id
stringlengths
36
36
source
stringclasses
15 values
formatted_source
stringclasses
13 values
text
stringlengths
2
7.55M
99c7ab0f-8947-4349-9acf-df4dd9010534
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Aligning the Aligners: Ensuring Aligned AI acts for the common good of all mankind The argument I’m going to make here is obviously not original with me, but I do think that while AI governance is talked about, and one of the career paths that 80,000 hours recommends, it is not sufficiently pointed to as an important ...
0a557eee-57a9-42a2-b67f-db85a35570c5
trentmkelly/LessWrong-43k
LessWrong
Whistleblowing Twitter Bot In this post, I propose an idea that could improve whistleblowing efficiency, thus hopefully improving AI Safety by making unsafe practices discovered marginally faster. I'm looking for feedback, ideas for improvement, and people interested in making it happen. It has been proposed before,...
ba6b3c23-a5bf-495e-913b-d24a486ec072
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
My current thoughts on MIRI's "highly reliable agent design" work Interpreting this writeup: -------------------------- I lead the Open Philanthropy Project's work on technical AI safety research. In our MIRI [grant writeup](http://www.openphilanthropy.org/focus/global-catastrophic-risks/potential-risks-advanced-art...
4e17d171-ca8c-40b4-b8da-213a7c4c11d6
trentmkelly/LessWrong-43k
LessWrong
Looking for opinions of people like Nick Bostrom or Anders Sandberg on current cryo techniques In June 2012, Robin Hanson wrote a post promoting plastination as a superior to cryopreservation as an approach to preserving people for later uploading. His post included a paragraph which said: > We don’t actually know th...
62551584-f7a9-494e-80d7-6c8a8b6351c2
trentmkelly/LessWrong-43k
LessWrong
The consistent guessing problem is easier than the halting problem The halting problem is the problem of taking as input a Turing machine M, returning true if it halts, false if it doesn't halt. This is known to be uncomputable. The consistent guessing problem (named by Scott Aaronson) is the problem of taking as inpu...
4898435d-39c3-4195-8ae9-b259fa1ba1a6
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Evolutions Building Evolutions: Layers of Generate and Test *[Epistemic Status: Early layers and the general framework seems firm (and substantially* [*Dennett-inspired*](https://everything2.com/title/Tower+of+generate-and-test)*), later parts are more speculative, but I've found them to be useful thinking tools.]* ...
323776a4-570b-4b24-b80a-292d0c0ef429
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
Current AI Misalignment and Potential Long Term Implications hello everyone and welcome to the second and final day of the inaugural stanford existential risks conference as mentioned in our earlier email we highly recommend prioritizing setting up meetings with other attendees over the course of the day our first ses...
6b98898a-0dfc-4f5a-a409-25a113988cb8
LDJnr/LessWrong-Amplify-Instruct
LessWrong
"Human beings are all crazy. And if you tap on our brains just a little, we get so crazy that even other humans notice. Anosognosics are one of my favorite examples of this; people with right-hemisphere damage whose left arms become paralyzed, and who deny that their left arms are paralyzed, coming up with excuses when...
1577bde1-ca83-4c08-ba35-bafe76559ed4
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Progress on Causal Influence Diagrams *By Tom Everitt, Ryan Carey, Lewis Hammond, James Fox, Eric Langlois, and Shane Legg* *Crossposted from* [*DeepMind Safety Research*](https://deepmindsafetyresearch.medium.com/progress-on-causal-influence-diagrams-a7a32180b0d1) About 2 years ago, we released the [first](https://...
78bae546-a592-4eb9-a45d-f03951232750
trentmkelly/LessWrong-43k
LessWrong
flowing like water; hard like stone SilverFlame I noticed in your recent post that you described "rhetorical aikido", which you called a Daoist technique. I have studied the Tao Te Ching's text a fair bit over the past few years, and I find its philosophy to be quite useful as a "theme" or "virtue". I've not had the ...
82516689-6928-41c9-92c2-ef4f4243b876
trentmkelly/LessWrong-43k
LessWrong
Thou Art Godshatter Before the 20th century, not a single human being had an explicit concept of "inclusive genetic fitness", the sole and absolute obsession of the blind idiot god. We have no instinctive revulsion of condoms or oral sex. Our brains, those supreme reproductive organs, don't perform a check for reprodu...
d8df35be-bb88-4502-bcd9-42d9cb548690
trentmkelly/LessWrong-43k
LessWrong
D&D.Sci Divination: Nine Black Doves This is an entry in the 'Dungeons & Data Science' series, a set of puzzles where players are given a dataset to analyze and an objective to pursue using information from that dataset.  STORY (skippable) The omens are troubling for we Romans this year!  Black doves have been spott...
