id
stringlengths
36
36
source
stringclasses
15 values
formatted_source
stringclasses
13 values
text
stringlengths
2
7.55M
18201389-b5a0-411b-98f8-56970e259e3f
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
AI Benefits Post 4: Outstanding Questions on Selecting Benefits This is a post in a series on "AI Benefits." It is cross-posted from my [personal blog](https://www.cullenokeefe.com/blog/ai-benefits-4). For other entries in this series, navigate to the [AI Benefits Blog Series Index page](https://www.cullenokeefe.com/a...
818d56bc-c0e2-4558-98f9-4d1f5168cf8d
trentmkelly/LessWrong-43k
LessWrong
My Kind of Moral Responsibility The following is an excerpt of an exchange between Julia Galef and Massimo Pigliucci, from the transcript for Rationally Speaking Podcast episode 132: > Massimo: [cultivating virtue and 'doing good' locally 'does more good' than directly eradicating malaria] > > Julia: [T]here's lower...
b1ccabad-4cf6-431a-95a6-1343cab294ae
trentmkelly/LessWrong-43k
LessWrong
Chronostasis: The Time-Capsule Conundrum of Language Models In the rapidly evolving world of artificial intelligence, large language models like GPT-4 have demonstrated remarkable capabilities in understanding and generating human-like text. These models owe their proficiency to massive datasets that capture a snapsho...
5eb838f7-5772-4053-8dfe-b9323f5a70c8
trentmkelly/LessWrong-43k
LessWrong
Rationality Quotes: June 2011 Y'all know the rules: * Please post all quotes separately, so that they can be voted up/down separately.  (If they are strongly related, reply to your own comments.  If strongly ordered, then go ahead and post them together.) * Do not quote yourself. * Do not quote comments/posts on L...
67364899-aee0-44a4-961e-0a677aa1fce7
trentmkelly/LessWrong-43k
LessWrong
Harry Potter and the Methods of Rationality discussion thread, February 2015, chapter 108 New long chapter! Since I expect its discussion to generate more than 160 comments (which would push the previous thread over the 500 comment limit) before the next chapter is posted, here is a new thread. This is a new thread t...
e4e2b9ae-714f-410d-85e2-1c27c259d13b
LDJnr/LessWrong-Amplify-Instruct
LessWrong
"Related to: Truly Part of You, A Technical Explanation of Technical Explanation Partly because of LessWrong discussions about what really counts as understanding (some typical examples), I came up with a scheme to classify different levels of understanding so that posters can be more precise about what they mean when ...
11e658eb-d686-488f-96b0-c416c8810033
trentmkelly/LessWrong-43k
LessWrong
Meetup : Effective Altruism Dinner Party Discussion article for the meetup : Effective Altruism Dinner Party WHEN: 22 February 2013 02:13:22PM (-0500) WHERE: Winterfell House, 316 W 138th street, New York NY, 10030 This Friday at Winterfell House, NY will be an Effective Altruism dinner party. Effective Altruism is...
74996930-9246-43b3-820e-b23474adaad9
trentmkelly/LessWrong-43k
LessWrong
Research Agenda v0.9: Synthesising a human's preferences into a utility function I'm now in a position where I can see a possible route to a safe/survivable/friendly Artificial Intelligence being developed. I'd give a 10+% chance of it being possible this way, and a 95% chance that some of these ideas will be very use...
d0ad64b3-f6b9-4d4a-88ea-08c1273449ef
trentmkelly/LessWrong-43k
LessWrong
Problems with predictive history classes Here I collect and (partly) address problems noted by commenters on "Predictive history classes", in no particular order. I do not expect to refute all of them. Some of them may actually cripple the proposal/show it to be impractical. I figure this is more honest than leaving ...
2ed6971d-bdce-4477-ac0d-0a30f3a78590
trentmkelly/LessWrong-43k
LessWrong
Behaviorism: Beware Anthropomorphizing Humans Related to: The Comedy of Behaviorism Behaviorism's gotten a bad rap. It's gone down in history as the school founded upon the idea that there's no such thing as mental phenomena or cognitive processing, and if there are we can't ever know anything about them, and if we ...
1124a458-4c2a-49b9-a997-16d952c773f0
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
AI strategy nearcasting This is the first in a series of pieces taking a stab at dealing with a conundrum: * I believe this could be the [most important century ever for humanity](https://www.cold-takes.com/most-important-century/), via the development of advanced AI systems that could dramatically speed up scienti...
97225f8a-1730-417d-8f78-fb7e07007f9d
trentmkelly/LessWrong-43k
LessWrong
Biofuels a climate mistake
3e3843b1-ac64-44a1-a772-b2d6349034f2
trentmkelly/LessWrong-43k
LessWrong
Oracle predictions don't apply to non-existent worlds In this post, I hope to persuade you of what I consider to be an important principle when dealing with decision theory and counterfactuals. Joe Carlsmith describes the Yankees vs. Red Sox Problem as below: > In this case, the Yankees win 90% of games, and you fac...
