id
stringlengths
36
36
source
stringclasses
15 values
formatted_source
stringclasses
13 values
text
stringlengths
2
7.55M
f6bd243d-4930-40e4-82dc-0e8181620b5b
trentmkelly/LessWrong-43k
LessWrong
'Theories of Values' and 'Theories of Agents': confusions, musings and desiderata Meta: * Content signposts: we talk about limits to expected utility theory; what values are (and ways in which we're confused about what values are); the need for a "generative"/developmental logic of agents (and their values); types o...
9fff7b95-8d0f-41bf-aae0-cc06d4801baf
trentmkelly/LessWrong-43k
LessWrong
Ads are everywhere, and it's not okay [Crossposted from damiensnyder.com.] Author's note: Thanks to the commenters who have pointed out some weaknesses in my argument I didn't think of. I still endorse the majority of this argument, but I don't think this was the best way to express it. (An earlier post I drafted on ...
eea8c5be-e97d-4f76-b362-6db6932d0b09
trentmkelly/LessWrong-43k
LessWrong
[AN #79]: Recursive reward modeling as an alignment technique integrated with deep RL Find all Alignment Newsletter resources here. In particular, you can sign up, or look through this spreadsheet of all summaries that have ever been in the newsletter. I'm always happy to hear feedback; you can send it to me by replyi...
e4f38c83-40a2-482f-bf03-acbcfc6bc410
trentmkelly/LessWrong-43k
LessWrong
Wanted: Executive Assistant to help build the progress movement I’m hiring an Executive Assistant to be part of the founding team at The Roots of Progress, and to work closely with me on all of our projects to build the progress movement. This role is perfect for someone who wants to combine dedication to an ambitiou...
42381aa1-c386-4bcb-8cc5-3750ea60b7de
trentmkelly/LessWrong-43k
LessWrong
Figuring out what Alice wants: non-human Alice I’ve shown that we cannot deduce the preferences of a potentially irrational agent. Even simplicity priors don’t help. We need to make extra ‘normative’ assumptions in order to be able to say anything about these preferences. I then presented a more intuitive example, in...
c69c51bf-a244-4165-a794-7fb2e9cb9407
trentmkelly/LessWrong-43k
LessWrong
To what ethics is an AGI actually safely alignable? Today, AI is thought to be alignable to nearly every task except for obviously unethical ones like writing insecure code without a noble reason, which was found to induce broad misalignment or lower the AI's inhibitions. The question of whether alignment of superinte...
4dc96c13-807c-497c-af05-6368086b1ead
trentmkelly/LessWrong-43k
LessWrong
Reward uncertainty In my last post, I argued that interaction between the human and the AI system was necessary in order for the AI system to “stay on track” as we encounter new and unforeseen changes to the environment. The most obvious implementation of this would be to have an AI system that keeps an estimate of th...
73f26d62-464b-4b65-bd75-e0830cc95dc0
trentmkelly/LessWrong-43k
LessWrong
Ghosts in the Machine People hear about Friendly AI and say - this is one of the top three initial reactions: "Oh, you can try to tell the AI to be Friendly, but if the AI can modify its own source code, it'll just remove any constraints you try to place on it." And where does that decision come from? Does it enter...
71a25a9e-243c-4928-bdd3-572ad348d197
trentmkelly/LessWrong-43k
LessWrong
ACI#9: What is Intelligence Abstract This article is a brief summary of ACI's new definition of intelligence: Doing the same thing in new situations as the examples of the right thing to do, by making predictions based on these examples. In other words, intelligence makes decisions by stare decisis with Solomonoff ind...
b0b2a5d1-a66c-419e-abc3-ae3a9ca69ce7
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
David Krueger on AI Alignment in Academia and Coordination [David Krueger](https://www.davidscottkrueger.com/) is an assistant professor at the University of Cambridge and got his PhD from Mila. His research group focuses on aligning deep learning systems, but he is also interested in governance and global coordinatio...
bc5680ee-4bae-49a7-9593-f30100709c38
trentmkelly/LessWrong-43k
LessWrong
Branding Biodiversity http://www.futerra.co.uk/downloads/Branding_Biodiversity.pdf I'm quite interested in seeing if this works. I have sent this to several wildlife-guides and conservationists and will monitor their reactions. It talks about what emotions drive people to actually do something to protect biodiversit...
b52bc8ec-a152-4b4b-9ba9-8e26c0e28642
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] Lawrence Watt-Evans's Fiction Today's post, Lawrence Watt-Evans's Fiction was originally published on 15 July 2008. A summary (taken from the LW wiki): > A review of Lawrence Watt-Evans's fiction. Discuss the post here (rather than in the comments to the original post). This post is part of the Rerunni...
0aaac58d-025e-443b-9d3c-416b6bdef83d
trentmkelly/LessWrong-43k
LessWrong
Monthly Roundup #15: February 2024 Another month. More things. Much roundup. BAD NEWS Jesse Smith writes in Asterisk that our HVAC workforce is both deeply incompetent and deeply corrupt. This certainly matches my own experience. Calculations are almost always flubbed when they are done at all, outright fraudulent p...