52d08061-89a3-4dca-a9ed-f9dfee724756
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Consciousness as a conflationary alliance term *Tl;dr: In this post, I argue that the concept of 'consciousness' is more conflated than people realize, in that there's a lot of divergence in what people mean by "consciousness", and people are unaware of the degree of divergence. This confusion allows the formation of ...
a4a54b49-8cb8-4ac3-8f8b-881f969047b7
StampyAI/alignment-research-dataset/arxiv
Arxiv
Collaborative Decision Making Using Action Suggestions
9a286dee-faae-4c95-adde-ee4a6c59d2bc
trentmkelly/LessWrong-43k
LessWrong
Prisoner's Dilemma vs the Afterlife I've had a thought I don't recall having encountered before described quite this way; but, given my past experiences with such thoughts, and the fact that it involves evo-psych, I currently peg my confidence in this idea at around 10%. But just in case this particular idea rose to m...
ee52f77d-c70c-4247-a729-bc79cec8114f
trentmkelly/LessWrong-43k
LessWrong
Previous Work on Recreating Neural Network Input from Intermediate Layer Activations Recently I've been experimenting with recreating a neural network's input layer from intermediate layer activations. The possibility has implications for interpretability. For example, if certain neurons are activated on certain inpu...
9d01567e-7e63-48ea-9b45-92e3dbd37fd1
trentmkelly/LessWrong-43k
LessWrong
Answering questions honestly given world-model mismatches (This post is superseded by our writeup on Eliciting Latent Knowledge.) In a recent post I discussed one reason that a naive alignment strategy might go wrong, by learning to “predict what humans would say” rather than “answer honestly.” In this post I want to...
bd54af73-bb15-4dcf-afed-18eda20a9ec3
trentmkelly/LessWrong-43k
LessWrong
Effective Altruism QALY workshop materials & outline (and Jan 13 '19 meetup notes) What the workshop is & a brief overview Hi all! This is my first post on LessWrong, but I'll be posting here more often with conversation notes & resources from the Effective Altruism Kansas City meetup group. This workshop was an expe...
9b9d9ff1-e365-431f-8925-50b9c076a698
trentmkelly/LessWrong-43k
LessWrong
Metaknowledge - Improving on the Wisdom of Crowds
5046ed45-948c-4d90-8206-f92a853435cc
trentmkelly/LessWrong-43k
LessWrong
Are smart contracts AI-complete? Many people are probably aware of the hack at DAO, using a bug in their smart contract system to steal millions of dollars worth of the crypto currency Ethereum. There's various arguments as to whether this theft was technically allowed or not, and what should be done about it, and so...
bf9dc7de-24b2-486e-a588-cd34299129f3
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Re-Define Intent Alignment? I think Evan's [Clarifying Inner Alignment Terminology](https://www.lesswrong.com/posts/SzecSPYxqRa5GCaSF/clarifying-inner-alignment-terminology) is quite clever; more well-optimized than it may at first appear. However, do think there are a couple of things which don't work as well as they...
691ab722-8f59-4d62-bac1-c81c6ac0d419
LDJnr/LessWrong-Amplify-Instruct
LessWrong
"I have previously spoken of the ancient, pulp-era magazine covers that showed a bug-eyed monster carrying off a girl in a torn dress; and about how people think as if sexiness is an inherent property of a sexy entity, without dependence on the admirer. "Of course the bug-eyed monster will prefer human females to its o...
e3e5b21b-248b-4a2f-9c93-be8d05772852
trentmkelly/LessWrong-43k
LessWrong
Is the average ethical review board ethical from an utilitarian standpoint? Many people argue that Facebook's study of how the emotions of it's users changed depending on the emotional content of messages in their facebook feed wouldn't have been approved by the average ethical review board because facebook didn't se...
ebb791f8-8556-4300-909d-5cb7e643f978
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
AI Research Considerations for Human Existential Safety (ARCHES) *Originally published in May 2020, cross-posted by* [*Aaron Gertler*](https://forum.effectivealtruism.org/users/aarongertler) *on 9/21/21.* **Abstract** ------------ Framed in positive terms, this report examines how technical AI research might be stee...
