous space embeddings help to alleviate the curse of dimensionality, which is the consequence of the number of possible sequences of words increasing exponentially with the size of the vocabulary, further causing a data sparsity problem. Neural networks avoid this problem by representing words as non-linear combinations of weights in a neural net. A large language model (LLM) is a language model trained with self-supervised machine learning on a vast amount of text, designed for natural language processing tasks, especially language generation. The largest and most capable LLMs are generative pre-trained transformers (GPTs) that provide the core capabilities of modern chatbots. LLMs can be fine-tuned for specific tasks or guided by prompt engineering. These models acquire predictive power regarding syntax, semantics, and ontologies inherent in human language corpora, but they also inherit inaccuracies and biases present in the data they are trained on. They consist of billions to trillions of parameters and operate as general-purpose sequence models, generating, summarizing, translating, and reasoning over text. LLMs represent a significant new technology in their ability to generalize across tasks with minimal task-specific supervision, enabling capabilities like conversational agents, code generation, knowledge retrieval, and automated reasoning that previously required bespoke systems. LLMs evolved from earlier statistical and recurrent neural network approaches to language modeling. The transformer architecture, introduced in 2017, replaced recurrence with self-attention, allowing efficient parallelization, longer context handling, and scalable training on unprecedented data volumes. This innovation enabled models like GPT, BERT, and their successors, which demonstrated emergent behaviors at scale, such as few-shot learning and compositional reasoning. Reinforcement learning, particularly policy gradient algorithms, has been adapted to fine-tune LLMs for desired behaviors beyond raw next-token prediction. Reinforcement learning from human feedback (RLHF) applies these methods to optimize a policy, the LLM's output distribution, against reward signals derived from human or automated preference judgments. This has been critical for aligning model outputs with user expectations, improving factuality, reducing harmful responses, and enhancing task performance. Benchmark evaluations for LLMs have evolved from narrow linguistic assessments toward comprehensive, multi-task evaluations measuring reasoning, factual accuracy, alignment, and safety. Hill climbing, iteratively optimizing models against benchmarks, has emerged as a dominant strategy, producing rapid incremental performance gains but raising concerns of overfitting to benchmarks rather than achieving genuine generalization or robust capability improvements. Although sometimes matching human performance, it is not clear whether they are plausible cognitive models. At least for recurrent neural networks, it has been shown that they sometimes learn patterns that humans do not, but fail to learn patterns that humans typically do. Evaluation and benchmarks Evaluation of the quality of language models is mostly done by comparison to human created sample benchmarks created from typical language-oriented tasks. Other, less established, quality tests examine the intrinsic character of a language model or compare two such models. Since language models are typically intended to be dynamic and to learn from data they see, some proposed models investigate the rate of learning, e.g., through inspection of learning curves. Various data sets have been developed for use in evaluating language processing systems. These include: See also References Further reading |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-07-28-nbdev2.html] | [TOKENS: 2772] |
nbdev+Quarto: A new secret weapon for productivity Hamel Husain and Jeremy Howard July 28, 2022 On this page Our new secret weapon for productivity Today we’re excited to announce that we’ve teamed up with Quarto to give nbdev superpowers. nbdev offers Python programmers a common set of tools for using Jupyter notebooks to: Although notebooks are already widely used for once-off exploratory work, it’s less well-known that they are perfectly capable of writing quality software. In fact, we’ve used nbdev for a wide range of software projects over the last three years, including deep learning libraries, API clients, Python language extensions, terminal user interfaces, and more. We discovered that it is not only capable of writing great software but that it has also increased our productivity by 300% or more. With nbdev, developers simply write notebooks with lightweight markup and get high-quality documentation, tests, continuous integration, and packaging for free! Nbdev has allowed us to maintain and scale manyopen source projects. Pull requests are often accompanied by detailed documentation and tests–contributors simply write notebooks. This is why we’re excited to share nbdev v2. It’s rewritten from the ground up, with much-anticipated features including: nbdev in industry We have piloted nbdev at several companies. We were delighted to receive the following feedback, which fits our own experience using and developing nbdev: David Berg, on using nbdev for internal documentation at Netflix: “Prior to using nbdev, documentation was the most cumbersome aspect of our software development process… Using nbdev allows us to spend more time creating rich prose around the many code snippets guaranteeing the whole experience is robust. nbdev has turned what was once a chore into a natural extension of the notebook-based testing we were already doing.” Erik Gaasedelen, on using nbdev in production at Lyft: “I use this in production at my company. It’s an awesome tool… nbdev streamlines everything so I can write docs, tests, and code all in one place… The packaging is also really well thought out. From my point of view it is close to a Pareto improvement over traditional Python library development.” Hugo Bowne-Anderson, on using nbdev for Outerbounds: “nbdev has transformed the way we write documentation. Gone are the days of worrying about broken code examples when our API changes or [due to] human errors associated with copying & pasting code into markdown files. The authoring experience of nbdev… [allows] us to write prose and live code in a unified interface, which allows more experimentation… On top of this, nbdev allows us to include unit tests in our documentation which mitigates the burden of maintaining the docs over time.” Roxanna Pourzand, on using nbdev for Transform: “We’re so excited about using nbdev. Our product is technical so our resulting documentation includes a lot of code-based examples. Before nbdev, we had no way of maintaining our code examples and ensuring that it was up-to-date for both command inputs and outputs. It was all manual. With nbdev, we now have this under control in a sustainable way. Since we’ve deployed these docs, we also had a situation where we were able to identify a bug in one of our interfaces, which we found by seeing the error that was output in the documentation.” What’s nbdev? Nbdev embraces the dynamic nature of python and REPL-driven development in ways that traditional IDEs and software development workflows cannot. We thoroughly discussed the motivation, history, and goals of nbdev in this initial launch post three years ago. The creator of Jupyter, Fernando Pérez, told us: [Nbdev] should be celebrated and used a lot more - I have kept a tab with your original nbdev blog post open for months in Chrome because of how often I refer to it and point others to this work In short, nbdev embraces ideas from literate programming and exploratory programming. These paradigms have been revisited in platforms like XCode Playgrounds and languages like Smalltalk, LISP, and Mathematica. With nbdev, we sought to push these paradigms even further by enabling it for one of the most popular dynamic programming languages in the world: Python. Even though nbdev is most widely used in scientific computing communities due to its integration with Jupyter Notebooks, we’ve found that nbdev is well suited for a much wider range of software. We have used nbdev to write deep learning libraries, API clients, python language extensions,terminal user interfaces, and more! Hamel: When I use nbdev, my colleagues are often astounded by how quickly I can create and distribute high-quality python packages. I consider nbdev to be a superpower that allows me to create tests and documentation without any additional friction, which makes all of my projects more maintainable. I also find writing software with nbdev to be more fun and productive as I can iterate very fast on ideas relative to more traditional software engineering workflows. Lastly, with nbdev I can also use traditional text-based IDEs if I want to, so I get the best of both worlds. What we learned after three years of using nbdev While nbdev was originally developed to simplify the software development workflow for various fast.ai projects, we found that users wanted to extend nbdev to: While we created projects such as fastpages and fastdoc to accomplish some of these tasks, we realized that it would be better to have a single set of flexible tools to accomplish all of them. To this end, we were extremely excited to discover Quarto, an open-source technical publishing system built on pandoc. Hamel: The more I used nbdev for creating Python modules, the more I wanted to use it for writing blogs and documenting existing codebases. The ability to customize the way notebooks are rendered (hiding vs. showing cells, stripping output, etc.), along with the facilities for including unit tests, made it my go-to authoring tool for all technical content. I’m excited that nbdev2 unlocks all of these possibilities for everyone! Enter Quarto: A pandoc super-processor Quarto is a project that enables technical publishing with support for Jupyter Notebook, VSCode, Observable, and plaintext editors. Furthermore, Quarto enables the publishing of high-quality articles, reports, websites, and blogs in HTML, PDF, ePub, PowerPoint slides, and more. Quarto is maintained by RStudio, a company with a long history of products supporting literate programming, such as RMarkdown and RStudio. Quarto is built on top of Pandoc, a universal document converter that supports nearly any format you can think of. Pandoc achieves this seemingly magical feat by representing documents in a common abstract syntax tree (AST) that serves as the medium through which different formats can be translated. By extension, Quarto allows you to generate content in almost any format you wish! You can use pandoc filters to modify the AST and the output format, which allows you to use any static site generator you want, and programmatically modify and generate content. Quarto allows you to compose pandoc filters in a processing pipeline and apply them to specific documents or entire projects. You can also distribute filters as Quarto extensions, which makes Quarto extremely customizable. We also find Quarto compelling because user interfaces such as comment directives (comments that start with #|) correlate with nbdev. In fact, we even learned that nbdev inspired Quarto in this regard! In general, Quarto and nbdev share many goals, and the Quarto team has been incredibly responsive to our suggestions. For example, the ability to create notebook filters to modify notebooks before rendering. Below is a screenshot of a Jupyter notebook rendered with Quarto and nbdev. Finally, Quarto supports more programming languages than just Python and has been adding new features and fixing bugs at an impressive speed. This gives us confidence that we will be able to expand nbdev to support more use cases in the future. We discuss some of these future directions in the closing section. A blazing fast notebook kernel: execnb A core component of nbdev is executing and testing notebooks programmatically. It is important that this notebook runner executes with minimal overhead to maintain our goal of providing a delightful developer experience. This is why we built execnb, a lightweight notebook runner for Python kernels, which executes notebooks blazingly fast. Furthermore, execnb allows parameterized execution of notebooks. Hamel: I have been an enthusiastic user of tools like papermill that programmatically run notebooks for use-cases like creating dashboards or enabling new kinds of machine learning workflows. I believe execnb unlocks even more possibilities with its ability to inject arbitrary code at any place in a notebook, as well as the ability to pass callbacks that run before and/or after cells are executed. This opens up possibilities to create new types of workflows with notebooks that I am excited about exploring in the near future. Towards a dialect of python that embraces its dynamic nature One way to understand nbdev is part of an ecosystem that is designed to embrace Python’s dynamic properties for REPL-driven software engineering. Similar to Clojure, our goal is to provide tools that remove all friction from using the REPL in your programming workflow. We believe that the REPL enhances developer workflows thanks to context-sensitive auto-completion, signature inspection, and documentation–all based on the actual state of your code, and none of which are available in IDEs that depend solely on static analysis. We have found that for this reason, nbdev, with its Jupyter notebook foundation, makes programming significantly more productive and enjoyable. Our efforts to support REPL-driven development and literate programming are not limited to nbdev. We maintain a number of libraries that extend python to bolster this programming experience. The most notable of these libraries is fastcore, which extends Python in terms of testing, documenting code, metaprogramming, attribute helpers, enhanced representations of objects, and notebook-friendly patching. This blog post offers a gentle introduction to fastcore. In addition to literate programming, fastcore encourages conventions such as brevity and efficient use of vertical space so you can accomplish more with significantly less code. For example, below is a simple decorator that enables notebook-friendly patching: We believe that this combination of a new developer workflow (nbdev), Python extensions (fastcore), and associated norms form a new dialect of Python that is centered on leveraging its dynamic nature–in contrast to an ever-growing trend toward static analysis. We suspect that this dialect of Python will be more productive for programmers in many scenarios. We are framing this ecosystem as a “dialect” as it is still very much Python and is approachable by anyone who is familiar with the language. Furthermore, despite nbdev’s notebook workflow, our tools generate plaintext modules that can be navigated and edited with text-based IDEs, allowing programmers to experience the best of both worlds, if they desire. Hamel: I believe this framing of a Python dialect is key to properly understanding what nbdev is. While it may be tempting to get stuck on specific features or technical details of nbdev, it is useful to zoom out to understand the overall intent of creating a better workflow rather than conforming too rigidly to existing ones. A good analogy is TypeScript’s relationship with JavaScript: it is an extension of an existing programming language that supports a new way of programming. I encourage you to treat nbdev in a similar fashion: be willing to try new ways of programming and observe which tradeoffs resonate with you. At the very least, I believe nbdev is a fun way to experience a different way of writing software, which will broaden your horizons about programming in general, all without having to learn an entirely new programming language! The future of nbdev While we are excited about nbdev2, we believe we have only scratched the surface of what’s possible. We are considering the following features: If you have interesting ideas about how nbdev can be extended, please drop and chat with us on discord or post a message in the forums. How you can get started with nbdev Our project’s website is at nbdev.fast.ai, where we will be posting tutorials, examples, and more documentation in the coming days. Thank You This new version of nbdev was a team effort by many wonderful people. We want to highlight two people who have made outstanding contributions: Wasim Lorgat was instrumental across different areas, including significant contributions to fastcore, execnb, and nbdev, as well as the implementation of the new nbdev home page. With Wasim’s help, we were able to push nbdev to a new level of functionality and quality. JJ Allaire is not only the CEO of RStudio but also the steward of Quarto. JJ was incredibly responsive and eager to work with us on nbdev and added many features to Quarto specifically with nbdev in mind, such as notebook filters. We were also astounded by the attention to detail and the pace at which bugs are addressed. This new version of nbdev would not have been possible without JJ’s help, and we are excited to continue to work with him. We also want to thank the amazing fastai community, notably Isaac Flath, Benjamin Warner and Zach Mueller for their tireless work on this project. A conversation with JJ Allaire To celebrate the launch of nbdev v2 and Quarto, Jeremy sat down with the CEO of Posit (previously known as RStudio, the company behind Quarto), JJ Allaire, to talk about software development, scientific publishing, R, Python, literate programming, and much more. |
======================================== |
[SOURCE: https://en.wikipedia.org/w/index.php?title=Minecraft&action=info] | [TOKENS: 46] |
Contents Information for "Minecraft" Basic information Page protection Edit history Page properties This page is a member of 30 hidden categories (help): Pages transcluded onto the current version of this page (help): Lint errors External tools |
======================================== |
[SOURCE: https://en.wikipedia.org/wiki/Monism] | [TOKENS: 4161] |
Contents Monism Monism attributes oneness or singleness (Greek: μόνος) to a concept, such as to existence. Various kinds of monism can be distinguished: Definitions There are two sorts of definitions for monism: Although the term monism is derived from Western philosophy to typify positions in the mind–body problem, it has also been used to typify religious traditions. In modern Hinduism, the term "absolute monism" has been applied to Advaita Vedanta, though Philip Renard points out that this may be a Western interpretation, bypassing the intuitive understanding of a nondual reality. It is more generally categorized by scholars as a form of absolute nondualism. History Material monism can be traced back to the pre-Socratic philosophers who sought to understand the arche or basic principle of the universe in terms of different material causes. These included Thales, who argued that the basis of everything was water, Anaximenes, who claimed it was air, and Heraclitus, who believed it to be fire. Later, Parmenides described the world as "One", which could not change in any way. Zeno of Elea defended this view of everything being a single entity through his paradoxes, which aim to show the existence of time, motion and space to be illusionary. Baruch Spinoza argued that 'God or Nature' (Deus sive Natura) is the only substance of the universe, which can be referred to as either 'God' or 'Nature' (the two being interchangeable). This is because God/Nature has all the possible attributes and no two substances can share an attribute, which means there can be no other substances than God/Nature. Monism has been discussed thoroughly in Indian philosophy and Vedanta throughout their history starting as early as the Rig Veda. The term monism was introduced in the 18th century by Christian von Wolff in his work Logic (1728), to designate types of philosophical thought in which the attempt was made to eliminate the dichotomy of body and mind and explain all phenomena by one unifying principle, or as manifestations of a single substance. The mind–body problem in philosophy examines the relationship between mind and matter, and in particular the relationship between consciousness and the brain. The problem was addressed by René Descartes in the 17th century, resulting in Cartesian dualism, and by pre-Aristotelian philosophers, in Avicennian philosophy, and in earlier Asian and more specifically Indian traditions. Monism was later also applied to the theory of absolute identity set forth by Hegel and Schelling.[clarification needed] Thereafter the term was more broadly used, for any theory postulating a unifying principle. The opponent thesis of dualism also was broadened, to include pluralism. According to Urmson, as a result of this extended use, the term is "systematically ambiguous". According to Jonathan Schaffer, monism lost popularity due to the emergence of analytic philosophy in the early twentieth century, which revolted against the neo-Hegelians. Rudolf Carnap and A. J. Ayer, who were strong proponents of positivism, "ridiculed the whole question as incoherent mysticism". The mind–body problem has reemerged in social psychology and related fields, with the interest in mind–body interaction and the rejection of Cartesian mind–body dualism in the identity thesis, a modern form of monism. Monism is also still relevant to the philosophy of mind, where various positions are defended. Types Different types of monism include: Views contrasting with monism are: Monism in modern philosophy of mind can be divided into three broad categories:[clarification needed] Certain positions do not fit easily into the above categories, such as functionalism, anomalous monism, and reflexive monism. Moreover, they do not define the meaning of "real". Monistic philosophers While the lack of information makes it difficult in some cases to be sure of the details, the following pre-Socratic philosophers thought in monistic terms: Monistic neuroscientists Religion Pantheism is the belief that everything composes an all-encompassing, immanent God, or that the universe (or nature) is identical with divinity. Pantheists thus do or do not believe in a personal or anthropomorphic god, but believe that interpretations of the term differ. Pantheism was popularized in the modern era as both a theology and philosophy based on the work of the 17th-century philosopher Baruch Spinoza, whose Ethics was an answer to Descartes' famous dualist theory that the body and spirit are separate. Spinoza held that the two are the same, and this monism is a fundamental quality of his philosophy. He was described as a "God-intoxicated man," and used the word God to describe the unity of all substance. Although the term pantheism was not coined until after his death, Spinoza is regarded as its most celebrated advocate. H. P. Owen claimed that Pantheists are "monists" ... they believe that there is only one Being, and that all other forms of reality are either modes (or appearances) of it or identical with it. Pantheism is closely related to monism, as pantheists too believe all of reality is one substance, called Universe, God or Nature. Panentheism, a slightly different concept, is explained below in the next section. Some of the most famous pantheists are the Stoics, Giordano Bruno and Spinoza. Panentheism (from Greek πᾶν (pân) "all"; ἐν (en) "in"; and θεός (theós) "God"; "all-in-God") is a belief system that posits that the divine (be it a monotheistic God, polytheistic gods, or an eternal cosmic animating force) interpenetrates every part of nature, but is not one with nature. Panentheism differentiates itself from pantheism, which holds that the divine is synonymous with the universe. In panentheism, there are two types of substance, "pan" the universe and God. The universe and the divine are not ontologically equivalent. God is viewed as the eternal animating force within the universe. In some forms of panentheism, the cosmos exists within God, who in turn "transcends", "pervades" or is "in" the cosmos. While pantheism asserts that 'All is God', panentheism claims that God animates all of the universe, and also transcends the universe. In addition, some forms indicate that the universe is contained within God, like in the Judaic concept of Tzimtzum. Much Hindu thought is highly characterized by panentheism and pantheism. Paul Tillich has argued for such a concept within Christian theology, as has liberal biblical scholar Marcus Borg and mystical theologian Matthew Fox, an Episcopal priest.[note 2] Pandeism or pan-deism (from Ancient Greek: πᾶν, romanized: pan, lit. 'all' and Latin: deus meaning "god" in the sense of deism) is a term describing beliefs coherently incorporating or mixing logically reconcilable elements of pantheism (that "God", or a metaphysically equivalent creator deity, is identical to Nature) and classical deism (that the creator-god who designed the universe no longer exists in a status where it can be reached, and can instead be confirmed only by reason). It is therefore most particularly the belief that the creator of the universe actually became the universe, and so ceased to exist as a separate entity. Through this synergy pandeism claims to answer primary objections to deism (why would God create and then not interact with the universe?) and to pantheism (how did the universe originate and what is its purpose?). The central problem in Asian (religious) philosophy is not the body-mind problem, but the search for an unchanging Real or Absolute beyond the world of appearances and changing phenomena, and the search for liberation from dukkha and the liberation from the cycle of rebirth. In Hinduism, substance-ontology prevails, seeing Brahman as the unchanging real beyond the world of appearances. In Buddhism, process ontology is prevalent, seeing reality as empty of an unchanging essence. Characteristic for various Asian philosophy, technology and religions is the discernment of levels of truth, an emphasis on intuitive-experiential understanding of the Absolute such as jnana, bodhi and jianxing: (Chinese; 見性), and the technology of yin and yang used within East Asian medicine with an emphasis on the integration of these levels of truth and its understanding. Vedanta is the inquiry into and systematisation of the Vedas and Upanishads, to harmonise the various and contrasting ideas that can be found in those texts. Within Vedanta, different schools exist: The colonisation of India by the British had a major impact on Hindu society. In response, leading Hindu intellectuals started to study western culture and philosophy, integrating several western notions into Hinduism. This modernised Hinduism, at its turn, has gained popularity in the west. A major role was played in the 19th century by Swami Vivekananda in the revival of Hinduism, and the spread of Advaita Vedanta to the west via the Ramakrishna Mission. His interpretation of Advaita Vedanta has been called Neo-Vedanta. In Advaita, Shankara suggests meditation and Nirvikalpa Samadhi are means to gain knowledge of the already existing unity of Brahman and Atman, not the highest goal itself: [Y]oga is a meditative exercise of withdrawal from the particular and identification with the universal, leading to contemplation of oneself as the most universal, namely, Consciousness. This approach is different from the classical Yoga of complete thought suppression. Vivekananda, according to Gavin Flood, was "a figure of great importance in the development of a modern Hindu self-understanding and in formulating the West's view of Hinduism." Central to his philosophy is the idea that the divine exists in all beings, that all human beings can achieve union with this "innate divinity", and that seeing this divine as the essence of others will further love and social harmony. According to Vivekananda, there is an essential unity to Hinduism, which underlies the diversity of its many forms. According to Flood, Vivekananda's view of Hinduism is the most common among Hindus today. This monism, according to Flood, is at the foundation of earlier Upanishads, to theosophy in the later Vedanta tradition and in modern Neo-Hinduism. According to the Pāli Canon, both pluralism (nānatta) and monism (ekatta) are speculative views. A Theravada commentary notes that the former is similar to or associated with nihilism (ucchēdavāda), and the latter is similar to or associated with eternalism (sassatavada). Within Buddhism, a rich variety of philosophical and pedagogical models can be found. Various schools of Buddhism discern levels of truth: The Prajnaparamita-sutras and Madhyamaka emphasize the non-duality of form and emptiness: "form is emptiness, emptiness is form", as the Heart Sutra says. In Chinese Buddhism this was understood to mean that ultimate reality is not a transcendental realm, but equal to the daily world of relative reality. This idea was well-situated for the existing Chinese culture, which emphasized the mundane world and society. But this does not tell how the absolute is present in the relative world: To deny the duality of samsara and nirvana, as the Perfection of Wisdom does, or to demonstrate logically the error of dichotomizing conceptualization, as Nagarjuna does, is not to address the question of the relationship between samsara and nirvana -or, in more philosophical terms, between phenomenal and ultimate reality [...] What, then, is the relationship between these two realms? This question is answered in such schemata as the Five Ranks of Tozan, the Oxherding Pictures, and Hakuin's Four ways of knowing. Sikhism complies with the concept of Absolute Monism. Sikh philosophy advocates that all that our senses comprehend is an illusion; God is the ultimate reality. Forms being subject to time shall pass away. God's Reality alone is eternal and abiding. The thought is that Atma (soul) is born from, and a reflection of, ParamAtma (Supreme Soul), and "will again merge into it", in the words of the fifth guru of Sikhs, Guru Arjan, "just as water merges back into the water." God and Soul are fundamentally the same; identical in the same way as Fire and its sparks. "Atam meh Ram, Ram meh Atam" which means "The Ultimate Eternal reality resides in the Soul and the Soul is contained in Him". As from one stream, millions of waves arise and yet the waves, made of water, again become water; in the same way all souls have sprung from the Universal Being and would blend again into it. Jewish thought considers God as separate from all physical, created things and as existing outside of time.[note 3][note 4] According to Maimonides, God is an incorporeal being that caused all other existence; to admit corporeality to God is tantamount to admitting complexity to God, which is a contradiction to God as the first cause and constitutes heresy. While Hasidic mystics considered the existence of the physical world a contradiction to God's simpleness, Maimonides saw no contradiction.[note 5] According to Hasidic thought (particularly as propounded by the 18th century, early 19th-century founder of Chabad, Shneur Zalman of Liadi), God is held to be immanent within creation for two interrelated reasons: Christians maintain that God created the universe ex nihilo and not from his own substance, so that the creator is not to be confused with creation, but rather transcends it. There is a movement of "Christian Panentheism". In On Free Choice of the Will, Augustine argued, in the context of the problem of evil, that evil is not the opposite of good, but rather merely the absence of good, something that does not have existence in itself. Likewise, C. S. Lewis described evil as a "parasite" in Mere Christianity, as he viewed evil as something that cannot exist without good to provide it with existence. Lewis went on to argue against dualism from the basis of moral absolutism, and rejected the dualistic notion that God and Satan are opposites, arguing instead that God has no equal, hence no opposite. Lewis rather viewed Satan as the opposite of Michael the archangel. Due to this, Lewis instead argued for a more limited type of dualism. Other theologians, such as Greg Boyd, have argued in more depth that the Biblical authors held a "limited dualism", meaning that God and Satan do engage in real battle, but only due to free will given by God, for the duration that God allows. Latter Day Saint theology also expresses a form of dual-aspect monism via materialism and eternalism, claiming that creation was ex materia (as opposed to ex nihilo in conventional Christianity), as expressed by Parley Pratt and echoed in view by the movement's founder Joseph Smith, making no distinction between the spiritual and the material, these being not just similarly eternal, but ultimately two manifestations of the same reality or substance. Parley Pratt implies a vitalism paired with evolutionary adaptation noting, "these eternal, self-existing elements possess in themselves certain inherent properties or attributes, in a greater or less degree; or, in other words, they possess intelligence, adapted to their several spheres." Parley Pratt's view is also similar to Gottfried Leibniz's monadology, which holds that "reality consists of mind atoms that are living centers of force." Brigham Young anticipates a proto-mentality of elementary particles with his vitalist view, "there is life in all matter, throughout the vast extent of all the eternities; it is in the rock, the sand, the dust, in water, air, the gases, and in short, in every description and organization of matter; whether it be solid, liquid, or gaseous, particle operating with particle." The LDS conception of matter is "essentially dynamic rather than static, if indeed it is not a kind of living energy, and that it is subject at least to the rule of intelligence." John A. Widstoe held a similar, more vitalist view, that "Life is nothing more than matter in motion; that, therefore, all matter possess a kind of life... Matter... [is] intelligence... hence everything in the universe is alive." However, Widstoe resisted outright affirming a belief in panpsychism. Vincent Cornell argues that the Quran provides a monist image of God by describing reality as a unified whole, with God being a single concept that would describe or ascribe all existing things. But most argue that Abrahamic religious scriptures, especially the Quran, see creation and God as two separate existences. It explains that everything has been created by God and is under his control, but at the same time distinguishes creation as being dependent on the existence of God. Some Sufi mystics advocate monism. One of the most notable being the 13th-century Persian poet Rumi (1207–1273) in his didactic poem Masnavi espoused monism. Rumi says in the Masnavi, In the shop for Unity (wahdat); anything that you see there except the One is an idol. Other Sufi mystics however, such as Ahmad Sirhindi, upheld dualistic Monotheism (the separation of God and the Universe). The most influential of the Islamic monists was the Sufi philosopher Ibn Arabi (1165–1240). He developed the concept of 'unity of being' (Arabic: waḥdat al-wujūd), which some argue is a monistic philosophy.[citation needed] Born in al-Andalus, he made an enormous impact on the Muslim world, where he was crowned "the great Master". In the centuries following his death, his ideas became increasingly controversial. Ahmad Sirhindi criticised monistic understanding of 'unity of being', advocating the dualistic-compatible 'unity of witness' (Arabic: wahdat ash-shuhud), maintaining separation of creator and creation. Later, Shah Waliullah Dehlawi reconciled the two ideas maintaining that their differences are semantic differences, arguing that the universal existence (which is different in creation to creator) and the divine essence are different and that the universal existence emanates (in a non-platonic sense) from the divine essence and that the relationship between them is similar to the relationship between the number four and a number being even. The doctrine of waḥdat al-wujūd also enjoys considerable following in the rationalist philosophy of Twelver Shi'ism, with the most famous modern-day adherent being Ruhollah Khomeini. Although the teachings of the Baháʼí Faith have a strong emphasis on social and ethical issues, there exist a number of foundational texts that have been described as mystical. Some of these include statements of a monist nature (e.g., The Seven Valleys and the Hidden Words). The differences between dualist and monist views are reconciled by the teaching that these opposing viewpoints are caused by differences in the observers themselves, not in that which is observed. This is not a 'higher truth/lower truth' position. God is unknowable. For man it is impossible to acquire any direct knowledge of God or the Absolute, because any knowledge that one has, is relative. See also Notes References Sources Further reading External links |