9afbe868-0dbc-4edc-892f-12432cc039c9
trentmkelly/LessWrong-43k
LessWrong
[Valence series] 1. Introduction 1.1 Summary & Table of Contents This is the first of a series of five blog posts on valence. Here’s an overview of the whole series, and then we’ll jump right into the first post! 1.1.1 Summary & Table of Contents—for the whole Valence series Let’s say a thought pops into your mind: ...
8f1de810-13c1-4d0b-948a-1eff57a1a1bc
trentmkelly/LessWrong-43k
LessWrong
Learning with catastrophes A catastrophe is an event so bad that we are not willing to let it happen even a single time. For example, we would be unhappy if our self-driving car ever accelerates to 65 mph in a residential area and hits a pedestrian. Catastrophes present a theoretical challenge for traditional machine...
0b05f412-adf6-49d8-910c-a92b7bafbdf9
StampyAI/alignment-research-dataset/blogs
Blogs
Historic trends in land speed records Land speed records did not see any greater-than-10-year discontinuities relative to linear progress across all records. Considered as several distinct linear trends it saw discontinuities of 12, 13, 25, and 13 years, the first two corresponding to early (but not first) jet-propell...
53578c5c-cb7c-47db-8ba7-50d4822dadc5
trentmkelly/LessWrong-43k
LessWrong
Carbon dioxide, climate sensitivity, feedbacks, and the historical record: a cursory examination of the Anthropogenic Global Warming (AGW) hypothesis Note: In this blog post, I reference a number of blog posts and academic papers. Two caveats to these references: (a) I often reference them for a specific graph or calc...
71ad77f0-752a-4da3-b5e6-cb2c0534865f
trentmkelly/LessWrong-43k
LessWrong
Is there a convenient way to make "sealed" predictions? Many people in our community claim to have ideas for how to build AGI, or other things, that they deem infohazardous and so don't want to publish. It would be great if they could publicly register these ideas in an encrypted way, so that later when their predicti...
815449ae-6a0f-4c0a-a336-e082644d3b65
trentmkelly/LessWrong-43k
LessWrong
The Sheer Folly of Callow Youth > "There speaks the sheer folly of callow youth; the rashness of an ignorance so abysmal as to be possible only to one of your ephemeral race..." >         —Gharlane of Eddore Once upon a time, years ago, I propounded a mysterious answer to a mysterious question—as I've hinted on sever...
ee3e7644-2fb7-4cb6-8cdb-6548028fb44e
StampyAI/alignment-research-dataset/lesswrong
LessWrong
New article from Oren Etzioni (Cross-posted from [EA Forum](https://forum.effectivealtruism.org/posts/7x2BokkkemjnXD9B6/new-article-from-oren-etzioni).) This just appeared in this week’s MIT Technology Review: Oren Etzioni, “[How to know if AI is about to destroy civilization](https://www.technologyreview.com/s/61526...
2e29d0a7-1aca-4394-9432-fc8c5da7cc61
trentmkelly/LessWrong-43k
LessWrong
A summary of every Replacing Guilt post 1 I recently finished the Replacing Guilt series (by Nate Soares).  I don't struggle much with guilt. I consider myself a happy and energetic person. I also spent several years studying psychology, and I'm familiar with many techniques from cognitive-behavioral therapy. So I w...
a810c3d1-0806-4bf6-ad89-49646b2adb13
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Can "Reward Economics" solve AI Alignment? I think we can try to solve AI Alignment this way: Model human values and objects in the world as a "money system" (a system of meaningful trades). Make the AGI learn the correct "money system", specify some obviously incorrect "money systems". Basically, you ask the AI *"m...
a4af3d28-41bb-448d-8df0-7cf5840e5a27
StampyAI/alignment-research-dataset/blogs
Blogs
MIRI has a new COO: Malo Bourgon ![Malo Bourgon](https://intelligence.org/wp-content/uploads/2016/03/Malo-Bourgon.jpg)I’m happy to announce that Malo Bourgon, formerly a program management analyst at MIRI, has taken on a new leadership role as our chief operating officer. As MIRI’s second-in-command, Malo will be ta...
b33dd55e-2bc6-4f91-b1ca-8d7a4dbfa546
trentmkelly/LessWrong-43k
LessWrong
Local Speech Recognition with Whisper I've been a heavy user of dictation, off and on, as my wrists have gotten better and worse. I've mostly used the built-in Mac and Android recognition: Mac isn't great, Android is pretty good, neither has improved much over the past ~5y despite large improvements in what should be ...