318c82e7-a0df-42b2-bece-57baca3d0937
trentmkelly/LessWrong-43k
LessWrong
Meetup : Prague - setting up Discussion article for the meetup : Prague - setting up WHEN: 23 March 2015 07:30:00PM (+0100) WHERE: Václavské náměstí 778/14, Praha 110 00 Hello, this is second long-term planed meetup in Prague. We are currently in the beginning of creating a core group here hence be prepared for t...
bf2437f3-993b-47b6-be99-10648255bcab
trentmkelly/LessWrong-43k
LessWrong
Personals, anyone? Meet ups are great and all, but a lot of us live far away from the large masses of LW-participants.  I live in the Toledo, Ohio area, and I'm also one of those people who are looking for people who I can relate to.  Most people I can't relate to, intellectual conversation makes most people I know un...
442e115f-1bb7-479f-b794-dfbd5bed9b48
trentmkelly/LessWrong-43k
LessWrong
Timaeus in 2024 > TLDR: We made substantial progress in 2024: > > * We published a series of papers that verify key predictions of Singular Learning Theory (SLT) [1, 2, 3, 4, 5, 6]. > * We scaled key SLT-derived techniques to models with billions of parameters, eliminating our main concerns around tractability. > ...
be523da6-4484-471e-9d78-c9c3c493bebf
trentmkelly/LessWrong-43k
LessWrong
Most AI value will come from broad automation, not from R&D This is a linkpost to an article by Ege Erdil and I that we wrote to explain an important perspective that we share regarding AI automation. I'll quote the introduction: > A popular view about the future impact of AI on the economy is that it will be primari...
6c85f522-3f59-4c64-94ca-bd66cc0ab8e0
trentmkelly/LessWrong-43k
LessWrong
Rogue AGI Embodies Valuable Intellectual Property This post was written by Mark Xu based on interviews with Carl Shulman. It was paid for by Open Philanthropy but is not representative of their views. Summary: * Rogue AGI has access to its embodied IP. * This IP will be worth a moderate fraction of the total value...
9aef6378-f7c4-4c10-9310-2949d46d0cf0
trentmkelly/LessWrong-43k
LessWrong
Multi-Agent Overoptimization, and Embedded Agent World Models I think this expands on the points being made in the recently completed Garrabrant / Demski Embedded Agency sequence. It also serves to connect a paper I wrote recently that discusses mostly non-AI risks from multiple agents that expands on the work done la...
e3937a18-e615-445b-abdd-e02d47a49094
trentmkelly/LessWrong-43k
LessWrong
The "Loss Function of Reality" Is Not So Spiky and Unpredictable In other words, there are no "rabbit holes" or "dark tunnels of the mind." The great success of neural networks is due in large part to the surprising ease in which gradient-based methods are used to train them. Neural networks are often trained using a ...
67f677e5-196e-4d25-b5e1-6cebc38365c1
StampyAI/alignment-research-dataset/arxiv
Arxiv
Cycle-of-Learning for Autonomous Systems from Human Interaction Using Humans to Train Autonomous Systems ---------------------------------------- Reinforcement learning (RL) has been successfully applied to solve challenging problems from playing video games to robotics. In simple scenarios, a reward function or mo...
c1658e37-10dc-4c00-938d-e5703fa0afe7
trentmkelly/LessWrong-43k
LessWrong
The Promoted Posts and the Metaethics sequence now available in audio We are proud to announce audio versions of the Less Wrong Promoted Posts and the Metaethics major sequence, both now available via a Castify Podcast. The Less Wrong Promoted Posts feed will have every new promoted post which has been tagged with th...
38baf351-39ce-4552-91d0-ea087ff1c020
trentmkelly/LessWrong-43k
LessWrong
Counterfactuals on POMDP A putative new idea for AI control; index here. This is technical note explaining how to define counterfactuals on partially observable Markov decision processes (POMDP). The POMDP formalism is explained here. This note will just sketch out how counterfactuals are defined; the full details w...
eccdc58c-4f52-4737-ab07-c0339916305c
trentmkelly/LessWrong-43k
LessWrong
Expected number of tries Epistemic status: An idea I had a few days ago and shared with two or three friends. Suppose you want to install a new habit. For example, gratitude journaling.  You might try to give it a go, by pledging to write down 3 things you are grateful for every day right before you go to sleep.  Aft...