======================================== |
[SOURCE: https://en.wikipedia.org/wiki/Overfitting] | [TOKENS: 2342] |
Contents Overfitting In mathematical modeling, overfitting is the production of an analysis that corresponds too closely or exactly to a particular set of data, and may therefore fail to fit to additional data or predict future observations reliably. An overfitted model is a mathematical model that contains more parameters than can be justified by the data. In the special case of a model that consists of a polynomial function, these parameters represent the degree of a polynomial. The essence of overfitting is to unknowingly extract some of the residual variation (i.e., noise) as if that variation represents the underlying model structure.: 45 Underfitting occurs when a mathematical model cannot adequately capture the underlying structure of the data. An under-fitted model is a model that is missing some parameters or terms that would appear in a correctly specified model. Underfitting would occur, for example, when fitting a linear model to nonlinear data. Such a model will tend to have poor predictive performance. The possibility of over-fitting exists when the criterion used for selecting the model is not the same as the criterion used to judge the suitability of a model. For example, a model might be selected by maximizing its performance on some set of training data, yet its suitability might be determined by its ability to perform well on unseen data; overfitting occurs when a model begins to "memorize" training data rather than "learning" to generalize from a trend. As an extreme example, if the number of parameters is the same as or greater than the number of observations, then a model can perfectly predict the training data simply by memorizing the data in its entirety. (For an illustration, see Figure 2.) Such a model will typically fail severely when making predictions. Overfitting is directly related to approximation error of the selected function class and the optimization error of the optimization procedure. A function class that is too large, in a suitable sense, relative to the dataset size is likely to overfit. Even when the fitted model does not have an excessive number of parameters, it is to be expected that the fitted relationship will appear to perform less well on a new dataset than on the dataset used for fitting (a phenomenon sometimes known as shrinkage). In particular, the value of the coefficient of determination will shrink relative to the original data. To lessen the chance or amount of overfitting, several techniques are available (e.g., model comparison, cross-validation, regularization, early stopping, pruning, Bayesian priors, or dropout). The basis of some techniques is to either (1) explicitly penalize overly complex models or (2) test the model's ability to generalize by evaluating its performance on a set of data not used for training, which is assumed to approximate the typical unseen data that a model will encounter. Statistical inference In statistics, an inference is drawn from a statistical model, which has been selected via some procedure. Burnham & Anderson, in their much-cited text on model selection, argue that to avoid overfitting, we should adhere to the "Principle of Parsimony". The authors also state the following.: 32–33 Overfitted models ... are often free of bias in the parameter estimators, but have estimated (and actual) sampling variances that are needlessly large (the precision of the estimators is poor, relative to what could have been accomplished with a more parsimonious model). False treatment effects tend to be identified, and false variables are included with overfitted models. ... A best approximating model is achieved by properly balancing the errors of underfitting and overfitting. Overfitting is more likely to be a serious concern when there is little theory available to guide the analysis, in part because then there tend to be a large number of models to select from. The book Model Selection and Model Averaging (2008) puts it this way. Given a data set, you can fit thousands of models at the push of a button, but how do you choose the best? With so many candidate models, overfitting is a real danger. Is the monkey who typed Hamlet actually a good writer? In regression analysis, overfitting occurs frequently. As an extreme example, if there are p variables in a linear regression with p data points, the fitted line can go exactly through every point. For logistic regression or Cox proportional hazards models, there are a variety of rules of thumb (e.g. 5–9, 10 and 10–15 — the guideline of 10 observations per independent variable is known as the "one in ten rule"). In the process of regression model selection, the mean squared error of the random regression function can be split into random noise, approximation bias, and variance in the estimate of the regression function. The bias–variance tradeoff is often used to overcome overfit models. With a large set of explanatory variables that actually have no relation to the dependent variable being predicted, some variables will in general be falsely found to be statistically significant and the researcher may thus retain them in the model, thereby overfitting the model. This is known as Freedman's paradox. Machine learning Usually, a learning algorithm is trained using some set of "training data": exemplary situations for which the desired output is known. The goal is that the algorithm will also perform well on predicting the output when fed "validation data" that was not encountered during its training. Overfitting is the use of models or procedures that violate Occam's razor, for example by including more adjustable parameters than are ultimately optimal, or by using a more complicated approach than is ultimately optimal. For an example where there are too many adjustable parameters, consider a dataset where training data for y can be adequately predicted by a linear function of two independent variables. Such a function requires only three parameters (the intercept and two slopes). Replacing this simple function with a new, more complex quadratic function, or with a new, more complex linear function on more than two independent variables, carries a risk: Occam's razor implies that any given complex function is a priori less probable than any given simple function. If the new, more complicated function is selected instead of the simple function, and if there was not a large enough gain in training data fit to offset the complexity increase, then the new complex function "overfits" the data and the complex overfitted function will likely perform worse than the simpler function on validation data outside the training dataset, even though the complex function performed as well, or perhaps even better, on the training dataset. When comparing different types of models, complexity cannot be measured solely by counting how many parameters exist in each model; the expressivity of each parameter must be considered as well. For example, it is nontrivial to directly compare the complexity of a neural net (which can track curvilinear relationships) with m parameters to a regression model with n parameters. Overfitting is especially likely in cases where learning was performed too long or where training examples are rare, causing the learner to adjust to very specific random features of the training data that have no causal relation to the target function. In this process of overfitting, the performance on the training examples still increases while the performance on unseen data becomes worse. As a simple example, consider a database of retail purchases that includes the item bought, the purchaser, and the date and time of purchase. It's easy to construct a model that will fit the training set perfectly by using the date and time of purchase to predict the other attributes, but this model will not generalize at all to new data because those past times will never occur again. Generally, a learning algorithm is said to overfit relative to a simpler one if it is more accurate in fitting known data (hindsight) but less accurate in predicting new data (foresight). One can intuitively understand overfitting from the fact that information from all past experience can be divided into two groups: information that is relevant for the future, and irrelevant information ("noise"). Everything else being equal, the more difficult a criterion is to predict (i.e., the higher its uncertainty), the more noise exists in past information that needs to be ignored. The problem is determining which part to ignore. A learning algorithm that can reduce the risk of fitting noise is called "robust." The most obvious consequence of overfitting is poor performance on the validation dataset. Other negative consequences include: The optimal function usually needs verification on bigger or completely new datasets. There are, however, methods like minimum spanning tree or life-time of correlation that applies the dependence between correlation coefficients and time-series (window width). Whenever the window width is big enough, the correlation coefficients are stable and don't depend on the window width size anymore. Therefore, a correlation matrix can be created by calculating a coefficient of correlation between investigated variables. This matrix can be represented topologically as a complex network where direct and indirect influences between variables are visualized. Dropout regularisation (random removal of training set data) can also improve robustness and therefore reduce over-fitting by probabilistically removing inputs to a layer. Pruning is another technique that mitigates overfitting and enhances generalization by identifying a sparse, optimal neural network structure, while simultaneously reducing the computational cost of both training and inference. Underfitting Underfitting is the inverse of overfitting, meaning that the statistical model or machine learning algorithm is too simplistic to accurately capture the patterns in the data. A sign of underfitting is that there is a high bias and low variance detected in the current model or algorithm used (the inverse of overfitting: low bias and high variance). This can be gathered from the Bias-variance tradeoff, which is the method of analyzing a model or algorithm for bias error, variance error, and irreducible error. With a high bias and low variance, the result of the model is that it will inaccurately represent the data points and thus insufficiently be able to predict future data results (see Generalization error). As shown in Figure 5, the linear line could not represent all the given data points due to the line not resembling the curvature of the points. We would expect to see a parabola-shaped line as shown in Figure 6 and Figure 1. If we were to use Figure 5 for analysis, we would get false predictive results contrary to the results if we analyzed Figure 6. Burnham & Anderson state the following.: 32 ... an underfitted model would ignore some important replicable (i.e., conceptually replicable in most other samples) structure in the data and thus fail to identify effects that were actually supported by the data. In this case, bias in the parameter estimators is often substantial, and the sampling variance is underestimated, both factors resulting in poor confidence interval coverage. Underfitted models tend to miss important treatment effects in experimental settings. There are multiple ways to deal with underfitting: Benign overfitting Benign overfitting describes the phenomenon of a statistical model that seems to generalize well to unseen data, even when it has been fit perfectly on noisy training data (i.e., obtains perfect predictive accuracy on the training set). The phenomenon is of particular interest in deep neural networks, but is studied from a theoretical perspective in the context of much simpler models, such as linear regression. In particular, it has been shown that overparameterization is essential for benign overfitting in this setting. In other words, the number of directions in parameter space that are unimportant for prediction must significantly exceed the sample size. See also Notes References Further reading External links |
======================================== |
[SOURCE: https://en.wikipedia.org/wiki/Markus_Persson#cite_note-GAMASUTRA-14] | [TOKENS: 3525] |
Contents Markus Persson Markus Alexej Persson (/ˈpɪərsən/ ⓘ PEER-sən, Swedish: [ˈmǎrːkɵs ˈpæ̌ːʂɔn] ⓘ; born 1 June 1979), known by the pseudonym Notch, is a Swedish video game programmer and designer. He is the creator of Minecraft, the best-selling video game in history. He founded the video game development company Mojang Studios in 2009. Persson began developing video games at an early age. His commercial success began after he published an early version of Minecraft in 2009. Prior to the game's official retail release in 2011, it had sold over four million copies. After this point Persson stood down as the lead designer and transferred his creative authority to Jens Bergensten. In September 2014 Persson announced his intention to leave Mojang, and in November of that year the company was sold to Microsoft reportedly for US$2.5 billion, which made him a billionaire. Since 2016 several of Persson's posts on Twitter regarding feminism, race, and transgender rights have caused public controversies. He has been described as "an increasingly polarizing figure, tweeting offensive statements regarding race, the LGBTQ community, gender, and other topics." In an effort to distance itself from Persson, Microsoft removed mentions of his name from Minecraft (excluding one instance in the game's end credits) and did not invite him to the game's tenth anniversary celebration. In 2015 he co-founded a separate game studio called Rubberbrain, which was relaunched in 2024 as Bitshift Entertainment. Early life Markus Alexej Persson was born in Stockholm, Sweden, to a Finnish mother, Ritva, and a Swedish father, Birger, on 1 June 1979. He has one sister. He grew up in Edsbyn until he was seven years old, when his family moved back to Stockholm. In Edsbyn, Persson's father worked for the railroad, and his mother was a nurse. He spent much time outdoors in Edsbyn, exploring the woods with his friends. When Persson was about seven years old, his parents divorced, and he and his sister lived with their mother. His father moved to a cabin in the countryside. Persson said in an interview that they experienced food insecurity around once a month. Persson lost contact with his father for several years after the divorce. According to Persson, his father suffered from depression, bipolar disorder, alcoholism, and medication abuse, and went to jail for robberies. While his father had somewhat recovered during Persson's early life, his father relapsed, contributing to the divorce. His sister also experimented with drugs and ran away from home. He had gained interest in video games at an early age. His father was "a really big nerd", who built his own modem and taught Persson to use the family's Commodore 128. On it, Persson played bootleg games and loaded in various type-in programs from computer magazines with the help of his sister. The first game he purchased with his own money was The Bard's Tale. He began programming on his father's Commodore 128 home computer at the age of seven. He produced his first game at the age of eight, a text-based adventure game. By 1994 Persson knew he wanted to become a video game developer, but his teachers advised him to study graphic design, which he did from ages 15 to 18. Persson, although introverted, was well-liked by his peers, but after entering secondary school was a "loner" and reportedly had only one friend. He spent most of his spare time with games and programming at home. He managed to reverse-engineer the Doom engine, which he continued to take great pride in as of 2014[update]. He never finished high school, but was reportedly a good student. Career Persson started his career working as a web designer. He later found employment at Game Federation, where he met Rolf Jansson. The pair worked in their spare time to build the 2006 video game Wurm Online. The game was released through a new entity, "Mojang Specifications AB". Persson left the project in late 2007. As Persson wanted to reuse the name "Mojang", Jansson agreed to rename the company to Onetoofree AB. Between 2004 and 2009 Persson worked as a game developer for Midasplayer (later known as King). There, he worked as a programmer, mostly building browser games made in Flash. He later worked as a programmer for jAlbum. Prior to creating Minecraft, Persson developed multiple, small games. He also entered a number of game design competitions and participated in discussions on the TIGSource forums, a web forum for independent game developers. One of Persson's more notable personal projects was called RubyDung, an isometric three-dimensional base-building game like RollerCoaster Tycoon and Dwarf Fortress. While working on RubyDung, Persson experimented with a first-person view mode similar to that found in Dungeon Keeper. However, he felt the graphics were too pixelated and omitted this mode. In 2009 Persson found inspiration in Infiniminer, a block-based open-ended mining game. Infiniminer heavily influenced his future work on RubyDung, and was behind Persson's reasoning for returning the first-person mode, the "blocky" visual style and the block-building fundamentals to the game. RubyDung is the earliest known Minecraft prototype created by Persson. On 17 May 2009 Persson released the original edition (later called "Classic version") of Minecraft on the TIGSource forums. He regularly updated the game based on feedback from TIGSource users. Persson released several new versions of Minecraft throughout 2009 and 2010, going through several phases of development including Survival Test, Indev, and Infdev. On 30 June 2010 Persson released the game's Alpha version. While working on the pre-Alpha version of Minecraft, Persson continued working at jAlbum. In 2010, after the release and subsequent success of Minecraft's Alpha version, Persson moved from a full-time role to a part-time role at jAlbum. He left jAlbum later that same year. In September 2010 Persson travelled to Valve Corporation's headquarters in Bellevue, Washington, United States, where he took part in a programming exercise and met Gabe Newell. Persson was subsequently offered a job at Valve, which he turned down in order to continue work on Minecraft. On 20 December 2010 Minecraft moved into its beta phase and began expanding to other platforms, including mobile. In January 2011 Minecraft reached one million registered accounts. Six months afterwards, it reached ten million. The game has sold over four million copies by 7 November 2011. Mojang held the first Minecon from 18 to 19 November 2011 to celebrate its full release, and subsequently made it an annual event. Following this, on 11 December 2011, Persson transferred creative control of Minecraft to Jens Bergensten and began working on another game title, 0x10c, although he reportedly abandoned the project around 2013. In 2013 Mojang recorded revenues of $330 million and profits of $129 million. Persson has stated that, due to the intense media attention and public pressure, he became exhausted with running Minecraft and Mojang. In a September 2014 blog post he shared his realization that he "didn't have the connection to my fans I thought I had", that he had "become a symbol", and that he did not wish to be responsible for Mojang's increasingly large operation. In June 2014 Persson tweeted "Anyone want to buy my share of Mojang so I can move on with my life? Getting hate for trying to do the right thing is not my gig", reportedly partly as a joke. Persson controlled a 71% stake in Mojang at the time. The offer attracted significant interest from Activision Blizzard, EA, and Microsoft. Forbes later reported that Microsoft wanted to purchase the game as a "tax dodge" to turn their taxable excess liquid cash into other assets. In September 2014 Microsoft agreed to purchase Mojang for $2.5 billion, making Persson a billionaire. He then left the company after the deal was finalised in November. Since leaving Mojang, Persson has worked on several small projects. On 23 June 2014 he founded a company with Porsér called Rubberbrain AB; the company had no games by 2021, despite spending SEK 60 million. The company was relaunched as Bitshift Entertainment, LLC on 28 March 2024. Persson expressed interest in creating a new video game studio in 2020, and in developing virtual reality games. He has also since created a series of narrative-driven immersive events called ".party()", which uses extensive visual effects and has been hosted in multiple cities. At the beginning of 2025 Persson decided to create a spiritual successor to Minecraft, referred to as "Minecraft 2", in response to the results of a poll on X. However, after speaking to his team, he shortly went against this in favour of developing the other choice on his Twitter poll, a roguelike titled Levers and Chests. Games Persson's most popular creation is the survival sandbox game Minecraft, which was first publicly available on 17 May 2009 and fully released on 18 November 2011. Persson left his job as a game developer to work on Minecraft full-time until completion. In early 2011, Mojang AB sold the one millionth copy of the game, several months later their second, and several more their third. Mojang hired several new staff members for the Minecraft team, while Persson passed the lead developer role to Jens Bergensten. He stopped working on Minecraft after a deal with Microsoft to sell Mojang for $2.5 billion. This brought his net worth to US$1.5 billion. Persson and Jakob Porsér came up with the idea for Scrolls including elements from board games and collectible card games. Persson noted that he will not be actively involved in development of the game and that Porsér will be developing it. Persson revealed on his Tumblr blog on 5 August 2011 that he was being sued by a Swedish law firm representing Bethesda Softworks over the trademarked name of Scrolls, claiming that it conflicted with their The Elder Scrolls series of games. On 17 August 2011 Persson challenged Bethesda to a Quake 3 tournament to decide the outcome of the naming dispute. On 27 September 2011 Persson confirmed that the lawsuit was going to court. ZeniMax Media, owner of Bethesda Softworks, announced the lawsuit's settlement in March 2012. The settlement allowed Mojang to continue using the Scrolls trademark. In 2018, Scrolls was made available free of charge and renamed to Caller's Bane. Cliffhorse is a humorous game programmed in two hours using the Unity game engine and free assets. The game took inspiration from Skyrim's physics engine, "the more embarrassing minimum-effort Greenlight games", Goat Simulator, and Big Rigs: Over the Road Racing. The game was released to Microsoft Windows systems as an early access and honourware game on the first day of E3 2014, instructing users to donate Dogecoin to "buy" the game before downloading it. The game accumulated over 280,000 dogecoins. Following the end to his involvement with Minecraft, Persson began pre-production of an alternate reality space game set in the distant future in March 2012. On April Fools' Day Mojang launched a satirical website for Mars Effect (parody of Mass Effect), citing the lawsuit with Bethesda as an inspiration. However, the gameplay elements remained true and on 4 April, Mojang revealed 0x10c (pronounced "Ten to the C") as a space sandbox title. Persson officially halted game production in August 2013. However, C418, the composer of the game's soundtrack (as well as that of Minecraft), released an album of the work he had made for the game. In 2013, Persson made a free game called Shambles in the Unity game engine. Persson has also participated in several Ludum Dare 48-hour game making competitions. Personal life In 2011 Persson married Elin Zetterstrand, whom he had dated for four years before. Zetterstrand was a former moderator on the Minecraft forums. They had a daughter together, but by mid-2012, he began to see little of her. On 15 August 2012 he announced that he and his wife had filed for divorce. The divorce was finalised later that year. On 14 December 2011 Persson's father committed suicide with a handgun after drinking heavily. In an interview with The New Yorker, Persson said of his father: When I decided I wanted to quit my day job and work on my own games, he was the only person who supported my decision. He was proud of me and made sure I knew. When I added the monsters to Minecraft, he told me that the dark caves became too scary for him. But I think that was the only true criticism I ever heard from him. Persson later admitted that he himself suffered from depression and various highs and lows in his mood. Persson has criticised the stance of large game companies on piracy. He once stated that "piracy is not theft", viewing unauthorised downloads as potential future customers. Persson stated himself to be a member of the Pirate Party of Sweden in 2011. He is also a member of Mensa. He has donated to numerous charities, including Médecins Sans Frontières (Doctors Without Borders). Under his direction, Mojang spent a week developing Catacomb Snatch for the Humble Indie Bundle and raised US$458,248 for charity. He also donated $250,000 to the Electronic Frontier Foundation in 2012. In 2011 he gave $3 million in dividends back to Mojang employees. According to Forbes, his net worth in 2023 was around $1.2 billion. In 2014 Persson was one of the biggest taxpayers in Sweden. Around 2014, he lived in a multi-level penthouse in Östermalm, Stockholm, an area he described as "where the rich people live". In December 2014 Persson purchased a home in Trousdale Estates, a neighbourhood in Beverly Hills, California, in the United States, for $70 million, a record sales price for Beverly Hills at the time. Persson reportedly outbid Beyoncé and Jay-Z for the property. Persson began receiving criticism for political and social opinions he expressed on social media as early as 2016. November 30, 2017 In 2017, he proposed a heterosexual pride holiday, and wrote that those who opposed the idea "deserve to be shot." After facing backlash, he deleted the tweets and rescinded his statements, writing, "So yeah, it's about pride of daring to express, not about pride of being who you are. I get it now." Later in the year, he wrote that feminism is a "social disease" and called the video game developer and feminist Zoë Quinn a "cunt", although he was generally critical of the GamerGate movement. He has described intersectional feminism as a "framework for bigotry" and the use of the word mansplaining as being sexist. Also in 2017, Persson tweeted that "It's okay to be white". Later that year, he stated that he believed in the Pizzagate conspiracy theory. In 2019, he tweeted referencing QAnon, saying "Q is legit. Don't trust the media." Later in 2019, he tweeted in response to a pro-transgender internet meme that, "You are absolutely evil if you want to encourage delusion. What happened to not stigmatizing mental illness?" He then also promoted claims that people were fined for "using the wrong pronoun". However, after facing backlash, he tweeted a day afterwards that he had "no idea what [being trans is] like of course, but it's inspiring as hell when people open up and choose to actually be who they know themselves as. Not because it's a cool choice, because it's a big step. I gues [sic] that's actually cool nvm". Later that year, Microsoft removed two mentions of Persson's name in the "19w13a" snapshot of Minecraft and did not invite him to the 10-year anniversary celebration of the game. A spokesperson for Microsoft stated that his views "do not reflect those of Microsoft or Mojang". He is still mentioned in the End Poem ("a flat, infinite world created by a man called Markus").[citation needed] Awards References External links |
======================================== |
[SOURCE: https://en.wikipedia.org/w/index.php?title=Minecraft&oldid=1339498619] | [TOKENS: 12858] |
Contents Minecraft Minecraft is a sandbox game developed and published by Mojang Studios. Following its initial public alpha release in 2009, it was formally released in 2011 for personal computers. The game has since been ported to numerous platforms, including mobile devices and various video game consoles. In Minecraft, players explore a procedurally generated world with virtually infinite terrain made up of voxels (cubes). They can discover and extract raw materials, craft tools and items, build structures, fight hostile mobs, and cooperate with or compete against other players in multiplayer. The game's large community offers a wide variety of user-generated content, such as modifications, servers, player skins, texture packs, and custom maps, which add new game mechanics and possibilities. Originally created by Markus "Notch" Persson using the Java programming language, Jens "Jeb" Bergensten was handed control over the game's development following its full release. In 2014, Mojang and the Minecraft intellectual property were purchased by Microsoft for US$2.5 billion; Xbox Game Studios hold the publishing rights for the Bedrock Edition, the unified cross-platform version which evolved from the Pocket Edition codebase[i] and replaced the legacy console versions. Bedrock is updated concurrently with Mojang's original Java Edition, although with numerous, generally small, differences. Minecraft is the best-selling video game in history with over 350 million copies sold. It has received critical acclaim, winning several awards and being cited as one of the greatest video games of all time. Social media, parodies, adaptations, merchandise, and the annual Minecon conventions have played prominent roles in popularizing it. The wider Minecraft franchise includes several spin-off games, such as Minecraft: Story Mode, Minecraft Dungeons, and Minecraft Legends. A film adaptation, titled A Minecraft Movie, was released in 2025 and became the second highest-grossing video game film of all time. Gameplay Minecraft is a 3D sandbox video game that has no required goals to accomplish, giving players a large amount of freedom in choosing how to play the game. The game features an optional achievement system. Gameplay is in the first-person perspective by default, but players have the option of third-person perspectives. The game world is composed of rough 3D objects—mainly cubes, referred to as blocks—representing various materials, such as dirt, stone, ores, tree trunks, water, and lava. The core gameplay revolves around picking up and placing these objects. These blocks are arranged in a voxel grid, while players can move freely around the world. Players can break, or mine, blocks and then place them elsewhere, enabling them to build things. Very few blocks are affected by gravity, instead maintaining their voxel position in the air. Players can also craft a wide variety of items, such as armor, which mitigates damage from attacks; weapons (such as swords or bows and arrows), which allow monsters and animals to be killed more easily; and tools (such as pickaxes or shovels), which break certain types of blocks more quickly. Some items have multiple tiers depending on the material used to craft them, with higher-tier items being more effective and durable. They may also freely craft helpful blocks—such as furnaces which can cook food and smelt ores, and torches that produce light—or exchange items with villagers (NPC) through trading emeralds for different goods and vice versa. The game has an inventory system, allowing players to carry a limited number of items. The in-game time system follows a day and night cycle, with one full cycle lasting for 20 real-time minutes. The game also contains a material called redstone, which can be used to make primitive mechanical devices, electrical circuits, and logic gates, allowing for the construction of many complex systems. New players are given a randomly selected default character skin out of nine possibilities, including Steve or Alex, but are able to create and upload their own skins. Players encounter various mobs (short for mobile entities) including animals, villagers, and hostile creatures. Passive mobs, such as cows, pigs, and chickens, spawn during the daytime and can be hunted for food and crafting materials, while hostile mobs—including large spiders, witches, skeletons, and zombies—spawn during nighttime or in dark places such as caves. Some hostile mobs, such as zombies and skeletons, burn under the sun if they have no headgear and are not standing in water. Other creatures unique to Minecraft include the creeper (an exploding creature that sneaks up on the player) and the enderman (a creature with the ability to teleport as well as pick up and place blocks). There are also variants of mobs that spawn in different conditions; for example, zombies have husk and drowned variants that spawn in deserts and oceans, respectively. The Minecraft environment is procedurally generated as players explore it using a map seed that is randomly chosen at the time of world creation (or manually specified by the player). Divided into biomes representing different environments with unique resources and structures, worlds are designed to be effectively infinite in traditional gameplay, though technical limits on the player have existed throughout development, both intentionally and not. Implementation of horizontally infinite generation initially resulted in a glitch termed the "Far Lands" at over 12 million blocks away from the world center, where terrain generated as wall-like, fissured patterns. The Far Lands and associated glitches were considered the effective edge of the world until they were resolved, with the current horizontal limit instead being a special impassable barrier called the world border, located 30 million blocks away. Vertical space is comparatively limited, with an unbreakable bedrock layer at the bottom and a building limit several hundred blocks into the sky. Minecraft features three independent dimensions accessible through portals and providing alternate game environments. The Overworld is the starting dimension and represents the real world, with a terrestrial surface setting including plains, mountains, forests, oceans, caves, and small sources of lava. The Nether is a hell-like underworld dimension accessed via an obsidian portal and composed mainly of lava. Mobs that populate the Nether include shrieking, fireball-shooting ghasts, alongside anthropomorphic pigs called piglins and their zombified counterparts. Piglins in particular have a bartering system, where players can give them gold ingots and receive items in return. Structures known as Nether Fortresses generate in the Nether, containing mobs such as wither skeletons and blazes, which can drop blaze rods needed to access the End dimension. The player can also choose to build an optional