656d9a63-be75-4871-b40c-f238153470a4
trentmkelly/LessWrong-43k
LessWrong
New method predicts who will respond to lithium therapy
eba5b6d3-5f72-43dd-9c41-58f1666574da
trentmkelly/LessWrong-43k
LessWrong
Me, Myself, and AI: the Situational Awareness Dataset (SAD) for LLMs TLDR: We build a comprehensive benchmark to measure situational awareness in LLMs. It consists of 16 tasks, which we group into 7 categories and 3 aspects of situational awareness (self-knowledge, situational inferences, and taking actions). We test...
64306985-2bb5-4cbd-8af6-e01235abd5ee
trentmkelly/LessWrong-43k
LessWrong
Improving long-run civilisational robustness People trying to guard civilisation against catastrophe usually focus on one specific kind of catastrophe at a time. This can be useful for building concrete knowledge with some certainty in order for others to build on it. However, there are disadvantages to this catastrop...
54996fcb-3179-4ace-be83-8cb8c0542364
trentmkelly/LessWrong-43k
LessWrong
When will computer programming become an unskilled job (if ever)? I expect for there to be a delay in deployment, but I think ultimately OpenAI is aiming as a near term goal to automate intellectually difficult portions of computer programming. Personally, as someone just getting into the tech industry, this is basica...
dd7f782d-c671-4f8a-b3f0-42c58a3b433c
trentmkelly/LessWrong-43k
LessWrong
A useful level distinction Recently, I've constantly been noticing the importance of a certain sort of distinguishing between something happening in reality, and something happening in your model of reality. In part, I'm noticing it because I've been thinking about it, not vice versa, but I'll explain for a few specif...
1ba9938a-ceb1-47d2-ae4a-8a56eb7cd00a
trentmkelly/LessWrong-43k
LessWrong
Meetup : Atlanta Discussion article for the meetup : Atlanta WHEN: 04 February 2012 06:30:00PM (-0500) WHERE: 2094 North Decatur Road, Decatur, GA 30033-5367 All, Our next meeting is February 4th. From now on we will be meeting every 1st and 3rd Saturday of the month at 6:30pm at Chocolate Coffee in Decatur: http...
c3e07c42-af7b-4d2d-8837-c360d778aefe
trentmkelly/LessWrong-43k
LessWrong
UVic AI Ethics Conference Hello! I'm part of the AI club at the University of Victoria. We want to help expose students and professors at our university to the ideas of AI alignment and AI ethics, and so we are putting together an online conference this November 25th. Our club website Eventbrite page We are looking...
6ba3e5c7-76e4-4877-8c32-6e48635d09ac
trentmkelly/LessWrong-43k
LessWrong
Do Prediction Markets Work? TLDR: Prediction markets rely on efficiency, but efficiency is not guaranteed. Prediction market structures can work. However, they rely on so many different components being in place that they do not consistently create accurate probability. The systems rely on complete market efficiency...
642bcbae-78c9-4ac3-8b71-142df9b0b62f
trentmkelly/LessWrong-43k
LessWrong
[LINK] What Can Internet Ads Teach Us About Persuasion? How "one weird trick" conquered the Internet. Some excerpt I found interesting: > Research on persuasion shows the more arguments you list in favor of something, regardless of the quality of those arguments, the more that people tend to believe it,” Norton says....
8f3ac51f-6cf2-41a8-9490-2bb68cbe24a4
trentmkelly/LessWrong-43k
LessWrong
Prettified AI Safety Game Cards Donald Hobson's AI Safety Card Game already has simple cards you can print out on cardstock and cut out, but I wanted a nicer print-on-demand version with illustrated card backs.[1] You can order it here. Major thanks to Mati Roy for helping make this happen, along with Jim Babcock, and...
5102a253-05f4-44b9-af58-c1d467350248
trentmkelly/LessWrong-43k
LessWrong
Detecting What's Been Seen Sometimes it makes sense for sites want to treat things differently based on whether the user has seen them. For example, I like sites that highlight new comments (ex: EA Forum and LessWrong) and I'd like them even better if comments didn't lose their "highlighted" status in cases where I ha...
09394f82-de99-4d0c-8c0a-155de64de08d
trentmkelly/LessWrong-43k
LessWrong
Big Advance in Infinite Ethics ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, Summary It is possible that our universe is infinite in both time and space. We might therefore reasonably consider the following question: given some sequences u=(u1,u2,…) and u′=(u′1,u′2,…) (where each ut represents the welfare of persons living ...
86588551-abe8-4a45-91ee-128e4344c65f
trentmkelly/LessWrong-43k
LessWrong
Limits to the Controllability of AGI Roman Yampolskiy (Summary by Karl von Wendt and Remmelt Ellen) This is a summary of the paper “On the Controllability of Artificial Intelligence: An Analysis of Limitations”, which shows that advanced artificial general intelligence (AGI) with the ability to self-improve cannot ...