4c42a9f2-ea9a-4dbc-a8de-14ee847a8542
trentmkelly/LessWrong-43k
LessWrong
Brief Question about FAI approaches I've been reading through this to get a sense of the state of the art at the moment: http://lukeprog.com/SaveTheWorld.html Near the bottom, when discussing safe utility functions, the discussion seems to center on analyzing human values and extracting from them some sort of clean,...
05052b54-d4dc-46ef-8486-56ef6df12113
StampyAI/alignment-research-dataset/special_docs
Other
Suffering-focused AI safety: Why “fail-safe” measures might be our top intervention Report FRI-16-1 Suffering-focused AI safety: Why “fail-safe” measures might be our top intervention Lukas Gloor Foundational Research Institute lukas.gloor@foundational-research.org June 2016 Abstract AI-safety efforts focused on sufferin...
5c55f54c-e435-4124-9b49-8598cdba7838
trentmkelly/LessWrong-43k
LessWrong
Custom Audio Switch Box When I play live I have a bunch of instruments, including: * Mandolin: an electric mandolin * Computer: a custom MIDI mapper driven by keyboard, foot drums, and breath controller. * Bass whistle: a whistle-controlled bass synthesizer I also have some effects, primarily a talkbox and an...
d2949fbc-aee2-4386-bc9b-3475e4286d5b
trentmkelly/LessWrong-43k
LessWrong
What is "True Love"? Meta: I recently entered the dating market, so naturally I have lots of random thoughts on the subject which you all get to suffer through for a while. Your usual diet of dry math and agency theory will resume shortly. Obviously the phrase “true love” has been so thoroughly overdone in so much fi...
1a4541e9-f272-445e-8e9c-88ac0e16e933
trentmkelly/LessWrong-43k
LessWrong
"25 Lessons from 25 Years of Marriage" by honorary rationalist Ferrett Steinmetz The Ferrett isn't an official member of the Rationalist Diaspora, but he's been blogging for longer than LessWrong has existed and often has useful insights that align with what the LessWrong community likes to talk and think about, and ...
0e63a3c8-7508-48ec-977c-86e23f5b2d0a
trentmkelly/LessWrong-43k
LessWrong
Covid 12/10: Vaccine Approval Day in America Today, the FDA is meeting to discuss Pfizer’s Covid-19 vaccine. By the time many of you read this, they will have hopefully given the vaccine, and perhaps Moderna’s as well, emergency use authorization. If that happens, distributions to the states can begin, and some people...
b0c4c72d-9854-487e-b74a-c364b1fff621
trentmkelly/LessWrong-43k
LessWrong
Machine Learning Group After signing up for this post, those of us that want to study machine learning have made a team. In an effort to actually get high returns on our time we won't delay, and instead actually build the skills. First project: work through Python Machine Learning by Sebastian Raschka, with the mid-t...
9ad07746-bd9c-4c75-8dbe-d94c10bea264
trentmkelly/LessWrong-43k
LessWrong
Welcome to Warsaw Rationality Dojo [Edit With Your Details] (The following are our suggestions for what kind of information is best to include in the welcome post of your group, feel free to replace them with whatever you think is best) What kind of events does your group usually run? What does it usually do? How fr...
f8a00c13-a6bd-472e-a50c-166618dfef98
trentmkelly/LessWrong-43k
LessWrong
Just because 2 things are opposites, doesn't mean they're just the same but flipped The 2 Aspects There’s 2 Aspects to things in general. I will call them Mapping Out and Mapping In, in titlecase so you know they’re distinct concepts. warmup: 0 -> 1 Here, 0 is an initial object and 1 is a terminal object. 0 is Mappe...
da62fadd-a881-4c1f-a427-e81b21a73f8f
StampyAI/alignment-research-dataset/blogs
Blogs
"Why Tool AIs Want to Be Agent AIs" > Autonomous AI systems (Agent AIs) trained using [reinforcement learning](!W) can do harm when they take wrong actions, especially superintelligent Agent AIs. One solution would be to eliminate their agency by not giving AIs the ability to take actions, confining them to purely inf...
74d59d2b-1098-43dd-bc2b-bafc29253d25
trentmkelly/LessWrong-43k
LessWrong
How to not talk about probability estimates I have recently begun asking for people's credence in interesting beliefs. I was speaking with two of my peers at dinner yesterday, and our discussion chanced upon the topic of the existence of extraterrestrial life. Curious, I asked them what probability they would assign ...
06bde087-34b3-4b20-86f4-132fa0b6e9b0
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] Selecting Rationalist Groups Today's post, Selecting Rationalist Groups was originally published on 02 April 2009. A summary (taken from the LW wiki):   > Trying to breed e.g. egg-laying chickens by individual selection can produce odd side effects on the farm level, since a more dominant hen can produce...