61418072-ee37-4df9-9ee5-6234ee28f57b
trentmkelly/LessWrong-43k
LessWrong
Finding Research Papers So I'm all geared up and ready to go trawl through some academic journals so I can figure out how to tackle a problem or two, and I've realized I completely forgot how to go about it. Not going to a university with paid subscriptions to JSTOR, etc. doesn't help either. So where exactly does one...
e8fe8459-4ea1-4e3d-b2fb-ae44d53aab9a
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
Anders Sandberg - Answering the Fermi Question: Is AI our Great Filter? hey everyone my name is jeremy welcome back to the tour of state of science podcast and i'm really excited about today's episode because i get to talk to anders sandberg now anders is someone i've been angling to talk to for a long time because hi...
089382bc-5500-42f5-a4aa-d5629ed4d9ff
trentmkelly/LessWrong-43k
LessWrong
Evolution is sampling error This is a linkpost for a (somewhat old) post on my blog apomorphic.com. I'm posting this and other old posts here to get feedback and in hopeful preparation for some follow-on work[1]. See also: An addendum on effective population size A common mistake people make about evolution is to th...
86202567-2a0e-46fb-b51f-909aec8e6be8
trentmkelly/LessWrong-43k
LessWrong
26 March 2011 Southern California Meetup We're having another SoCal LessWrong meetup this Saturday, the 26th. It'll be held in the upstairs meeting area at this IHOP in Irvine. It will start at 2PM and probably run until 7 or so. The format for past meetups has varied based on the number of attendees and their intere...
544e6049-860e-4ae6-b151-5663ebaf4892
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Löb's Lemma: an easier approach to Löb's Theorem Related to: [Löb's Theorem for implicit reasoning in natural language: Löbian party invitations](https://www.lesswrong.com/posts/QsZ3ycfRYs2ps5sNA/loeb-s-theorem-for-implicit-reasoning-in-natural-language) tl;dr: Löb's Theorem is much easier to grok if you separate the...
aec923e8-372e-4dc4-82a4-ed311f5b6bc8
trentmkelly/LessWrong-43k
LessWrong
Raven paradox settled to my satisfaction The raven paradox, originated by Carl Gustav Hempel, is an apparent absurdity of inductive reasoning. Consider the hypothesis: H1: All ravens are black. Inductively, one might expect that seeing many black ravens and no non-black ones is evidence for this hypothesis. As you s...
3c6982ce-88c4-4f1b-a0ac-b9c13313cf77
trentmkelly/LessWrong-43k
LessWrong
The most important meta-skill Note: This article underwent a significant revision on 5/28/2015. Thank you to estimator for all your feedback. The most important meta-skill that anyone can learn is how to learn skills. With practice, you learn how to pick up new skills as they are needed, which is infinitely (quite...
42469092-43e5-48aa-b2e7-31594674c002
trentmkelly/LessWrong-43k
LessWrong
Walking Sue An Essay[1] PART I: Conjecture on The Development of Proto-Communication in Response to Complex Environmental Factors[2] I was walking with my Sue when I noticed a peculiar aspect of her nature: a dog is a stochastic agent. When you simply let the dog walk, she finds a random path, turning at intervals a...
1aed2562-5276-404e-9256-c6281003df22
trentmkelly/LessWrong-43k
LessWrong
What would the creation of aligned AGI look like for us? I'm having a failure of imagination here, I assume an aligned AGI would upload human minds, preventing them from dying, but beyond that it's not clear to me what would happen. Would we just get wireheaded? Is there a CEV from human values that results in choice ...
677f8835-1357-4906-8750-bf0cc436cafc
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Misc. questions about EfficientZero Perhaps these can be thought of as homework questions -- when I imagine us successfully making AI go well, I imagine us building expertise such that we can answer these questions quickly and easily. Before I read the answers I'm going to think for 10min or so about each one and post...
