id
stringlengths
36
36
source
stringclasses
15 values
formatted_source
stringclasses
13 values
text
stringlengths
2
7.55M
e5569405-a243-4479-9f0a-2bcef9f14a57
trentmkelly/LessWrong-43k
LessWrong
10/28/17 Development Update: New editor framework (markdown support!), speed improvements, style improvements Here is a quick summary of the changes I just pushed: * We now have a new editor framework (we moved from ory-editor to draft-js-plugins). The primary reason for this was performance, though we also had a fe...
570a6ac3-a1cb-4581-aa80-8df1ba8def0f
trentmkelly/LessWrong-43k
LessWrong
On attunement (Cross-posted from my website. Podcast version here, or search for "Joe Carlsmith Audio" on your podcast app. This essay is part of a series that I'm calling "Otherness and control in the age of AGI." I'm hoping that the individual essays can be read fairly well on their own, but see here for brief summ...
f1a190fa-e9ec-468b-8ab4-65a8895983d6
trentmkelly/LessWrong-43k
LessWrong
Twelve Virtues booklet printing? For a while now, I've been using a laser printer to print out a couple of hundred copies of the Twelve Virtues of Rationality (in its printable pamphlet version) and taking them with me to conferences, talks, and science fiction conventions.  Cut, staple in the middle (using a large-si...
55b9ec68-5248-4c5c-9e7c-7e413b501f28
trentmkelly/LessWrong-43k
LessWrong
AISN #9: Statement on Extinction Risks, Competitive Pressures, and When Will AI Reach Human-Level? Welcome to the AI Safety Newsletter by the Center for AI Safety. We discuss developments in AI and AI safety. No technical background required. Subscribe here to receive future versions. -------------------------------...
bdd0dc3d-461b-4a0d-92f5-eb16281fc3fc
trentmkelly/LessWrong-43k
LessWrong
My (naive) take on Risks from Learned Optimization This is my current understanding of Risks from Learned Optimization. I'm pretty sure there are mistakes here and there. I also list things I agree or not with in the paper as I think it might be too large for one comment. So feel free to point out where I'm wrong (and...
740285f6-ae1e-4e4d-a36d-9c69b41eb0ad
trentmkelly/LessWrong-43k
LessWrong
I'm creating a deep dive podcast episode about the original Leverage Research - would you like to take part? Hi all! I'm creating a deep dive podcast episode about (the original) Leverage Research (for my podcast, Clearer Thinking). If you worked at (or currently work at) Leverage, or you have had experiences connecte...
d0e5cdaf-4987-4e74-860b-b6fc813a5633
trentmkelly/LessWrong-43k
LessWrong
Privacy Tradeoffs There's a lot of privacy-related writing that looks like: 1. Introduce a risk due to insufficient privacy. 2. Describe things you can do to increase your privacy. For example, I recently came across "It's time to worry about Internet privacy again" (via LW), which talks about the risk of AI-acc...
45909a01-1f32-4045-87d2-78fc1f4b7d20
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Understanding and visualizing sycophancy datasets *Produced as part of the* [*SERI ML Alignment Theory Scholars Program*](https://serimats.org/) *- Summer 2023 Cohort, under the mentorship of Evan Hubinger.* Generating datasets that effectively test for and elicit sycophancy in LLMs is helpful for several purposes, s...
5cef9ab5-8278-4e2a-b9b9-4d67805b29f6
trentmkelly/LessWrong-43k
LessWrong
Eliezer is still ridiculously optimistic about AI risk All power to crypto channel Bankless for their interview with Eliezer: https://www.youtube.com/watch?v=gA1sNLL6yg4 https://www.lesswrong.com/posts/Aq82XqYhgqdPdPrBA/full-transcript-eliezer-yudkowsky-on-the-bankless-podcast They actually take his arguments seriou...
8cebc307-ea7f-4e13-b903-6ff72cb8a5cf
trentmkelly/LessWrong-43k
LessWrong
Meetup : Denver Area Meetup Discussion article for the meetup : Denver Area Meetup WHEN: 15 October 2014 06:00:00PM (-0600) WHERE: 1417 S Broadway, Denver, CO 80210 First time meet up. The date and location are open for revision if anyone would prefer. All conversation topics are welcome. Unless there are unexpecte...
9d620f52-30fc-44ae-88a8-1f529cef0615
trentmkelly/LessWrong-43k
LessWrong
Reward Is Not Enough Three case studies 1. Incentive landscapes that can’t feasibly be induced by a reward function You’re a deity, tasked with designing a bird brain. You want the bird to get good at singing, as judged by a black-box hardcoded song-assessing algorithm that you already built into the brain last week...
f201e752-580c-4a0a-bdb8-6c0f23b64228
trentmkelly/LessWrong-43k
LessWrong
Meetup : Washington, D.C.: Chesterton's Fence Discussion article for the meetup : Washington, D.C.: Chesterton's Fence WHEN: 23 August 2015 03:00:00PM (-0400) WHERE: National Portrait Gallery Crossposted from mailing list. Meeting in courtyard, gathering from 3:00pm, hard start 3:30pm - until closing We'll be meet...
ba06b587-03ad-4364-af68-a6a02d735401
StampyAI/alignment-research-dataset/arxiv
Arxiv
Developing a Fidelity Evaluation Approach for Interpretable Machine Learning 1 Introduction --------------- Modern machine learning and deep learning techniques have allowed for analytics of complex data and enable decision-making (Tan et al, [2020](#bib.bib21)). However, these advanced machine learning techniques ...