5aa3da88-5990-4fc6-bd42-d05073b0167a
StampyAI/alignment-research-dataset/lesswrong
LessWrong
"Normal" is the equilibrium state of past optimization processes Orienting around the ideas and conclusions involved with AI x-risk can be very difficult. The future possibilities can feel extreme and [far-mode](https://www.lesswrong.com/tag/near-far-thinking), even when we whole-heartedly affirm their plausibility. ...
f46d98b4-6795-4e82-8c66-2c3b0b4789a9
StampyAI/alignment-research-dataset/lesswrong
LessWrong
AI Governance: A Research Agenda Allan Dafoe, head of the AI Governance team at FHI has written up a research agenda. It's a very large document, and I haven't gotten around to reading it all, but one of the things I would be most excited about is people in the comments quoting the most important snippets that give a ...
ab02fe57-0bdd-437d-b6fb-a56446a304d3
trentmkelly/LessWrong-43k
LessWrong
Consciousness Actually Explained: EC Theory Greetings! This has taken too long to make, and I'd like to polish it more, but it's time to release and move on to other things. Apologies for not posting the text here, I was having formatting trouble.  Yes, I attempt to solve/explain consciousness. The main document is ...
1c640878-ada3-4b01-bafc-7a8b4e06d5d3
trentmkelly/LessWrong-43k
LessWrong
A rationalist My Little Pony fanfic For the past two months, I've been writing, and posting, roughly two thousand words a day of "Myou've Gotta be Kidding Me", a story set in a "My Little Pony: Friendship is Magic" fanfic universe, "Chess Game of the Gods". Outside of the sheer NaNoWriMo-like exercise of pushing out n...
f629cd29-f97f-4fc4-a9be-9297491aa9b1
trentmkelly/LessWrong-43k
LessWrong
Why do theists, undergrads, and Less Wrongers favor one-boxing on Newcomb? Follow-up to: Normative uncertainty in Newcomb's problem Philosophers and atheists break for two-boxing; theists and Less Wrong break for one-boxing Personally, I would one-box on Newcomb's Problem. Conditional on one-boxing for lawful reasons...
6224deaa-646b-4a42-a750-3f13b37434cd
trentmkelly/LessWrong-43k
LessWrong
The Parable of the Boy Who Cried 5% Chance of Wolf Epistemic status: a parable making a moderately strong claim about statistics Once upon a time, there was a boy who cried "there's a 5% chance there's a wolf!" The villagers came running, saw no wolf, and said "He said there was a wolf and there was not. Thus his pr...
50582676-3bac-47f4-939f-afc2176815bd
trentmkelly/LessWrong-43k
LessWrong
Slava! I want to begin with a musical example.  The link is the Coronation Scene from Mussorsky's opera Boris Godunov, in which Boris is crowned Tsar while courtiers sing his praises.  The tune is quoted from an old Russian hymn, "Slava Bogu" or "Glory to God."  And, if I can trust the English subtitles, it's an apt c...
fd9a7e61-8824-466b-8b27-682696dd80bc
trentmkelly/LessWrong-43k
LessWrong
Less Wrongers from Austin As far as I know, there is not a group of Less Wrongers from Austin that has formed yet. I haven't ever seen posts about meet-ups, etc. I would really like to get a group started, but I honestly do not know if there is even a single other member on this site who lives nearby, and I don't part...
bcde6051-5bdb-476b-940a-b31478d960b7
StampyAI/alignment-research-dataset/special_docs
Other
CHAI Newsletter #2 2022 07/04/2023, 07:54 CHAI Newsletter + Internship Application https://mailchi.mp/humancompatible.ai/chai-newsletter -internship-application?e=[UNIQID] 1/9 CHAI Newsletter This newsletter includes CHAI activities from April 2022 to mid-September 2022  Events CHAI 2022 annual W orkshop On the first w...
7ca1aff7-4f62-4503-b0b4-8949ace8416a
trentmkelly/LessWrong-43k
LessWrong
Running a Prediction Market Mafia Game In a normal game of mafia/werewolf, there are a few high-level strategic elements: * Night Play: During night phases, the mafia has to choose who to try to kill and the players with night actions (cops, doctors, and in some cases trackers, watchers, vigilantes, etc.) have to ch...
d5ecea6b-ac3b-4e4c-9134-49bcde4e9f63
StampyAI/alignment-research-dataset/arbital
Arbital
Flag the load-bearing premises If someone says, "I think your AI safety scheme is horribly flawed because X will go wrong," and you reply "Nah, X will be fine because of Y and Z", then good practice calls for highlighting that Y and Z are important propositions. Y and Z may need to be debated in their own right, espe...
77c04a6e-74ca-49cc-b8cd-498e4cfe3b93
trentmkelly/LessWrong-43k
LessWrong
How are you dealing with ontology identification? Not really novel, but I've made this point quite often in recent conversations, so I decided to make a short write-up. I think a wide range of alignment proposals will somehow have to deal with ontology identification, i.e. mapping an AI's beliefs about the world to a...
cc8ae89a-bc1a-43d2-b6b9-e6d7e35065a7
StampyAI/alignment-research-dataset/lesswrong
LessWrong
DELBERTing as an Adversarial Strategy There is an entire subgenre of "reverse scammer" videos on youtube.  The way these videos work is, a youtuber waits for a scam call from Pakistan or somewhere.  When it becomes obvious that it's a scam, like when the scammer starts asking for a credit card number in order to do so...