boss mob known as the Wither, using skulls obtained from wither skeletons and soul sand. The End can be reached through an end portal, consisting of twelve end portal frames. End portals are found in underground structures in the Overworld known as strongholds. To find strongholds, players must craft eyes of ender using an ender pearl and blaze powder. Eyes of ender can then be thrown, traveling in the direction of the stronghold. Once the player reaches the stronghold, they can place eyes of ender into each portal frame to activate the end portal. The dimension consists of islands floating in a dark, bottomless void. A boss enemy called the Ender Dragon guards the largest, central island. Killing the dragon opens access to an exit portal, which, when entered, cues the game's ending credits and the End Poem, a roughly 1,500-word work written by Irish novelist Julian Gough, which takes about nine minutes to scroll past, is the game's only narrative text, and the only text of significant length directed at the player.: 10–12 At the conclusion of the credits, the player is teleported back to their respawn point and may continue the game indefinitely. In Survival mode, players have to gather natural resources such as wood and stone found in the environment in order to craft certain blocks and items. Depending on the difficulty, monsters spawn in darker areas outside a certain radius of the character, requiring players to build a shelter in order to survive at night. The mode also has a health bar which is depleted by attacks from mobs, falls, drowning, falling into lava, suffocation, starvation, and other events. Players also have a hunger bar, which must be periodically refilled by eating food in-game unless the player is playing on peaceful difficulty. If the hunger bar is empty, the player starves. Health replenishes when players have a full hunger bar or continuously on peaceful. Upon losing all health, players die. The items in the players' inventories are dropped unless the game is reconfigured not to do so. Players then re-spawn at their spawn point, which by default is where players first spawn in the game and can be changed by sleeping in a bed or using a respawn anchor. Dropped items can be recovered if players can reach them before they despawn after 5 minutes. Players may acquire experience points (commonly referred to as "xp" or "exp") by killing mobs and other players, mining, smelting ores, animal breeding, and cooking food. Experience can then be spent on enchanting tools, armor and weapons. Enchanted items are generally more powerful, last longer, or have other special effects. The game features two more game modes based on Survival, known as Hardcore mode and Adventure mode. Hardcore mode plays identically to Survival mode, but with the game's difficulty setting locked to "Hard" and with permadeath, forcing them to delete the world or explore it as a spectator after dying. Adventure mode was added to the game in a post-launch update, and prevents the player from directly modifying the game's world. It was designed primarily for use in custom maps, allowing map designers to let players experience it as intended. In Creative mode, players have access to an infinite number of all resources and items in the game through the inventory menu and can place or mine them instantly. Players can toggle the ability to fly freely around the game world at will, and their characters usually do not take any damage nor are affected by hunger. The game mode helps players focus on building and creating projects of any size without disturbance. Multiplayer in Minecraft enables multiple players to interact and communicate with each other on a single world. It is available through direct game-to-game multiplayer, local area network (LAN) play, local split screen (console-only), and servers (player-hosted and business-hosted). Players can run their own server by making a realm, using a host provider, hosting one themselves or connect directly to another player's game via Xbox Live, PlayStation Network or Nintendo Switch Online. Single-player worlds have LAN support, allowing players to join a world on locally interconnected computers without a server setup. Minecraft multiplayer servers are guided by server operators, who have access to server commands such as setting the time of day and teleporting players. Operators can also set up restrictions concerning which usernames or IP addresses are allowed or disallowed to enter the server. Multiplayer servers have a wide range of activities, with some servers having their own unique rules and customs. The largest and most popular server is Hypixel, which has been visited by over 14 million unique players. Player versus player combat (PvP) can be enabled to allow fighting between players. In 2013, Mojang announced Minecraft Realms, a server hosting service intended to enable players to run server multiplayer games easily and safely without having to set up their own. Unlike a standard server, only invited players can join Realms servers, and these servers do not use server addresses. Minecraft: Java Edition Realms server owners can invite up to twenty people to play on their server, with up to ten players online at a time. Minecraft Realms server owners can invite up to 3,000 people to play on their server, with up to ten players online at one time. The Minecraft: Java Edition Realms servers do not support user-made plugins, but players can play custom Minecraft maps. Minecraft Bedrock Realms servers support user-made add-ons, resource packs, behavior packs, and custom Minecraft maps. At Electronic Entertainment Expo 2016, support for cross-platform play between Windows 10, iOS, and Android platforms was added through Realms starting in June 2016, with Xbox One and Nintendo Switch support to come later in 2017, and support for virtual reality devices. On 31 July 2017, Mojang released the beta version of the update allowing cross-platform play. Nintendo Switch support for Realms was released in July 2018. The modding community consists of fans, users and third-party programmers. Using a variety of application program interfaces that have arisen over time, they have produced a wide variety of downloadable content for Minecraft, such as modifications, texture packs and custom maps. Modifications of the Minecraft code, called mods, add a variety of gameplay changes, ranging from new blocks, items, and mobs to entire arrays of mechanisms. The modding community is responsible for a substantial supply of mods from ones that enhance gameplay, such as mini-maps, waypoints, and durability counters, to ones that add to the game elements from other video games and media. While a variety of mod frameworks were independently developed by reverse engineering the code, Mojang has also enhanced vanilla Minecraft with official frameworks for modification, allowing the production of community-created resource packs, which alter certain game elements including textures and sounds. Players can also create their own "maps" (custom world save files) that often contain specific rules, challenges, puzzles and quests, and share them for others to play. Mojang added an adventure mode in August 2012 and "command blocks" in October 2012, which were created specially for custom maps in Java Edition. Data packs, introduced in version 1.13 of the Java Edition, allow further customization, including the ability to add new achievements, dimensions, functions, loot tables, predicates, recipes, structures, tags, and world generation. The Xbox 360 Edition supported downloadable content, which was available to purchase via the Xbox Games Store; these content packs usually contained additional character skins. It later received support for texture packs in its twelfth title update while introducing "mash-up packs", which combined texture packs with skin packs and changes to the game's sounds, music and user interface. The first mash-up pack (and by extension, the first texture pack) for the Xbox 360 Edition was released on 4 September 2013, and was themed after the Mass Effect franchise. Unlike Java Edition, however, the Xbox 360 Edition did not support player-made mods or custom maps. A cross-promotional resource pack based on the Super Mario franchise by Nintendo was released exclusively for the Wii U Edition worldwide on 17 May 2016, and later bundled free with the Nintendo Switch Edition at launch. Another based on Fallout was released on consoles that December, and for Windows and Mobile in April 2017. In April 2018, malware was discovered in several downloadable user-made Minecraft skins for use with the Java Edition of the game. Avast stated that nearly 50,000 accounts were infected, and when activated, the malware would attempt to reformat the user's hard drive. Mojang promptly patched the issue, and released a statement stating that "the code would not be run or read by the game itself", and would run only when the image containing the skin itself was opened. In June 2017, Mojang released the "1.1 Discovery Update" to the Pocket Edition of the game, which later became the Bedrock Edition. The update introduced the "Marketplace", a catalogue of purchasable user-generated content intended to give Minecraft creators "another way to make a living from the game". Various skins, maps, texture packs and add-ons from different creators can be bought with "Minecoins", a digital currency that is purchased with real money. Additionally, users can access specific content with a subscription service titled "Marketplace Pass". Alongside content from independent creators, the Marketplace also houses items published by Mojang and Microsoft themselves, as well as official collaborations between Minecraft and other intellectual properties. By 2022, the Marketplace had over 1.7 billion content downloads, generating over $500 million in revenue. Development Before creating Minecraft, Markus "Notch" Persson was a game developer at King, where he worked until March 2009. At King, he primarily developed browser games and learned several programming languages. During his free time, he prototyped his own games, often drawing inspiration from other titles, and was an active participant on the TIGSource forums for independent developers. One such project was "RubyDung", a base-building game inspired by Dwarf Fortress, but with an isometric, three-dimensional perspective similar to RollerCoaster Tycoon. Among the features in RubyDung that he explored was a first-person view similar to Dungeon Keeper, though he ultimately discarded this idea, feeling the graphics were too pixelated at the time. Around March 2009, Persson left King and joined jAlbum, while continuing to work on his prototypes. Infiniminer, a block-based open-ended mining game first released in April 2009, inspired Persson's vision for RubyDung's future direction. Infiniminer heavily influenced the visual style of gameplay, including bringing back the first-person mode, the "blocky" visual style and the block-building fundamentals. However, unlike Infiniminer, Persson wanted Minecraft to have RPG elements. The first public alpha build of Minecraft was released on 17 May 2009 on TIGSource. Over the years, Persson regularly released test builds that added new features, including tools, mobs, and entire new dimensions. In 2011, partly due to the game's rising popularity, Persson decided to release a full 1.0 version—a second part of the "Adventure Update"—on 18 November 2011. Shortly after, Persson stepped down from development, handing the project's lead to Jens "Jeb" Bergensten. On 15 September 2014, Microsoft, the developer behind the Microsoft Windows operating system and Xbox video game console, announced a $2.5 billion acquisition of Mojang, which included the Minecraft intellectual property. Persson had suggested the deal on Twitter, asking a corporation to buy his stake in the game after receiving criticism for enforcing terms in the game's end-user license agreement (EULA), which had been in place for the past three years. According to Persson, Mojang CEO Carl Manneh received a call from a Microsoft executive shortly after the tweet, asking if Persson was serious about a deal. Mojang was also approached by other companies including Activision Blizzard and Electronic Arts. The deal with Microsoft was arbitrated on 6 November 2014 and led to Persson becoming one of Forbes' "World's Billionaires". After 2014, Minecraft's primary versions received usually annual major updates—free to players who have purchased the game— each primarily centered around a specific theme. For instance, version 1.13, the Update Aquatic, focused on ocean-related features, while version 1.16, the Nether Update, introduced significant changes to the Nether dimension. However, in late 2024, Mojang announced a shift in their update strategy; rather than releasing large updates annually, they opted for a more frequent release schedule with smaller, incremental updates, stating, "We know that you want new Minecraft content more often." The Bedrock Edition has also received regular updates, now matching the themes of the Java Edition updates. Other versions of the game, such as various console editions and the Pocket Edition, were either merged into Bedrock or discontinued and have not received further updates. On 7 May 2019, coinciding with Minecraft's 10th anniversary, a JavaScript recreation of an old 2009 Java Edition build named Minecraft Classic was made available to play online for free. On 16 April 2020, a Bedrock Edition-exclusive beta version of Minecraft, called Minecraft RTX, was released by Nvidia. It introduced physically-based rendering, real-time path tracing, and DLSS for RTX-enabled GPUs. The public release was made available on 8 December 2020. Path tracing can only be enabled in supported worlds, which can be downloaded for free via the in-game Minecraft Marketplace, with a texture pack from Nvidia's website, or with compatible third-party texture packs. It cannot be enabled by default with any texture pack on any world. Initially, Minecraft RTX was affected by many bugs, display errors, and instability issues. On 22 March 2025, a new visual mode called Vibrant Visuals, an optional graphical overhaul similar to Minecraft RTX, was announced. It promises modern rendering features—such as dynamic shadows, screen space reflections, volumetric fog, and bloom—without the need of RTX-capable hardware. Vibrant Visuals was released as a part of the Chase the Skies update on 17 June 2025 for Bedrock Edition and is planned to release on Java Edition at a later date. Development began for the original edition of Minecraft—then known as Cave Game, and now known as the Java Edition—in May 2009,[k] and ended on 13 May, when Persson released a test video on YouTube of an early version of the game, dubbed the "Cave game tech test" or the "Cave game tech demo". The game was named Minecraft: Order of the Stone the next day, after a suggestion made by a player. "Order of the Stone" came from the webcomic The Order of the Stick, and "Minecraft" was chosen "because it's a good name". The title was later shortened to just Minecraft, omitting the subtitle. Persson completed the game's base programming over a weekend in May 2009, and private testing began on TigIRC on 16 May. The first public release followed on 17 May 2009 as a developmental version shared on the TIGSource forums. Based on feedback from forum users, Persson continued updating the game. This initial public build later became known as Classic. Further developmental phases—dubbed Survival Test, Indev, and Infdev—were released throughout 2009 and 2010. The first major update, known as Alpha, was released on 30 June 2010. At the time, Persson was still working a day job at jAlbum but later resigned to focus on Minecraft full-time as sales of the alpha version surged. Updates were distributed automatically, introducing new blocks, items, mobs, and changes to game mechanics such as water flow. With revenue generated from the game, Persson founded Mojang, a video game studio, alongside former colleagues Jakob Porser and Carl Manneh. On 11 December 2010, Persson announced that Minecraft would enter its beta phase on 20 December. He assured players that bug fixes and all pre-release updates would remain free. As development progressed, Mojang expanded, hiring additional employees to work on the project. The game officially exited beta and launched in full on 18 November 2011. On 1 December 2011, Jens "Jeb" Bergensten took full creative control over Minecraft, replacing Persson as lead designer. On 28 February 2012, Mojang announced the hiring of the developers behind Bukkit, a popular developer API for Minecraft servers, to improve Minecraft's support of server modifications. This move included Mojang taking apparent ownership of the CraftBukkit server mod, though this apparent acquisition later became controversial, and its legitimacy was questioned due to CraftBukkit's open-source nature and licensing under the GNU General Public License and Lesser General Public License. In August 2011, Minecraft: Pocket Edition was released as an early alpha for the Xperia Play via the Android Market, later expanding to other Android devices on 8 October 2011. The iOS version followed on 17 November 2011. A port was made available for Windows Phones shortly after Microsoft acquired Mojang. Unlike Java Edition, Pocket Edition initially focused on Minecraft's creative building and basic survival elements but lacked many features of the PC version. Bergensten confirmed on Twitter that the Pocket Edition was written in C++ rather than Java, as iOS does not support Java. On 10 December 2014, a port of Pocket Edition was released for Windows Phone 8.1. In July 2015, a port of the Pocket Edition to Windows 10 was released as the Windows 10 Edition, with full crossplay to other Pocket versions. In January 2017, Microsoft announced that it would no longer maintain the Windows Phone versions of Pocket Edition. On 20 September 2017, with the "Better Together Update", the Pocket Edition was ported to the Xbox One, and was renamed to the Bedrock Edition. The console versions of Minecraft debuted with the Xbox 360 edition, developed by 4J Studios and released on 9 May 2012. Announced as part of the Xbox Live Arcade NEXT promotion, this version introduced a redesigned crafting system, a new control interface, in-game tutorials, split-screen multiplayer, and online play via Xbox Live. Unlike the PC version, its worlds were finite, bordered by invisible walls. Initially, the Xbox 360 version resembled outdated PC versions but received updates to bring it closer to Java Edition before eventually being discontinued. The Xbox One version launched on 5 September 2014, featuring larger worlds and support for more players. Minecraft expanded to PlayStation platforms with PlayStation 3 and PlayStation 4 editions released on 17 December 2013 and 4 September 2014, respectively. Originally planned as a PS4 launch title, it was delayed before its eventual release. A PlayStation Vita version followed in October 2014. Like the Xbox versions, the PlayStation editions were developed by 4J Studios. Nintendo platforms received Minecraft: Wii U Edition on 17 December 2015, with a physical release in North America on 17 June 2016 and in Europe on 30 June. The Nintendo Switch version launched via the eShop on 11 May 2017. During a Nintendo Direct presentation on 13 September 2017, Nintendo announced that Minecraft: New Nintendo 3DS Edition, based on the Pocket Edition, would be available for download immediately after the livestream, and a physical copy available on a later date. The game is compatible only with the New Nintendo 3DS or New Nintendo 2DS XL systems and does not work with the original 3DS or 2DS systems. On 20 September 2017, the Better Together Update introduced Bedrock Edition across Xbox One, Windows 10, VR, and mobile platforms, enabling cross-play between these versions. Bedrock Edition later expanded to Nintendo Switch and PlayStation 4, with the latter receiving the update in December 2019, allowing cross-platform play for users with a free Xbox Live account. The Bedrock Edition released a native version for PlayStation 5 on 22 October 2024, while the Xbox Series X/S version launched on 17 June 2025. On 18 December 2018, the PlayStation 3, PlayStation Vita, Xbox 360, and Wii U versions of Minecraft received their final update and would later become known as "Legacy Console Editions". On 15 January 2019, the New Nintendo 3DS version of Minecraft received its final update, effectively becoming discontinued as well. An educational version of Minecraft, designed for use in schools, launched on 1 November 2016. It is available on Android, ChromeOS, iPadOS, iOS, MacOS, and Windows. On 20 August 2018, Mojang announced that it would bring Education Edition to iPadOS in Autumn 2018. It was released to the App Store on 6 September 2018. On 27 March 2019, it was announced that it would be operated by JD.com in China. On 26 June 2020, a public beta for the Education Edition was made available to Google Play Store compatible Chromebooks. The full game was released to the Google Play Store for Chromebooks on 7 August 2020. On 20 May 2016, China Edition (also known as My World) was announced as a localized edition for China, where it was released under a licensing agreement between NetEase and Mojang. The PC edition was released for public testing on 8 August 2017. The iOS version was released on 15 September 2017, and the Android version was released on 12 October 2017. The PC edition is based on the original Java Edition, while the iOS and Android mobile versions are based on the Bedrock Edition. The edition is free-to-play and had over 700 million registered accounts by September 2023. This version of Bedrock Edition is exclusive to Microsoft's Windows 10 and Windows 11 operating systems. The beta release for Windows 10 launched on the Windows Store on 29 July 2015. After nearly a year and a half in beta, Microsoft fully released the version on 19 December 2016. Called the "Ender Update", this release implemented new features to this version of Minecraft like world templates and add-on packs. On 7 June 2022, the Java and Bedrock Editions of Minecraft were merged into a single bundle for purchase on Windows; those who owned one version would automatically gain access to the other version. Both game versions would otherwise remain separate. Around 2011, prior to Minecraft's full release, Mojang collaborated with The Lego Group to create a Lego brick-based Minecraft game called Brickcraft. This would have modified the base Minecraft game to use Lego bricks, which meant adapting the basic 1×1 block to account for larger pieces typically used in Lego sets. Persson worked on an early version called "Project Rex Kwon Do", named after the character of the same name from the film Napoleon Dynamite. Although Lego approved the project and Mojang assigned two developers for six months, it was canceled due to the Lego Group's demands, according to Mojang's Daniel Kaplan. Lego considered buying Mojang to complete the game, but when Microsoft offered over $2 billion for the company, Lego stepped back, unsure of Minecraft's potential. On 26 June 2025, a build of Brickcraft dated 28 June 2012 was published on a community archive website Omniarchive. Initially, Markus Persson planned to support the Oculus Rift with a Minecraft port. However, after Facebook acquired Oculus in 2013, he abruptly canceled the plans, stating, "Facebook creeps me out." In 2016, a community-made mod, Minecraft VR, added VR support for Java Edition, followed by Vivecraft for HTC Vive. Later that year, Microsoft introduced official Oculus Rift support for Windows 10 Edition, leading to the discontinuation of the Minecraft VR mod due to trademark complaints. Vivecraft was endorsed by Minecraft VR contributors for its Rift support. Also available is a Gear VR version, titled Minecraft: Gear VR Edition. Windows Mixed Reality support was added in 2017. On 7 September 2020, Mojang Studios announced that the PlayStation 4 Bedrock version would receive PlayStation VR support later that month. In September 2024, the Minecraft team announced they would no longer support PlayStation VR, which received its final update in March 2025. Music and sound design Minecraft's music and sound effects were produced by German musician Daniel Rosenfeld, better known as C418. To create the sound effects for the game, Rosenfeld made extensive use of Foley techniques. On learning the processes for the game, he remarked, "Foley's an interesting thing, and I had to learn its subtleties. Early on, I wasn't that knowledgeable about it. It's a whole trial-and-error process. You just make a sound and eventually you go, 'Oh my God, that's it! Get the microphone!' There's no set way of doing anything at all." He reminisced on creating the in-game sound for grass blocks, stating "It turns out that to make grass sounds you don't actually walk on grass and record it, because grass sounds like nothing. What you want to do is get a VHS, break it apart, and just lightly touch the tape." According to Rosenfeld, his favorite sound to design for the game was the hisses of spiders. He elaborates, "I like the spiders. Recording that was a whole day of me researching what a spider sounds like. Turns out, there are spiders that make little screeching sounds, so I think I got this recording of a fire hose, put it in a sampler, and just pitched it around until it sounded like a weird spider was talking to you." Many of the sound design decisions by Rosenfeld were done accidentally or spontaneously. The creeper notably lacks any specific noises apart from a loud fuse-like sound when about to explode; Rosenfeld later recalled "That was just a complete accident by Markus and me [sic]. We just put in a placeholder sound of burning a matchstick. It seemed to work hilariously well, so we kept it." On other sounds, such as those of the zombie, Rosenfeld remarked, "I actually never wanted the zombies so scary. I intentionally made them sound comical. It's nice to hear that they work so well [...]." Rosenfeld remarked that the sound engine was "terrible" to work with, remembering "If you had two song files at once, it [the game engine] would actually crash. There were so many more weird glitches like that the guys never really fixed because they were too busy with the actual game and not the sound engine." The background music in Minecraft consists of instrumental ambient music. To compose the music of Minecraft, Rosenfeld used the package from Ableton Live, along with several additional plug-ins. Speaking on them, Rosenfeld said "They can be pretty much everything from an effect to an entire orchestra. Additionally, I've got some synthesizers that are attached to the computer. Like a Moog Voyager, Dave Smith Prophet 08 and a Virus TI." On 4 March 2011, Rosenfeld released a soundtrack titled Minecraft – Volume Alpha; it includes most of the tracks featured in Minecraft, as well as other music not featured in the game. Kirk Hamilton of Kotaku chose the music in Minecraft as one of the best video game soundtracks of 2011. On 9 November 2013, Rosenfeld released the second official soundtrack, titled Minecraft – Volume Beta, which included the music that was added in a 2013 "Music Update" for the game. A physical release of Volume Alpha, consisting of CDs, black vinyl, and limited-edition transparent green vinyl LPs, was issued by indie electronic label Ghostly International on 21 August 2015. On 14 August 2020, Ghostly released Volume Beta on CD and vinyl, with alternate color LPs and lenticular cover pressings released in limited quantities. The final update Rosenfeld worked on was 2018's 1.13 Update Aquatic. His music remained the only music in the game until 2020's "Nether Update", introducing pieces from Lena Raine. Since then, other composers have made contributions, including Kumi Tanioka, Samuel Åberg, Aaron Cherof, and Amos Roddy, with Raine remaining as the new primary composer. Ownership of all music besides Rosenfeld's independently released albums has been retained by Microsoft, with their label publishing all of the other artists' releases. Gareth Coker also composed some of the music for the game's mini games from the Legacy Console editions. Rosenfeld had stated his intent to create a third album of music for the game in a 2015 interview with Fact, and confirmed its existence in a 2017 tweet, stating that his work on the record as of then had tallied up to be longer than the previous two albums combined, which in total clocks in at over 3 hours and 18 minutes. However, due to licensing issues with Microsoft, the third volume has since not seen release. On 8 January 2021, Rosenfeld was asked in an interview with Anthony Fantano whether or not there was still a third volume of his music intended for release. Rosenfeld responded, saying, "I have something—I consider it finished—but things have become complicated, especially as Minecraft is now a big property, so I don't know." Reception Minecraft has received critical acclaim, with praise for the creative freedom it grants players in-game, as well as the ease of enabling emergent gameplay. Critics have expressed enjoyment in Minecraft's complex crafting system, commenting that it is an important aspect of the game's open-ended gameplay. Most publications were impressed by the game's "blocky" graphics, with IGN describing them as "instantly memorable". Reviewers also liked the game's adventure elements, noting that the game creates a good balance between exploring and building. The game's multiplayer feature has been generally received favorably, with IGN commenting that "adventuring is always better with friends". Jaz McDougall of PC Gamer said Minecraft is "intuitively interesting and contagiously fun, with an unparalleled scope for creativity and memorable experiences". It has been regarded as having introduced millions of children to the digital world, insofar as its basic game mechanics are logically analogous to computer commands. IGN was disappointed about the troublesome steps needed to set up multiplayer servers, calling it a "hassle". Critics also said that visual glitches occur periodically. Despite its release out of beta in 2011, GameSpot said the game had an "unfinished feel", adding that some game elements seem "incomplete or thrown together in haste". A review of the alpha version, by Scott Munro of the Daily Record, called it "already something special" and urged readers to buy it. Jim Rossignol of Rock Paper Shotgun also recommended the alpha of the game, calling it "a kind of generative 8-bit Lego Stalker". On 17 September 2010, gaming webcomic Penny Arcade began a series of comics and news posts about the addictiveness of the game. The Xbox 360 version was generally received positively by critics, but did not receive as much praise as the PC version. Although reviewers were disappointed by the lack of features such as mod support and content from the PC version, they acclaimed the port's addition of a tutorial and in-game tips and crafting recipes, saying that they make the game more user-friendly. The Xbox One Edition was one of the best received ports, being praised for its relatively large worlds. The PlayStation 3 Edition also received generally favorable reviews, being compared to the Xbox 360 Edition and praised for its well-adapted controls. The PlayStation 4 edition was the best received port to date, being praised for having 36 times larger worlds than the PlayStation 3 edition and described as nearly identical to the Xbox One edition. The PlayStation Vita Edition received generally positive reviews from critics but was noted for its technical limitations. The Wii U version received generally positive reviews from critics but was noted for a lack of GamePad integration. The 3DS version received mixed reviews, being criticized for its high price, technical issues, and lack of cross-platform play. The Nintendo Switch Edition received fairly positive reviews from critics, being praised, like other modern ports, for its relatively larger worlds. Minecraft: Pocket Edition initially received mixed reviews from critics. Although reviewers appreciated the game's intuitive controls, they were disappointed by the lack of content. The inability to collect resources and craft items, as well as the limited types of blocks and lack of hostile mobs, were especially criticized. After updates added more content, Pocket Edition started receiving more positive reviews. Reviewers complimented the controls and the graphics, but still noted a lack of content. Minecraft surpassed over a million purchases less than a month after entering its beta phase in early 2011. At the same time, the game had no publisher backing and has never been commercially advertised except through word of mouth, and various unpaid references in popular media such as the Penny Arcade webcomic. By April 2011, Persson estimated that Minecraft had made €23 million (US$33 million) in revenue, with 800,000 sales of the alpha version of the game, and over 1 million sales of the beta version. In November 2011, prior to the game's full release, Minecraft beta surpassed 16 million registered users and 4 million purchases. By March 2012, Minecraft had become the 6th best-selling PC game of all time. As of 10 October 2014[update], the game had sold 17 million copies on PC, becoming the best-selling PC game of all time. On 25 February 2014, the game reached 100 million registered users. By May 2019, 180 million copies had been sold across all platforms, making it the single best-selling video game of all time. The free-to-play Minecraft China version had over 700 million registered accounts by September 2023. By 2023, the game had sold over 300 million copies. As of April 2025, Minecraft has sold over 350 million copies. The Xbox 360 version of Minecraft became profitable within the first day of the game's release in 2012, when the game broke the Xbox Live sales records with 400,000 players online. Within a week of being on the Xbox Live Marketplace, Minecraft sold a million copies. GameSpot announced in December 2012 that Minecraft sold over 4.48 million copies since the game debuted on Xbox Live Arcade in May 2012. In 2012, Minecraft was the most purchased title on Xbox Live Arcade; it was also the fourth most played title on Xbox Live based on average unique users per day. As of 4 April 2014[update], the Xbox 360 version has sold 12 million copies. In addition, Minecraft: Pocket Edition has reached a figure of 21 million in sales. The PlayStation 3 Edition sold one million copies in five weeks. The release of the game's PlayStation Vita version boosted Minecraft sales by 79%, outselling both PS3 and PS4 debut releases and becoming the largest Minecraft launch on a PlayStation console. The PS Vita version sold 100,000 digital copies in Japan within the first two months of release, according to an announcement by SCE Japan Asia. By January 2015, 500,000 digital copies of Minecraft were sold in Japan across all PlayStation platforms, with a surge in primary school children purchasing the PS Vita version. As of 2022, the Vita version has sold over 1.65 million physical copies in Japan, making it the best-selling Vita game in the country. Minecraft helped improve Microsoft's total first-party revenue by $63 million for the 2015 second quarter. The game, including all of its versions, had over 112 million monthly active players by September 2019. On its 11th anniversary in May 2020, the company announced that Minecraft had reached over 200 million copies sold across platforms with over 126 million monthly active players. By April 2021, the number of active monthly users had climbed to 140 million. In July 2010, PC Gamer listed Minecraft as the fourth-best game to play at work. In December of that year, Good Game selected Minecraft as their choice for Best Downloadable Game of 2010, Gamasutra named it the eighth best game of the year as well as the eighth best indie game of the year, and