73a8f18b-21cd-4263-b9ed-2bd0ae35e978
StampyAI/alignment-research-dataset/arxiv
Arxiv
Retrieval Augmented Classification for Long-Tail Visual Recognition 1 Introduction --------------- Large Transformer [transformer] models have arrived in Computer Vision, with parameter counts and pretraining dataset size increasing rapidly [vit, clip, cvtransformers, jft, resnext, scalingvits]. The distributed rep...
c61ca4aa-6dc9-4acb-8a18-2bf128023248
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
AI? Just Sandbox it... - Computerphile we see a lot of comments on your videos about people who just say oh just simply do this that will be the answer for all of these problems yeah and I admire them for getting stuck in and getting involved but one thing that always strikes me is people say just change this bit of c...
2cc271a7-0799-4592-94ef-e553200d2e86
trentmkelly/LessWrong-43k
LessWrong
Legibility: Notes on "Man as a Rationalist Animal" Cross-posted on my roam-blog. This post is meant to be part of a series that explores the ideas in samzdat's The Uruk Series, and it also serves as a standalone introduction to the idea of legibility.   The first book that Lou rings in is Seeing Like a State by Jame...
3e954469-e0cf-42e9-bb9e-66e47af06810
trentmkelly/LessWrong-43k
LessWrong
Growing Independence Note: this is based on my experience with my two kids, currently four and six. It may not generalize as much as I think it does. People start out dependent on their parents for food, changing, contact, motion, and even sleep timing. Typically they end up as adults, no longer dependent on their pa...
a3ac2ae2-736e-4ed2-bbd8-0a1fc08ceee0
trentmkelly/LessWrong-43k
LessWrong
Kling, Probability, and Economics Related to: Beautiful Probability, Probability is in the Mind Arnold Kling ponders probability: > How one thinks about probability affects how one thinks about economics. Consider the use of the word "probability" in each of the following sentences: > > 1. What is the probability t...
c36b5204-c6d5-4dd2-bbee-8b2303b9a570
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Conjecture: Internal Infohazard Policy *This post benefited from feedback and comments from the whole Conjecture team, as well as others including Steve Byrnes, Paul Christiano, Leo Gao, Evan Hubinger, Daniel Kokotajlo, Vanessa Kosoy, John Wentworth, Eliezer Yudkowsky. Many others also kindly shared their feedback and...
e43442a4-28d6-4c92-af77-4e376a6ad252
trentmkelly/LessWrong-43k
LessWrong
Sam Harris - TED Talk on AI
fd22f705-63a2-4136-8c29-bf64d288b4c5
trentmkelly/LessWrong-43k
LessWrong
Weekly LW Meetups: Berlin, Cambridge UK, Durham, Frankfurt, London, NYC, Vancouver, Washington DC This summary was posted to LW main on February 22nd. The following week's summary is here. There are upcoming irregularly scheduled Less Wrong meetups in: * First meetup in Frankfurt (Main) : 22 February 2013 06:30PM ...
ce6e3eab-bab9-489b-b992-3d9822365677
trentmkelly/LessWrong-43k
LessWrong
Did social desirability effects mask Trump's true support?
13e773a4-9aaf-4a72-b42a-7c395bd8736b
trentmkelly/LessWrong-43k
LessWrong
Voting for people harms people So, look, The Culture War totally sucks, right? I mean, really, it more than sucks. I've put it this way before: The Culture War is The Nothing. (If you're an American and you're not sure what The Nothing is, I'm probably envious of your youth. You should know the 80s were cool and The ...
b19cd242-d3b9-4750-8a20-e6d482617034
StampyAI/alignment-research-dataset/lesswrong
LessWrong
ChatGPT seems overconfident to me I've talked to ChatGPT about physics, a topic I don't know much about, except that a few things like air lift, kinetic vs static friction, osmosis, and the stuff listed in <https://xkcd.com/2682/> (how paracetamol/tylenol works? why baloons electrify hair?) are somewhat "controversial...
5b95615b-b874-4146-bc00-505d4e530052
trentmkelly/LessWrong-43k
LessWrong
A Healthy News Diet [Note: This post is a reaction to my drastic increase in Twitter consumption this year. All uses of the second person "you" and "your" in this post are in fact me talking to myself.] If you aren’t careful, you’ll spend too much of your life scrolling a feed. Might be Twitter, HN, Reddit, NYT, etc....