6f10451b-1b1b-4c62-aa85-4bafdc6d9f0b
trentmkelly/LessWrong-43k
LessWrong
Bazant: An alternate covid calculator Most of what I see people use Microcovid.org for now is estimating risk for large gatherings, which it was not designed for and thus doesn’t handle very well. I spent a few hours going through every covid calculator I could find and this calculator from the Bazant lab at MIT, whil...
6cb61b3b-ff3a-443c-ade1-9cbe76167447
trentmkelly/LessWrong-43k
LessWrong
What working on AI safety taught me about B2B SaaS sales Subtitle: you're too poor to use AGI. WTF is WTP? In Econ 101, you learn about something called willingness to pay (WTP). WTP is the highest price at which you're willing to buy some good or service. As long as the equilibrium price is less than your WTP, you'...
8f346930-838d-49a2-b2a6-494ccf037fec
trentmkelly/LessWrong-43k
LessWrong
Meetup : Bath, UK: Agreement, practical meetups, and report from last meetup Discussion article for the meetup : Bath, UK: Agreement, practical meetups, and report from last meetup WHEN: 02 November 2014 02:00:00PM (+0000) WHERE: 5-10 James St W, Avon, Bath BA1 2BX Bath, UK will be having its second meetup this Sun...
515540c7-8fda-4651-a426-d54c8e44f3bb
trentmkelly/LessWrong-43k
LessWrong
Require contributions in advance If you are a person who finds it difficult to tell "no" to their friends, this one weird trick may save you a lot of time!   Scenario 1 Alice: "Hi Bob! You are a programmer, right?" Bob: "Hi Alice! Yes, I am." Alice: "I have this cool idea, but I need someone to help me. I am not ...
983da280-164c-4c9f-879f-8d007e4eed9d
StampyAI/alignment-research-dataset/blogs
Blogs
Overcoming Bias From August 2007 through May 2009, I blogged daily on the topic of human rationality at the econblog [Overcoming Bias](http://www.overcomingbias.com/) by Robin Hanson, getting around a quarter-million monthly pageviews. This then forked off the community blog [Less Wrong](http://lesswrong.com/) , and I...
26a06178-9388-4eab-b6ef-68b600200cd8
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Oracle paper Available on the arXiv, my paper on [two types of Oracles](https://arxiv.org/abs/1711.05541) (AIs constrained to answering questions only), and how to use them more safely. > An Oracle is a design for potentially high power artificial intelligences (AIs), where the AI is made safe by restricting it to o...
61077537-6ff6-477d-a2cc-6d7b9adb3780
trentmkelly/LessWrong-43k
LessWrong
Metaculus Predicts Weak AGI in 2 Years and AGI in 10 Just a quick update on predicted timelines. Obviously, there’s no guarantee that Metaculus is 100% reliable + you should look at other sources as well, but I find this concerning. Weak AGI is now predicted in a little over two years: https://www.metaculus.com/ques...
1876b62c-d731-4ac9-ae32-c9fc064f2192
trentmkelly/LessWrong-43k
LessWrong
Gettier cases, Rigid Designators, and Referential Opacity NB: I don't spellcheck these things until I've completed them. This is a WIP. I also find it hard to focus on the writing, while I'm thinking about the argumentation. I would be grateful if members could overlook some of the errors in the text -- to try to fol...
78c995d3-4b59-41fd-8b2b-ae72fecd42d1
trentmkelly/LessWrong-43k
LessWrong
Deeply Cover Car Crashes? When there is a train, plane, or bus crash, it's newsworthy: it doesn't happen very often, lots of lives at stake, lots of people are interested in it. Multiple news outlets will send out reporters, and we will hear a lot of details. On the other hand, a car crash does not get this treatment ...
21a73afc-7542-4091-88a1-d81766bc12cd
trentmkelly/LessWrong-43k
LessWrong
Is Anyone Else Seeking to Help Community Members in Ukraine Who Make Refugee Claims? Some rationalists and effective altruists in British Columbia have begun mulling over the option of whether they would be able to sponsor a fellow community member living in Ukraine to stay as a refugee in Canada. There is a process b...
da8a224d-d3d2-4dd4-909e-6d2335a2fa4a
trentmkelly/LessWrong-43k
LessWrong
Meetup : Predictions Chi September 15 (September 8 -- cancelled) Discussion article for the meetup : Predictions Chi September 15 (September 8 -- cancelled) WHEN: 15 September 2012 01:05:00PM (-0500) WHERE: 1 S. Franklin St, Chicago, IL Bring your extrapolations, intuitions, wild guesses, and most importantly some ...