Rock, Paper, Shotgun named it the "game of the year". Indie DB awarded the game the 2010 Indie of the Year award as chosen by voters, in addition to two out of five Editor's Choice awards for Most Innovative and Best Singleplayer Indie. It was also awarded Game of the Year by PC Gamer UK. The game was nominated for the Seumas McNally Grand Prize, Technical Excellence, and Excellence in Design awards at the March 2011 Independent Games Festival and won the Grand Prize and the community-voted Audience Award. At Game Developers Choice Awards 2011, Minecraft won awards in the categories for Best Debut Game, Best Downloadable Game and Innovation Award, winning every award for which it was nominated. It also won GameCity's video game arts award. On 5 May 2011, Minecraft was selected as one of the 80 games that would be displayed at the Smithsonian American Art Museum as part of The Art of Video Games exhibit that opened on 16 March 2012. At the 2011 Spike Video Game Awards, Minecraft won the award for Best Independent Game and was nominated in the Best PC Game category. In 2012, at the British Academy Video Games Awards, Minecraft was nominated in the GAME Award of 2011 category and Persson received The Special Award. In 2012, Minecraft XBLA was awarded a Golden Joystick Award in the Best Downloadable Game category, and a TIGA Games Industry Award in the Best Arcade Game category. In 2013, it was nominated as the family game of the year at the British Academy Video Games Awards. During the 16th Annual D.I.C.E. Awards, the Academy of Interactive Arts & Sciences nominated the Xbox 360 version of Minecraft for "Strategy/Simulation Game of the Year". Minecraft Console Edition won the award for TIGA Game Of The Year in 2014. In 2015, the game placed 6th on USgamer's The 15 Best Games Since 2000 list. In 2016, Minecraft placed 6th on Time's The 50 Best Video Games of All Time list. Minecraft was nominated for the 2013 Kids' Choice Awards for Favorite App, but lost to Temple Run. It was nominated for the 2014 Kids' Choice Awards for Favorite Video Game, but lost to Just Dance 2014. The game later won the award for the Most Addicting Game at the 2015 Kids' Choice Awards. In addition, the Java Edition was nominated for "Favorite Video Game" at the 2018 Kids' Choice Awards, while the game itself won the "Still Playing" award at the 2019 Golden Joystick Awards, as well as the "Favorite Video Game" award at the 2020 Kids' Choice Awards. Minecraft also won "Stream Game of the Year" at inaugural Streamer Awards in 2021. The game later garnered a Nickelodeon Kids' Choice Award nomination for Favorite Video Game in 2021, and won the same category in 2022 and 2023. At the Golden Joystick Awards 2025, it won the Still Playing Award - PC and Console. Minecraft has been subject to several notable controversies. In June 2014, Mojang announced that it would begin enforcing the portion of Minecraft's end-user license agreement (EULA) which prohibits servers from giving in-game advantages to players in exchange for donations or payments. Spokesperson Owen Hill stated that servers could still require players to pay a fee to access the server and could sell in-game cosmetic items. The change was supported by Persson, citing emails he received from parents of children who had spent hundreds of dollars on servers. The Minecraft community and server owners protested, arguing that the EULA's terms were more broad than Mojang was claiming, that the crackdown would force smaller servers to shut down for financial reasons, and that Mojang was suppressing competition for its own Minecraft Realms subscription service. The controversy contributed to Notch's decision to sell Mojang. In 2020, Mojang announced an eventual change to the Java Edition to require a login from a Microsoft account rather than a Mojang account, the latter of which would be sunsetted. This also required Java Edition players to create Xbox network Gamertags. Mojang defended the move to Microsoft accounts by saying that improved security could be offered, including two-factor authentication, blocking cyberbullies in chat, and improved parental controls. The community responded with intense backlash, citing various technical difficulties encountered in the process and how account migration would be mandatory, even for those who do not play on servers. As of 10 March 2022, Microsoft required that all players migrate in order to maintain access the Java Edition of Minecraft. Mojang announced a deadline of 19 September 2023 for account migration, after which all legacy Mojang accounts became inaccessible and unable to be migrated. In June 2022, Mojang added a player-reporting feature in Java Edition. Players could report other players on multiplayer servers for sending messages prohibited by the Xbox Live Code of Conduct; report categories included profane language,[l] substance abuse, hate speech, threats of violence, and nudity. If a player was found to be in violation of Xbox Community Standards, they would be banned from all servers for a specific period of time or permanently. The update containing the report feature (1.19.1) was released on 27 July 2022. Mojang received substantial backlash and protest from community members, one of the most common complaints being that banned players would be forbidden from joining any server, even private ones. Others took issue to what they saw as Microsoft increasing control over its player base and exercising censorship, leading some to start a hashtag #saveminecraft and dub the version "1.19.84", a reference to the dystopian novel Nineteen Eighty-Four. The "Mob Vote" was an online event organized by Mojang in which the Minecraft community voted between three original mob concepts; initially, the winning mob was to be implemented in a future update, while the losing mobs were scrapped, though after the first mob vote this was changed, and losing mobs would now have a chance to come to the game in the future. The first Mob Vote was held during Minecon Earth 2017 and became an annual event starting with Minecraft Live 2020. The Mob Vote was often criticized for forcing players to choose one mob instead of implementing all three, causing divisions and flaming within the community, and potentially allowing internet bots and Minecraft content creators with large fanbases to conduct vote brigading. The Mob Vote was also blamed for a perceived lack of new content added to Minecraft since Microsoft's acquisition of Mojang in 2014. The 2023 Mob Vote featured three passive mobs—the crab, the penguin, and the armadillo—with voting scheduled to start on 13 October. In response, a Change.org petition was created on 6 October, demanding that Mojang eliminate the Mob Vote and instead implement all three mobs going forward. The petition received approximately 445,000 signatures by 13 October and was joined by calls to boycott the Mob Vote, as well as a partially tongue-in-cheek "revolutionary" propaganda campaign in which sympathizers created anti-Mojang and pro-boycott posters in the vein of real 20th century propaganda posters. Mojang did not release an official response to the boycott, and the Mob Vote otherwise proceeded normally, with the armadillo winning the vote. In September 2024, as part of a blog post detailing their future plans for Minecraft's development, Mojang announced the Mob Vote would be retired. Cultural impact In September 2019, The Guardian classified Minecraft as the best video game of the 21st century to date, and in November 2019, Polygon called it the "most important game of the decade" in its 2010s "decade in review". In June 2020, Minecraft was inducted into the World Video Game Hall of Fame. Minecraft is recognized as one of the first successful games to use an early access model to draw in sales prior to its full release version to help fund development. As Minecraft helped to bolster indie game development in the early 2010s, it also helped to popularize the use of the early access model in indie game development. Social media sites such as YouTube, Facebook, and Reddit have played a significant role in popularizing Minecraft. Research conducted by the Annenberg School for Communication at the University of Pennsylvania showed that one-third of Minecraft players learned about the game via Internet videos. In 2010, Minecraft-related videos began to gain influence on YouTube, often made by commentators. The videos usually contain screen-capture footage of the game and voice-overs. Common coverage in the videos includes creations made by players, walkthroughs of various tasks, and parodies of works in popular culture. By May 2012, over four million Minecraft-related YouTube videos had been uploaded. The game would go on to be a prominent fixture within YouTube's gaming scene during the entire 2010s; in 2014, it was the second-most searched term on the entire platform. By 2018, it was still YouTube's biggest game globally. Some popular commentators have received employment at Machinima, a now-defunct gaming video company that owned a highly watched entertainment channel on YouTube. The Yogscast is a British company that regularly produces Minecraft videos; their YouTube channel has attained billions of views, and their panel at Minecon 2011 had the highest attendance. Another well-known YouTube personality is Jordan Maron, known online as CaptainSparklez, who has also created many Minecraft music parodies, including "Revenge", a parody of Usher's "DJ Got Us Fallin' in Love". Minecraft's popularity on YouTube was described by Polygon as quietly dominant, although in 2019, thanks in part to PewDiePie's playthroughs of the game, Minecraft experienced a visible uptick in popularity on the platform. Longer-running series include Far Lands or Bust, dedicated to reaching the obsolete "Far Lands" glitch by foot on an older version of the game. YouTube announced that on 14 December 2021 that the total amount of Minecraft-related views on the website had exceeded one trillion. Minecraft has been referenced by other video games, such as Torchlight II, Team Fortress 2, Borderlands 2, Choplifter HD, Super Meat Boy, The Elder Scrolls V: Skyrim, The Binding of Isaac, The Stanley Parable, and FTL: Faster Than Light. Minecraft is officially represented in downloadable content for the crossover fighter Super Smash Bros. Ultimate, with Steve as a playable character with a moveset including references to building, crafting, and redstone, alongside an Overworld-themed stage. It was also referenced by electronic music artist Deadmau5 in his performances. The game is also referenced heavily in "Informative Murder Porn", the second episode of the seventeenth season of the animated television series South Park. In 2025, A Minecraft Movie was released. It made $313 million in the box office in the first week, a record-breaking opening for a video game adaptation. Minecraft has been noted as a cultural touchstone for Generation Z, as many of the generation's members played the game at a young age. The possible applications of Minecraft have been discussed extensively, especially in the fields of computer-aided design (CAD) and education. In a panel at Minecon 2011, a Swedish developer discussed the possibility of using the game to redesign public buildings and parks, stating that rendering using Minecraft was much more user-friendly for the community, making it easier to envision the functionality of new buildings and parks. In 2012, a member of the Human Dynamics group at the MIT Media Lab, Cody Sumter, said: "Notch hasn't just built a game. He's tricked 40 million people into learning to use a CAD program." Various software has been developed to allow virtual designs to be printed using professional 3D printers or personal printers such as MakerBot and RepRap. In September 2012, Mojang began the Block by Block project in cooperation with UN Habitat to create real-world environments in Minecraft. The project allows young people who live in those environments to participate in designing the changes they would like to see. Using Minecraft, the community has helped reconstruct the areas of concern, and citizens are invited to enter the Minecraft servers and modify their own neighborhood. Carl Manneh, Mojang's managing director, called the game "the perfect tool to facilitate this process", adding "The three-year partnership will support UN-Habitat's Sustainable Urban Development Network to upgrade 300 public spaces by 2016." Mojang signed Minecraft building community, FyreUK, to help render the environments into Minecraft. The first pilot project began in Kibera, one of Nairobi's informal settlements and is in the planning phase. The Block by Block project is based on an earlier initiative started in October 2011, Mina Kvarter (My Block), which gave young people in Swedish communities a tool to visualize how they wanted to change their part of town. According to Manneh, the project was a helpful way to visualize urban planning ideas without necessarily having a training in architecture. The ideas presented by the citizens were a template for political decisions. In April 2014, the Danish Geodata Agency generated all of Denmark in fullscale in Minecraft based on their own geodata. This is possible because Denmark is one of the flattest countries with the highest point at 171 meters (ranking as the country with the 30th smallest elevation span), where the limit in default Minecraft was around 192 meters above in-game sea level when the project was completed. Taking advantage of the game's accessibility where other websites are censored, the non-governmental organization Reporters Without Borders has used an open Minecraft server to create the Uncensored Library, a repository within the game of journalism by authors from countries (including Egypt, Mexico, Russia, Saudi Arabia and Vietnam) who have been censored and arrested, such as Jamal Khashoggi. The neoclassical virtual building was created over about 250 hours by an international team of 24 people. Despite its unpredictable nature, Minecraft speedrunning, where players time themselves from spawning into a new world to reaching The End and defeating the Ender Dragon boss, is popular. Some speedrunners use a combination of mods, external programs, and debug menus, while other runners play the game in a more vanilla or more consistency-oriented way. Minecraft has been used in educational settings through initiatives such as MinecraftEdu, founded in 2011 to make the game affordable and accessible for schools in collaboration with Mojang. MinecraftEdu provided features allowing teachers to monitor student progress, including screenshot submissions as evidence of lesson completion, and by 2012 reported that approximately 250,000 students worldwide had access to the platform. Mojang also developed Minecraft: Education Edition with pre-built lesson plans for up to 30 students in a closed environment. Educators have used Minecraft to teach subjects such as history, language arts, and science through custom-built environments, including reconstructions of historical landmarks and large-scale models of biological structures such as animal cells. The introduction of redstone blocks enabled the construction of functional virtual machines such as a hard drive and an 8-bit computer. Mods have been created to use these mechanics for teaching programming. In 2014, the British Museum announced a project to reproduce its building and exhibits in Minecraft in collaboration with the public. Microsoft and Code.org have offered Minecraft-based tutorials and activities designed to teach programming, reporting by 2018 that more than 85 million children had used their resources. In 2025, the Musée de Minéralogie in Paris held a temporary exhibition titled "Minerals in Minecraft." Following the initial surge in popularity of Minecraft in 2010, other video games were criticised for having various similarities to Minecraft, and some were described as being "clones", often due to a direct inspiration from Minecraft, or a superficial similarity. Examples include Ace of Spades, CastleMiner, CraftWorld, FortressCraft, Terraria, BlockWorld 3D, Total Miner, and Luanti (formerly Minetest). David Frampton, designer of The Blockheads, reported that one failure of his 2D game was the "low resolution pixel art" that too closely resembled the art in Minecraft, which resulted in "some resistance" from fans. A homebrew adaptation of the alpha version of Minecraft for the Nintendo DS, titled DScraft, has been released; it has been noted for its similarity to the original game considering the technical limitations of the system. In response to Microsoft's acquisition of Mojang and their Minecraft IP, various developers announced further clone titles developed specifically for Nintendo's consoles, as they were the only major platforms not to officially receive Minecraft at the time. These clone titles include UCraft (Nexis Games), Cube Life: Island Survival (Cypronia), Discovery (Noowanda), Battleminer (Wobbly Tooth Games), Cube Creator 3D (Big John Games), and Stone Shire (Finger Gun Games). Despite this, the fears of fans were unfounded, with official Minecraft releases on Nintendo consoles eventually resuming. Markus Persson made another similar game, Minicraft, for a Ludum Dare competition in 2011. In 2025, Persson announced through a poll on his X account that he was considering developing a spiritual successor to Minecraft. He later clarified that he was "100% serious", and that he had "basically announced Minecraft 2". Within days, however, Persson cancelled the plans after speaking to his team. In November 2024, artificial intelligence companies Decart and Etched released Oasis, an artificially generated version of Minecraft, as a proof of concept. Every in-game element is completely AI-generated in real time and the model does not store world data, leading to "hallucinations" such as items and blocks appearing that were not there before. In January 2026, indie game developer Unomelon announced that their voxel sandbox game Allumeria would be playable in Steam Next Fest that year. On 10 February, Mojang issued a DMCA takedown of Allumeria on Steam through Valve, alleging the game was infringing on Minecraft's copyright. Some reports suggested that the takedown may have used an automatic AI copyright claiming service. The DMCA was later withdrawn. Minecon was an annual official fan convention dedicated to Minecraft. The first full Minecon was held in November 2011 at the Mandalay Bay Hotel and Casino in Las Vegas. The event included the official launch of Minecraft; keynote speeches, including one by Persson; building and costume contests; Minecraft-themed breakout classes; exhibits by leading gaming and Minecraft-related companies; commemorative merchandise; and autograph and picture times with Mojang employees and well-known contributors from the Minecraft community. In 2016, Minecon was held in-person for the last time, with the following years featuring annual "Minecon Earth" livestreams on minecraft.net and YouTube instead. These livestreams, later rebranded to "Minecraft Live", included the mob/biome votes, and announcements of new game updates. In 2025, "Minecraft Live" became a biannual event as part of Minecraft's changing update schedule.[citation needed] Notes References External links |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-07-04-updated-masks-evidence.html] | [TOKENS: 5276] |
Masks for COVID: Updating the evidence Jeremy Howard July 4, 2022 On this page These are notes I took whilst preparing a paper on mask efficacy from Nov 2021 to Jan 2022. In the end, I gave up on the paper, because I felt like people had given up on masks, so there wasn’t much point in finishing it. I’ve decided to publish these notes in the hope some people will find them a useful starting point for their own research, and since I’ve noticed some signs in recent weeks that people might be open to avoiding COVID again. My previous paper on this topic, in which I led a team of 19 experts, was written in April 2020, and published here in the Proceedings of the National Academy of Science. The rise of better masks In the US, 400 million N95 masks are being distributed for free, coming from the 750 million stored in the US’ Strategic National Stockpile. A similar campaign to distribute 650 millions masks in the US in 2020 was cancelled. KN95 masks are being given to US congressional staff, and masks are required for federal workers and whilst in federal buildings. The Los Angeles school district has required students to upgrade from cloth masks to “well-fitting, non-cloth masks with a nose wire”. Masks work A review paper discussed both lab evidence and empirical evidence for the importance of face masks, with eight “seminal studies” showing a reduction in transmission when masks are used, and one Danish study of surgical masks with “several design limitations” which “demonstrated only a modest benefit in limiting COVID-19 transmission”. The authors note that “laboratory studies have demonstrated the ability of surgical masks to block SARS-COV-2 and other viruses”, with the masks “60%–70% effective at protecting others and 50% effective at protecting the wearer”. An evidence review from early in the pandemic concluded that “given the current shortages of medical masks, we recommend the adoption of public cloth mask wearing, as an effective form of source control”. It noted that “by the end of June 2020, nearly 90% of the global population lived in regions that had nearly universal mask use, or had laws requiring mask use in some public locations.” The review said that “There has been one controlled trial of mask use for influenza control in the general community. The study looked at Australian households, was not done during a pandemic, and was done without any enforcement of compliance” – and yet still found “masks as a group had protective efficacy in excess of 80% against clinical influenza-like illness.” An observational study of Beijing households analyzed the impact of mask use in the community on COVID-19 transmission, finding that face masks were 79% effective in preventing transmission, if used by all household members prior to symptoms occurring. One study used a multiple regression of policy interventions and country and population characteristics to infer the relationship between mask use and SARS-CoV-2 transmission. It found that transmission was around 7.5 times higher in countries that did not have a mask mandate or universal mask use, a result similar to that found in an analogous study of fewer countries. Similar results were found by numerous other papers. A mathematical model of mask use estimates that mask wearing reduces the reproduction number R by (1−mp)^2, where m is the efficacy of trapping viral particles inside the mask, and p is the percentage of the population that wears masks. A report in Nature explained that researchers running a randomized controlled trial (RCT) of community mask use in Bangladesh “began by developing a strategy to promote mask wearing, with measures such as reminders from health workers in public places. This ultimately tripled mask usage, from only 13% in control villages to 42% in villages where it was encouraged”, and “then compared numbers of COVID-19 cases in control villages and the treatment communities”. They found that the number of infections in mask wearing communities decreased, with a reduction of COVID symptoms using surgical masks to 0.87 times the incidence in unmasked communities, and 0.91 times when using cloth masks. The report noted that “the researchers suggest that the true risk reduction is probably much greater, in part because they did no SARS-CoV-2 testing of people without symptoms or whose symptoms did not meet the World Health Organization’s definition of the disease.” The researchers concluded that “promoting community mask-wearing can improve public health”. The Johns Hopkins School of Public Health reviewed the work and concluded that “This study is the largest and best-designed randomized controlled trial to date of a realistic non-pharmaceutical intervention on SARS-CoV-2 transmission.” A paper investigating an upper bound on one-to-one exposure to infectious human respiratory particles concludes that “face masks significantly reduce the risk of SARS-CoV-2 infection compared to social distancing. We find a very low risk of infection when everyone wears a face mask, even if it doesn’t fit perfectly on the face.” They calculate that “social distancing alone, even at 3.0 m between two speaking individuals, leads to an upper bound of 90% for risk of infection after a few minutes”, but that when both source and susceptible wear a well-fitting FFP2 mask, there is only 0.4% after one hour of contact. They found that to achieve good fit it is important to mold the nose piece wire to the size of the nose, rather than leaving it in a sharp folded position. A similar study “quantifies the extent to which transmission risk is reduced in large rooms with high air exchange rates, increased for more vigorous respiratory activities, and dramatically reduced by the use of face masks.” The authors describe the six-foot rule widely used to ensure social distancing as “a guideline that offers little protection from pathogen-bearing aerosol droplets sufficiently small to be continuously mixed through an indoor space.” Instead, they develop a safety guideline based on cumulative exposure time,” the product of the number of occupants and their time in an enclosed space. In particular, they identify that the greatest risk comes in places where people are speaking (other than quietly) or singing, and that “the benefit of face masks is immediately apparent”, due to the multiplicative effect when both source and susceptible wear a mask. They further note that “Air filtration has a less dramatic effect than face mask use in increasing the CET bound. Nevertheless, it does offer a means of mitigating indoor transmission with greater comfort, albeit at greater cost.” Another study of the combined impacts of ventilation and mask effective filtration efficiency in classroom settings found that “ventilation alone is not able to achieve probabilities <0.01 (1%)” of transmitting COVID in a classroom. However, they found that good masks reduce infection probability by >5× in some cases, and that “reductions provided by ventilation and masks are synergistic and multiplicative”. However they also noted that “most masks fit poorly”, recommending that work be done to ensure that high quality masks are used. Similar results were found in a study of community public health interventions, which concluded that “control the pandemic, our models suggest that high adherence to social distance is necessary to curb the spread of the disease, and that wearing face masks provides optimal protection even if only a small portion of the population comply with social distance”. Guidance from the independent scientific advisory group OzSAGE points out “that school children are able to wear masks. As an example, all children over two years of age in San Francisco are required to wear masks at school”. Omicron changes the game An analysis of fine aerosol emissions found that, compared to the original wild type (WT) virus: “Delta and Omicron both also have increased transmissibility: the number of cells infected for a given number of ribonucleic acid (RNA) virus copies was found to be doubled and quadrupled respectively. Furthermore, Omicron also seems to be better at evading the immune system. This implies that the critical dose of virus copies above which a situation is potentially infectious needs to be lowered. For the WT, we had proposed a critical dose of 500 virus copies. If the above-mentioned capacity to infect cells translates into an infection risk, this would imply a critical dose of around 300 virus copies for Delta and around 100 virus copies for Omicron.” The study finds that “surgical masks are no longer sufficient in most public settings, while correctly fitted FFP2 respirators still provide sufficient protection, except in high aerosol producing situations such as singing or shouting.” Data from Hong Kong shows that “Omicron SARS-CoV-2 infects and multiplies 70 times faster than the Delta variant and original SARS-CoV-2 in human bronchus”. A study of transmission in Danish households estimated the secondary attack rate (SAR) of omicron compared to delta, finding it 1.2 times higher for unvaccinated people, 2.6 times higher for double-dosed, and 3.7 times higher for boosted. The authors conclude that “the rapid spread of the Omicron VOC primarily can be ascribed to the immune evasiveness”. According to UK statistics, the risk of hospitalization from omicron when unvaccinated is about the same as the wildtype virus, which is about half the risk of the delta variant. The journal Infection Control Today reported that many experts are concerned that “‘Omicron the Pandemic Killer’ Idea Ignores Dangers of Long COVID”: “Linda Spaulding, RN-BC, CIC, CHEC, CHOP, a member of Infection Control Today®’s Editorial Advisory Board (EAB), says that she’s “seen athletes in their 20s on the wait list for double lung transplants because of long COVID. That’s something that has long-term consequences. Some people talk of COVID fog. They just can’t put their thoughts together.” In addition, even the treatments for those with long COVID can put toll on a patient’s body.” “As noted by Kevin Kavanagh, MD, another member of ICT®’s EAB, a core difficulty in society’s attempt to guide COVID-19 from pandemic to endemic is that COVID is not just a respiratory virus. Kavanagh wrote in October that SARS-CoV-2 is similar to HIV because it can “silently spread throughout the host’s body and attack almost every organ.”” Better masks work better The US Centers for Disease Control and Prevention (CDC) explains that: “Loosely woven cloth products provide the least protection, layered finely woven products offer more protection, well-fitting disposable surgical masks and KN95s offer even more protection, and well-fitting NIOSH-approved respirators (including N95s) offer the highest level of protection.” Unfortunately “well-fitting disposable surgical masks” do not exist out of the box, since there are large gaps on each side of the mask. Surgical masks require modifications to achive a good fit. That’s because they are made to stop liquid splashes during surgery, rather than made to stop airborne transmission. There are two methods shown by the CDC to improve fit: Research shows that both of these approaches dramatically reduce exposure to aerosols emitted during a period of breathing: “…adding a cloth mask over the source headform’s medical procedure mask or knotting and tucking the medical procedure mask reduced the cumulative exposure of the unmasked receiver by 82.2% (SD = 0.16) and 62.9% (SD = 0.08), respectively (Figure 2). When the source was unmasked and the receiver was fitted with the double mask or the knotted and tucked medical procedure mask, the receiver’s cumulative exposure was reduced by 83.0% (SD = 0.15) and 64.5% (SD = 0.03), respectively. When the source and receiver were both fitted with double masks or knotted and tucked masks, the cumulative exposure of the receiver was reduced 96.4% (SD = 0.02) and 95.9% (SD = 0.02), respectively.” An airborne transmission simulator was used to estimate the ability of various types of face masks to block COVID-19 transmission. In this experiment, “cotton mask led to an approximately 20% to 40% reduction in virus uptake compared to no mask. The N95 mask had the highest protective efficacy (approximately 80% to 90% reduction)”. All of the masks were much more effective at source control than at protecting the wearer, with the N95 stopping all detectable transmission. The American Conference of Governmental Industrial Hygienists (ACGIH) say that “workers need respirators”, noting that a worker with an “N95 filtering facepiece respirator… has 1-10% inward leakage and outward leakage”, but with a surgical mask “has 50% inward leakage and outward leakage”, and with a cloth face covering “has 75% inward leakage and outward leakage”. They explain that “N95 FFRs have an assigned protection factor of 10 (10% inward leakage) but must receive a fit factor of 100 (1% inward leakage) on an individual worker.” ACGIH created a table showing how, if we start with an assumption that it takes on average 15 minutes to get infected if no-one is wearing a mask (based on CDC contact tracing premises), we can calculate the time it would take on average to get infected if one or both of source and receiver are wearing various types of mask. This is calculated by simply dividing the base time of 15 minutes by the leakage factor for the source’s mask (if any), and then dividing that by the leakage factor for the receiver’s mask (if any). This approach is, however, an over-simplification. Reseach based on a a single-hit model of infection shows that the probability of infection “shows a highly nonlinear sensitivity” to inhaled virus number. Therefore, “In a virus-rich regime… wearing a mask may not suffice to prevent infection.” Research undertaken by the National Personal Protective Technology Laboratory (NPPTL) found that respirators with an exhalation valve “reduce particle emissions to levels similar to or better than those provided by surgical masks, procedure masks, or cloth face coverings”. Furthermore, “surgical tape secured over the valve from the inside of the FFR can provide source control similar to that of an FFR with no exhalation valve”. Pushing back against masks Professor Alison McMillan, Commonwealth Chief Nursing and Midwifery Officer in Australia claims that “there is no evidence to suggest that we should be moving towards… N95 respirators in the community setting.” She added “I am aware that there are some publications out there suggesting a move to N95 (masks). But that’s not supported in the empirical evidence”. According to Norman Swan, host of the ABC’s Coronacast, “If you’re wearing an N95 that hasn’t been fit tested – and it’s not an easy process to do yourself at home – there’s no guarantee that it’s an awful lot more effective than wearing a surgical mask. Professor Catherine Bennett, chair in epidemiology at Deakin University, claims that”Technically, the instructions say you shouldn’t reuse” respirators, and that “If you’re not particularly checking its fit, you’re probably wasting your time”. Occupational environment physician Malcolm Sim agrees: “If you put it [an N95 mask] off and put it on, they’re not meant for that purpose… They’re easily damaged in somebody’s handbag,” adding that the integrity of the masks can be compromised. He says that “If you’re handling them a lot, taking them on and off, there’s much more potential for you to get it [the virus] on your hands, your face, different parts of your body.” University of New South Wales epidemiologist Mary-Louise McLaws claimed that “There’s no evidence yet that a N95 mask will protect you more than a surgical mask for Omicron.” An opinion piece in Newsweek claims that “the effectiveness of respirators is vastly overestimated, and there is scant evidence that they stop community transmission. Moreover, NIOSH-approved respirators are tight, uncomfortable, and can impede breathing.” The article further claims that “For respirators to work, they must be well fitting, must be tested by OSHA, and must be used for only short time windows as their effectiveness diminishes as they get wet from breathing.” Recently there has been particular pushback against the use of masks by children, with the Newsweek article alleging that “Respirators are not necessary to protect children from COVID-19 because of the astoundingly low risk COVID-19 presents to them”, and that in fact wearing masks involves “existing well-documented harms”. There hasn’t been any documented harms to children from wearing masks, Respirators can be reused According to mask manufacturer 3M, respirators (which they refer to as “Filtering Facepiece Respirators (FFRs)”) “can be used many times.” They say