55226c05-16b8-4a7b-9af5-06f4dfdc9523
trentmkelly/LessWrong-43k
LessWrong
Presidents, asteroids, natural categories, and reduced impact A putative new idea for AI control; index here. EDIT: I feel this post is unclear, and will need to be redone again soon. This post attempts to use the ideas developed about natural categories in order to get high impact from reduced impact AIs.   Exte...
face0bbb-0f89-44e5-aded-161acd02eccb
trentmkelly/LessWrong-43k
LessWrong
Rationality Reading Group: Part E: Overly Convenient Excuses This is part of a semi-monthly reading group on Eliezer Yudkowsky's ebook, Rationality: From AI to Zombies. For more information about the group, see the announcement post. ---------------------------------------- Welcome to the Rationality reading group. ...
bb9de0dc-4e95-427f-b820-98bd29acac0d
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] Third Alternatives for Afterlife-ism Today's post, Third Alternatives for Afterlife-ism was originally published on May 8, 2007. A summary (from the LW wiki): > One source of hope against death is Afterlife-ism. Some say that this justifies it as a Noble Lie. But there are better (because more plausible) ...
b06f55a7-55ed-4726-971e-4a7b0cf220a0
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Apply to the Cavendish Labs Fellowship (by 4/15) [Cavendish Labs](https://forum.effectivealtruism.org/posts/xBeqaWEJfWZv8ALWn/announcing-cavendish-labs) is a new research organization in Vermont focused on technical work on existential risks. We'd like to invite you to [apply to our fellowships](https://cavendishlabs....
80b7c988-63a7-4b33-9e4b-9e369ef35c5e
StampyAI/alignment-research-dataset/arxiv
Arxiv
Synthesizing the preferred inputs for neurons in neural networks via deep generator networks 1 Introduction and Related Work -------------------------------- Understanding how the human brain works has been a long-standing quest in human history. Neuroscientists have discovered neurons in human brains that selectiv...
35807a03-cc92-4056-b81a-1b4fd35e92da
trentmkelly/LessWrong-43k
LessWrong
Is ruthlessness in business executives ever useful? We have a tradition of treating ruthlessness in businesspeople as something of a virtue. Certainly, ruthlessness can help one get ahead in the business world, and companies often benefit from executives who're willing to put aside scruples while devising means of tur...
44786481-d309-4aee-8ccc-c0a735ddf243
trentmkelly/LessWrong-43k
LessWrong
4 Common Prediction Failures and How to Fix Them [LINK] A post I wrote on CFAR's blog about the errors we make in hedonic prospecting, based on Gilbert and Wilson's 2007 review article of the topic. Let me know what you think! Reposting from the discussion in case anyone missed it. :)  Link. 
e4a6bdfe-fd3b-4dff-bff2-4c9e7bdc7ba8
trentmkelly/LessWrong-43k
LessWrong
West LA Weekly Meetups: Wednesday, June 8th ETA: Meetups are now every other Tuesday. When: 7pm - 9pm, Wednesday, June 8th. Where: The Westside Pavillion - on the bridge, which connects Nordstrom 3rd floor with Barnes & Noble / Landmark Theatres 3rd floor. Parking is free for 3 hours. Whether you're a regular re...
19d46868-ebad-4ee9-b35e-11ae6a56c274
StampyAI/alignment-research-dataset/arxiv
Arxiv
Autonomous Agents Modelling Other Agents: A Comprehensive Survey and Open Problems 1 Introduction --------------- A core area of research in modern artificial intelligence (AI) is the development of autonomous agents that can interact effectively with other agents. An important aspect of such agents is the ability ...
1c0bb03d-00ce-4337-9319-1cbe2ac0ae6a
trentmkelly/LessWrong-43k
LessWrong
Shouldn't there be a Chinese translation of Human Compatible? Considering that China seems pretty serious about investing heavily in AI in the near future, it may be kind of important that Stuart Russel's work of AI alignment advocacy Human Compatible is made as accessible as possible to a Chinese audience, but there ...
7a815881-4a15-460d-a33c-e2f06e97c1b3
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Acausal trade: being unusual .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; di...
9b1ced94-c1a2-4152-a0ab-064b41f126ba
StampyAI/alignment-research-dataset/blogs
Blogs
August 2017 Newsletter | | | --- | | **Research updates** * “[A Formal Approach to the Problem of Logical Non-Omniscience](https://arxiv.org/abs/1707.08747v1)”: We presented our work on [logical induction](https://intelligence.org/2016/09/12/new-paper-logical-induction/) at the [16th Conference on Theoretical Aspec...
e2d12211-6a17-467e-8ab1-bb351511d24a
trentmkelly/LessWrong-43k
LessWrong
Group Rationality Diary, June 1-30 This is the public group instrumental rationality diary for June 1-30; it's going up pretty late (sorry!) but entries for earlier in June are welcome. > It's a place to record and chat about it if you have done, or are actively doing, things like: > > * Established a useful new h...