01c5ac05-b6e0-46dd-8a47-46590804a6d0
trentmkelly/LessWrong-43k
LessWrong
Jan Kulveit's Corrigibility Thoughts Distilled Many thanks to Jan for commenting on a draft of this post. There were a lot of great comments on "Let's see you write that corrigibility tag". This is my attempt at expanding Jan Kulveit's comment[1], because I thought it was useful, and should be read more widely. This ...
f159d138-2b7c-49ae-9d07-8ee012e02120
trentmkelly/LessWrong-43k
LessWrong
Solstice 2015: What Memes May Come? (Part I) Winter is coming, and so is Solstice season. There'll be large rationality-centric-or-adjaecent events in NYC, the Bay Area, and Seattle (and possibly other places - if you're interested in running a Solstice event or learning what that involves, send me a PM). In NYC, ther...
aaa5fb57-4255-4e1d-80ff-3607f3fe1821
trentmkelly/LessWrong-43k
LessWrong
Open thread, Jul. 25 - Jul. 31, 2016 If it's worth saying, but not worth its own post, then it goes here. ---------------------------------------- Notes for future OT posters: 1. Please add the 'open_thread' tag. 2. Check if there is an active Open Thread before posting a new one. (Immediately before; refresh the ...
b6546602-a226-41c4-a504-7979de0c2deb
trentmkelly/LessWrong-43k
LessWrong
Meetup : San Francisco Meetup: Board Games Discussion article for the meetup : San Francisco Meetup: Board Games WHEN: 16 December 2015 06:15:00PM (-0800) WHERE: 1597 Howard St. San Francisco, CA THIS IS ON WEDNESDAY! We'll be meeting to hang out and play games! You can call 301-458-0764 to be let in. Discussio...
03efeb59-8cac-49f9-838b-b5300f5cc1c7
trentmkelly/LessWrong-43k
LessWrong
The EA case for Trump There’s a decent chance Trump is our next president. In that case, it would be great for him to be AI-alignment-pilled.  The good news is that, according to Samuel Hammond, he is: > "Unfortunately, the EA vs e/acc debate on Twitter has created the perception that AI safety is left-right polariz...
a14142fb-2f87-4432-8d3e-9a1970304102
trentmkelly/LessWrong-43k
LessWrong
War Mode vs. Peace Mode Cross-posted from konstell.com [Epistemic status: rough heuristics, probably needs refinement, comments are always welcome.] I first got this dichotomy by reading someone's comment on how Mao was a wartime general but not a peacetime general. Of course, it's possible that a more accurate mode...
0fc67331-9b94-42af-b234-9947d0dd3062
trentmkelly/LessWrong-43k
LessWrong
The Defence production act and AI policy Quick Summary * Gives the President wide-ranging powers to strengthen the US industrial base  * Has been around without changing that much since 1953 * Has provisions which allow firms to make voluntary agreements that would normally be illegal under antitrust law  * Provi...
19387c22-039f-4aae-b669-566c49033ced
trentmkelly/LessWrong-43k
LessWrong
The case for value learning This post is mainly fumbling around trying to define a reasonable research direction for contributing to FAI research. I've found that laying out what success looks like in the greatest possible detail is a personal motivational necessity. Criticism is strongly encouraged.  The power and i...
7baf1a8d-863e-4b54-a14d-c76835e09db3
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
185. If I were a Well-intentioned AI 3,4 all right so welcome to the 185th session of the AI safety reading group will be discussing parts three and four or if I will were well-intentioned AI so last time we saw that good hearts law is not quite a law good Harting is only a problem for values with certain properties a...
8ae5e63c-020e-48dc-9c63-e46e2bc8a022
trentmkelly/LessWrong-43k
LessWrong
Follow-up survey: inositol Two months ago I wrote about inositol as a treatment that occasionally works for anxiety and depression, especially when the user also has weird digestive issues (not medical advice, I am not a doctor). If that inspired you to try inositol I would love if you would fill out this 5-7 question...
46c2ffad-b241-493a-8220-bc9e9a6ad897
trentmkelly/LessWrong-43k
LessWrong
Meetup : Washington, D.C.: Performance of Masculinity Discussion article for the meetup : Washington, D.C.: Performance of Masculinity WHEN: 01 February 2015 03:00:00PM (-0500) WHERE: Donald W. Reynolds Center for American Art and Portraiture We will be meeting in the Kogod Courtyard of the Donald W. Reynolds Cente...
62619148-6823-475d-921a-cbc074b4e051
LDJnr/LessWrong-Amplify-Instruct
LessWrong
"In the original When None Dare Urge Restraint post, Eliezer discusses the dangers of the "spiral of hate" that can develop when saying negative things about the Hated Enemy trumps saying accurate things. Specifically, he uses the example of how the 9/11 hijackers were widely criticized as "cowards," even though this v...