that “There is no time limit to wearing an FFR. Respirators can be worn until they are dirty, damaged or difficult to breathe through.” In reporting from CNN, Linsey Marr, a professor of civil and environmental engineering at Virginia Tech, explained that an N95 mask’s material and filtration ability aren’t “going to degrade unless you physically rub it or poke holes in it.”You’d have to be in really polluted air … for several days before it lost its ability to filter out particles. So, you can really wear them for a long time. People have been talking about 40 hours – I think that’s fine. Really, it’s going to get gross from your face or the straps will get too loose or maybe break before you’re going to lose filtration ability… One of the first indicators of being able to change it if it looks nice and clean is that it just feels a little harder to breathe through. There appears to be more resistance with every breath.” She also noted that the contamination risk in reusing N95 masks is “lower, much lower, than the risk of you not wearing an N95 and breathing in particles”. The CDC has prepared guidelines for optimizing the supply of respirators which recommend reusing respirators at most five times. This guidelines were created for people “implementing policies and procedures for preventing pathogen transmission in healthcare settings”. They have been widely shared, incorrectly, by reporters as being recommendations for community use. The inventor of N95 mask material, Peter Tsai, says that “N95 masks can be rotated, 1 mask every 3–4 days”, and that in doing this “there is no change in the mask’s properties.” According to the NIOSH Guide to the Selection and Use of Particulate Respirators N95 respirators must maintain at least 95% filtration after a total mass loading of 200mg. This is designed to ensure they continue to work in sites with high particulate matter, such as some construction environment. However in normal use, even outside in a city with high levels of population, it would take over 200 days of 24 hour per day use to get to this level. The guide says that “generally, the use and reuse of N-series lters would also be subject only to considerations of hygiene, damage, and increased breathing resistance”. The NIOSH guidelines are well supported by research. Fit tests are not required for respirators to be effective In one study non-experts were asked to read the instructions that come with a respirator, and then to don the respirator without assistance and complete a fit test. The average fit factor achieved was 88, and the lowest fit factor of the subjects was 15, with nearly half achieving a fit factor greater than 100. Surgical masks have been found to have a much poorer fit in practice. One study showed that for surgical masks “quantitative fit factors ranged from 2.5 to 9.6”, and another found an average fit factor of 3.0. Guidance from the US Food and Drug Administration (FDA) explains that: “Fit Factor is a means of expressing the difference in particle concentration inside the mask and outside the mask during use. For example, a fit factor of 2 means that the concentration of particles within the mask is ½ or 50% of the concentration outside the mask; a fit factor of 5 means the concentration of particles within the mask is 1/5 th or 20% of the concentration outside the mask.” The guidance says that failing to achieve a fit factor of 2 “may suggest that respirator fit will not be sufficient to assure that the device will help reduce wearer exposure to pathogenic biological airborne particulates.” An analysis of the fitted filtration efficiency (FFE) of surgical masks found that, unmodified, they only achieved an FFE of 38.5%. The “knot and tuck” technique improved that to 60.3%, and a DIY mask fitter consisting of three rubber bands increased it to 78.2%. A 3-layer cotton mask had an FFE of just 26.5%. An N95, on the other hand, achieved an FFE of 98.4%. Furthermore, the N95 FFE had a standard deviation of only 0.5% — that is, it was effective for multiple tests during “a series of repeated movements of the torso, head, and facial muscles”. Interestingly, a 2-layer nylon mask had an FFE of 79.0% (standard devatiation 4.3%), showing that some cloth masks can be quite effective. These findings were replicated in a study of numerous types of cloth mask, which found that hybrids of 600 TPI cotton with silk, chiffon, or polypropelene achieved 72-96% filtration efficiency. Researchers have calculated that “the particle size most likely to deposit in the respiratory tract when wearing a mask is ∼2μm”. Unfortunately, this particle size is not considered in N95 or similar standards. Instead, 0.3 μm particles are used. A 2010 study of fit testing respirators for public health medical emergencies found that 98% of non-experts wearing masks without training achieved a fit factor of over 5 (20% leakage) and 75% of them achieved a fit factor of over 10 (10% leakage). Donning and doffing masks is not complex or risky Analysis by the CDC concludes that the risk of infection through surfaces (fomites) “is generally considered to be low”, a view that was supported by the evidence as early as July 2020. An analysis of “418 samples from mask fronts, cell phones, paper money, card machines, sewage, air and bedding” during a COVID surge “did not detect any trace of SARS-CoV-2 in all samples analyzed”. We should not reserve respirators for healthcare workers According to Anne Miller, executive director of Project N95, there are many U.S. manufacturers of N95 masks and an ample supply. The Economist reported that in Europe “at the start of the pandemic, FFP2 masks were scarce and costly. Even governments fell victim to price gouging, paying more than €4 ($4.50) per mask. Demand had previously been low, so stockpiles and production capacity could not satisfy the sudden surge. Governments wanted to reserve supplies for those most at risk of contracting the virus, such as health-care workers.” However they reported that by the end of 2021 “FFP2 masks are in healthy supply, and as the highly transmissible Omicron variant spreads across the world, updating guidance to recommend their wider use could be one way to help reduce transmission.” In the first 6 months of 2020, over 70,000 new face mask companies were registered in China, many run by people with no previous experience and no registration or licensing. The Chinese government stepped in to make licensing more stringent, shutting down many companies, and international demand fell over quality concerns. Due to “a dramatic reduction in demand for N95s”, US mask factories are closing. In June 2021 the American Mask Manufacturer’s Association said that “we have 28 members who are going to go out of business in the next 60 to 90 days.” By July 2021 they estimated “that 5,000 workers have been laid off across its member companies”. However following school mask mandates and demand during the omicron surge, demand in the US spiked in early 2022. The CDC has found that 60% of KN95s are counterfeit. In Australia it has been reported that “general practitioners have been left without highly protective N95 masks as consumers rush to stock up after a sharp rise in COVID-19 cases.” In May 2021 the CDC stated that “The supply and availability of NIOSH-approved respirators have increased significantly over the last several months. Healthcare facilities should not be using crisis capacity strategies at this time and should promptly resume conventional practices.” Demand distortions can increase as we proceed up the supply chain, creating inefficiencies for upstream firms. This is known as the Bullwhip Effect. Respirators need not be uncomfortable In an analysis of the physiological impact of the N95 filtering facepiece respirator (FFR) “in healthy healthcare workers, FFR did not impose any important physiological burden during 1 hour of use, at realistic clinical work rates”. A study of KF80, KF94, KF99, N95, and N99 masks found that self-reported comfort levels were nearly perfectly correlated with the ease of inhalation. |
======================================== |
[SOURCE: https://en.wikipedia.org/wiki/Video_game_publisher] | [TOKENS: 933] |
Contents Video game publisher A video game publisher is a company that publishes video games that have been developed either internally by the publisher or externally by a video game developer. They often finance the development, sometimes by paying a video game developer (the publisher calls this external development) and sometimes by paying an internal staff of developers called a studio. The large video game publishers also distribute the games they publish, while some smaller publishers instead hire distribution companies (or larger video game publishers) to distribute the games they publish. Other functions usually performed by the publisher include deciding on and paying for any licenses that are used by the game; paying for localization; layout, printing, and possibly the writing of the user manual; and the creation of graphic design elements such as the box design. Some large publishers with vertical structure also own publishing subsidiaries (labels). Large publishers also attempt to boost efficiency across all internal and external development teams by providing services such as sound design and code packages for commonly needed functionality. Because the publisher often finances development, they usually try to manage development risk along with a staff of producers or project managers to monitor the developer's progress, critique ongoing development, and assist as necessary. Most video games created by an external video game developer are paid for with periodic advances on royalties. These advances are paid when the developer reaches certain stages of development, called milestones. In recent years, the rise of digital distribution platforms such as Steam and console-based online stores has somewhat reduced the impact of seasonal sales cycles, allowing publishers to release titles throughout the year rather than focusing solely on the holiday period. Consumers tend to purchase the most heavily marketed titles rather than those of highest quality, resulting in fewer sales for other games within the same genre. This dynamic has contributed to rising development budgets, as publishers compete to dominate key market segments. It has also encouraged the prioritization of sequels to successful franchises over new intellectual properties, a trend for which publishers such as Activision Blizzard and Electronic Arts have faced criticism. Types of game publishers AAA game publishers produce and create games that are high budget and groundbreaking. They are advanced in technology and forward the boundaries of technology and creativity in the video game world. AAA game publishers often produce popular and blockbuster games. These publishers have the financial resource and means to fund large game development projects. These publishers implement and fund marketing and distribution to guarantee reach and exposure for their games. With their funds to market they are able to advertise and reach a wider consumer pool and have access to distribute to a big network. Although they have creative constraints within game development and marketing, they often focus and follow market trends. They have a higher demand to attain commercial success. Examples of AAA video game publishers are Electronic Arts, Ubisoft, and Activision. Indie game publishers are companies that work with independent developers. Their focus is on developing games that promotes creativity and originality. Developers have creative control over their games. These publishers implement intimate collaborations between the publishers and the developers. Often stand out in the video game market due to the more unique genres. Indie game publishers have restrict marketing budgets and have small audience reach and visibility. Examples of Indie video game publishers are Devolver Digital, Annapurna Interactive and Raw Fury. Mobile game publishers produce and specialize in video games on smartphones and tablet devices. They take advantage of the widespread appeal and rise of mobile gaming. These publishers enhance games for touch based interfaces and devices. They are proficient in designing monetization tactics for mobile platforms. Mobile game publishers have a comprehensive understanding of the mobile gaming market. They have proficiency in strategies for engagement and user acquisition for mobile sites. For mobile gaming there is access through app stores for distribution channels. There are obstacles with monetization due to lack of in-app purchase and free-to-play(F2P) models. Examples of Mobile game publishers are Supercell, King, and Zynga. When engaging a publisher of any of these types as a studio or developer, there will be a publishing agreement set in place. Receiving an offer from a publisher needs careful reviewing and understanding the publishing agreement before signing. A publishing agreement defines your entire working relationship with the publisher. Since the draft is usually prepared by the publisher’s lawyer, the contract will naturally be written to protect the publisher’s interests first. Investor interest Numerous video game publishers are traded publicly on stock markets. As a group, they have had mixed performance. At present, Electronic Arts is the only third-party publisher present in the S&P 500 diversified list of large U.S. corporations; in April 2010, it entered the Fortune 500 for the first time. Hype over video game publisher stocks has been breathless at two points: Publishers References |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-07-04-updated-masks-evidence.html] | [TOKENS: 5276] |
Masks for COVID: Updating the evidence Jeremy Howard July 4, 2022 On this page These are notes I took whilst preparing a paper on mask efficacy from Nov 2021 to Jan 2022. In the end, I gave up on the paper, because I felt like people had given up on masks, so there wasn’t much point in finishing it. I’ve decided to publish these notes in the hope some people will find them a useful starting point for their own research, and since I’ve noticed some signs in recent weeks that people might be open to avoiding COVID again. My previous paper on this topic, in which I led a team of 19 experts, was written in April 2020, and published here in the Proceedings of the National Academy of Science. The rise of better masks In the US, 400 million N95 masks are being distributed for free, coming from the 750 million stored in the US’ Strategic National Stockpile. A similar campaign to distribute 650 millions masks in the US in 2020 was cancelled. KN95 masks are being given to US congressional staff, and masks are required for federal workers and whilst in federal buildings. The Los Angeles school district has required students to upgrade from cloth masks to “well-fitting, non-cloth masks with a nose wire”. Masks work A review paper discussed both lab evidence and empirical evidence for the importance of face masks, with eight “seminal studies” showing a reduction in transmission when masks are used, and one Danish study of surgical masks with “several design limitations” which “demonstrated only a modest benefit in limiting COVID-19 transmission”. The authors note that “laboratory studies have demonstrated the ability of surgical masks to block SARS-COV-2 and other viruses”, with the masks “60%–70% effective at protecting others and 50% effective at protecting the wearer”. An evidence review from early in the pandemic concluded that “given the current shortages of medical masks, we recommend the adoption of public cloth mask wearing, as an effective form of source control”. It noted that “by the end of June 2020, nearly 90% of the global population lived in regions that had nearly universal mask use, or had laws requiring mask use in some public locations.” The review said that “There has been one controlled trial of mask use for influenza control in the general community. The study looked at Australian households, was not done during a pandemic, and was done without any enforcement of compliance” – and yet still found “masks as a group had protective efficacy in excess of 80% against clinical influenza-like illness.” An observational study of Beijing households analyzed the impact of mask use in the community on COVID-19 transmission, finding that face masks were 79% effective in preventing transmission, if used by all household members prior to symptoms occurring. One study used a multiple regression of policy interventions and country and population characteristics to infer the relationship between mask use and SARS-CoV-2 transmission. It found that transmission was around 7.5 times higher in countries that did not have a mask mandate or universal mask use, a result similar to that found in an analogous study of fewer countries. Similar results were found by numerous other papers. A mathematical model of mask use estimates that mask wearing reduces the reproduction number R by (1−mp)^2, where m is the efficacy of trapping viral particles inside the mask, and p is the percentage of the population that wears masks. A report in Nature explained that researchers running a randomized controlled trial (RCT) of community mask use in Bangladesh “began by developing a strategy to promote mask wearing, with measures such as reminders from health workers in public places. This ultimately tripled mask usage, from only 13% in control villages to 42% in villages where it was encouraged”, and “then compared numbers of COVID-19 cases in control villages and the treatment communities”. They found that the number of infections in mask wearing communities decreased, with a reduction of COVID symptoms using surgical masks to 0.87 times the incidence in unmasked communities, and 0.91 times when using cloth masks. The report noted that “the researchers suggest that the true risk reduction is probably much greater, in part because they did no SARS-CoV-2 testing of people without symptoms or whose symptoms did not meet the World Health Organization’s definition of the disease.” The researchers concluded that “promoting community mask-wearing can improve public health”. The Johns Hopkins School of Public Health reviewed the work and concluded that “This study is the largest and best-designed randomized controlled trial to date of a realistic non-pharmaceutical intervention on SARS-CoV-2 transmission.” A paper investigating an upper bound on one-to-one exposure to infectious human respiratory particles concludes that “face masks significantly reduce the risk of SARS-CoV-2 infection compared to social distancing. We find a very low risk of infection when everyone wears a face mask, even if it doesn’t fit perfectly on the face.” They calculate that “social distancing alone, even at 3.0 m between two speaking individuals, leads to an upper bound of 90% for risk of infection after a few minutes”, but that when both source and susceptible wear a well-fitting FFP2 mask, there is only 0.4% after one hour of contact. They found that to achieve good fit it is important to mold the nose piece wire to the size of the nose, rather than leaving it in a sharp folded position. A similar study “quantifies the extent to which transmission risk is reduced in large rooms with high air exchange rates, increased for more vigorous respiratory activities, and dramatically reduced by the use of face masks.” The authors describe the six-foot rule widely used to ensure social distancing as “a guideline that offers little protection from pathogen-bearing aerosol droplets sufficiently small to be continuously mixed through an indoor space.” Instead, they develop a safety guideline based on cumulative exposure time,” the product of the number of occupants and their time in an enclosed space. In particular, they identify that the greatest risk comes in places where people are speaking (other than quietly) or singing, and that “the benefit of face masks is immediately apparent”, due to the multiplicative effect when both source and susceptible wear a mask. They further note that “Air filtration has a less dramatic effect than face mask use in increasing the CET bound. Nevertheless, it does offer a means of mitigating indoor transmission with greater comfort, albeit at greater cost.” Another study of the combined impacts of ventilation and mask effective filtration efficiency in classroom settings found that “ventilation alone is not able to achieve probabilities <0.01 (1%)” of transmitting COVID in a classroom. However, they found that good masks reduce infection probability by >5× in some cases, and that “reductions provided by ventilation and masks are synergistic and multiplicative”. However they also noted that “most masks fit poorly”, recommending that work be done to ensure that high quality masks are used. Similar results were found in a study of community public health interventions, which concluded that “control the pandemic, our models suggest that high adherence to social distance is necessary to curb the spread of the disease, and that wearing face masks provides optimal protection even if only a small portion of the population comply with social distance”. Guidance from the independent scientific advisory group OzSAGE points out “that school children are able to wear masks. As an example, all children over two years of age in San Francisco are required to wear masks at school”. Omicron changes the game An analysis of fine aerosol emissions found that, compared to the original wild type (WT) virus: “Delta and Omicron both also have increased transmissibility: the number of cells infected for a given number of ribonucleic acid (RNA) virus copies was found to be doubled and quadrupled respectively. Furthermore, Omicron also seems to be better at evading the immune system. This implies that the critical dose of virus copies above which a situation is potentially infectious needs to be lowered. For the WT, we had proposed a critical dose of 500 virus copies. If the above-mentioned capacity to infect cells translates into an infection risk, this would imply a critical dose of around 300 virus copies for Delta and around 100 virus copies for Omicron.” The study finds that “surgical masks are no longer sufficient in most public settings, while correctly fitted FFP2 respirators still provide sufficient protection, except in high aerosol producing situations such as singing or shouting.” Data from Hong Kong shows that “Omicron SARS-CoV-2 infects and multiplies 70 times faster than the Delta variant and original SARS-CoV-2 in human bronchus”. A study of transmission in Danish households estimated the secondary attack rate (SAR) of omicron compared to delta, finding it 1.2 times higher for unvaccinated people, 2.6 times higher for double-dosed, and 3.7 times higher for boosted. The authors conclude that “the rapid spread of the Omicron VOC primarily can be ascribed to the immune evasiveness”. According to UK statistics, the risk of hospitalization from omicron when unvaccinated is about the same as the wildtype virus, which is about half the risk of the delta variant. The journal Infection Control Today reported that many experts are concerned that “‘Omicron the Pandemic Killer’ Idea Ignores Dangers of Long COVID”: “Linda Spaulding, RN-BC, CIC, CHEC, CHOP, a member of Infection Control Today®’s Editorial Advisory Board (EAB), says that she’s “seen athletes in their 20s on the wait list for double lung transplants because of long COVID. That’s something that has long-term consequences. Some people talk of COVID fog. They just can’t put their thoughts together.” In addition, even the treatments for those with long COVID can put toll on a patient’s body.” “As noted by Kevin Kavanagh, MD, another member of ICT®’s EAB, a core difficulty in society’s attempt to guide COVID-19 from pandemic to endemic is that COVID is not just a respiratory virus. Kavanagh wrote in October that SARS-CoV-2 is similar to HIV because it can “silently spread throughout the host’s body and attack almost every organ.”” Better masks work better The US Centers for Disease Control and Prevention (CDC) explains that: “Loosely woven cloth products provide the least protection, layered finely woven products offer more protection, well-fitting disposable surgical masks and KN95s offer even more protection, and well-fitting NIOSH-approved respirators (including N95s) offer the highest level of protection.” Unfortunately “well-fitting disposable surgical masks” do not exist out of the box, since there are large gaps on each side of the mask. Surgical masks require modifications to achive a good fit. That’s because they are made to stop liquid splashes during surgery, rather than made to stop airborne transmission. There are two methods shown by the CDC to improve fit: Research shows that both of these approaches dramatically reduce exposure to aerosols emitted during a period of breathing: “…adding a cloth mask over the source headform’s medical procedure mask or knotting and tucking the medical procedure mask reduced the cumulative exposure of the unmasked receiver by 82.2% (SD = 0.16) and 62.9% (SD = 0.08), respectively (Figure 2). When the source was unmasked and the receiver was fitted with the double mask or the knotted and tucked medical procedure mask, the receiver’s cumulative exposure was reduced by 83.0% (SD = 0.15) and 64.5% (SD = 0.03), respectively. When the source and receiver were both fitted with double masks or knotted and tucked masks, the cumulative exposure of the receiver was reduced 96.4% (SD = 0.02) and 95.9% (SD = 0.02), respectively.” An airborne transmission simulator was used to estimate the ability of various types of face masks to block COVID-19 transmission. In this experiment, “cotton mask led to an approximately 20% to 40% reduction in virus uptake compared to no mask. The N95 mask had the highest protective efficacy (approximately 80% to 90% reduction)”. All of the masks were much more effective at source control than at protecting the wearer, with the N95 stopping all detectable transmission. The American Conference of Governmental Industrial Hygienists (ACGIH) say that “workers need respirators”, noting that a worker with an “N95 filtering facepiece respirator… has 1-10% inward leakage and outward leakage”, but with a surgical mask “has 50% inward leakage and outward leakage”, and with a cloth face covering “has 75% inward leakage and outward leakage”. They explain that “N95 FFRs have an assigned protection factor of 10 (10% inward leakage) but must receive a fit factor of 100 (1% inward leakage) on an individual worker.” ACGIH created a table showing how, if we start with an assumption that it takes on average 15 minutes to get infected if no-one is wearing a mask (based on CDC contact tracing premises), we can calculate the time it would take on average to get infected if one or both of source and receiver are wearing various types of mask. This is calculated by simply dividing the base time of 15 minutes by the leakage factor for the source’s mask (if any), and then dividing that by the leakage factor for the receiver’s mask (if any). This approach is, however, an over-simplification. Reseach based on a a single-hit model of infection shows that the probability of infection “shows a highly nonlinear sensitivity” to inhaled virus number. Therefore, “In a virus-rich regime… wearing a mask may not suffice to prevent infection.” Research undertaken by the National Personal Protective Technology Laboratory (NPPTL) found that respirators with an exhalation valve “reduce particle emissions to levels similar to or better than those provided by surgical masks, procedure masks, or cloth face coverings”. Furthermore, “surgical tape secured over the valve from the inside of the FFR can provide source control similar to that of an FFR with no exhalation valve”. Pushing back against masks Professor Alison McMillan, Commonwealth Chief Nursing and Midwifery Officer in Australia claims that “there is no evidence to suggest that we should be moving towards… N95 respirators in the community setting.” She added “I am aware that there are some publications out there suggesting a move to N95 (masks). But that’s not supported in the empirical evidence”. According to Norman Swan, host of the ABC’s Coronacast, “If you’re wearing an N95 that hasn’t been fit tested – and it’s not an easy process to do yourself at home – there’s no guarantee that it’s an awful lot more effective than wearing a surgical mask. Professor Catherine Bennett, chair in epidemiology at Deakin University, claims that”Technically, the instructions say you shouldn’t reuse” respirators, and that “If you’re not particularly checking its fit, you’re probably wasting your time”. Occupational environment physician Malcolm Sim agrees: “If you put it [an N95 mask] off and put it on, they’re not meant for that purpose… They’re easily damaged in somebody’s handbag,” adding that the integrity of the masks can be compromised. He says that “If you’re handling them a lot, taking them on and off, there’s much more potential for you to get it [the virus] on your hands, your face, different parts of your body.” University of New South Wales epidemiologist Mary-Louise McLaws claimed that “There’s no evidence yet that a N95 mask will protect you more than a surgical mask for Omicron.” An opinion piece in Newsweek claims that “the effectiveness of respirators is vastly overestimated, and there is scant evidence that they stop community transmission. Moreover, NIOSH-approved respirators are tight, uncomfortable, and can impede breathing.” The article further claims that “For respirators to work, they must be well fitting, must be tested by OSHA, and must be used for only short time windows as their effectiveness diminishes as they get wet from breathing.” Recently there has been particular pushback against the use of masks by children, with the Newsweek article alleging that “Respirators are not necessary to protect children from COVID-19 because of the astoundingly low risk COVID-19 presents to them”, and that in fact wearing masks involves “existing well-documented harms”. There hasn’t been any documented harms to children from wearing masks, Respirators can be reused According to mask manufacturer 3M, respirators (which they refer to as “Filtering Facepiece Respirators (FFRs)”) “can be used many times.” They say that “There is no time limit to wearing an FFR. Respirators can be worn until they are dirty, damaged or difficult to breathe through.” In reporting from CNN, Linsey Marr, a professor of civil and environmental engineering at Virginia Tech, explained that an N95 mask’s material and filtration ability aren’t “going to degrade unless you physically rub it or poke holes in it.”You’d have to be in really polluted air … for several days before it lost its ability to filter out particles. So, you can really wear them for a long time. People have been talking about 40 hours – I think that’s fine. Really, it’s going to get gross from your face or the straps will get too loose or maybe break before you’re going to lose filtration ability… One of the first indicators of being able to change it if it looks nice and clean is that it just feels a little harder to breathe through. There appears to be more resistance with every breath.” She also noted that the contamination risk in reusing N95 masks is “lower, much lower, than the risk of you not wearing an N95 and breathing in particles”. The CDC has prepared guidelines for optimizing the supply of respirators which recommend reusing respirators at most five times. This guidelines were created for people “implementing policies and procedures for preventing pathogen transmission in healthcare settings”. They have been widely shared, incorrectly, by reporters as being recommendations for community use. The inventor of N95 mask material, Peter Tsai, says that “N95 masks can be rotated, 1 mask every 3–4 days”, and that in doing this “there is no change in the mask’s properties.” According to the NIOSH Guide to the Selection and Use of Particulate Respirators N95 respirators must maintain at least 95% filtration after a total mass loading of 200mg. This is designed to ensure they continue to work in sites with high particulate matter, such as some construction environment. However in normal use, even outside in a city with high levels of population, it would take over 200 days of 24 hour per day use to get to this level. The guide says that “generally, the use and reuse of N-series lters would also be subject only to considerations of hygiene, damage, and increased breathing resistance”. The NIOSH guidelines are well supported by research. Fit tests are not required for respirators to be effective In one study non-experts were asked to read the instructions that come with a respirator, and then to don the respirator without assistance and complete a fit test. The average fit factor achieved was 88, and the lowest fit factor of the subjects was 15, with nearly half achieving a fit factor greater than 100. Surgical masks have been found to have a much poorer fit in practice. One study showed that for surgical masks “quantitative fit factors ranged from 2.5 to 9.6”, and another found an average fit factor of 3.0. Guidance from the US Food and Drug Administration (FDA) explains that: “Fit Factor is a means of expressing the difference in particle concentration inside the mask and outside the mask during use. For example, a fit factor of 2 means that the concentration of particles within the mask is ½ or 50% of the concentration outside the mask; a fit factor of 5 means the concentration of particles within the mask is 1/5 th or 20% of the concentration outside the mask.” The guidance says that failing to achieve a fit factor of 2 “may suggest that respirator fit will not be sufficient to assure that the device will help reduce wearer exposure to pathogenic biological airborne particulates.” An analysis of the fitted filtration efficiency (FFE) of surgical masks found that, unmodified, they only achieved an FFE of 38.5%. The “knot and tuck” technique improved that to 60.3%, and a DIY mask fitter consisting of three rubber bands increased it to 78.2%. A 3-layer cotton mask had an FFE of just 26.5%. An N95, on the other hand, achieved an FFE of 98.4%. Furthermore, the N95 FFE had a standard deviation of only 0.5% — that is, it was effective for multiple tests during “a series of repeated movements of the torso, head, and facial muscles”. Interestingly, a 2-layer nylon mask had an FFE of 79.0% (standard devatiation 4.3%), showing that some cloth masks can be quite effective. These findings were replicated in a study of numerous types of cloth mask, which found that hybrids of 600 TPI cotton with silk, chiffon, or polypropelene achieved 72-96% filtration efficiency. Researchers have calculated that “the particle size most likely to deposit in the respiratory tract when wearing a mask is ∼2μm”. Unfortunately, this particle size is not considered in N95 or similar standards. Instead, 0.3 μm particles are used. A 2010 study of fit testing respirators for public health medical emergencies found that 98% of non-experts wearing masks without training achieved a fit factor of over 5 (20% leakage) and 75% of them achieved a fit factor of over 10 (10% leakage). Donning and doffing masks is not complex or risky Analysis by the CDC concludes that the risk of infection through surfaces (fomites) “is generally considered to be low”, a view that was supported by the evidence as early as July 2020. An analysis of “418 samples from mask fronts, cell phones, paper money, card machines, sewage, air and bedding” during a COVID surge “did not detect any trace of SARS-CoV-2 in all samples analyzed”. We should not reserve respirators for healthcare workers According to Anne Miller, executive director of Project N95, there are many U.S. manufacturers of N95 masks and an ample supply. The Economist reported that in Europe “at the start of the pandemic, FFP2 masks were scarce and costly. Even governments fell victim to price gouging, paying more than €4 ($4.50) per mask. Demand had previously been low, so stockpiles and production capacity could not satisfy the sudden surge. Governments wanted to reserve supplies for those most at risk of contracting the virus, such as health-care workers.” However they reported that by the end of 2021 “FFP2 masks are in healthy supply, and as the highly transmissible Omicron variant spreads across the world, updating guidance to recommend their wider use could be one way to help reduce transmission.” In the first 6 months of 2020, over 70,000 new face mask companies were registered in China, many run by people with no previous experience and no registration or licensing. The Chinese government stepped in to make licensing more stringent, shutting down many companies, and international demand fell over quality concerns. Due to “a dramatic reduction in demand for N95s”, US mask factories are closing. In June 2021 the American Mask Manufacturer’s Association said that “we have 28 members who are going to go out of business in the next 60 to 90 days.” By July 2021 they estimated “that 5,000 workers have been laid off across its member companies”. However following school mask mandates and demand during the omicron surge, demand in the US spiked in early 2022. The CDC has found that 60% of KN95s are counterfeit. In Australia it has been reported that “general practitioners have been left without highly protective N95 masks as consumers rush to stock up after a sharp rise in COVID-19 cases.” In May 2021 the CDC stated that “The supply and availability of NIOSH-approved respirators have increased significantly over the last several months. Healthcare facilities should not be using crisis capacity strategies at this time and should promptly resume conventional practices.” Demand distortions can increase as we proceed up the supply chain, creating inefficiencies for upstream firms. This is known as the Bullwhip Effect. Respirators need not be uncomfortable In an analysis of the physiological impact of the N95 filtering facepiece respirator (FFR) “in healthy healthcare workers, FFR did not impose any important physiological burden during 1 hour of use, at realistic clinical work rates”. A study of KF80, KF94, KF99, N95, and N99 masks found that self-reported comfort levels were nearly perfectly correlated with the ease of inhalation. |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-06-01-qualitative.html] | [TOKENS: 2835] |