4041923a-54d8-4170-bddf-9a8761062c68
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
[Linkpost] Existential Risk Analysis in Empirical Research Papers I’ve noted before that it can be difficult to separate progress in safety from progress in capabilities. However, doing so is important, as we want to ensure that we are making differential progress on safety, rather than just advancing safety as a cons...
f56e3afb-cad0-4194-bbd9-a2afdd7c0c13
trentmkelly/LessWrong-43k
LessWrong
A reason to see the future I just learned of The Future Library project. In short, famous authors will be asked to write new, original fiction that will not be released until 2114. First one announced was Margaret Atwood, of The Handmaiden's Tale fame. I learned of this when a friend posted on Facebook that "I'm offi...
2dba6755-602d-49b4-9ee2-5b7494f0299e
trentmkelly/LessWrong-43k
LessWrong
[Review] Two People Smoking Behind the Supermarket When I was a teenager, I worked as a street performer on Seattle's waterfront pier. A good street performer can make a good income. I wasn't good. I performed near a fish & chips restaurant. I couldn't afford the fish & chips, but if I packed a peanut butter sandwich,...
8481d4ad-b718-4bd7-a8ae-5a69fb6a57ac
awestover/filtering-for-misalignment
Redwood Research: Alek's Filtering Results
id: post138 We are releasing a new paper called “The Elicitation Game: Evaluating Capability Elicitation Techniques” . See tweet thread here . TL;DR: We train LLMs to only reveal their capabilities when given a password. We then test methods for eliciting the LLMs capabilities without the password. Fine-tuning works b...
a3c4f064-9980-4c5d-891e-301a5f19549d
trentmkelly/LessWrong-43k
LessWrong
Focus Your Uncertainty Will bond yields go up, or down, or remain the same? If you’re a TV pundit and your job is to explain the outcome after the fact, then there’s no reason to worry. No matter which of the three possibilities comes true, you’ll be able to explain why the outcome perfectly fits your pet market theor...
3f45da2b-b299-4816-9aa5-24c2de8abb8f
StampyAI/alignment-research-dataset/arxiv
Arxiv
Adversarial Motion Priors Make Good Substitutes for Complex Reward Functions. I Introduction --------------- Developing controllers for high-dimensional continuous control systems such as legged robots has long been an area of study. Early work in this field focused on developing approximate dynamics models of a sy...
1cc52c05-ff89-46f5-99c6-facfac063016
trentmkelly/LessWrong-43k
LessWrong
Meetup : Washington DC meetup- new sequence? Discussion article for the meetup : Washington DC meetup- new sequence? WHEN: 23 December 2012 03:00:00PM (-0500) WHERE: National Portrait Gallery Plaza, Washington, DC 20001, USA Topic is tentative. Discussion article for the meetup : Washington DC meetup- new sequenc...
4729b9a1-5494-417c-a96a-9b747dd866da
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] Practical Advice Backed By Deep Theories Today's post, Practical Advice Backed By Deep Theories was originally published on 25 April 2009. A summary (taken from the LW wiki):   > Practical advice is genuinely much, much more useful when it's backed up by concrete experimental results, causal models that ...
b3f92324-3844-4930-a49a-6b81258ff761
trentmkelly/LessWrong-43k
LessWrong
[LINK] Prizes and open source for drug research (proposed, and some politics) Trying to get drug research to happen in spite of the drug companies: The article also has rather a lot about US government opposition to the proposed treaty, which includes a requirement that member nations spend 0.01% of GDP annually on n...
a520336e-df01-47b7-b201-f8093724b33c
trentmkelly/LessWrong-43k
LessWrong
A Bird's Eye View of the ML Field [Pragmatic AI Safety #2] This is the second post in a sequence of posts that describe our models for Pragmatic AI Safety.  The internal dynamics of the ML field are not immediately obvious to the casual observer. This post will present some important high-level points that are critic...
0e9b61a2-4d15-4d5a-8698-3cc52b05e8f8
trentmkelly/LessWrong-43k
LessWrong
Arguing Absolute Velocities CW: Analogies You and a friend are arguing over a physics question: "You're on a train speeding seventy miles per hour east, while you run 5 miles per hour west. How fast are you moving?" That's obvious, you think to yourself. The question says I'm running 5 miles per hour west, so I must...