197750d8-c1e7-486e-be95-6537e4537034
trentmkelly/LessWrong-43k
LessWrong
Idea: build alignment dataset for very capable models In Training language models to follow instructions with human feedback, OpenAI showed that we can significantly improve the alignment of language models through reinforcement learning targeting a model of human preferences. I.e., they train GPT-3 to generate text t...
833e0c25-d3ac-473f-9bdd-45a622afa64d
trentmkelly/LessWrong-43k
LessWrong
"How To Become Less Wrong" - Feedback on Article Request Would appreciate feedback on this article I plan to submit to a broad media publication as part of my broader project of promoting rationality and raising the sanity waterline. Can't make it much longer as I'm at word limit, so if you suggest adding something, a...
31309fd0-5fed-4af0-afe1-d30f5128dd85
trentmkelly/LessWrong-43k
LessWrong
Motivating Alignment of LLM-Powered Agents: Easy for AGI, Hard for ASI? TL;DR: If we manage to develop an AGI or even ASI agent in the near-term, then it is likely to be powered by, or at very least include, an LLM (probably supplemented with additional cognitive scaffolding, long-term memory, and logic). LLMs have a ...
24c2be03-67f6-4804-9adf-fd1dcc787bec
trentmkelly/LessWrong-43k
LessWrong
Continuous Improvement When is it adaptive for an organism to be satisfied with what it has?  When does an organism have enough children and enough food?  The answer to the second question, at least, is obviously "never" from an evolutionary standpoint.  The first proposition might be true if the reproductive risks of...
3676912d-b8a1-44bc-9ba5-41b87efd8c8f
StampyAI/alignment-research-dataset/arxiv
Arxiv
Ethics and Governance of Artificial Intelligence: Evidence from a Survey of Machine Learning Researchers 1 Introduction --------------- Tech companies and governments alike see the potential for AI and have moved to develop machine learning, particularly deep learning, applications across a variety of sectors — fro...
a1f01841-1e1b-496e-a781-ab63a01dd7fb
trentmkelly/LessWrong-43k
LessWrong
The True Story of How GPT-2 Became Maximally Lewd This video recounts an incident that occurred at OpenAI in which flipping a single minus sign led the RLHF process to make GPT-2 only output sexually explicit continuations.  The incident is described in OpenAI's paper "Fine-Tuning Language Models from Human Preferenc...
c0e4f153-a162-42d4-a8a4-dafd3bf478fa
StampyAI/alignment-research-dataset/arxiv
Arxiv
Toward a Research Agenda in Adversarial Reasoning: Computational Approaches to Anticipating the Opponent's Intent and Actions Toward a Research Agenda in Adversarial Reasoning : Computational Approaches to Anticipating the Opponent ’s Intent and Actions Alexander Kotta, Michael Ownbyb aDefense Advanced Research ...
c49e326b-9c03-4fd1-afe6-b3f6380b558a
trentmkelly/LessWrong-43k
LessWrong
Dutch-Booking CDT [This post is now superseded by a much better version of the argument.] In a previous post, I speculated that you might be able to Dutch-Book CDT agents if their counterfactual expectations differed from the conditional expectations of EDT. The answer turns out to be yes. I'm going to make this a s...
55034d40-8b11-48a7-9312-6b78118b9176
trentmkelly/LessWrong-43k
LessWrong
Musings on the Speed Prior Thanks to Paul Christiano, Mark Xu, Abram Demski, Kate Woolverton, and Beth Barnes for some discussions which informed this post. In the ELK report, Paul, Mark, and Ajeya express optimism about penalizing computation time as a potentially viable way to select the direct translator over the ...
0a2be269-5db6-4f23-8bc3-17f25968ea99
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Separating the "control problem" from the "alignment problem" I'm often confused by how "loss of control" is somewhat similar in meaning to "misalignment" in the context of AGI risk. An average person might reasonably equate the two, or feel like there's an large overlap of meaning between the two words. Here's my att...
2924e8cb-a2de-4d6a-8561-05cf60b2db4e
trentmkelly/LessWrong-43k
LessWrong
Conjuring An Evolution To Serve You GreyThumb.blog offers an interesting analogue between research on animal breeding and the fall of Enron.  Before 1995, the way animal breeding worked was that you would take the top individual performers in each generation and breed from them, or their parents.  A cockerel doesn't l...
91b667be-40d4-4858-b9c5-034d7671666f
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] The Dilemma: Science or Bayes? Today's post, The Dilemma: Science or Bayes? was originally published on 13 May 2008. A summary (taken from the LW wiki):   > The failure of first-half-of-20th-century-physics was not due to straying from the scientific method. Science and rationality - that is, Science and...
75e5a6bf-0024-47cc-8aec-03858d46c798
trentmkelly/LessWrong-43k
LessWrong
A failure of conservation of evidence From Bloomberg: U.S. stocks rallied, driving the Standard & Poor’s 500 Index up from the cheapest valuations since 2009, as weaker-than-estimated economic data reinforced optimism the Federal Reserve will act to spur growth. The S&P 500 rose 3.4 percent to 1,162.35 at 4 p.m. in ...