Qualitative humanities research is crucial to AI Louisa Bartolo and Rachel Thomas June 1, 2022 On this page “All research is qualitative; some is also quantitative” Harvard Social Scientist and Statistician Gary King Suppose you wanted to find out whether a machine learning system being adopted - to recruit candidates, lend money, or predict future criminality - exhibited racial bias. You might calculate model performance across groups with different races. But how was race categorised– through a census record, a police officer’s guess, or by an annotator? Each possible answer raises another set of questions. Following the thread of any seemingly quantitative issue around AI ethics quickly leads to a host of qualitative questions. Throughout AI, qualitative decisions are made about what metrics to optimise for, which categories to use, how to define their bounds, who applies the labels. Similarly, qualitative research is necessary to understand AI systems operating in society: evaluating system performance beyond what can be captured in short term metrics, understanding what is missed by large-scale studies (which can elide details and overlook outliers), and shedding light on the circumstances in which data is produced (often by crowd-sourced or poorly paid workers). Unfortunately, there is often a large divide between computer scientists and social scientists, with over-simplified assumptions and fundamental misunderstandings of one another. Even when cross-disciplinary partnerships occur, they often fall into “normal disciplinary divisions of labour: social scientists observe, data scientists make; social scientists do ethics, data scientists do science; social scientists do the incalculable, data scientists do the calculable.” The solution is not for computer scientists to absorb a shallow understanding of the social sciences, but for deeper collaborations. In a paper on exclusionary practices in AI ethics, an interdisciplinary team wrote of the “indifference, devaluation, and lack of mutual support between CS and humanistic social science (HSS), [which elevates] the myth of technologists as ‘ethical unicorns’ that can do it all, though their disciplinary tools are ultimately limited.” This is further reflected in an increasing number of job ads for AI ethicists that list a computer science degree as a requirement, “prioritising technical computer science infrastructure over the social science skills that can evaluate AI’s social impact. In doing so, we are building the field of AI Ethics to replicate the very flaws this field is trying to fix.” Interviews with 26 responsible AI practitioners working in industry highlighted a number of challenges, including that qualitative work was not prioritised. Not only is it impossible to fully understand ethics issues solely through quantitative metrics, inappropriate and misleading quantitative metrics are used to evaluate the responsible AI practitioners themselves. Interviewees reported that their fairness work was evaluated on metrics related to generating revenue, in a stark misalignment of goals. Qualitative research helps us evaluate AI systems beyond short term metrics When companies like Google and YouTube want to test whether the recommendations they are making (in the form of search engine results or YouTube videos, for example) are “good” - they will often focus quite heavily on “engagement” or “dwell time” - the time a user spent looking at or watching the item recommended to them. But it turns out, unsurprisingly, that a focus on engagement and dwell time, narrowly understood, raises all sorts of problems. Demographics can impact dwell time (e.g. older users may spend longer on websites than younger users, just as part of the way they use the internet). A system that ‘learns’ from a user’s behavioural cues (rather than their ‘stated preferences’) might lock them into a limiting feedback loop, appealing to that user’s short term interests rather than those of their ‘Better Selves.’ Scholars have called for more qualitative research to understand user experience and build this into the development of metrics. This is the part where people will point out, rightly, that companies like Google and YouTube rely on a complex range of metrics and signals in their machine learning systems - and that where a website ranks on Google, or how a YouTube video performs in recommendation does not boil down to simple popularity metrics, like engagement. Google employs an extensive process to determine “relevance” and “usefulness” for search results. In its 172-page manual for search result ‘Quality’ evaluation, for example, the company explains how evaluators should assess a website’s ‘Expertise/ Authoritativeness/ Trustworthiness’ or ‘E-A-T’; and what types of content, by virtue of its harmful nature (e.g., to protected groups), should be given a ‘low’ ranking. YouTube has identified specific categories of content (such as news, scientific subjects, and historical information) for which ‘authoritativeness’ should be considered especially important. It has also determined that dubious-but-not-quite-rule-breaking information (what it calls ‘borderline content’) should not be recommended, regardless of the video’s engagement levels. Irrespective of how successful we consider the existing approaches of Google Search and YouTube to be (and partly, the issue is that evaluating their implementation from the outside is frustratingly difficult), the point here is that there are constant qualitative judgments being made, about what makes a search result or recommendation “good” and of how to define and quantify expertise, authoritativeness, trustworthiness, borderline content, and other values. This is true of all machine learning evaluation, even when it isn’t explicit. In a paper guiding companies about how to carry out internal audits of their AI systems, Inioluwa Deborah Raji and colleagues emphasise the importance of interviews with management and engineering teams to “capture and pay attention to what falls outside the measurements and metrics, and to render explicit the assumptions and values the metrics apprehend.” (p.40). The importance of thoughtful humanities research is heightened if we are serious about grappling with the potential broader social effects of machine learning systems (both good and bad), which are often delayed, distributed and cumulative. Small-scale qualitative studies tell an important story even (and perhaps especially) when they seem to contradict large-scale ‘objective’ studies Hypothetically, let’s say you wanted to find out whether the use of AI technologies by doctors during a medical appointment would make doctors less attentive to patients - what do you think the best way of doing it would be? You could find some criteria and method for measuring ‘attentiveness’, say tracking the amount of eye contact between the doctor and patient, and analyse this across a representative sample of medical appointments where AI technologies were being used, compared to a control group of medical appointments where AI technologies weren’t being used. Or would you interview doctors about their experiences using the technology during appointments? Or talk to patients about how they felt the technology did, or didn’t, impact their experience? In research circles, we describe these as ‘epistemological’ choices - your judgement of what constitutes the ‘best’ approach is inextricably linked to your judgement about how we can claim to ‘know’ something. These are all valid methods for approaching the question, but you can imagine how they might result in different, even conflicting, insights. For example, you might end up with the following results: - The eye contact tracking experiment suggests that overall, there is no significant difference in doctors’ attentiveness to the patient when the AI tech is introduced. - The interviews with doctors and patients reveal that some doctors and patients feel that the AI technology reduces doctors’ attentiveness to patients, and others feel that it makes no difference or even increases doctors’ attention to the patient. Even if people are not negatively impacted by something ‘on average’ (e.g., in our hypothetical eye contact tracking experiment above), there will remain groups of people who will experience negative impacts, perhaps acutely so. “Many of people’s most pressing questions are about effects that vary for different people,” write Matias, Pennington and Chan in a recent paper on the idea of N-of-one trials. To tell people that their experiences aren’t real or valid because they don’t meet some threshold for statistical significance across a large population doesn’t help us account for the breadth and nature of AI’s impacts on the world. Examples of this tension between competing claims to knowledge about AI systems’ impacts abound. Influencers who believe they are being systematically downranked (‘shadowbanned’) by Instagram’s algorithmic systems are told by Instagram that this simply isn’t true. Given the inscrutability of these proprietary algorithmic systems, it is impossible for influencers to convincingly dispute Instagram’s claims. Kelley Cotter refers to this as a form of “black box gaslighting”: platforms can “leverage perceptions of their epistemic authority on their algorithms to undermine users’ confidence in what they know about algorithms and destabilise credible criticism.” Her interviews with influencers give voice to stakeholder concerns and perspectives that are elided in Instagram’s official narrative about its systems. The mismatch between different stakeholders’ accounts of ‘reality’ is instructive. For example, a widely-cited paper by Netflix employees claims that Netflix recommendation “influences choice for about 80% of hours streamed at Netflix.” But this claim stands in stark contrast to Mattias Frey’s mixed-methods research (representative survey plus small sample for interviews) run with UK and US adults, in which less than 1 in 5 adults said they primarily relied on Netflix recommendations when deciding what films to watch. Even if this is because users underestimate their reliance on recommender systems, that’s a critically important finding - particularly when we’re trying to regulate recommendation and so many are advocating providing better user-level controls as a check on platform power. Are people really going to go to the trouble of changing their settings if they don’t think they rely on algorithmic suggestions that much anyway? Qualitative research sheds light on the context of data annotation Machine learning systems rely on vast amounts of data. In many cases, for that data to be useful, it needs to be labelled/ annotated. For example, a hate speech classifier (an AI-enabled tool used to identify and flag potential cases of hate speech on a website) relies on huge datasets of text labelled as ‘hate speech’ or ‘not hate speech’ to ‘learn’ how to spot hate speech. But it turns out that who is doing the annotating and in what context they’re doing it, matters. AI-powered content moderation is often held up as the solution to harmful content online. What has continued to be underplayed is the extent to which those automated systems are and most likely will remain dependent on the manual work of human content moderators sifting through some of the worst and most traumatic online material to power the machine learning datasets on which automated content moderation depends. Emily Denton and her colleagues highlight the significance of annotators’ social identity (e.g., race, gender) and their expertise when it comes to annotation tasks, and they point out the risks associated with overlooking these factors and simply ‘aggregating’ results as ‘ground truth’ rather than properly exploring disagreements between annotators and the important insights that this kind of disagreement might offer. Human commercial content moderators (such as the people that identify and remove violent and traumatic imagery on Facebook) often labour in terrible conditions, lacking psychological support or appropriate financial compensation. The interview-based research of Sarah T. Roberts has been pioneering in highlighting these conditions. Most demand for crowdsourced digital labour comes from the Global North, yet the majority of these workers are based in the Global South and receive low wages. Semi-structured interviews reveal the extent to which workers feel unable to bargain effectively for better pay in the current regulatory environment. As Mark Graham and his colleagues point out, these findings are hugely important in a context where several governments and supranational development organisations like the World Bank are holding up digital work as a promising tool to fight poverty. The decision of how to measure ‘race’ in machine learning systems is highly consequential, especially in the context of existing efforts to evaluate these systems for their “fairness.” Alex Hanna, Emily Denton, Andrew Smart and Jamila Smith-Loud have done crucial work highlighting the limitation of machine learning systems that rely on official records of race or their proxies (e.g. census records), noting that the racial categories provided by such records are “unstable, contingent, and rooted in racial inequality.” The authors emphasise the importance of conducting research in ways that prioritise the perspectives of the marginalised racial communities that fairness metrics are supposed to protect. Qualitative research is ideally placed to contribute to a consideration of “race” in machine learning systems that is grounded in the lived experiences and needs of the racially subjugated. What next? Collaborations between quantitative and qualitative researchers are valuable in understanding AI ethics from all angles. Consider reading more broadly, outside your particular area. Perhaps using the links and researchers listed here as starting points. They’re just a sliver of the wealth that’s out there. You could also check out the Social Media Collective’s Critical Algorithm Studies reading list, the reading list provided by the LSE Digital Ethnography Collective, and Catherine Yeo’s suggestions. Strike up conversations with researchers in other fields, and consider the possibility of collaborations. Find a researcher slightly outside your field but whose work you broadly understand and like, and follow them on Twitter. With any luck, they will share more of their work and help you identify other researchers to follow. Collaboration can be an incremental process: Consider inviting the researcher to form part of a discussion panel, reach out to say what you liked and appreciated about their work and why, and share your own work with them if you think it’s aligned with their interests. Within your university or company, is there anything you could do to better reward or facilitate interdisciplinary work? As Humanities Computing Professor Willard McCarty notes, somewhat discouragingly, “professional reward for genuinely interdisciplinary research is rare.” To be sure, individual researchers and practitioners have to be prepared to put themselves out there, compromise and challenge themselves - but carefully tailored institutional incentives and enablers matter. |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-07-04-updated-masks-evidence.html] | [TOKENS: 5276] |
Masks for COVID: Updating the evidence Jeremy Howard July 4, 2022 On this page These are notes I took whilst preparing a paper on mask efficacy from Nov 2021 to Jan 2022. In the end, I gave up on the paper, because I felt like people had given up on masks, so there wasn’t much point in finishing it. I’ve decided to publish these notes in the hope some people will find them a useful starting point for their own research, and since I’ve noticed some signs in recent weeks that people might be open to avoiding COVID again. My previous paper on this topic, in which I led a team of 19 experts, was written in April 2020, and published here in the Proceedings of the National Academy of Science. The rise of better masks In the US, 400 million N95 masks are being distributed for free, coming from the 750 million stored in the US’ Strategic National Stockpile. A similar campaign to distribute 650 millions masks in the US in 2020 was cancelled. KN95 masks are being given to US congressional staff, and masks are required for federal workers and whilst in federal buildings. The Los Angeles school district has required students to upgrade from cloth masks to “well-fitting, non-cloth masks with a nose wire”. Masks work A review paper discussed both lab evidence and empirical evidence for the importance of face masks, with eight “seminal studies” showing a reduction in transmission when masks are used, and one Danish study of surgical masks with “several design limitations” which “demonstrated only a modest benefit in limiting COVID-19 transmission”. The authors note that “laboratory studies have demonstrated the ability of surgical masks to block SARS-COV-2 and other viruses”, with the masks “60%–70% effective at protecting others and 50% effective at protecting the wearer”. An evidence review from early in the pandemic concluded that “given the current shortages of medical masks, we recommend the adoption of public cloth mask wearing, as an effective form of source control”. It noted that “by the end of June 2020, nearly 90% of the global population lived in regions that had nearly universal mask use, or had laws requiring mask use in some public locations.” The review said that “There has been one controlled trial of mask use for influenza control in the general community. The study looked at Australian households, was not done during a pandemic, and was done without any enforcement of compliance” – and yet still found “masks as a group had protective efficacy in excess of 80% against clinical influenza-like illness.” An observational study of Beijing households analyzed the impact of mask use in the community on COVID-19 transmission, finding that face masks were 79% effective in preventing transmission, if used by all household members prior to symptoms occurring. One study used a multiple regression of policy interventions and country and population characteristics to infer the relationship between mask use and SARS-CoV-2 transmission. It found that transmission was around 7.5 times higher in countries that did not have a mask mandate or universal mask use, a result similar to that found in an analogous study of fewer countries. Similar results were found by numerous other papers. A mathematical model of mask use estimates that mask wearing reduces the reproduction number R by (1−mp)^2, where m is the efficacy of trapping viral particles inside the mask, and p is the percentage of the population that wears masks. A report in Nature explained that researchers running a randomized controlled trial (RCT) of community mask use in Bangladesh “began by developing a strategy to promote mask wearing, with measures such as reminders from health workers in public places. This ultimately tripled mask usage, from only 13% in control villages to 42% in villages where it was encouraged”, and “then compared numbers of COVID-19 cases in control villages and the treatment communities”. They found that the number of infections in mask wearing communities decreased, with a reduction of COVID symptoms using surgical masks to 0.87 times the incidence in unmasked communities, and 0.91 times when using cloth masks. The report noted that “the researchers suggest that the true risk reduction is probably much greater, in part because they did no SARS-CoV-2 testing of people without symptoms or whose symptoms did not meet the World Health Organization’s definition of the disease.” The researchers concluded that “promoting community mask-wearing can improve public health”. The Johns Hopkins School of Public Health reviewed the work and concluded that “This study is the largest and best-designed randomized controlled trial to date of a realistic non-pharmaceutical intervention on SARS-CoV-2 transmission.” A paper investigating an upper bound on one-to-one exposure to infectious human respiratory particles concludes that “face masks significantly reduce the risk of SARS-CoV-2 infection compared to social distancing. We find a very low risk of infection when everyone wears a face mask, even if it doesn’t fit perfectly on the face.” They calculate that “social distancing alone, even at 3.0 m between two speaking individuals, leads to an upper bound of 90% for risk of infection after a few minutes”, but that when both source and susceptible wear a well-fitting FFP2 mask, there is only 0.4% after one hour of contact. They found that to achieve good fit it is important to mold the nose piece wire to the size of the nose, rather than leaving it in a sharp folded position. A similar study “quantifies the extent to which transmission risk is reduced in large rooms with high air exchange rates, increased for more vigorous respiratory activities, and dramatically reduced by the use of face masks.” The authors describe the six-foot rule widely used to ensure social distancing as “a guideline that offers little protection from pathogen-bearing aerosol droplets sufficiently small to be continuously mixed through an indoor space.” Instead, they develop a safety guideline based on cumulative exposure time,” the product of the number of occupants and their time in an enclosed space. In particular, they identify that the greatest risk comes in places where people are speaking (other than quietly) or singing, and that “the benefit of face masks is immediately apparent”, due to the multiplicative effect when both source and susceptible wear a mask. They further note that “Air filtration has a less dramatic effect than face mask use in increasing the CET bound. Nevertheless, it does offer a means of mitigating indoor transmission with greater comfort, albeit at greater cost.” Another study of the combined impacts of ventilation and mask effective filtration efficiency in classroom settings found that “ventilation alone is not able to achieve probabilities <0.01 (1%)” of transmitting COVID in a classroom. However, they found that good masks reduce infection probability by >5× in some cases, and that “reductions provided by ventilation and masks are synergistic and multiplicative”. However they also noted that “most masks fit poorly”, recommending that work be done to ensure that high quality masks are used. Similar results were found in a study of community public health interventions, which concluded that “control the pandemic, our models suggest that high adherence to social distance is necessary to curb the spread of the disease, and that wearing face masks provides optimal protection even if only a small portion of the population comply with social distance”. Guidance from the independent scientific advisory group OzSAGE points out “that school children are able to wear masks. As an example, all children over two years of age in San Francisco are required to wear masks at school”. Omicron changes the game An analysis of fine aerosol emissions found that, compared to the original wild type (WT) virus: “Delta and Omicron both also have increased transmissibility: the number of cells infected for a given number of ribonucleic acid (RNA) virus copies was found to be doubled and quadrupled respectively. Furthermore, Omicron also seems to be better at evading the immune system. This implies that the critical dose of virus copies above which a situation is potentially infectious needs to be lowered. For the WT, we had proposed a critical dose of 500 virus copies. If the above-mentioned capacity to infect cells translates into an infection risk, this would imply a critical dose of around 300 virus copies for Delta and around 100 virus copies for Omicron.” The study finds that “surgical masks are no longer sufficient in most public settings, while correctly fitted FFP2 respirators still provide sufficient protection, except in high aerosol producing situations such as singing or shouting.” Data from Hong Kong shows that “Omicron SARS-CoV-2 infects and multiplies 70 times faster than the Delta variant and original SARS-CoV-2 in human bronchus”. A study of transmission in Danish households estimated the secondary attack rate (SAR) of omicron compared to delta, finding it 1.2 times higher for unvaccinated people, 2.6 times higher for double-dosed, and 3.7 times higher for boosted. The authors conclude that “the rapid spread of the Omicron VOC primarily can be ascribed to the immune evasiveness”. According to UK statistics, the risk of hospitalization from omicron when unvaccinated is about the same as the wildtype virus, which is about half the risk of the delta variant. The journal Infection Control Today reported that many experts are concerned that “‘Omicron the Pandemic Killer’ Idea Ignores Dangers of Long COVID”: “Linda Spaulding, RN-BC, CIC, CHEC, CHOP, a member of Infection Control Today®’s Editorial Advisory Board (EAB), says that she’s “seen athletes in their 20s on the wait list for double lung transplants because of long COVID. That’s something that has long-term consequences. Some people talk of COVID fog. They just can’t put their thoughts together.” In addition, even the treatments for those with long COVID can put toll on a patient’s body.” “As noted by Kevin Kavanagh, MD, another member of ICT®’s EAB, a core difficulty in society’s attempt to guide COVID-19 from pandemic to endemic is that COVID is not just a respiratory virus. Kavanagh wrote in October that SARS-CoV-2 is similar to HIV because it can “silently spread throughout the host’s body and attack almost every organ.”” Better masks work better The US Centers for Disease Control and Prevention (CDC) explains that: “Loosely woven cloth products provide the least protection, layered finely woven products offer more protection, well-fitting disposable surgical masks and KN95s offer even more protection, and well-fitting NIOSH-approved respirators (including N95s) offer the highest level of protection.” Unfortunately “well-fitting disposable surgical masks” do not exist out of the box, since there are large gaps on each side of the mask. Surgical masks require modifications to achive a good fit. That’s because they are made to stop liquid splashes during surgery, rather than made to stop airborne transmission. There are two methods shown by the CDC to improve fit: Research shows that both of these approaches dramatically reduce exposure to aerosols emitted during a period of breathing: “…adding a cloth mask over the source headform’s medical procedure mask or knotting and tucking the medical procedure mask reduced the cumulative exposure of the unmasked receiver by 82.2% (SD = 0.16) and 62.9% (SD = 0.08), respectively (Figure 2). When the source was unmasked and the receiver was fitted with the double mask or the knotted and tucked medical procedure mask, the receiver’s cumulative exposure was reduced by 83.0% (SD = 0.15) and 64.5% (SD = 0.03), respectively. When the source and receiver were both fitted with double masks or knotted and tucked masks, the cumulative exposure of the receiver was reduced 96.4% (SD = 0.02) and 95.9% (SD = 0.02), respectively.” An airborne transmission simulator was used to estimate the ability of various types of face masks to block COVID-19 transmission. In this experiment, “cotton mask led to an approximately 20% to 40% reduction in virus uptake compared to no mask. The N95 mask had the highest protective efficacy (approximately 80% to 90% reduction)”. All of the masks were much more effective at source control than at protecting the wearer, with the N95 stopping all detectable transmission. The American Conference of Governmental Industrial Hygienists (ACGIH) say that “workers need respirators”, noting that a worker with an “N95 filtering facepiece respirator… has 1-10% inward leakage and outward leakage”, but with a surgical mask “has 50% inward leakage and outward leakage”, and with a cloth face covering “has 75% inward leakage and outward leakage”. They explain that “N95 FFRs have an assigned protection factor of 10 (10% inward leakage) but must receive a fit factor of 100 (1% inward leakage) on an individual worker.” ACGIH created a table showing how, if we start with an assumption that it takes on average 15 minutes to get infected if no-one is wearing a mask (based on CDC contact tracing premises), we can calculate the time it would take on average to get infected if one or both of source and receiver are wearing various types of mask. This is calculated by simply dividing the base time of 15 minutes by the leakage factor for the source’s mask (if any), and then dividing that by the leakage factor for the receiver’s mask (if any). This approach is, however, an over-simplification. Reseach based on a a single-hit model of infection shows that the probability of infection “shows a highly nonlinear sensitivity” to inhaled virus number. Therefore, “In a virus-rich regime… wearing a mask may not suffice to prevent infection.” Research undertaken by the National Personal Protective Technology Laboratory (NPPTL) found that respirators with an exhalation valve “reduce particle emissions to levels similar to or better than those provided by surgical masks, procedure masks, or cloth face coverings”. Furthermore, “surgical tape secured over the valve from the inside of the FFR can provide source control similar to that of an FFR with no exhalation valve”. Pushing back against masks Professor Alison McMillan, Commonwealth Chief Nursing and Midwifery Officer in Australia claims that “there is no evidence to suggest that we should be moving towards… N95 respirators in the community setting.” She added “I am aware that there are some publications out there suggesting a move to N95 (masks). But that’s not supported in the empirical evidence”. According to Norman Swan, host of the ABC’s Coronacast, “If you’re wearing an N95 that hasn’t been fit tested – and it’s not an easy process to do yourself at home – there’s no guarantee that it’s an awful lot more effective than wearing a surgical mask. Professor Catherine Bennett, chair in epidemiology at Deakin University, claims that”Technically, the instructions say you shouldn’t reuse” respirators, and that “If you’re not particularly checking its fit, you’re probably wasting your time”. Occupational environment physician Malcolm Sim agrees: “If you put it [an N95 mask] off and put it on, they’re not meant for that purpose… They’re easily damaged in somebody’s handbag,” adding that the integrity of the masks can be compromised. He says that “If you’re handling them a lot, taking them on and off, there’s much more potential for you to get it [the virus] on your hands, your face, different parts of your body.” University of New South Wales epidemiologist Mary-Louise McLaws claimed that “There’s no evidence yet that a N95 mask will protect you more than a surgical mask for Omicron.” An opinion piece in Newsweek claims that “the effectiveness of respirators is vastly overestimated, and there is scant evidence that they stop community transmission. Moreover, NIOSH-approved respirators are tight, uncomfortable, and can impede breathing.” The article further claims that “For respirators to work, they must be well fitting, must be tested by OSHA, and must be used for only short time windows as their effectiveness diminishes as they get wet from breathing.” Recently there has been particular pushback against the use of masks by children, with the Newsweek article alleging that “Respirators are not necessary to protect children from COVID-19 because of the astoundingly low risk COVID-19 presents to them”, and that in fact wearing masks involves “existing well-documented harms”. There hasn’t been any documented harms to children from wearing masks, Respirators can be reused According to mask manufacturer 3M, respirators (which they refer to as “Filtering Facepiece Respirators (FFRs)”) “can be used many times.” They say that “There is no time limit to wearing an FFR. Respirators can be worn until they are dirty, damaged or difficult to breathe through.” In reporting from CNN, Linsey Marr, a professor of civil and environmental engineering at Virginia Tech, explained that an N95 mask’s material and filtration ability aren’t “going to degrade unless you physically rub it or poke holes in it.”You’d have to be in really polluted air … for several days before it lost its ability to filter out particles. So, you can really wear them for a long time. People have been talking about 40 hours – I think that’s fine. Really, it’s going to get gross from your face or the straps will get too loose or maybe break before you’re going to lose filtration ability… One of the first indicators of being able to change it if it looks nice and clean is that it just feels a little harder to breathe through. There appears to be more resistance with every breath.” She also noted that the contamination risk in reusing N95 masks is “lower, much lower, than the risk of you not wearing an N95 and breathing in particles”. The CDC has prepared guidelines for optimizing the supply of respirators which recommend reusing respirators at most five times. This guidelines were created for people “implementing policies and procedures for preventing pathogen transmission in healthcare settings”. They have been widely shared, incorrectly, by reporters as being recommendations for community use. The inventor of N95 mask material, Peter Tsai, says that “N95 masks can be rotated, 1 mask every 3–4 days”, and that in doing this “there is no change in the mask’s