4bf7455f-6cff-41a7-9263-47e877824320
trentmkelly/LessWrong-43k
LessWrong
Exploiting Crypto Prediction Markets for Fun and Profit I believe that exploiting inefficiencies in crypto prediction markets is an easy way to make market beating returns. This guide is a step by step walkthrough of how to make bets on these markets. I think this is useful because the process is somewhat complicated ...
c42d5d0e-0e3f-4c5b-b563-49cd29d1d847
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Seeking social science students / collaborators interested in AI existential risks ***tldr:** I'm looking for undergraduate research assistants / collaborators to work on research questions at the intersection of social science and long-term risks from AI. I've collected some research questions* [*here*](https://docs....
a8ff0f8f-b1be-4137-b6b5-297b16cc11a8
trentmkelly/LessWrong-43k
LessWrong
Evolutionary Psychology Like "IRC chat" or "TCP/IP protocol", the phrase "reproductive organ" is redundant. All organs are reproductive organs. Where do a bird's wings come from? An Evolution-of-Birds Fairy who thinks that flying is really neat? The bird's wings are there because they contributed to the bird's ancesto...
57f7f3a6-f612-4f14-8336-3f08e310b181
trentmkelly/LessWrong-43k
LessWrong
Is Bjorn Lomborg roughly right about climate change policy? I recently came across Bjorn Lomborg’s 2020 article Welfare in the 21st century: Increasing development, reducing inequality, the impact of climate change, and the cost of climate policies. It runs counter to many things I have heard about climate change. For...
1d2bb459-88f5-41af-88b2-16bdb297e147
trentmkelly/LessWrong-43k
LessWrong
Freewill vs. Determinism This topic has been covered ad nauseum on nearly every page on the internet. But it still generates so much good conversation that I will stir he pot again.  Imagine you are walking down the street and come upon a woman lying on the ground crying. Suddenly you realize she has obviously been a...
1f32f080-47d8-40c5-867e-2777e1659252
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Responses to Christiano on takeoff speeds? It's been over two and a half years since Paul put [this blog post on takeoff speeds](https://sideways-view.com/2018/02/24/takeoff-speeds/) online. In particular, it argues that the "fast takeoff" undergone by humans is not very strong evidence that AIs will also undergo a fa...
fe05f129-8410-401e-ab52-0bd95e37db61
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
De Dicto and De Se Reference Matters for Alignment I submitted an entry in the current [AI Fables Contest](https://forum.effectivealtruism.org/posts/gYxY5Mr2srBnrbuaT/announcing-the-ai-fables-writing-contest), "[The Fable of the Humans](https://docs.google.com/document/d/1ofDjFNWSdree1h4WzIg27fI7pKoMWmMxjeDGyXviwoQ/ed...
de2a1b92-7b0d-4098-a379-70cc96530793
trentmkelly/LessWrong-43k
LessWrong
What product are you building? Epistemic status: It appears like it works for me. Not meant to be a hard/fast rule. A frame I frequently examine conversations through is "are we building a product together, or not?" Many conversations (online or in person) are "just sorta hanging out." Just-sorta-hanging-out can be ...
b18c9cc1-c6b2-45a0-ac68-5a04351b2f28
trentmkelly/LessWrong-43k
LessWrong
Apply to be a Safety Engineer at Lockheed Martin! Are you passionate about ensuring the safety and reliability of the world’s most lethal and cutting-edge weaponry? Does the idea of creating technology and then working out its impacts excite you? Do you thrive in dynamic environments where innovation meets rigorous sa...
0c1af644-f12b-42b5-a301-98f7ea2080c0
trentmkelly/LessWrong-43k
LessWrong
Polymarket Covid-19 1/17/2022 Polymarket’s prediction markets are back and they are more plentiful than ever, both on Covid-19 and otherwise. Americans are not eligible to trade, but plenty of my readers reside elsewhere. And even if you can’t trade, the information component is still free. Liquidity in any given mar...
706f759e-40a9-4180-a34f-6b91bbb4cc45
trentmkelly/LessWrong-43k
LessWrong
Studies On Slack I. Imagine a distant planet full of eyeless animals. Evolving eyes is hard: they need to evolve Eye Part 1, then Eye Part 2, then Eye Part 3, in that order. Each of these requires a separate series of rare mutations. Here on Earth, scientists believe each of these mutations must have had its own bene...
6666c908-213a-4238-890a-fe2bc87b571f
trentmkelly/LessWrong-43k
LessWrong
We might be missing some key feature of AI takeoff; it'll probably seem like "we could've seen this coming" Predicting the future is hard, so it’s no surprise that we occasionally miss important developments. However, several times recently, in the contexts of Covid forecasting and AI progress, I noticed that I misse...