71ed1983-d7a1-45dd-b106-8d344f8a9aa7
trentmkelly/LessWrong-43k
LessWrong
A Bit For You This button will send a single bit. This is no mindgame, no weird trolley-problem-monkey's-paw-dilemma. * This page, this post, pressing this button, are meant to be whatever they need to be for you in this moment. The purpose of this singular bit is entirely up to you. Take a second, and Consider T...
ea3f4a94-3ae1-4238-9c8f-a854000dbc43
trentmkelly/LessWrong-43k
LessWrong
Reflections on Premium Poker Tools: Part 3 - What I've learned Previous posts: * Part 1 - My journey * Part 2 - Deciding to call it quits I finally made the decision to call it quits. Now I think it would be a good time to reflect on my experiences and see if I could learn something from them. Market size As I t...
bfca64a4-f436-4de4-8020-88098f195f46
trentmkelly/LessWrong-43k
LessWrong
Potential Meetup: Ann Arbor, MI, USA Tentative date: either Saturday, May 11 or Sunday, May 12 (2013).  Is anyone interested?
9eae4cef-7f92-4fd1-955f-78ede9ac3297
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Cruxes for overhang Suppose a policy regime artificially restricts training compute.[[1]](#fnb57t6bcuh2g) Will this lead to rapid progress[[2]](#fnmlh644axnwa) later? Yes if and only if both the policy is suddenly reversed/evaded and progress can largely 'catch back up to where it would have been.' This post aims to i...
97117683-a46e-48c7-a7c8-daa688af0201
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
How Smart Are Humans? Epistemic status: free speculation How intelligent should we expect AI to be, compared to humans, and how quickly should we expect it to reach this level of intelligence? This is of course an important strategic question, which has a large impact on our AI threat models. At the moment, the most ...
167bcb5a-b1ff-44cd-94cc-b9aa8e8eaa87
trentmkelly/LessWrong-43k
LessWrong
Meetup : Berlin Meetup Discussion article for the meetup : Berlin Meetup WHEN: 08 February 2013 07:30:00PM (+0100) WHERE: Ming Dynastie, Brückenstraße 6, 10179 Berlin We're meeting to chat and talk about practical rationality. This time we'll * discuss the first few items of the checklist of rationality habits, ...
ef6e049e-1630-4eed-830e-6c56d3bde382
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Recruit the World’s best for AGI Alignment ***Offer the world's most intelligent/capable/qualified people a ridiculous amount of money to work on AGI Alignment as part of a prestigious fellowship*** [Greg Colbourn](https://www.linkedin.com/in/greg-colbourn-3399a320/), Nov 2021 - Nov 2022 **[Publication note (30 Mar ...
ac91de8a-b8a5-4611-8066-036b157be238
trentmkelly/LessWrong-43k
LessWrong
Request: Sequences book reading group The book version of the Sequences is supposed to be published in the next month or two, if I understand correctly. I would really enjoy an online reading group to go through the book together. Reasons for a reading group: * It would give some of us the motivation to actually go...
cbb28cb9-35dd-4712-9339-1d0c6bb0341e
trentmkelly/LessWrong-43k
LessWrong
Implications of living within a Simulation Disclaimer: It is important to already understand the simulation argument given by Nick Bostrom. This post assumes that we are in an ancestor simulation, so if you do not know what that is, definitely read this first. Reader discretion advised, as this rabbit hole is pretty...
f1b0650f-8b8f-45be-8395-2233e2d42a5a
trentmkelly/LessWrong-43k
LessWrong
Pessimistic Shard Theory It is often said that a partial alignment solution will bring about an S-risk, as your agent cares enough about humans to keep them around, but not enough about them to allow them to flourish. This is usually not worried about because the thought is that we are currently very far away from eve...
ebe3d617-0437-444c-bf90-e2cdd9279f83
trentmkelly/LessWrong-43k
LessWrong
Anthropic: Three Sketches of ASL-4 Safety Case Components > The cleanest argument that current-day AI models will not cause a catastrophe is probably that they lack the capability to do so.  However, as capabilities improve, we’ll need new tools for ensuring that AI models won’t cause a catastrophe even if we can’t ru...
eabf18e1-1bed-4bbc-bf46-87e19f63b4e7
StampyAI/alignment-research-dataset/arxiv
Arxiv
What does Newcomb's paradox teach us? What does Newcomb’s paradox teach us? David H. Wolpert, Gregory Benfordy February 26, 2010 Abstract In Newcomb’s paradox you choose to receive either the contents of a particular closed box, or the contents of both that closed box and another one. Before you choose, a prediction ...
c1ef1a50-c701-4c12-b515-6499cb24b5c1
trentmkelly/LessWrong-43k
LessWrong
Why is AGI/ASI Inevitable? Hello! My name is Amy. This is my first LessWrong post. I'm about somewhat certain it will be deleted, but I'm giving it a shot anyway, because I've seen this argument thrown around a few places and I still don't understand. I've read a few chunks of the Sequences, and the fundamentals of r...