808f9515-85b1-421f-a18e-9a3a7e94f838
trentmkelly/LessWrong-43k
LessWrong
What will be the big-picture implications of the coronavirus, assuming it eventually infects >10% of the world? Brainstorming some possibilities, not taking myself too seriously: --2020 will be remembered as a year of death and disruption; the economy will suffer massively, perhaps enough to count as a recession. --...
7bd70b5c-df0a-4ee3-bb39-e19e4e0df488
StampyAI/alignment-research-dataset/lesswrong
LessWrong
What’s the likelihood of only sub exponential growth for AGI? A possible scenario could be that the first confirmed AGI(s) are completely unimpressive,  i.e. with equivalent or less capacities than an average 10 year child. And then tremendous effort is put into ’growing’ its capacities with the best results yielding ...
a763ba47-37fb-4596-8142-008db5dc9741
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
What I would do if I wasn’t at ARC Evals **In which:**I list 9 projects that I would work on if I wasn’t busy working on safety standards at ARC Evals, and explain why they might be good to work on.  **Epistemic status:**I’m prioritizing getting this out fast as opposed to writing it carefully. I’ve thought for at le...
4bd67367-3904-4c68-bd53-3a1378220a93
trentmkelly/LessWrong-43k
LessWrong
Metaculus Launches Chinese AI Chips Tournament, Supporting Institute for AI Policy and Strategy Research Metaculus’s Chinese AI Chips Tournament is live: Forecast and support research by the Institute for AI Policy and Strategy (IAPS), a think tank that works to reduce risks from the development and deployment of fron...
c5fe1f7e-3a5a-4341-b271-41acf44f3859
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Alignment Might Never Be Solved, By Humans or AI I think it's possible that neither we, nor any future AI system, will manage to solve alignment. By "solving alignment" I mean having a cheap easily-scalable method for aligning AIs of arbitrary capability level to arbitrary goals. Here I will briefly present some argum...
93d5f595-23f1-47de-b779-5742f0d2886f
StampyAI/alignment-research-dataset/lesswrong
LessWrong
An additional problem with Solomonoff induction Let's continue from my [previous post](/r/discussion/lw/jhm/understanding_and_justifying_solomonoff_induction/) and look how Solomonoff induction fails to adequately deal with hypercomputation.   You may have heard of the Physical Church-Turing thesis. It's the idea ...
314b4854-7ff5-4159-b563-874b760c6d50
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Open technical problem: A Quinean proof of Löb's theorem, for an easier cartoon guide **Motivation** Löb's theorem is pretty counterintuitive.  Speaking informally about one or more agents engaging in logically valid reasoning, the theorem says something like this: * "If it's believed that believing a particular pro...
34d48b05-0122-4d3f-8065-3f242a7492e3
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
Ben Goertzel - The unorthodox path to AGI hey everyone welcome back to the podcast today we're talking to ben gertzel who's actually one of the founders of the agi research community he's been interested in this area for decades really long before most people were and he's done a whole bunch of interesting thinking ar...
91ed3cc1-e132-4fd9-8b57-a92f8f229765
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Introduction This series of posts presents our work on the frameworks of “Positive Attractors” and “Inherently Interpretable Architectures” during the AI Safety Camp 2023.  (Note: Sections written by the different team members were marked with their initials in square brackets, since the investigated research strand...
17374d98-b419-4ba0-a117-153aec1c6789
trentmkelly/LessWrong-43k
LessWrong
Should I treat pain differently if it’s “all in my head?” I sometimes suffer from RSI. I’ve heard several people I respect say that they think most RSI is a psychological phenomenon. My understanding of their view is that they think there’s a feedback loop where you notice your wrists hurting and then you are anxious ...
9748d9e3-3d26-47f6-8d0d-1c98d6fc22a1
trentmkelly/LessWrong-43k
LessWrong
Preface to CLR's Research Agenda on Cooperation, Conflict, and TAI The Center on Long-Term Risk (CLR) is focused on reducing risks of astronomical suffering, or s-risks, from transformative artificial intelligence (TAI). S-risks are defined as risks of cosmically significant amounts of suffering[1]. As has been discu...
1bb2c468-af44-4af8-8c49-1a7b671b86bb
trentmkelly/LessWrong-43k
LessWrong
Request for input: draft of my "coming out" statement on religious deconversion Edited 3/4/2012: I shortened up the summary a bit and add the following update: Thanks for the lively comments. As a preliminary summary of things I've found quite useful/helpful: * Shorten/transform the document (David_Gerard) * Remov...
7b1e82dd-0053-4fca-aa50-f2709cf194a7
trentmkelly/LessWrong-43k
LessWrong
What OpenAI Told California's Attorney General In a previously unreported letter, the AI company defends its restructuring plan while attacking critics and making surprising admissions This is the full text of a post first published on Obsolete, a Substack that I write about the intersection of capitalism, geopolitics...
fb533e8b-5e8a-4b53-b98c-ed6e508d3ca2
trentmkelly/LessWrong-43k
LessWrong
Constraining narrow AI in a corporate setting I have a question that is extremely practical. My company is working with AI that is frankly not all that capable. A key example includes text classification that humans do, but that would let us avoid the expense of having a human do it. Off-the-shelf machine learning wo...