347a3b0c-aec9-4d59-8624-5638bcd0c247
trentmkelly/LessWrong-43k
LessWrong
Meetup : Washington, D.C.: Fun & Games Discussion article for the meetup : Washington, D.C.: Fun & Games WHEN: 22 January 2017 03:30:00PM (-0500) WHERE: Donald W. Reynolds Center for American Art and Portraiture We will be meeting in the courtyard to hang out, play games, and engage in fun conversation. Upcoming m...
ce72c0b4-e6d0-48f5-a142-ea38329c227d
trentmkelly/LessWrong-43k
LessWrong
Thoughts on How Consciousness Can Affect the World Overview In "The Generalized Anti-Zombie Principle" Eliezer mentioned that one aspect of consciousness is that it can affect the world, e.g. by making us say out loud "I feel conscious", or deciding to read (or write) this article. However my current working hypothesi...
f280e600-45bd-4d3c-aae1-2d9935eba6f1
StampyAI/alignment-research-dataset/blogs
Blogs
Video Games Needs A Platform Video Games Needs A Platform ---------------------------- i feel like, for video games to feel legitimate as an artform, it needs a platform that has the following properties: * open source, rather than controlled by one or even several companies * standard, such that every artist know...
9ce87ad3-d2e0-49a9-9e4f-53aab5f45889
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Why Bet Kelly? The Kelly criterion is an elegant, but often misunderstood, result in decision theory. To begin with, suppose you have some amount of some resource, which you would like to increase. (For example, the resource might be monetary wealth.) You are given the opportunity to make a series of identical bets. Y...
3dc5faa1-e002-4c23-9e3c-e59dbc49bdcb
trentmkelly/LessWrong-43k
LessWrong
Implications of large language model diffusion for AI governance This post is one part of the sequence Understanding the diffusion of large language models.  As context for this post, I strongly recommend reading at least the 5-minute summary of the sequence. The primary aim of this research project was to be descrip...
ebcbfdfc-586f-487d-b0d0-8d0c4bc5d5f4
StampyAI/alignment-research-dataset/lesswrong
LessWrong
The reward engineering problem Today we usually train reinforcement learning agents to perform narrow tasks with simple goals. We may eventually want to train RL agents to behave “well” in open-ended environments where there is no simple goal. Suppose that we are trying to train an RL agent **A**. In each episode, *...
a94383af-3742-4f01-940a-b6996ce0ed41
trentmkelly/LessWrong-43k
LessWrong
Help Ukraine Hi, my name is Dmytro. I'm from Ukraine, Kyiv. You've probably heard about what's going on here now. It's not "crisis" or "conflict" it's a WAR. People are dying, thousands, cities are bombed, millions of refugees, millions left their homes. The government is trying to control the situation and it's doin...
6dbc219a-36c2-4098-a9d9-ff92f0a86780
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Categorizing failures as “outer” or “inner” misalignment is often confused Pop quiz: Are the following failures examples of outer or inner misalignment? Or is it ambiguous? (I’ve spoilered the scenarios so you look at each one separately and answer before moving on to the next. Even if you don’t want to actually answ...
fec7aca5-9f53-4c5e-90cc-250580cca8bd
trentmkelly/LessWrong-43k
LessWrong
AI #54: Clauding Along The big news this week was of course the release of Claude 3.0 Opus, likely in some ways the best available model right now. Anthropic now has a highly impressive model, impressive enough that it seems as if it breaks at least the spirit of their past commitments on how far they will push the fr...
d9971c89-db7b-4563-a1be-09409f28a295
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] Rationality and the English Language Today's post, Rationality and the English Language was originally published on 12 September 2007. A summary (taken from the LW wiki):   > George Orwell's writings on language and totalitarianism are critical to understanding rationality. Orwell was an opponent of the ...
7a9ea4a8-9738-447a-bd6c-ddf4437b58cc
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Alignment Newsletter #27 Dan Hendrycks has now joined, and will likely write summaries primarily on adversarial examples and robustness. As with Richard, his summaries are marked as such; I'm reviewing some of them now but expect to review less over time. Highlights ---------- **[80K podcast with Paul Christiano](ht...
96541e92-e809-47eb-b3c8-adf0e0521f77
trentmkelly/LessWrong-43k
LessWrong
Welcome to LW-Cologne LW-Cologne runs a monthly social meetup for people interested in LW / SSC style rationalist philosophy. We also have some overlap with the local EA community. Sometimes there are additional events and new suggestions especially suggestions by new people are allways welcome (If you want something...
6d8101d3-3c53-425a-bfe7-3be5ee3e4c34
trentmkelly/LessWrong-43k
LessWrong
Subagents and impact measures, full and fully illustrated 0. Introduction: why yet another post about subagents? I’ve recently been writing a sequence on how subagents can undermine impact penalties such as attainable utility preservation. I’m not happy with that sequence; it’s messy and without examples (apart from i...
3cc30d8e-7ab1-4a22-8eae-1e58dada2915
trentmkelly/LessWrong-43k
LessWrong
Significance of Compression Rate Method Summary: The significance of the Compression Rate Method (CRM) is that it justifies a form of empirical inquiry into aspects of reality that have previously resisted systematic interrogation. Some examples of potential investigations are described. A key hypothesis is discussed,...
c4944bad-4e58-41cc-839b-41807da89f97
trentmkelly/LessWrong-43k
LessWrong
How to Evaluate Data? What I'm trying to figure out is, how to I determine whether a source I'm looking at is telling the truth? For an example, let's take this page from Metamed: http://www.metamed.com/vital-facts-and-statistics At first glance, I see some obvious things I ought to consider. It often gives numbers f...