b0c8ccfa-4d30-452a-b0dd-de25bc1fd242
trentmkelly/LessWrong-43k
LessWrong
Many Weak Arguments vs. One Relatively Strong Argument My epistemic framework has recently undergone some major shifts, and I believe that my current epistemic framework is better than my previous one. In the past, I tended to try to discover and rely on a single relatively strong argument in favor or against a positi...
e901bae4-553c-464a-9789-7fbf5baf48c0
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
A Comprehensive Mechanistic Interpretability Explainer & Glossary *This is a linkpost for a [very long doc](https://neelnanda.io/glossary) defining, explaining, and giving intuitions and conceptual frameworks for all the concepts I think you should know about when engaging with mechanistic interpretability. If you fin...
c998b2e2-dbed-4075-b240-72c3310a959d
trentmkelly/LessWrong-43k
LessWrong
My Alignment "Plan": Avoid Strong Optimisation and Align Economy Summary: Many people seem to put their hopes on something like the following "plan": * (Optional) Step 0: Hope, or assume, that sharp left turn will not happen. Proceed with capabilities until we get ~human-level AI. Then jump to Step 3. * Step 1: Con...
a95a8153-e49c-40ac-a7bb-a9ffe65ae5e1
trentmkelly/LessWrong-43k
LessWrong
MDPs and the Bellman Equation, Intuitively Explained Imagine you're a selfish taxi driver who takes passengers around between 3 towns. Each trip has a different probability of catastrophe and different earnings potential. You want to find a repeatable path of trips to maximise profit. No, you can’t just greedily pick ...
169cd8ca-12e5-49df-a3e1-85360e4bbd54
trentmkelly/LessWrong-43k
LessWrong
Running a Basic ACX Everywhere Meetup This is a guide for running a basic Astral Codex Ten Everywhere meetup, also known as Schelling Meetups. These are a roughly twice a year coordination by Scott Alexander, the author of Astral Codex Ten, to have meetups for people who like his blog.  This guide comes in three main...
e3b977fb-f8ee-4c61-b7c2-ab8d32594df1
trentmkelly/LessWrong-43k
LessWrong
Levels of safety for AI and other technologies What does it mean for AI to be “safe”? Right now there is a lot of debate about AI safety. But people often end up talking past each other because they’re not using the same definitions or standards. For the sake of productive debates, let me propose some distinctions t...
0c1d611f-9032-48bb-a2bb-0b3bdce2796d
trentmkelly/LessWrong-43k
LessWrong
Case Story: Lack of Consumer Protection Procedures AI Manipulation and the Threat of Fund Concentration in Crypto Seeking Assistance to Fund a Civil Case to Establish Facts and Protect Vulnerable Consumers from Damage Caused by Automated Systems Good time of the day. And thank you for the wonderful work you do! I ...
1857478b-717f-4a6a-8274-1eaa1d41a3a7
trentmkelly/LessWrong-43k
LessWrong
Katja Grace's Anthropic Reasoning in the Great Filter The wonderful Katja Grace has just uploaded her thesis, Anthropic Reasoning in the Great Filter, linked to from her blog (which contains a convientient summary)   She discusses how deal with indexical updating; updating on information not simply based on the like...
2e076ab2-b906-4196-9c9f-654e0e435f81
trentmkelly/LessWrong-43k
LessWrong
Interested in learning Linux? Need hosting? Free shells! Sign up form: Click Here   I own a personal server running Debian Squeeze which has a 1Gb/s symmetric connection and 15TB per month bandwidth. I am offering free shell accounts to lesswrongers, with one contingency: You'll be placed in a usergroup, 'lw', as ...
3ab416d7-d1ce-4aba-bb42-e1251d4832c3
trentmkelly/LessWrong-43k
LessWrong
Feelings are Good, Actually This is another casually written post in a series of posts written in the spirit of advice to my past self. It's hard to remember exactly when I started to try to control my emotions. Probably I could not have been very old, only 4 or 5. Learning to regulate your own emotions is an import...
10edd395-9beb-4786-b15b-3bafa1867264
trentmkelly/LessWrong-43k
LessWrong
Long Live the Usurper The notion of boxing –particularly failed boxing– is much more ubiquitous than people seem to give it credit for. It appears to me to be a common shape in systems which attempt to implement some sort of shape-agnostic competence-search at all. Intentionally or not. It is a relation between two ga...
462a72d9-d3f9-450a-b566-fcef1dc8b0c2
trentmkelly/LessWrong-43k
LessWrong
My Mid-Career Transition into Biosecurity After working as a professional programmer for fourteen years, primarily in ads and web performance, I switched careers to biosecurity. It's now been a bit over a year: how has it gone? In terms of my day-to-day work it's very different. I'd been at Google for a decade [1] an...