65ea4dec-e23e-4191-a42b-3161addf00e3
trentmkelly/LessWrong-43k
LessWrong
Seriously, what goes wrong with "reward the agent when it makes you smile"? Suppose you're training a huge neural network with some awesome future RL algorithm with clever exploration bonuses and a self-supervised pretrained multimodal initialization and a recurrent state. This NN implements an embodied agent which ta...
d505f275-570c-40f5-829e-83ec44f4bb57
trentmkelly/LessWrong-43k
LessWrong
Why long-term AI safety is not an issue (IMHO) Hello I wanted to share this thought for a long time and now, that I absolutely do not have the time, I will: I think that an artificial general intelligence (AGI) is not an existential risk for human kind. In other words, long-term artificial intelligence (AI) safety i...
66250ead-2c9c-4b43-9417-789085094ad5
trentmkelly/LessWrong-43k
LessWrong
The Fallacy of Dressing Like a Winner Imagine you are a sprinter, and your one goal in life is to win the 100m sprint in the Olympics. Naturally, you watch the 100m sprint winners of the past in the hope that you can learn something from them, and it doesn't take you long to spot a pattern.   Every one of them can b...
1603d9e7-8783-4128-a807-8ef881d3708b
trentmkelly/LessWrong-43k
LessWrong
Since the "Will COVID-19 survivors suffer lasting disability at a high rate?" thread was posted, has new information emerged that may update those estimates? https://www.lesswrong.com/posts/h4GFHbhxE2pfiadhi/will-covid-19-survivors-suffer-lasting-disability-at-a-high
6e674e00-ce5f-4f9a-a8f7-88d7ad617098
trentmkelly/LessWrong-43k
LessWrong
What exactly did that great AI future involve again? I've been trying my best to think of something that AGI could do which I really love and deeply endorse. I can think of some nice things. New kinds of animals. Non-habit-forming heroine. Stupid-pointless-disagreements-between-family-members fixomatic maybe. Turn ev...
b270f20a-8679-4d54-82f7-319b352dd46f
trentmkelly/LessWrong-43k
LessWrong
Nyoom Let me tell you about my scooter. I have a foldable electric trike scooter slightly heavier than my toddler. It has no pedal option, just footrests. It can't reverse. It corners badly. It has a little nubbin that's supposed to keep it from unfolding till I deliberately unfold it, but that usually doesn't work....
b260344d-7076-443b-b330-01019862a851
trentmkelly/LessWrong-43k
LessWrong
Weekly LW Meetups This summary was posted to LW Main on November 7th. The following week's summary is here. New meetups (or meetups with a hiatus of more than a year) are happening in: * Bangalore Meetup: 22 November 2014 03:56AM Irregularly scheduled Less Wrong meetups are taking place in: * Bratislava: 10 Nove...
806d0986-4caa-4bc2-aace-2eee664e5a60
trentmkelly/LessWrong-43k
LessWrong
The "hard" problem of consciousness is the least interesting problem of consciousness Cross-posted from my roam-blog. Nothing new to anyone who's read about these ideas, meant to be a reference for how I think about things. If you come up with what you think is a decent and useful theory/partial-theory of consciousne...
13b42be8-08ba-435f-8921-a1492da5f359
StampyAI/alignment-research-dataset/special_docs
Other
Fast reinforcement learning with generalized policy updates The combination of reinforcement learning with deep learning is a promising approach to tackle important sequential decisionmaking problems that are currently intractable. One obstacle to overcome is the amount of data needed by learning systems of this type....
d24732cc-b474-450e-8f4c-3acfc5baa33d
trentmkelly/LessWrong-43k
LessWrong
Mud and Despair (Part 4 of "The Sense Of Physical Necessity") This is the fourth post in a sequence that demonstrates a complete naturalist study, specifically a study of query hugging (sort of), as described in The Nuts and Bolts of Naturalism. For context on this sequence, see the intro post. “Mud and Despair” is ...
5ea280d1-2807-45e3-9fb5-4cf8fd8feb14
trentmkelly/LessWrong-43k
LessWrong
Thoughts on the Repugnant Conclusion When we accept the repugnant conclusion, we trade lower average utility for greater total utility. This is unavoidable. But the image the repugnant conclusion evokes in many people's minds  (of trading higher average toil and squalor for an economic surplus that is spent creating m...
4b20f1b7-cb5f-4caf-a778-8449bae797f2
trentmkelly/LessWrong-43k
LessWrong
White Sphere Hypothesis Imagine an existence where the only perspective for your entire life (40 years) has been from the inside of a white sphere. In this white sphere, you have no physical body, and there is nothing else in the white sphere with you. You only exist in this white sphere as a brain with perfectly work...