8b02686a-1d15-4966-92af-b507128a2bfb
trentmkelly/LessWrong-43k
LessWrong
Calculated Clairvoyance (Crosspost: spacelutt.com/cal) In Steven Universe, there is a character called Sapphire who has the ability to see into the future. For the entirety of the show, it is presented to us as an innate ability which Sapphire can do naturally. But in one of the final episodes of the show, Sapphire,...
896a7155-173d-4924-9aa2-0d7d53b102dc
trentmkelly/LessWrong-43k
LessWrong
The danger of wishful thinking Or "The problems inherent in making a goal maximiser with a changing world model." No paper clips were created or destroyed in the making of this script. *This is an experimental post to try and get this point across. I'll write something similar for the type of systems I would like to...
fba4cb25-b7ec-4de3-9355-fdcbe110d6c9
trentmkelly/LessWrong-43k
LessWrong
Addresses in the Multiverse Abstract: If we assume that any universe can be modeled as a computer program which has been running for finitely many steps, then we can assign a multiverse-address to every event by combining its world-program with the number of steps into the world-program where it occurs. We define a pr...
b82244aa-8291-4d44-9d18-37ebc1c63e65
trentmkelly/LessWrong-43k
LessWrong
Brain storm: What is the theory behind a good political mechanism? Patrissimo argue that we should try to design good mechanisms for governance rather than try and use the current broken mechanisms. I agree, however we don't have a theoretical framework that we can use to evaluate different systems that are proposed....
1aa7ba5f-5a0e-4568-b382-7ce7a13fc089
trentmkelly/LessWrong-43k
LessWrong
Spreading messages to help with the most important century Click lower right to download or find on Apple Podcasts, Spotify, Stitcher, etc. In the most important century series, I argued that the 21st century could be the most important century ever for humanity, via the development of advanced AI systems that coul...
d41c31cc-ff6d-4946-90cc-7f712a3fdac6
trentmkelly/LessWrong-43k
LessWrong
I bet $500 on AI winning the IMO gold medal by 2026 The bet was arranged on Twitter between @MichaelVassar and I (link). Conditions are similar to this question on Metaculus, except for the open-source condition (I win even if the AI is closed-source, and in fact I would very much prefer it to be closed-source). @Zv...
774e1be6-0348-4701-843a-f9a0b2b62d9d
trentmkelly/LessWrong-43k
LessWrong
A new (?) anthropic principle, and motivation thereof The discussions of anthropic principles that I've seen on LW and, say, Katja Grace's list, are limited to SIA and SSA. These both give some results that most people find unintuitive (Doomsday for SIA, presumptuous philosopher for SIA), and the debate between them s...
f708591f-dc1c-40a1-9370-03023275929d
trentmkelly/LessWrong-43k
LessWrong
The World is Continuous, Not Discrete We have a bias towards using discrete mental models over continuous ones because they're more convenient. But they're not always the best choice. I've heard a lot of different definitions for "dead". Some people say living things have souls, and they die when the soul leaves. Oth...
67552604-a4bb-4aa9-ab86-dc6f0e568bb6
StampyAI/alignment-research-dataset/arbital
Arbital
The rocket alignment problem %todo: reference specific contributions by others or oneself that led to emergent discoveries at NASA et. al. -- proceed to display how these specific contributions may relate to current state of the art developments in artificial intelligence% (*Somewhere in a not-very-near neighboring w...
ccfc747c-80ac-4387-90ee-efda656f9bf6
StampyAI/alignment-research-dataset/blogs
Blogs
Embedded World-Models An agent which is larger than its environment can:   * Hold an exact model of the environment in its head. * Think through the [consequences](https://www.lesswrong.com/posts/tKwJQbo6SfWF2ifKh/toward-a-new-technical-explanation-of-technical-explanation) of every potential course of action. * I...
a9a89bc6-6e0e-4b0a-93a9-006d7290c03f
trentmkelly/LessWrong-43k
LessWrong
AI Benefits Post 1: Introducing “AI Benefits” This is a post in a series on "AI Benefits." It is cross-posted from my personal blog. For other entries in this series, navigate to the AI Benefits Blog Series Index page. This post is also discussed on the Effective Altruism Forum. Links to those cross-posts are availab...
5808c60c-f082-4dd2-9ecf-7bf059131451
trentmkelly/LessWrong-43k
LessWrong
Can you only realize objects in your mind after they appear? When meditating, you try to “be present”. But, say a background noise happens. You can only be aware of the noise after the fact. Does this take you away from the present? Or is it that you are actually observing the mind, what the mind is doing in the pres...