23fc33c6-513e-4cc3-b42b-9f72cb711936
trentmkelly/LessWrong-43k
LessWrong
Predict - "Log your predictions" app As an exercise on programming Android, I've made an app to log predictions you make and keep score of your results. Like PredictionBook, but taking more of a personal daily exercise feel, in line with this post. The "statistics" right now are only a score I copied from the old Cre...
3013441d-948e-44ba-9f75-d37921dcef20
trentmkelly/LessWrong-43k
LessWrong
The case for pay-on-results coaching Thanks to Ruby, Stag Lynn, Brian Toomey, Kaj Sotala, Anna Salmon, Damon Sasi, Ethan Kuntz, Alex Zhu, and others for conversations that helped develop these ideas. Most coaches charge hourly (~$125-300). This makes sense: predictable income, easy scheduling, matches industry norms....
e88ef6d8-5068-4c4a-8495-01590cd6c2d6
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Sandboxing by Physical Simulation? This is a simple idea. I do not remember if I've seen it anywhere. It is probably not original, but I am mildly surprised that I haven't come across it, even if only to see it refuted. If this is an old/dumb idea, please let me know and I'll delete it. People have built universal Tu...
80e0dd0f-02ea-40be-af22-42f92cf50d74
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Shah and Yudkowsky on alignment failures This is the final discussion log in the [Late 2021 MIRI Conversations](https://forum.effectivealtruism.org/s/QtBPgszyK4yXwduKS) sequence, featuring Rohin Shah and Eliezer Yudkowsky, with additional comments from Rob Bensinger, Nate Soares, Richard Ngo, and Jaan Tallinn. The di...
3aa4e273-cc03-4684-b483-19695e2e85f2
trentmkelly/LessWrong-43k
LessWrong
Dimensional decoupling This is also posted at brianlui.dog. ---------------------------------------- Cornelius has a problem in his life. Cornelius: Emma, I know you’re really good with people. I feel silly saying this, but maybe you could help me with a problem I have? Emma: That’s a very nice thing for you to sa...
9ccb41ae-3e21-4d30-bd71-60f347cb8cb1
trentmkelly/LessWrong-43k
LessWrong
Forecasting Newsletter: February 2021 Highlights * Biden orders the creation of a National Center for Epidemic Forecasting and Outbreak Analytics * Americans get a clone of Betfair/Smarkets, expected to be marginally better than PredictIt * Hypermind has a new forecasting tournament on the future of AI in 2023 In...
769f6aa9-f333-4444-bb23-6b20b173f712
trentmkelly/LessWrong-43k
LessWrong
Different time horizons for forecasting Disclaimer: This post contains some very preliminary thoughts on a topic that I believe would be of interest to some people here. There are probably better expositions on the subject that I haven't been able to find. If you know of such expositions, I'd appreciate being pointed ...
a08b77af-dfa9-42b9-a829-45b1b1c67a71
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Harry Potter and the Data Centers of Doom The atmosphere was thick with intellectual tension in a small, concealed room at Hogwarts School of Witchcraft and Wizardry. Three prominent figures, each a beacon of knowledge in their own right, were engaged in an intense discussion. Harry Potter, the young prodigy, Alastor ...
e485399f-1819-4809-9ada-8fd4529c19fa
trentmkelly/LessWrong-43k
LessWrong
The Puce Tribe In 2014 Scott Alexander introduced us to the Red/Blue/Grey tribe model, with his post I Can Tolerate Anything Except the Outgroup: > The Red Tribe is most classically typified by conservative political beliefs, strong evangelical religious beliefs, creationism, opposing gay marriage, owning guns, eatin...
ac11ded9-9344-4c2a-a5c4-356bb247bb4a
trentmkelly/LessWrong-43k
LessWrong
Contra Common Knowledge This post assumes some background knowledge about "common knowledge", a technical term in philosophy. See the tag page for an introduction and a list of LessWrong essays on the subject. Epistemic effort: basically reading the SEP article on common knowledge and forming my own conclusions. I've...
e1398b96-794f-488d-b905-73bd10b8e140
trentmkelly/LessWrong-43k
LessWrong
D&D.Sci: The Mad Tyrant's Pet Turtles [Evaluation and Ruleset] This is a followup to the D&D.Sci post I made ten days ago; if you haven’t already read it, you should do so now before spoiling yourself. Here is the web interactive I built to let you evaluate your solution; below is an explanation of the rules used to ...
21498b5e-b50f-4879-8cd1-3937bc7d05b2
trentmkelly/LessWrong-43k
LessWrong
Things I learned talking to the new breed of scientific institution Note: this article is sponsored by and cross-posted to the Good Science Project. They also write a fair bit, and their articles were essential reading for writing this essay! Also, this article would not be possible without the hours of discussion/ed...