properties.” According to the NIOSH Guide to the Selection and Use of Particulate Respirators N95 respirators must maintain at least 95% filtration after a total mass loading of 200mg. This is designed to ensure they continue to work in sites with high particulate matter, such as some construction environment. However in normal use, even outside in a city with high levels of population, it would take over 200 days of 24 hour per day use to get to this level. The guide says that “generally, the use and reuse of N-series lters would also be subject only to considerations of hygiene, damage, and increased breathing resistance”. The NIOSH guidelines are well supported by research. Fit tests are not required for respirators to be effective In one study non-experts were asked to read the instructions that come with a respirator, and then to don the respirator without assistance and complete a fit test. The average fit factor achieved was 88, and the lowest fit factor of the subjects was 15, with nearly half achieving a fit factor greater than 100. Surgical masks have been found to have a much poorer fit in practice. One study showed that for surgical masks “quantitative fit factors ranged from 2.5 to 9.6”, and another found an average fit factor of 3.0. Guidance from the US Food and Drug Administration (FDA) explains that: “Fit Factor is a means of expressing the difference in particle concentration inside the mask and outside the mask during use. For example, a fit factor of 2 means that the concentration of particles within the mask is ½ or 50% of the concentration outside the mask; a fit factor of 5 means the concentration of particles within the mask is 1/5 th or 20% of the concentration outside the mask.” The guidance says that failing to achieve a fit factor of 2 “may suggest that respirator fit will not be sufficient to assure that the device will help reduce wearer exposure to pathogenic biological airborne particulates.” An analysis of the fitted filtration efficiency (FFE) of surgical masks found that, unmodified, they only achieved an FFE of 38.5%. The “knot and tuck” technique improved that to 60.3%, and a DIY mask fitter consisting of three rubber bands increased it to 78.2%. A 3-layer cotton mask had an FFE of just 26.5%. An N95, on the other hand, achieved an FFE of 98.4%. Furthermore, the N95 FFE had a standard deviation of only 0.5% — that is, it was effective for multiple tests during “a series of repeated movements of the torso, head, and facial muscles”. Interestingly, a 2-layer nylon mask had an FFE of 79.0% (standard devatiation 4.3%), showing that some cloth masks can be quite effective. These findings were replicated in a study of numerous types of cloth mask, which found that hybrids of 600 TPI cotton with silk, chiffon, or polypropelene achieved 72-96% filtration efficiency. Researchers have calculated that “the particle size most likely to deposit in the respiratory tract when wearing a mask is ∼2μm”. Unfortunately, this particle size is not considered in N95 or similar standards. Instead, 0.3 μm particles are used. A 2010 study of fit testing respirators for public health medical emergencies found that 98% of non-experts wearing masks without training achieved a fit factor of over 5 (20% leakage) and 75% of them achieved a fit factor of over 10 (10% leakage). Donning and doffing masks is not complex or risky Analysis by the CDC concludes that the risk of infection through surfaces (fomites) “is generally considered to be low”, a view that was supported by the evidence as early as July 2020. An analysis of “418 samples from mask fronts, cell phones, paper money, card machines, sewage, air and bedding” during a COVID surge “did not detect any trace of SARS-CoV-2 in all samples analyzed”. We should not reserve respirators for healthcare workers According to Anne Miller, executive director of Project N95, there are many U.S. manufacturers of N95 masks and an ample supply. The Economist reported that in Europe “at the start of the pandemic, FFP2 masks were scarce and costly. Even governments fell victim to price gouging, paying more than €4 ($4.50) per mask. Demand had previously been low, so stockpiles and production capacity could not satisfy the sudden surge. Governments wanted to reserve supplies for those most at risk of contracting the virus, such as health-care workers.” However they reported that by the end of 2021 “FFP2 masks are in healthy supply, and as the highly transmissible Omicron variant spreads across the world, updating guidance to recommend their wider use could be one way to help reduce transmission.” In the first 6 months of 2020, over 70,000 new face mask companies were registered in China, many run by people with no previous experience and no registration or licensing. The Chinese government stepped in to make licensing more stringent, shutting down many companies, and international demand fell over quality concerns. Due to “a dramatic reduction in demand for N95s”, US mask factories are closing. In June 2021 the American Mask Manufacturer’s Association said that “we have 28 members who are going to go out of business in the next 60 to 90 days.” By July 2021 they estimated “that 5,000 workers have been laid off across its member companies”. However following school mask mandates and demand during the omicron surge, demand in the US spiked in early 2022. The CDC has found that 60% of KN95s are counterfeit. In Australia it has been reported that “general practitioners have been left without highly protective N95 masks as consumers rush to stock up after a sharp rise in COVID-19 cases.” In May 2021 the CDC stated that “The supply and availability of NIOSH-approved respirators have increased significantly over the last several months. Healthcare facilities should not be using crisis capacity strategies at this time and should promptly resume conventional practices.” Demand distortions can increase as we proceed up the supply chain, creating inefficiencies for upstream firms. This is known as the Bullwhip Effect. Respirators need not be uncomfortable In an analysis of the physiological impact of the N95 filtering facepiece respirator (FFR) “in healthy healthcare workers, FFR did not impose any important physiological burden during 1 hour of use, at realistic clinical work rates”. A study of KF80, KF94, KF99, N95, and N99 masks found that self-reported comfort levels were nearly perfectly correlated with the ease of inhalation. |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-06-01-qualitative.html] | [TOKENS: 2835] |
Qualitative humanities research is crucial to AI Louisa Bartolo and Rachel Thomas June 1, 2022 On this page “All research is qualitative; some is also quantitative” Harvard Social Scientist and Statistician Gary King Suppose you wanted to find out whether a machine learning system being adopted - to recruit candidates, lend money, or predict future criminality - exhibited racial bias. You might calculate model performance across groups with different races. But how was race categorised– through a census record, a police officer’s guess, or by an annotator? Each possible answer raises another set of questions. Following the thread of any seemingly quantitative issue around AI ethics quickly leads to a host of qualitative questions. Throughout AI, qualitative decisions are made about what metrics to optimise for, which categories to use, how to define their bounds, who applies the labels. Similarly, qualitative research is necessary to understand AI systems operating in society: evaluating system performance beyond what can be captured in short term metrics, understanding what is missed by large-scale studies (which can elide details and overlook outliers), and shedding light on the circumstances in which data is produced (often by crowd-sourced or poorly paid workers). Unfortunately, there is often a large divide between computer scientists and social scientists, with over-simplified assumptions and fundamental misunderstandings of one another. Even when cross-disciplinary partnerships occur, they often fall into “normal disciplinary divisions of labour: social scientists observe, data scientists make; social scientists do ethics, data scientists do science; social scientists do the incalculable, data scientists do the calculable.” The solution is not for computer scientists to absorb a shallow understanding of the social sciences, but for deeper collaborations. In a paper on exclusionary practices in AI ethics, an interdisciplinary team wrote of the “indifference, devaluation, and lack of mutual support between CS and humanistic social science (HSS), [which elevates] the myth of technologists as ‘ethical unicorns’ that can do it all, though their disciplinary tools are ultimately limited.” This is further reflected in an increasing number of job ads for AI ethicists that list a computer science degree as a requirement, “prioritising technical computer science infrastructure over the social science skills that can evaluate AI’s social impact. In doing so, we are building the field of AI Ethics to replicate the very flaws this field is trying to fix.” Interviews with 26 responsible AI practitioners working in industry highlighted a number of challenges, including that qualitative work was not prioritised. Not only is it impossible to fully understand ethics issues solely through quantitative metrics, inappropriate and misleading quantitative metrics are used to evaluate the responsible AI practitioners themselves. Interviewees reported that their fairness work was evaluated on metrics related to generating revenue, in a stark misalignment of goals. Qualitative research helps us evaluate AI systems beyond short term metrics When companies like Google and YouTube want to test whether the recommendations they are making (in the form of search engine results or YouTube videos, for example) are “good” - they will often focus quite heavily on “engagement” or “dwell time” - the time a user spent looking at or watching the item recommended to them. But it turns out, unsurprisingly, that a focus on engagement and dwell time, narrowly understood, raises all sorts of problems. Demographics can impact dwell time (e.g. older users may spend longer on websites than younger users, just as part of the way they use the internet). A system that ‘learns’ from a user’s behavioural cues (rather than their ‘stated preferences’) might lock them into a limiting feedback loop, appealing to that user’s short term interests rather than those of their ‘Better Selves.’ Scholars have called for more qualitative research to understand user experience and build this into the development of metrics. This is the part where people will point out, rightly, that companies like Google and YouTube rely on a complex range of metrics and signals in their machine learning systems - and that where a website ranks on Google, or how a YouTube video performs in recommendation does not boil down to simple popularity metrics, like engagement. Google employs an extensive process to determine “relevance” and “usefulness” for search results. In its 172-page manual for search result ‘Quality’ evaluation, for example, the company explains how evaluators should assess a website’s ‘Expertise/ Authoritativeness/ Trustworthiness’ or ‘E-A-T’; and what types of content, by virtue of its harmful nature (e.g., to protected groups), should be given a ‘low’ ranking. YouTube has identified specific categories of content (such as news, scientific subjects, and historical information) for which ‘authoritativeness’ should be considered especially important. It has also determined that dubious-but-not-quite-rule-breaking information (what it calls ‘borderline content’) should not be recommended, regardless of the video’s engagement levels. Irrespective of how successful we consider the existing approaches of Google Search and YouTube to be (and partly, the issue is that evaluating their implementation from the outside is frustratingly difficult), the point here is that there are constant qualitative judgments being made, about what makes a search result or recommendation “good” and of how to define and quantify expertise, authoritativeness, trustworthiness, borderline content, and other values. This is true of all machine learning evaluation, even when it isn’t explicit. In a paper guiding companies about how to carry out internal audits of their AI systems, Inioluwa Deborah Raji and colleagues emphasise the importance of interviews with management and engineering teams to “capture and pay attention to what falls outside the measurements and metrics, and to render explicit the assumptions and values the metrics apprehend.” (p.40). The importance of thoughtful humanities research is heightened if we are serious about grappling with the potential broader social effects of machine learning systems (both good and bad), which are often delayed, distributed and cumulative. Small-scale qualitative studies tell an important story even (and perhaps especially) when they seem to contradict large-scale ‘objective’ studies Hypothetically, let’s say you wanted to find out whether the use of AI technologies by doctors during a medical appointment would make doctors less attentive to patients - what do you think the best way of doing it would be? You could find some criteria and method for measuring ‘attentiveness’, say tracking the amount of eye contact between the doctor and patient, and analyse this across a representative sample of medical appointments where AI technologies were being used, compared to a control group of medical appointments where AI technologies weren’t being used. Or would you interview doctors about their experiences using the technology during appointments? Or talk to patients about how they felt the technology did, or didn’t, impact their experience? In research circles, we describe these as ‘epistemological’ choices - your judgement of what constitutes the ‘best’ approach is inextricably linked to your judgement about how we can claim to ‘know’ something. These are all valid methods for approaching the question, but you can imagine how they might result in different, even conflicting, insights. For example, you might end up with the following results: - The eye contact tracking experiment suggests that overall, there is no significant difference in doctors’ attentiveness to the patient when the AI tech is introduced. - The interviews with doctors and patients reveal that some doctors and patients feel that the AI technology reduces doctors’ attentiveness to patients, and others feel that it makes no difference or even increases doctors’ attention to the patient. Even if people are not negatively impacted by something ‘on average’ (e.g., in our hypothetical eye contact tracking experiment above), there will remain groups of people who will experience negative impacts, perhaps acutely so. “Many of people’s most pressing questions are about effects that vary for different people,” write Matias, Pennington and Chan in a recent paper on the idea of N-of-one trials. To tell people that their experiences aren’t real or valid because they don’t meet some threshold for statistical significance across a large population doesn’t help us account for the breadth and nature of AI’s impacts on the world. Examples of this tension between competing claims to knowledge about AI systems’ impacts abound. Influencers who believe they are being systematically downranked (‘shadowbanned’) by Instagram’s algorithmic systems are told by Instagram that this simply isn’t true. Given the inscrutability of these proprietary algorithmic systems, it is impossible for influencers to convincingly dispute Instagram’s claims. Kelley Cotter refers to this as a form of “black box gaslighting”: platforms can “leverage perceptions of their epistemic authority on their algorithms to undermine users’ confidence in what they know about algorithms and destabilise credible criticism.” Her interviews with influencers give voice to stakeholder concerns and perspectives that are elided in Instagram’s official narrative about its systems. The mismatch between different stakeholders’ accounts of ‘reality’ is instructive. For example, a widely-cited paper by Netflix employees claims that Netflix recommendation “influences choice for about 80% of hours streamed at Netflix.” But this claim stands in stark contrast to Mattias Frey’s mixed-methods research (representative survey plus small sample for interviews) run with UK and US adults, in which less than 1 in 5 adults said they primarily relied on Netflix recommendations when deciding what films to watch. Even if this is because users underestimate their reliance on recommender systems, that’s a critically important finding - particularly when we’re trying to regulate recommendation and so many are advocating providing better user-level controls as a check on platform power. Are people really going to go to the trouble of changing their settings if they don’t think they rely on algorithmic suggestions that much anyway? Qualitative research sheds light on the context of data annotation Machine learning systems rely on vast amounts of data. In many cases, for that data to be useful, it needs to be labelled/ annotated. For example, a hate speech classifier (an AI-enabled tool used to identify and flag potential cases of hate speech on a website) relies on huge datasets of text labelled as ‘hate speech’ or ‘not hate speech’ to ‘learn’ how to spot hate speech. But it turns out that who is doing the annotating and in what context they’re doing it, matters. AI-powered content moderation is often held up as the solution to harmful content online. What has continued to be underplayed is the extent to which those automated systems are and most likely will remain dependent on the manual work of human content moderators sifting through some of the worst and most traumatic online material to power the machine learning datasets on which automated content moderation depends. Emily Denton and her colleagues highlight the significance of annotators’ social identity (e.g., race, gender) and their expertise when it comes to annotation tasks, and they point out the risks associated with overlooking these factors and simply ‘aggregating’ results as ‘ground truth’ rather than properly exploring disagreements between annotators and the important insights that this kind of disagreement might offer. Human commercial content moderators (such as the people that identify and remove violent and traumatic imagery on Facebook) often labour in terrible conditions, lacking psychological support or appropriate financial compensation. The interview-based research of Sarah T. Roberts has been pioneering in highlighting these conditions. Most demand for crowdsourced digital labour comes from the Global North, yet the majority of these workers are based in the Global South and receive low wages. Semi-structured interviews reveal the extent to which workers feel unable to bargain effectively for better pay in the current regulatory environment. As Mark Graham and his colleagues point out, these findings are hugely important in a context where several governments and supranational development organisations like the World Bank are holding up digital work as a promising tool to fight poverty. The decision of how to measure ‘race’ in machine learning systems is highly consequential, especially in the context of existing efforts to evaluate these systems for their “fairness.” Alex Hanna, Emily Denton, Andrew Smart and Jamila Smith-Loud have done crucial work highlighting the limitation of machine learning systems that rely on official records of race or their proxies (e.g. census records), noting that the racial categories provided by such records are “unstable, contingent, and rooted in racial inequality.” The authors emphasise the importance of conducting research in ways that prioritise the perspectives of the marginalised racial communities that fairness metrics are supposed to protect. Qualitative research is ideally placed to contribute to a consideration of “race” in machine learning systems that is grounded in the lived experiences and needs of the racially subjugated. What next? Collaborations between quantitative and qualitative researchers are valuable in understanding AI ethics from all angles. Consider reading more broadly, outside your particular area. Perhaps using the links and researchers listed here as starting points. They’re just a sliver of the wealth that’s out there. You could also check out the Social Media Collective’s Critical Algorithm Studies reading list, the reading list provided by the LSE Digital Ethnography Collective, and Catherine Yeo’s suggestions. Strike up conversations with researchers in other fields, and consider the possibility of collaborations. Find a researcher slightly outside your field but whose work you broadly understand and like, and follow them on Twitter. With any luck, they will share more of their work and help you identify other researchers to follow. Collaboration can be an incremental process: Consider inviting the researcher to form part of a discussion panel, reach out to say what you liked and appreciated about their work and why, and share your own work with them if you think it’s aligned with their interests. Within your university or company, is there anything you could do to better reward or facilitate interdisciplinary work? As Humanities Computing Professor Willard McCarty notes, somewhat discouragingly, “professional reward for genuinely interdisciplinary research is rare.” To be sure, individual researchers and practitioners have to be prepared to put themselves out there, compromise and challenge themselves - but carefully tailored institutional incentives and enablers matter. |
======================================== |
[SOURCE: https://www.fast.ai/posts/2022-07-04-updated-masks-evidence.html] | [TOKENS: 5276] |
Masks for COVID: Updating the evidence Jeremy Howard July 4, 2022 On this page These are notes I took whilst preparing a paper on mask efficacy from Nov 2021 to Jan 2022. In the end, I gave up on the paper, because I felt like people had given up on masks, so there wasn’t much point in finishing it. I’ve decided to publish these notes in the hope some people will find them a useful starting point for their own research, and since I’ve noticed some signs in recent weeks that people might be open to avoiding COVID again. My previous paper on this topic, in which I led a team of 19 experts, was written in April 2020, and published here in the Proceedings of the National Academy of Science. The rise of better masks In the US, 400 million N95 masks are being distributed for free, coming from the 750 million stored in the US’ Strategic National Stockpile. A similar campaign to distribute 650 millions masks in the US in 2020 was cancelled. KN95 masks are being given to US congressional staff, and masks are required for federal workers and whilst in federal buildings. The Los Angeles school district has required students to upgrade from cloth masks to “well-fitting, non-cloth masks with a nose wire”. Masks work A review paper discussed both lab evidence and empirical evidence for the importance of face masks, with eight “seminal studies” showing a reduction in transmission when masks are used, and one Danish study of surgical masks with “several design limitations” which “demonstrated only a modest benefit in limiting COVID-19 transmission”. The authors note that “laboratory studies have demonstrated the ability of surgical masks to block SARS-COV-2 and other viruses”, with the masks “60%–70% effective at protecting others and 50% effective at protecting the wearer”. An evidence review from early in the pandemic concluded that “given the current shortages of medical masks, we recommend the adoption of public cloth mask wearing, as an effective form of source control”. It noted that “by the end of June 2020, nearly 90% of the global population lived in regions that had nearly universal mask use, or had laws requiring mask use in some public locations.” The review said that “There has been one controlled trial of mask use for influenza control in the general community. The study looked at Australian households, was not done during a pandemic, and was done without any enforcement of compliance” – and yet still found “masks as a group had protective efficacy in excess of 80% against clinical influenza-like illness.” An observational study of Beijing households analyzed the impact of mask use in the community on COVID-19 transmission, finding that face masks were 79% effective in preventing transmission, if used by all household members prior to symptoms occurring. One study used a multiple regression of policy interventions and country and population characteristics to infer the relationship between mask use and SARS-CoV-2 transmission. It found that transmission was around 7.5 times higher in countries that did not have a mask mandate or universal mask use, a result similar to that found in an analogous study of fewer countries. Similar results were found by numerous other papers. A mathematical model of mask use estimates that mask wearing reduces the reproduction number R by (1−mp)^2, where m is the efficacy of trapping viral particles inside the mask, and p is the percentage of the population that wears masks. A report in Nature explained that researchers running a randomized controlled trial (RCT) of community mask use in Bangladesh “began by developing a strategy to promote mask wearing, with measures such as reminders from health workers in public places. This ultimately tripled mask usage, from only 13% in control villages to 42% in villages where it was encouraged”, and “then compared numbers of COVID-19 cases in control villages and the treatment communities”. They found that the number of infections in mask wearing communities decreased, with a reduction of COVID symptoms using surgical masks to 0.87 times the incidence in unmasked communities, and 0.91 times when using cloth masks. The report noted that “the researchers suggest that the true risk reduction is probably much greater, in part because they did no SARS-CoV-2 testing of people without symptoms or whose symptoms did not meet the World Health Organization’s definition of the disease.” The researchers concluded that “promoting community mask-wearing can improve public health”. The Johns Hopkins School of Public Health reviewed the work and concluded that “This study is the largest and best-designed randomized controlled trial to date of a realistic non-pharmaceutical intervention on SARS-CoV-2 transmission.” A paper investigating an upper bound on one-to-one exposure to infectious human respiratory particles concludes that “face masks significantly reduce the risk of SARS-CoV-2 infection compared to social distancing. We find a very low risk of infection when everyone wears a face mask, even if it doesn’t fit perfectly on the face.” They calculate that “social distancing alone, even at 3.0 m between two speaking individuals, leads to an upper bound of 90% for risk of infection after a few minutes”, but that when both source and susceptible wear a well-fitting FFP2 mask, there is only 0.4% after one hour of contact. They found that to achieve good fit it is important to mold the nose piece wire to the size of the nose, rather than leaving it in a sharp folded position. A similar study “quantifies the extent to which transmission risk is reduced in large rooms with high air exchange rates, increased for more vigorous respiratory activities, and dramatically reduced by the use of face masks.” The authors describe the six-foot rule widely used to ensure social distancing as “a guideline that offers little protection from pathogen-bearing aerosol droplets sufficiently small to be continuously mixed through an indoor space.” Instead, they develop a safety guideline based on cumulative exposure time,” the product of the number of occupants and their time in an enclosed space. In particular, they identify that the greatest risk comes in places where people are speaking (other than quietly) or singing, and that “the benefit of face masks is immediately apparent”, due to the multiplicative effect when both source and susceptible wear a mask. They further note that “Air filtration has a less dramatic effect than face mask use in increasing the CET bound. Nevertheless, it does offer a means of mitigating indoor transmission with greater comfort, albeit at greater cost.” Another study of the combined impacts of ventilation and mask effective filtration efficiency in classroom settings found that “ventilation alone is not able to achieve probabilities <0.01 (1%)” of transmitting COVID in a classroom. However, they found that good masks reduce infection probability by >5× in some cases, and that “reductions provided by ventilation and masks are synergistic and multiplicative”. However they also noted that “most masks fit poorly”, recommending that work be done to ensure that high quality masks are used. Similar results were found in a study of community public health interventions, which concluded that “control the pandemic, our models suggest that high adherence to social distance is necessary to curb the spread of the disease, and that wearing face masks provides optimal protection even if only a small portion of the population comply with social distance”. Guidance from the independent scientific advisory group OzSAGE points out “that school children are able to wear masks. As an example, all children over two years of age in San Francisco are required to wear masks at school”. Omicron changes the game An analysis of fine aerosol emissions found that, compared to the original wild type (WT) virus: “Delta and Omicron both also have increased transmissibility: the number of cells infected for a given number of ribonucleic acid (RNA) virus copies was found to be doubled and quadrupled respectively. Furthermore, Omicron also seems to be better at evading the immune system. This implies that the critical dose of virus copies above which a situation is potentially infectious needs to be lowered. For the WT, we had proposed a critical dose of 500 virus copies. If the above-mentioned capacity to infect cells translates into an infection risk, this would imply a critical dose of around 300 virus copies for Delta and around 100 virus copies for Omicron.” The study finds that “surgical masks are no longer sufficient in most public settings, while correctly fitted FFP2 respirators still provide sufficient protection, except in high aerosol producing situations such as singing or shouting.” Data from Hong Kong shows that “Omicron SARS-CoV-2 infects and multiplies 70 times faster than the Delta variant and original SARS-CoV-2 in human bronchus”. A study of transmission in Danish households estimated the secondary attack rate (SAR) of omicron compared to delta, finding it 1.2 times higher for unvaccinated people, 2.6 times higher for double-dosed, and 3.7 times higher for boosted. The authors conclude that “the rapid spread of the Omicron VOC primarily can be ascribed to the immune evasiveness”. According to UK statistics, the risk of hospitalization from omicron when unvaccinated is about the same as the wildtype virus, which is about half the risk of the delta variant. The journal Infection Control Today reported that many experts are concerned that “‘Omicron the Pandemic Killer’ Idea Ignores Dangers of Long COVID”: “Linda Spaulding, RN-BC, CIC, CHEC, CHOP, a member of Infection Control Today®’s Editorial Advisory Board (EAB), says that she’s “seen athletes in their 20s on the wait list for double lung transplants because of long COVID. That’s something that has long-term consequences. Some people talk of COVID fog. They just can’t put their thoughts together.” In addition, even the treatments for those with long COVID can put toll on a patient’s body.” “As noted by Kevin Kavanagh, MD, another member of ICT®’s EAB, a core difficulty in society’s attempt to guide COVID-19 from pandemic to endemic is that COVID is not just a respiratory virus. Kavanagh wrote in October that SARS-CoV-2 is similar to HIV because it can “silently spread throughout the host’s body and attack almost every organ.”” Better masks work better The US Centers for Disease Control and Prevention (CDC) explains that: “Loosely woven cloth products provide the least protection, layered finely woven products offer more protection, well-fitting disposable surgical masks and KN95s offer even more protection, and well-fitting NIOSH-approved respirators (including N95s) offer the highest level of protection.” Unfortunately “well-fitting disposable surgical masks” do not exist out of the box, since there are large gaps on each side of the mask. Surgical masks require modifications to achive a good fit. That’s because they are made to stop liquid splashes during surgery, rather than made to stop airborne transmission. There are two methods shown by the CDC to improve fit: Research shows that both of these approaches dramatically reduce exposure to aerosols emitted during a period of breathing: “…adding a cloth mask over the source headform’s medical procedure mask or knotting and tucking the medical procedure mask reduced the cumulative exposure of the unmasked receiver by 82.2% (SD = 0.16) and 62.9% (SD = 0.08), respectively (Figure 2). When the source was unmasked and the receiver was fitted with the double mask or the knotted and tucked medical procedure mask, the receiver’s cumulative exposure was reduced by 83.0% (SD = 0.15) and 64.5% (SD = 0.03), respectively. When the source and receiver were both fitted with double masks or knotted and tucked masks, the cumulative exposure of the receiver was reduced 96.4% (SD = 0.02) and 95.9% (SD = 0.02), respectively.” An airborne transmission simulator was used to estimate the ability of various types of face masks to block COVID-19 transmission. In this experiment, “cotton mask led to an approximately 20% to 40% reduction in virus uptake compared to no mask. The N95 mask had the highest protective efficacy (approximately 80% to 90% reduction)”. All of the masks were much more effective at source control than at protecting the wearer, with the N95 stopping all detectable transmission. The American Conference of Governmental Industrial Hygienists (ACGIH) say that “workers need respirators”, noting that a worker with an “N95 filtering facepiece respirator… has 1-10% inward leakage and outward leakage”, but with a surgical mask “has 50% inward leakage and outward leakage”, and with a cloth face covering “has 75% inward leakage and outward leakage”. They explain that “N95 FFRs have an assigned protection factor of 10 (10% inward leakage) but must receive a fit factor of 100 (1% inward leakage) on an individual worker.” ACGIH created a table showing how, if we start with an assumption that it takes on average 15 minutes to get infected if no-one is wearing a mask (based on CDC contact tracing premises), we can calculate the time it would take on average to get infected if one or both of source and receiver are wearing various types of mask. This is calculated by simply dividing the base time of 15 minutes by the leakage factor for the source’s mask (if any), and then dividing that by the leakage factor for the receiver’s mask (if any). This approach is, however, an over-simplification. Reseach based on a a single-hit model of infection shows that the probability of infection “shows a highly nonlinear sensitivity” to inhaled virus number. Therefore, “In a virus-rich regime… wearing a mask may not suffice to prevent infection.” Research undertaken by the National Personal Protective Technology Laboratory (NPPTL) found that respirators with an exhalation valve “reduce particle emissions to levels similar to or better than those provided by surgical masks, procedure masks, or cloth face coverings”. Furthermore, “surgical tape secured over the valve from the inside of the FFR can provide source control similar to that of an FFR with no exhalation valve”. Pushing back against masks