203c17dd-0e0c-4b09-b5e9-e56805a8ddf2
trentmkelly/LessWrong-43k
LessWrong
Navigation by Moonlight Previously in sequence: You Are not a Thought Experiment Cross-posted from SecondPerson.dating. ---------------------------------------- I want to conclude our interplanetary exploration of sex differences with one that’s of great interest to me personally: there seems to be little explicit ...
295f0964-59c4-4217-bbd7-29555804408c
trentmkelly/LessWrong-43k
LessWrong
Wisdom of the Crowd: not always so wise I have a confession to make: I have been not "publishing" my results to an experiment because the results were uninteresting. You may recall some time ago that I made a post asking people to take a survey so that I could look at a small variation of the typical "Wisdom of the Cr...
7841bbae-7f8c-4491-8537-b1577d008f61
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] Class Project Today's post, Class Project was originally published on 31 May 2008. A summary (taken from the LW wiki):   > From the world of Initiation Ceremony. Brennan and the others are faced with their midterm exams. Discuss the post here (rather than in the comments to the original post). This po...
994356b5-c60a-4796-9f56-a8db38fbddcf
trentmkelly/LessWrong-43k
LessWrong
Are the other Rationality: A-Z sequences coming out as books? I found a website that purports that the remaining 4 books in the series are “coming soon.” But it doesn’t indicate a date, and I don’t know if they’re perpetually “coming soon” or if this is still a project being pursued. Anyone know about the status of th...
b4a7bf77-dc87-422f-b57c-443153afa237
trentmkelly/LessWrong-43k
LessWrong
Actually Practicing Rationality and the 5-Second Level [I first posted this as a link to my blog post, but I'm reposting as a focused article here that trims some fat of the original post, which was less accessible] ---------------------------------------- I think a lot about heuristics and biases, and I admit that ...
312985a9-9bd8-4d87-937a-d49e8edcc85e
trentmkelly/LessWrong-43k
LessWrong
Simply locate yourself Imagine I offer you the following bet: I'll roll a fair ten-sided die. If it comes up 1-9, you win a million dollars. If it comes up 0, you lose $10,000. (If you're significantly richer or poorer than the median person, adjust the numbers up or down accordingly, such that winning is very great a...
2e22fbdd-3198-4f2c-8c08-7226ce921ed2
trentmkelly/LessWrong-43k
LessWrong
Thoughts on Hardware limits to Prevent AGI? Summary: An individual Commodore 64 is almost certainly safe, Top 10 super computers could almost certainly run a superpowerful AGI, but where is the safe line, and how would we get to the safe side? I started thinking about this topic when I realized that we can safely use...
83ba6095-28bb-4aff-b58b-352b8b9f8e19
trentmkelly/LessWrong-43k
LessWrong
Toy model of human values This is just a summary via analogy where I human values come from, as far as I understand it. The expanded version would be Eli's http://lesswrong.com/lw/l3/thou_art_godshatter/. The basic analogy is to chess-playing programs (at least the basic ones from 40 years ago, the art has progressed...
ec4d729c-c19a-411b-9f15-c935a8b40eb9
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Dual-Useness is a Ratio A lot of AI-risk-concerned people are struggling with how to relate to dual-use research, and relatedly, to doing alignment research inside of AI orgs. There's a pretty simple concept that seems, to me, to be key to thinking about this coherently: the dual-useness ratio. Most prosaic alignment ...
a3236f2b-6a27-4488-9d20-c3bc65687074
trentmkelly/LessWrong-43k
LessWrong
Covid-19 5/7: Fighting Limbo Last week: Covid-19 4/30: Stuck in Limbo Recently: Covid-19: New York’s Antibody Tests 2, On “COVID-19 Superspreader Events in 28 Countries: Critical Patterns and Lessons” Background Assumptions: On R0, Taking Initial Viral Load Seriously, On New York’s Antibody Tests, My Covid-19 Thinki...
86986612-07a4-44f6-b58c-5cbfffd98b82
trentmkelly/LessWrong-43k
LessWrong
Vaniver's thoughts on Anthropic's RSP Announcement, Policy v1.0, evhub's argument in favor on LW. These are my personal thoughts; in the interest of full disclosure, one of my housemates and several of my friends work at Anthropic; my spouse and I hold OpenAI units (but are financially secure without them). This post ...
dc661c98-41f2-41bd-92d6-d912ab865d3a
trentmkelly/LessWrong-43k
LessWrong
What to include in a guest lecture on existential risks from AI? A professor I'm friendly with has been teaching a course on AI ethics this semester, and he asked me if I could come give a guest lecture on "AI apocalypse" scenarios. What should I include in the lecture? Details: * Audience is mostly graduate engine...