361139dc-421b-4752-b0e8-128edeabbf99
trentmkelly/LessWrong-43k
LessWrong
Type 2 as an aggregation of Type 1 processes This post assumes basic knowledge of Type 1/Type 2 (System 1/System 2) categorization of mental processes. Background (safe to skip) After my first reaction of surprise (consuming perhaps a few months) to the topic of heuristics and biases, and after a few more readings o...
6397ec07-6b2c-4339-a2da-ad86fc0c77c7
StampyAI/alignment-research-dataset/arxiv
Arxiv
Examining the Differential Risk from High-level Artificial Intelligence and the Question of Control 1 © 2022 . This manuscript version is made available under the CC -BY-NC-ND 4.0 license https://creativecommons.org/licenses/by -nc-nd/4.0/ EXAMINING THE DIFFERENTIAL RISK FROM HIGH -LEVEL ARTIFICIAL ...
857f6681-963a-47b9-b545-a71d63d99e7b
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Why you must maximize expected utility *This post explains von Neumann-Morgenstern (VNM) axioms for decision theory*, *and what follows from them: that if you have a consistent direction in which you are trying to steer the future, you must be an expected utility maximizer. I'm writing this post in preparation for a s...
d5e29b74-c9c4-412c-97f0-20ee9bac5661
trentmkelly/LessWrong-43k
LessWrong
Amplituhedron? I recently ran across a rather interesting result while browsing the Internet: Physics Discover Geometry Underlying Particle Physics > Physicists have discovered a jewel-like geometric object that dramatically simplifies calculations of particle interactions and challenges the notion that space and ti...
d2cdd633-a7a1-40d8-89a5-9e0e3223ce1a
trentmkelly/LessWrong-43k
LessWrong
How would free prediction markets have altered the pandemic? In WebMD and the Tragedy of Legible Expertise, Scott Alexander ends by saying that we should be grateful for the experts we have, but that prediction markets would still be better. Is he right? If so, how much better would prediction markets be? Let's do so...
639cf4ca-d36f-4a68-ab28-23b8b01626f6
trentmkelly/LessWrong-43k
LessWrong
What is SB 1047 *for*? Emmett Shear asked on twitter: > I think SB 1047 has gotten much better from where it started. It no longer appears actively bad. But can someone who is pro-SB 1047 explain the specific chain of causal events where they think this bill becoming law results in an actual safer world? What’s the t...
29624733-71fa-4400-b5e3-3303f08522e4
StampyAI/alignment-research-dataset/arbital
Arbital
Subtraction of rational numbers (Math 0) So far, we have met [the idea of a rational number](https://arbital.com/p/4zx), treating them as chunks of apples, and [how to add them together](https://arbital.com/p/55m). Now we will discover how the idea of the anti-apple (by analogy with the [integers' anti-cow](https://ar...
e4a20d9f-735b-4edc-bed3-071067e22e60
trentmkelly/LessWrong-43k
LessWrong
Meetup : Moscow social meetup: Party games! Discussion article for the meetup : Moscow social meetup: Party games! WHEN: 04 December 2016 02:00:00PM (+0300) WHERE: Russia, Moscow, Strelbishensky pereulok, 10 Welcome to the social meetup! Main theme of this meetup is party games; and some of them are party versions ...
a4a7f3ef-8ac3-42dc-a4ae-a6e10a838005
trentmkelly/LessWrong-43k
LessWrong
Wireheading is in the eye of the beholder tl;dr: there is no natural category called "wireheading", only wireheading relative to some desired ideal goal. Suppose that we have a built an AI, and have invited a human H to help test it. The human H is supposed to press a button B if the AI seems to be behaving well. The...
a1987623-2960-436c-aa06-d90fe04fbfd0
trentmkelly/LessWrong-43k
LessWrong
Infinity is an adjective like positive rather than an amount I often see the belief that infinity is a quantity, when I think it's a quality. This post introduces two qualitative notions on a field where people are likely to have a good quantitative understanding in order to show how the qualitative understanding isn'...
820bd365-4328-4cbb-a7d1-f2eede0c6327
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
AI Alignment Podcast: An Overview of Technical AI Alignment in 2018 and 2019 with Buck Shlegeris and Rohin Shah Just a year ago we released a two part episode titled [An Overview of Technical AI Alignment with Rohin Shah](https://futureoflife.org/2019/04/11/an-overview-of-technical-ai-alignment-with-rohin-shah-part-1/...
16501481-5cb4-4da0-b96e-e6b8b6567cde
trentmkelly/LessWrong-43k
LessWrong
LW in BusinessInsider http://www.businessinsider.com/ten-things-you-should-learn-from-lesswrongcom-2011-7  
e8ca2dc3-c3d7-433d-ad98-58d0dee3375f
StampyAI/alignment-research-dataset/arxiv
Arxiv
Off-Policy Deep Reinforcement Learning without Exploration 1 Introduction --------------- Batch reinforcement learning, the task of learning from a fixed dataset without further interactions with the environment, is a crucial requirement for scaling reinforcement learning to tasks where the data collection procedur...