7c062066-dc9a-4a4f-b395-7f12843c5d9f
StampyAI/alignment-research-dataset/blogs
Blogs
Updates to the research team, and a major donation We have several major announcements to make, covering new developments in the two months since our [2017 strategy update](https://intelligence.org/2017/04/30/2017-updates-and-strategy/): 1. On May 30th, we received a surprise **$1.01 million donation** from an [Ethe...
2737f0fe-9e7a-4be3-83c4-6149b46302fa
trentmkelly/LessWrong-43k
LessWrong
Weekly LW Meetups This summary was posted to LW main on June 14th. The following week's summary is here. Irregularly scheduled Less Wrong meetups are taking place in: * Atlanta LessWrong June Meetup: Effective Altruism: 15 June 2013 07:00PM * Berlin Social Meetup: 15 June 2013 05:00PM * Bratislava Meetup IV.: 24 ...
4a4dadec-4179-49ac-86f6-d7f723052847
trentmkelly/LessWrong-43k
LessWrong
Scott Adams mentions Prediction Markets and explains Cognitive Blindness bias
ef2fb1d5-4afa-4ca0-bd53-e93369b59061
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Decision theory does not imply that we get to have nice things (*Note: I wrote this with editing help from Rob and Eliezer. Eliezer's responsible for a few of the paragraphs.*) A common confusion I see in the tiny fragment of the world that knows about [logical decision theory](https://arbital.com/p/logical_dt/) (...
a08d2140-99aa-4e48-b5f3-3f4b8e05e602
trentmkelly/LessWrong-43k
LessWrong
Longevity: A critical look at "Loss of epigenetic information as a cause of mammalian aging" Introduction: Epigenetic information loss as a cause of mammalian aging I have generally been uncompelled by theories of aging which focus on accumulation of error at the cellular level, NAD+ decline, mutation accumulation, fr...
97be3915-049b-4376-bbc0-8153f04eada1
trentmkelly/LessWrong-43k
LessWrong
Estimating training compute of Deep Learning models by Jaime Sevilla, Lennart Heim, Marius Hobbhahn, Tamay Besiroglu, and Anson Ho You can find the complete article here. We provide a short summary below. In short: To estimate the compute used to train a Deep Learning model we can either: 1) directly count the numbe...
0781889e-4f91-4bc0-813f-19d7185cac9b
trentmkelly/LessWrong-43k
LessWrong
Diet Advice? So, I know a number of friends on Paleo who recommend it. I recently read through a lot of bulletproofexec, who recommends his own variant of paleo. I care about my health, and so I need to resolve my diet and their advice somehow. Summarized data points: 1. My diet is about 80% wheat, and the rest is s...
2769d432-db25-47d2-9330-c1cebf6c934c
trentmkelly/LessWrong-43k
LessWrong
Reaching out to people with the problems of friendly AI There have been a few attempts to reach out to broader audiences in the past, but mostly in very politically/ideologically loaded topics. After seeing several examples of how little understanding people have about the difficulties in creating a friendly AI, I'm ...
1d6341ee-ca08-461f-b9de-60a69e5488f4
trentmkelly/LessWrong-43k
LessWrong
AISafety.info: What is the "natural abstractions hypothesis"? AISafety.info writes AI safety intro content. We'd appreciate any feedback.  Introduction The natural abstraction hypothesis (NAH) claims: * Our physical world “abstracts well” into high-level abstractions of low-level systems. * These abstractions are...
c7782ac9-e876-44e1-af2f-79e7303e379d
trentmkelly/LessWrong-43k
LessWrong
Rationalistic Losing Playing to learn I like losing. I don't even think that losing is necessarily evil. Personally, I believe this has less to do with a desire to lose and more to do with curiosity about the game-space. Technically, my goals are probably shifted into some form of meta-winning — I like to understand ...
1ce32f74-4bcb-49bf-ae47-887f3caf5ac0
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Delegative Inverse Reinforcement Learning We introduce a reinforcement-like learning setting we call *Delegative Inverse Reinforcement Learning* (DIRL). In DIRL, the agent can, at any point of time, delegate the choice of action to an "advisor". The agent knows neither the environment nor the reward function, whereas ...
9c355b68-88ac-474d-ac8b-ae0d26185ad2
trentmkelly/LessWrong-43k
LessWrong
Infinite Ethics: Infinite Problems 1 Introduction (Crossposted from my blog--the formatting is probably a bit better there.) > With entity, dementedly meant to be infinite —Eminem. (Crossposted from my blog).  The formatting is better there.   I never had a Nietzsche phase. You know the way lots of people get obs...
bbb4a680-6700-403d-95da-57be202f1e3d
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Importance of foresight evaluations within ELK This post makes a few basic observations regarding the importance of foresight-based evaluations within ELK [[Christiano et al 2021](https://www.alignmentforum.org/posts/qHCDysDnvhteW7kRd/arc-s-first-technical-report-eliciting-latent-knowledge)]: 1. Without foresight-bas...