1578b0bd-3819-436d-8610-29a0bd43bdc2
trentmkelly/LessWrong-43k
LessWrong
Bay Area Meetup Saturday 6/12 There will be a meetup at 7PM this Saturday at the SIAI house (3755 Benton St, Santa Clara). (More info on the official page.) The usual set of people will be present, as well as new SIAI Visiting Fellows. (Sorry for the short notice.)
c316f0bb-b2b7-4ace-b5e1-06b742f289ef
trentmkelly/LessWrong-43k
LessWrong
Covid 5/27: The Final Countdown (Personal note: My Facebook account was hacked. Attacker has not contacted me, and deleted my name from the profile. Since I am already Against Facebook I do not view this as a great loss unless it leads to further trouble, but readers should be aware that for now I have no Facebook acc...
41ad7971-fbe7-4e96-9aa9-d8b1f683b88d
trentmkelly/LessWrong-43k
LessWrong
Rational vs Reasonable This post draws ideas from Personhood: A Game for Two or More Players on Melting Asphalt. I've been lax in my attempt to write something for LW once weekly, but I hope to approximately continue nonetheless. I still have many posts planned -- the next one after this will likely be a rationality ...
ad8220aa-9f23-49eb-a051-9c29aa9e8afb
trentmkelly/LessWrong-43k
LessWrong
Our society lacks good self-preservation mechanisms The prospect of a dangerous collection of existential risks and risks of major civilizational-level catastrophes in the 21st century, combined with a distinct lack of agencies whose job it is to mitigate against such risks probably indicates that the world might be i...
33140d3b-c62b-4445-8b3b-981a38b0cbe6
trentmkelly/LessWrong-43k
LessWrong
Machines vs Memes Part 3: Imitation and Memes This is the last in a series of three posts on the interlinkages between Memetics and AI Alignment. It was written as an output from the 2022 AI Safety Camp, for the ‘Impact of Memetics on Alignment’ team, coached by Daniel Kokotajlo and comprising Harriet Farlow, Nate Rus...
9cf863bf-72de-4632-a8ab-6f1dcebab86f
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Stable Pointers to Value: An Agent Embedded in Its Own Utility Function (This post is largely a write-up of a conversation with Scott Garrabrant.) Stable Pointers to Value ------------------------ How do we build stable pointers to values? As a first example, consider the wireheading problem for AIXI-like agents...
eba6cd55-7236-45f7-b816-80ba5ace2b18
StampyAI/alignment-research-dataset/distill
Distill Scientific Journal
Sequence Modeling with CTC Introduction ------------ Consider speech recognition. We have a dataset of audio clips and corresponding transcripts. Unfortunately, we don’t know how the characters in the transcript align to the audio. This makes training a speech recognizer harder than it might at first seem. Without...
8eef9317-7b96-426d-849a-fd66c1def839
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Risks from Learned Optimization: Introduction *This is the first of five posts in the [Risks from Learned Optimization Sequence](https://www.alignmentforum.org/s/r9tYkB2a8Fp4DN8yB) based on the paper “[Risks from Learned Optimization in Advanced Machine Learning Systems](https://arxiv.org/abs/1906.01820)” by Evan Hubi...
7474b619-6956-4ba9-90ae-45454281423a
trentmkelly/LessWrong-43k
LessWrong
Mathematical models of Ethics Hello, I am currently trying to find all the attempts at mathematical modeling of ethics (utilitarianism, axiologies, etc.) I have found several, do you have any ideas of keywords to type or even articles to suggest here ? An example of a (simple) model is the notion of utility in the Vo...
e2d67f8f-d40c-4b23-b5c8-57db9cb57b46
trentmkelly/LessWrong-43k
LessWrong
A mind needn't be curious to reap the benefits of curiosity Context: Stating a point that is obvious in local circles, but that I regularly run into among economists and longevity researchers and more general technologists and on twitter and so on. ---------------------------------------- Short version: To learn thi...
432e1117-daa9-46bc-9259-f13a396eb166
trentmkelly/LessWrong-43k
LessWrong
AI #14: A Very Good Sentence “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” That is the entire text of the one-line open letter signed this week by what one could reasonably call ‘everyone,’ including the CEOs of all thre...
1421056a-8094-4d56-8cb7-4ea659f7611b
trentmkelly/LessWrong-43k
LessWrong
Series of absurd upgrades in nature's great search So you want to find that special thing that replicates best and lasts longest? Just vibrate a bunch of molecules a long time! You might reasonably assume that molecules wouldn't practically ever randomly assemble themselves into anything worth looking at. It happens ...
d4782524-8fb1-4689-ad85-e7b7ffc7c85a
trentmkelly/LessWrong-43k
LessWrong
[LINK] Cliffs Notes: "Probability Theory: The Logic of Science", part 1 http://reasonableapproximation.net/2014/02/02/cliffs-notes-pttlos-part-1.html > A book sometimes cited on LessWrong as recommended reading is E.T. Jaynes' Probability Theory: The Logic of Science. I intend to write a series of posts reading this ...