Professor Alison McMillan, Commonwealth Chief Nursing and Midwifery Officer in Australia claims that “there is no evidence to suggest that we should be moving towards… N95 respirators in the community setting.” She added “I am aware that there are some publications out there suggesting a move to N95 (masks). But that’s not supported in the empirical evidence”. According to Norman Swan, host of the ABC’s Coronacast, “If you’re wearing an N95 that hasn’t been fit tested – and it’s not an easy process to do yourself at home – there’s no guarantee that it’s an awful lot more effective than wearing a surgical mask. Professor Catherine Bennett, chair in epidemiology at Deakin University, claims that”Technically, the instructions say you shouldn’t reuse” respirators, and that “If you’re not particularly checking its fit, you’re probably wasting your time”. Occupational environment physician Malcolm Sim agrees: “If you put it [an N95 mask] off and put it on, they’re not meant for that purpose… They’re easily damaged in somebody’s handbag,” adding that the integrity of the masks can be compromised. He says that “If you’re handling them a lot, taking them on and off, there’s much more potential for you to get it [the virus] on your hands, your face, different parts of your body.” University of New South Wales epidemiologist Mary-Louise McLaws claimed that “There’s no evidence yet that a N95 mask will protect you more than a surgical mask for Omicron.” An opinion piece in Newsweek claims that “the effectiveness of respirators is vastly overestimated, and there is scant evidence that they stop community transmission. Moreover, NIOSH-approved respirators are tight, uncomfortable, and can impede breathing.” The article further claims that “For respirators to work, they must be well fitting, must be tested by OSHA, and must be used for only short time windows as their effectiveness diminishes as they get wet from breathing.” Recently there has been particular pushback against the use of masks by children, with the Newsweek article alleging that “Respirators are not necessary to protect children from COVID-19 because of the astoundingly low risk COVID-19 presents to them”, and that in fact wearing masks involves “existing well-documented harms”. There hasn’t been any documented harms to children from wearing masks, Respirators can be reused According to mask manufacturer 3M, respirators (which they refer to as “Filtering Facepiece Respirators (FFRs)”) “can be used many times.” They say that “There is no time limit to wearing an FFR. Respirators can be worn until they are dirty, damaged or difficult to breathe through.” In reporting from CNN, Linsey Marr, a professor of civil and environmental engineering at Virginia Tech, explained that an N95 mask’s material and filtration ability aren’t “going to degrade unless you physically rub it or poke holes in it.”You’d have to be in really polluted air … for several days before it lost its ability to filter out particles. So, you can really wear them for a long time. People have been talking about 40 hours – I think that’s fine. Really, it’s going to get gross from your face or the straps will get too loose or maybe break before you’re going to lose filtration ability… One of the first indicators of being able to change it if it looks nice and clean is that it just feels a little harder to breathe through. There appears to be more resistance with every breath.” She also noted that the contamination risk in reusing N95 masks is “lower, much lower, than the risk of you not wearing an N95 and breathing in particles”. The CDC has prepared guidelines for optimizing the supply of respirators which recommend reusing respirators at most five times. This guidelines were created for people “implementing policies and procedures for preventing pathogen transmission in healthcare settings”. They have been widely shared, incorrectly, by reporters as being recommendations for community use. The inventor of N95 mask material, Peter Tsai, says that “N95 masks can be rotated, 1 mask every 3–4 days”, and that in doing this “there is no change in the mask’s properties.” According to the NIOSH Guide to the Selection and Use of Particulate Respirators N95 respirators must maintain at least 95% filtration after a total mass loading of 200mg. This is designed to ensure they continue to work in sites with high particulate matter, such as some construction environment. However in normal use, even outside in a city with high levels of population, it would take over 200 days of 24 hour per day use to get to this level. The guide says that “generally, the use and reuse of N-series lters would also be subject only to considerations of hygiene, damage, and increased breathing resistance”. The NIOSH guidelines are well supported by research. Fit tests are not required for respirators to be effective In one study non-experts were asked to read the instructions that come with a respirator, and then to don the respirator without assistance and complete a fit test. The average fit factor achieved was 88, and the lowest fit factor of the subjects was 15, with nearly half achieving a fit factor greater than 100. Surgical masks have been found to have a much poorer fit in practice. One study showed that for surgical masks “quantitative fit factors ranged from 2.5 to 9.6”, and another found an average fit factor of 3.0. Guidance from the US Food and Drug Administration (FDA) explains that: “Fit Factor is a means of expressing the difference in particle concentration inside the mask and outside the mask during use. For example, a fit factor of 2 means that the concentration of particles within the mask is ½ or 50% of the concentration outside the mask; a fit factor of 5 means the concentration of particles within the mask is 1/5 th or 20% of the concentration outside the mask.” The guidance says that failing to achieve a fit factor of 2 “may suggest that respirator fit will not be sufficient to assure that the device will help reduce wearer exposure to pathogenic biological airborne particulates.” An analysis of the fitted filtration efficiency (FFE) of surgical masks found that, unmodified, they only achieved an FFE of 38.5%. The “knot and tuck” technique improved that to 60.3%, and a DIY mask fitter consisting of three rubber bands increased it to 78.2%. A 3-layer cotton mask had an FFE of just 26.5%. An N95, on the other hand, achieved an FFE of 98.4%. Furthermore, the N95 FFE had a standard deviation of only 0.5% — that is, it was effective for multiple tests during “a series of repeated movements of the torso, head, and facial muscles”. Interestingly, a 2-layer nylon mask had an FFE of 79.0% (standard devatiation 4.3%), showing that some cloth masks can be quite effective. These findings were replicated in a study of numerous types of cloth mask, which found that hybrids of 600 TPI cotton with silk, chiffon, or polypropelene achieved 72-96% filtration efficiency. Researchers have calculated that “the particle size most likely to deposit in the respiratory tract when wearing a mask is ∼2μm”. Unfortunately, this particle size is not considered in N95 or similar standards. Instead, 0.3 μm particles are used. A 2010 study of fit testing respirators for public health medical emergencies found that 98% of non-experts wearing masks without training achieved a fit factor of over 5 (20% leakage) and 75% of them achieved a fit factor of over 10 (10% leakage). Donning and doffing masks is not complex or risky Analysis by the CDC concludes that the risk of infection through surfaces (fomites) “is generally considered to be low”, a view that was supported by the evidence as early as July 2020. An analysis of “418 samples from mask fronts, cell phones, paper money, card machines, sewage, air and bedding” during a COVID surge “did not detect any trace of SARS-CoV-2 in all samples analyzed”. We should not reserve respirators for healthcare workers According to Anne Miller, executive director of Project N95, there are many U.S. manufacturers of N95 masks and an ample supply. The Economist reported that in Europe “at the start of the pandemic, FFP2 masks were scarce and costly. Even governments fell victim to price gouging, paying more than €4 ($4.50) per mask. Demand had previously been low, so stockpiles and production capacity could not satisfy the sudden surge. Governments wanted to reserve supplies for those most at risk of contracting the virus, such as health-care workers.” However they reported that by the end of 2021 “FFP2 masks are in healthy supply, and as the highly transmissible Omicron variant spreads across the world, updating guidance to recommend their wider use could be one way to help reduce transmission.” In the first 6 months of 2020, over 70,000 new face mask companies were registered in China, many run by people with no previous experience and no registration or licensing. The Chinese government stepped in to make licensing more stringent, shutting down many companies, and international demand fell over quality concerns. Due to “a dramatic reduction in demand for N95s”, US mask factories are closing. In June 2021 the American Mask Manufacturer’s Association said that “we have 28 members who are going to go out of business in the next 60 to 90 days.” By July 2021 they estimated “that 5,000 workers have been laid off across its member companies”. However following school mask mandates and demand during the omicron surge, demand in the US spiked in early 2022. The CDC has found that 60% of KN95s are counterfeit. In Australia it has been reported that “general practitioners have been left without highly protective N95 masks as consumers rush to stock up after a sharp rise in COVID-19 cases.” In May 2021 the CDC stated that “The supply and availability of NIOSH-approved respirators have increased significantly over the last several months. Healthcare facilities should not be using crisis capacity strategies at this time and should promptly resume conventional practices.” Demand distortions can increase as we proceed up the supply chain, creating inefficiencies for upstream firms. This is known as the Bullwhip Effect. Respirators need not be uncomfortable In an analysis of the physiological impact of the N95 filtering facepiece respirator (FFR) “in healthy healthcare workers, FFR did not impose any important physiological burden during 1 hour of use, at realistic clinical work rates”. A study of KF80, KF94, KF99, N95, and N99 masks found that self-reported comfort levels were nearly perfectly correlated with the ease of inhalation. |
======================================== |
[SOURCE: https://en.wikipedia.org/wiki/Language_model#cite_note-37] | [TOKENS: 1793] |
Contents Language model A language model is a computational model that predicts sequences in natural language. Language models are useful for a variety of tasks, including speech recognition, machine translation, natural language generation (generating more human-like text), optical character recognition, route optimization, handwriting recognition, grammar induction, and information retrieval. Large language models (LLMs), currently their most advanced form as of 2019, are predominantly based on transformers trained on larger datasets (frequently using texts scraped from the public internet). They have superseded recurrent neural network-based models, which had previously superseded the purely statistical models, such as the word n-gram language model. History Noam Chomsky did pioneering work on language models in the 1950s by developing a theory of formal grammars. In 1980, statistical approaches were explored and found to be more useful for many purposes than rule-based formal grammars. Discrete representations like word n-gram language models, with probabilities for discrete combinations of words, made significant advances. In the 2000s, continuous representations for words, such as word embeddings, began to replace discrete representations. Typically, the representation is a real-valued vector that encodes a word’s meaning such that words closer in vector space are similar in meaning and common relationships between words, such as plurality or gender, are preserved. Pure statistical models In 1980, the first significant statistical language model was proposed, and during the decade IBM performed 'Shannon-style' experiments, in which potential sources for language modeling improvement were identified by observing and analyzing the performance of human subjects in predicting or correcting text. A word n-gram language model is a statistical model of language which calculates the probability of the next word in a sequence from a fixed size window of previous words. If one previous word is considered, it is a bigram model; if two words, a trigram model; if n − 1 words, an n-gram model. Special tokens are introduced to denote the start and end of a sentence ⟨ s ⟩ {\displaystyle \langle s\rangle } and ⟨ / s ⟩ {\displaystyle \langle /s\rangle } . To prevent a zero probability being assigned to unseen words, the probability of each seen word is slightly lowered to make room for the unseen words in a given corpus. To achieve this, various smoothing methods are used, from simple "add-one" smoothing (assigning a count of 1 to unseen n-grams, as an uninformative prior) to more sophisticated techniques, such as Good–Turing discounting or back-off models. Word n-gram models have largely been superseded by recurrent neural network–based models, which in turn have been superseded by Transformer-based models often referred to as large language models. Maximum entropy language models encode the relationship between a word and the n-gram history using feature functions. The equation is P ( w m ∣ w 1 , … , w m − 1 ) = 1 Z ( w 1 , … , w m − 1 ) exp ( a T f ( w 1 , … , w m ) ) {\displaystyle P(w_{m}\mid w_{1},\ldots ,w_{m-1})={\frac {1}{Z(w_{1},\ldots ,w_{m-1})}}\exp(a^{T}f(w_{1},\ldots ,w_{m}))} where Z ( w 1 , … , w m − 1 ) {\displaystyle Z(w_{1},\ldots ,w_{m-1})} is the partition function, a {\displaystyle a} is the parameter vector, and f ( w 1 , … , w m ) {\displaystyle f(w_{1},\ldots ,w_{m})} is the feature function. In the simplest case, the feature function is just an indicator of the presence of a certain n-gram. It is helpful to use a prior on a {\displaystyle a} or some form of regularization. The log-bilinear model is another example of an exponential language model. Skip-gram language model is an attempt at overcoming the data sparsity problem that the preceding model (i.e. word n-gram language model) faced. Words represented in an embedding vector were not necessarily consecutive anymore, but could leave gaps that are skipped over (thus the name "skip-gram"). Formally, a k-skip-n-gram is a length-n subsequence where the components occur at distance at most k from each other. For example, in the input text: the set of 1-skip-2-grams includes all the bigrams (2-grams), and in addition the subsequences In skip-gram model, semantic relations between words are represented by linear combinations, capturing a form of compositionality. For example, in some such models, if v is the function that maps a word w to its n-d vector representation, then v ( k i n g ) − v ( m a l e ) + v ( f e m a l e ) ≈ v ( q u e e n ) {\displaystyle v(\mathrm {king} )-v(\mathrm {male} )+v(\mathrm {female} )\approx v(\mathrm {queen} )} where ≈ is made precise by stipulating that its right-hand side must be the nearest neighbor of the value of the left-hand side. Neural models Continuous representations or embeddings of words are produced in recurrent neural network-based language models (known also as continuous space language models). Such continuous space embeddings help to alleviate the curse of dimensionality, which is the consequence of the number of possible sequences of words increasing exponentially with the size of the vocabulary, further causing a data sparsity problem. Neural networks avoid this problem by representing words as non-linear combinations of weights in a neural net. A large language model (LLM) is a language model trained with self-supervised machine learning on a vast amount of text, designed for natural language processing tasks, especially language generation. The largest and most capable LLMs are generative pre-trained transformers (GPTs) that provide the core capabilities of modern chatbots. LLMs can be fine-tuned for specific tasks or guided by prompt engineering. These models acquire predictive power regarding syntax, semantics, and ontologies inherent in human language corpora, but they also inherit inaccuracies and biases present in the data they are trained on. They consist of billions to trillions of parameters and operate as general-purpose sequence models, generating, summarizing, translating, and reasoning over text. LLMs represent a significant new technology in their ability to generalize across tasks with minimal task-specific supervision, enabling capabilities like conversational agents, code generation, knowledge retrieval, and automated reasoning that previously required bespoke systems. LLMs evolved from earlier statistical and recurrent neural network approaches to language modeling. The transformer architecture, introduced in 2017, replaced recurrence with self-attention, allowing efficient parallelization, longer context handling, and scalable training on unprecedented data volumes. This innovation enabled models like GPT, BERT, and their successors, which demonstrated emergent behaviors at scale, such as few-shot learning and compositional reasoning. Reinforcement learning, particularly policy gradient algorithms, has been adapted to fine-tune LLMs for desired behaviors beyond raw next-token prediction. Reinforcement learning from human feedback (RLHF) applies these methods to optimize a policy, the LLM's output distribution, against reward signals derived from human or automated preference judgments. This has been critical for aligning model outputs with user expectations, improving factuality, reducing harmful responses, and enhancing task performance. Benchmark evaluations for LLMs have evolved from narrow linguistic assessments toward comprehensive, multi-task evaluations measuring reasoning, factual accuracy, alignment, and safety. Hill climbing, iteratively optimizing models against benchmarks, has emerged as a dominant strategy, producing rapid incremental performance gains but raising concerns of overfitting to benchmarks rather than achieving genuine generalization or robust capability improvements. Although sometimes matching human performance, it is not clear whether they are plausible cognitive models. At least for recurrent neural networks, it has been shown that they sometimes learn patterns that humans do not, but fail to learn patterns that humans typically do. Evaluation and benchmarks Evaluation of the quality of language models is mostly done by comparison to human created sample benchmarks created from typical language-oriented tasks. Other, less established, quality tests examine the intrinsic character of a language model or compare two such models. Since language models are typically intended to be dynamic and to learn from data they see, some proposed models investigate the rate of learning, e.g., through inspection of learning curves. Various data sets have been developed for use in evaluating language processing systems. These include: See also References Further reading |
======================================== |
[SOURCE: https://en.wikipedia.org/wiki/King_(company)] | [TOKENS: 3584] |
Contents King (company) King.com Limited is a Swedish video game developer and publisher that specialises in social games. Since 2016, it is one of 3 publishing businesses of Activision Blizzard. Headquartered in Stockholm and London, and incorporated in St. Julian's, Malta, King rose to prominence after releasing the cross-platform title Candy Crush Saga in 2012. It is considered as one of the most financially successful games utilising the freemium model. King was acquired by Activision Blizzard in February 2016 for US$5.9 billion, and operates as its own entity within that company. King is led by Todd Green, who holds the position of President. Gerhard Florin took over Melvyn Morris's role as chairman in November 2014. As of 2017, King employs 2,000 people. In October 2023, Microsoft acquired parent company Activision Blizzard, maintaining that the company will continue to operate as a separate business. While part of the larger Microsoft Gaming division, King retains its function as the publisher of games developed by themselves. History Prior to founding King, Riccardo Zacconi and Toby Rowland, the latter of whom is the only son of British businessman Tiny Rowland, had worked together on uDate.com, a dating website created by Melvyn Morris which, by 2003, was the second-largest such site in the world. Morris opted to sell the site to the leading dating website Match.com (a subsidiary of IAC) for $150 million in 2003. Zacconi and Rowland joined with Thomas Hartwig, Sebastian Knutsson, Lars Markgren and Patrik Stymne, all of whom had worked previously with Zacconi at the failed dot-com web portal Spray, to create a new company with angel investment provided by Morris, who became the company's chairman. The company was initially based out of Stockholm, Sweden, and started with the development of browser-based video games. The site, Midasplayer.com, was then launched in August of that year. Initially, Midasplayer.com was not profitable, and nearly went bankrupt until a cash infusion from Morris on Christmas Eve of 2003 helped to finance the company. By 2005, the company had been able to turn a profit. During this year, the company raised $43 million by selling a large stake to Apax Partners and Index Ventures. This investment was the last one that the company received before its initial public offering in 2014. Midasplayer.com was rebranded King.com in November 2005. King.com continued to develop games for its web portal, which it would also share to other web portals like Yahoo! Overall, King had developed about 200 games for their portal. By 2009, the company was making about $60 million annually. Rowland departed the company in 2008 to found Mangahigh, a web portal aimed for educational math games, and sold his stake back to the company for $3 million in 2011. Angel investor and former board member Klaus Hommels sold his similar stake at the same time. Around 2009, social network games on Facebook began to gain popularity, led primarily through games developed by Zynga. King.com saw a significant drop in players on their portal games as a result, and started to develop their own Facebook-based games using the games already developed on the King.com portal, with their first such game released in 2010. King.com used their web portal as a testing ground for new game ideas and determine which ones to bring to Facebook, as well as determining how to implement various microtransactions for tournament-style play into the Facebook games. Their first cross-platform web portal/Facebook game, Miner Speed, which allowed sharing of player information between platforms, was released in 2011, and was a simple match-3 tile game inspired by Bejeweled. Following this model, in October 2011, the company released Bubble Witch Saga to both platforms. Bubble Witch Saga introduced the nature of a "saga" game: instead of playing the same gameboard for as long as the player could continue to match matches, the game offered individual levels that would challenge the player to complete certain goals in a limited number of turns. These saga elements allowed for the basics of social gameplay, but did not require the time investment that then-popular titles like Zynga's Farmville required; players could play just for a few minutes each day through the saga model. The formula proved extremely successful, and January 2012, Bubble Witch Saga had over 10 million players and was one of the most-played Facebook games. By April 2012, King.com had the second largest player count, around 30 million unique users, second only to Zynga on the Facebook platform. Facebook's director of games partnerships Sean Ryan described King.com's growth on the platform as "They were not a flash in the pan – they've been around seven years. But they came out of no where in an area that was unexpected." King.com next released Candy Crush Saga in April 2012, based on the popularity of its Candy Crush web-portal game and following the saga model from Bubble Witch Saga. The game attracted more than 4 million players within a few weeks. The popularity of Bubble Witch Saga and Candy Crush Saga led King.com to start a new strategy into developing for the growing mobile game market, in a manner that would allow players to synchronise with the Facebook platform. Zacconi said that "As consumers and the industry focus more on games for mobile devices, launching a truly cross-platform Facebook game has been a top priority for King.com." A mobile version for iOS device of Bubble Witch Saga was released in July 2012, while the iOS mobile version of Candy Crush Saga was released in October 2012. Both games saw boosts in the number of unique players with the mobile introduction; King.com saw that previously-declining player counts for Bubble Witch Saga become steady with the mobile version's release, while Candy Crush Saga saw more than 5.2 million unique players on Facebook in November 2012 and which were continuing to climb. Additionally, in-game advertising, which factored into about 15% of King.com's revenues, had increased ten-fold from 2011 into 2012. Users jumped to 408 million by the end of 2013. Revenues for King.com increased from a little over $62 million in 2011 to $1.88 billion in 2013. In March 2013, on the ten-year anniversary of its founding, the company announced it was dropping the ".com" part of its branding and would continue on as just "King". In November 2014, King sued Korean company Avocado Entertainment for copying its Farm Hero Saga game in distributed game Forest Mania. In mid-2013, King.com had considered filing an initial public offering (IPO) in the United States. Zacconi had said that "The IPO is an option...We are building the company and part of that is investigating options." The company applied for IPO in September 2013. Its filing was made using allowances in the Jumpstart Our Business Startups Act to keep details of the IPO secret until it was to be offered. The IPO was backed by Bank of America, Merrill Lynch, Credit Suisse Group AG and JPMorgan Chase & Co. The IPO gained great interest, as it followed Zynga's $1 billion IPO in 2011 and Twitter's IPO earlier in the month. King completed its IPO on 26 March 2014. Priced at $22.50 a share, the middle of its projected price range, the IPO valued the company at US$7.08 billion. About $500 million was raised through the sale of 22.2 million shares. Of that, 15.3 million shares came from the company and the rest from Apax and other stakeholders. It was the largest ever IPO for a mobile/social gaming company in the US, eclipsing Zynga's 2011 offering. To celebrate the debut, Candy Crush mascots took to the New York Stock Exchange. Morris was the company's largest shareholder with approximately 35.6 million shares valued at $821 million. The company began trading under the "KING" symbol on the New York Stock Exchange. Shares of King fell 15.6% on the first day of trading, closing at $19. By June, the company's valuation had dropped by $2 billion, though otherwise was still profitable. Zacconi noted that their strategy from this point was not to find another "mega-hit" like Candy Crush Saga, but to "build a portfolio of games", carrying King's game design approach to other genres. Revenue following the IPO were over $2.6 billion in 2014, with Candy Crush Saga generating nearly half of that amount. King acquired Seattle-based mobile studio Z2Live in February 2015. In November 2015, Activision Blizzard announced its plans to acquire King for $5.9 billion. Upon announcement of the news, USA Today reported that the deal "gives Activision immediate access to the growing mobile gaming audience, the fastest-rising sector in video games". On 23 February 2016, Activision Blizzard closed its acquisition of King for a deal of $5.9 billion. Activision Blizzard as a result operates the world's largest game network, reaching around 500 million users in 196 countries. About the King acquisition, the CEO of Activision Blizzard explained that "we see great opportunities to create new ways for audiences to experience their favorite franchises, from Candy Crush to World of Warcraft to Call of Duty and more, across mobile devices, consoles and personal computers." In January 2019, Humam Sakhnini was installed as president of King, reporting directly to Zacconi. As part of a large workforce reduction announced in February 2019 across the whole of Activision Blizzard, King's Z2Live studio in Seattle was shuttered. Zacconi stepped down as CEO on 1 July 2019, remaining as chairman until August 2020, when he left the company entirely. Tjodolf Sommestad, the former chief development officer, replaced Sakhnini in February 2022. King's games portal site King.com had been rebranded to Royalgames.com, through which they offered paid-entry tournaments for a chance at cash prizes up until 2019, after which this feature was disabled for new accounts. During the first half of 2021, King had been forced to hold back on payout withdrawals by users over an investigation launched by PayPal over these withdrawals, eventually unfreezing accounts by June 2021 once the investigation was complete. King announced in October 2021 that the portal would be shuttered in December 2021 in a phased removal of the available games. Players that still had funds available on the site would be able to continue to withdraw these funds for some indefinite time after games from the site had been removed. In June 2022, King acquired the Swedish AI company Peltarion. Stockholm employees voted to form a "union club" (Swedish: Fackklubb) with Unionen, a Swedish trade union in October 2024. As of 2025, they have 217 members and meet with management to negotiate for a collective agreement. In May 2025, it was announced that Sommestad will be stepping down as president on June 1 and will be succeeded by Todd Green, the general manager of the Candy Crush franchise. In July 2025, it was reported that Microsoft had eliminated 10% of King's total workforce as part of the company's effort to cut 9,000 jobs. in November 2025, Mojang Studios and King have announced Minecraft Blast, a match-three style puzzle game such as Candy Crush Saga. Revenue model King's games, prior to June 2013, made revenue for the company through a combination of in-game advertising and microtransactions. These microtransactions allow for players to use funds to purchase in-game booster items that could be used to help clear certain levels, additional lives, and immediate access to new levels instead of having to wait for a few days. In June 2013, the company opted to remove all in-game advertising from their games, relying solely on microtransactions. The company stated that due to their "focus around delivering an uninterrupted entertainment experience for our network of loyal players across web, tablet and mobile has unfortunately led to the difficult decision of removing advertising as a core element of King's overall strategy". Advertising revenue had only made up 10% of the company's earnings in 2012, and only 1% within 2013; the company in its IPO files stated they do not anticipate any further earnings from advertising revenue. While King relies heavily on in-game purchases, it is estimated that only single-digit percentages of all players of their games have spent money on their titles. In Q4 2014, King had 356 million monthly unique users, with 8.3 million of them spending money. The 2.3% that pay spent an average of $23.42 a month within the games. King stated that their model is aimed to continue to draw existing and new players to all of their games: "If the cost to acquire players is greater than the revenue we generate over time from those players and if we cannot successfully migrate our current players to new games and new platforms as we have historically done so, our business and operating results will be harmed". The removal of in-game advertising was rolled back in 2018. Games King games offer asynchronous play, enabling users to connect to their Facebook account whilst playing on their smartphone or tablet device. This means that the user's progress is updated across all platforms, allowing the player to switch from smartphone, to tablet, to Facebook without losing their progress in the game. Bubble Witch Saga was King's first mobile game, released in July 2012 after its launch on Facebook in September 2011. Papa Pear Saga was released in March 2013 on Facebook, it is a Peggle variation. Around 2012, Pyramid Solitaire Saga was soft launched on Facebook. It was released on mobile in May 2014. In late 2012 Pet Rescue Saga was launched on Facebook, then on iOS and Android In June 2013, Candy Crush Soda Saga was soft launched on Facebook and mobile and Bubble Witch 2 Saga was widely released for Android and iOS devices. In November 2014, Candy Crush Soda Saga was widely released on Android and iOS. Alpha Betty Saga launched on Facebook in April 2015. This game is a variation of Bookworm. In 2013, King acquired the Defold game engine, developed by Ragnar Svensson and Christian Murray in 2007 as a lightweight 2D game engine. The two had offered the engine to King as well as their services as contractors to support it, and later bought the engine, using it first for the game Blossom Blast Saga. In March 2016, King released the Defold engine as a free development tool for any user, and by May 2020, it ceded control of the engine to the Defold Foundation, which made the engine open source with plans to continue to support it with additional investment from King. King announced in April 2017 that they will be developing a mobile Call of Duty game, a property owned by Activision; the game would be one of the first ones outside of the casual mobile space for the company. King's most popular game is Candy Crush Saga, a tile-matching game which was launched on King's website in March 2011. It launched on Facebook in April 2012 and quickly gained popularity. Following its success on Facebook, King launched Candy Crush Saga on mobile (iOS and Android) in November 2012. The game was downloaded over 10 million times in its first month. In January 2013, it became the number one most played game on Facebook. It had over 45 million monthly users in March 2013. By January 2014, it had over 150 million monthly users. While King continues to release other titles, the company's principle focus as of November 2017 are on its four most popular series: Candy Crush Saga, Bubble Witch Saga, Pet Rescue Saga, and Farm Heroes Saga. Controversies In January 2014, King attracted controversy after attempting to trademark the words "Candy" and "Saga" in game titles. This directly impacted Stoic's trademark request for The Banner Saga, to which King filed an opposition, calling the name "deceptively similar" to King games. Stoic said that the dispute hindered work on a planned sequel to their game. On 17 April 2014, it was reported that King has settled its disputes with Stoic Studio and Runsome Apps. Also in January 2014, game developer Matthew Cox accused King of ripping off his game Scamperghost, saying King's Pac-Avoid was a clone of it. According to Cox, he was in talks with King about licensing Scamperghost, but when the deal fell through the company released the game Pac-Avoid. Cox said Epicshadows, the developer of Pac-Avoid, told him that King had approached them to "clone the game very quickly". King removed the game from its website, but denied the cloning allegation, stating that they were removing the game "for the avoidance of doubt". References External links |
======================================== |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.