97506e5b-a9c6-4822-a0ea-b3702e9fce45
trentmkelly/LessWrong-43k
LessWrong
LDL 7: I wish I had a map When I was an undergraduate considering going to graduate school in mathematics, one thing that I knew a reasonably amount about and spent a good amount of time thinking about was what the field of math looked like overall. In very, very broad strokes I knew that geometry, algebra, and analy...
39cb8167-3854-4848-8178-7b9aa2b3f86a
trentmkelly/LessWrong-43k
LessWrong
Creating a “Conscience Calculator” to Guard-Rail an AGI [Crossposted to the EA Forum here.] TL;DR: I present initial work towards creating a “conscience calculator” that could be used to guard-rail an AGI to make decisions in pursuing its goal(s) as if it had a human-like conscience. A list of possible conscience bre...
c09f351d-1b0f-416c-9688-e30c432534e4
StampyAI/alignment-research-dataset/special_docs
Other
Future progress in artificial intelligence: A survey of expert opinion Müller, Vincent C. and Bostrom, Nick (forthcoming 2014), ‘Future progress in artifi-cial intelligence: A Survey of Expert Opinion, in Vincent C. Müller (ed.), Fundamental Issues of Artificial Intelligence (Synthese Library; Berlin: Springer). Futur...
f6abd9b1-827f-4d9e-9cb1-bdff038208d3
trentmkelly/LessWrong-43k
LessWrong
CFAR, responsibility and bureaucracy Status: I wrote the draft for this article a few months before the more recent discussions. While I'm not sure that I get everything right, I believe that simply publishing my post is better then to keep it in my drawer. Introduction CFAR was founded in the spirit of being a plac...
2995c745-1182-4144-8e68-4c7be00e4cdd
trentmkelly/LessWrong-43k
LessWrong
The "best" mathematically-informed topics? Recently, I asked LessWrong about the important math of rationality. I found the responses extremely helpful, but thinking about it, I think there’s a better approach. I come from a new-age-y background. As such, I hear a lot about “quantum physics.” Accordingly, I have ...
d0784836-980e-47a0-a3c6-feb62fcf6b19
trentmkelly/LessWrong-43k
LessWrong
Dialogue on the Claim: "OpenAI's Firing of Sam Altman (And Shortly-Subsequent Events) On Net Reduced Existential Risk From AGI" johnswentworth I've seen/heard a bunch of people in the LW-o-sphere saying that the OpenAI corporate drama this past weekend was clearly bad. And I'm not really sure why people think that? To...
1acd3c05-2556-4c49-a0f5-0d725ba1a4ee
trentmkelly/LessWrong-43k
LessWrong
How many people are working (directly) on reducing existential risk from AI? Summary I've updated my estimate of the number of FTE (full-time equivalent) working (directly) on reducing existential risks from AI from 300 FTE to 400 FTE. Below I've pasted some slightly edited excepts of the relevant sections of the 80,...
d081f190-0793-4b15-8d07-126b70484774
trentmkelly/LessWrong-43k
LessWrong
Quantilizers and Generative Models Thanks to Evan Hubinger for discussions about quantilizers, and to James Lucassen for discussions about conditioned generative models. Many of these ideas are discussed in Jessica Taylor's Quantilizers: A Safer Alternative to Maximizers for Limited Optimization: this post just expand...
bc34da6f-b266-4f0d-acce-998999c85b45
StampyAI/alignment-research-dataset/lesswrong
LessWrong
What are the relative speeds of AI capabilities and AI safety? If you want to solve AI safety before AI capabilities become too great, then it seems that AI safety must have some of the following: * More researchers * Better researchers * Less necessary insights * Easier necessary insights * Ability to steal insight...
565acead-0856-4557-9ff7-f5844a24768b
trentmkelly/LessWrong-43k
LessWrong
The anti-psychotic Q-learner trick
151ce667-e8bd-4101-b141-ec23ccab3ede
trentmkelly/LessWrong-43k
LessWrong
Instrumental Rationality 7: Closing Disclaimer [Instrumental Rationality Sequence 7/7] [A disclaimer that instrumental rationality as presented in this sequence is incomplete. Your feelings are also important! Pay attention to them.] After reading through this sequence, you might be feeling very excited to go out an...
76859a86-3402-4b4d-a099-583914d18e89
trentmkelly/LessWrong-43k
LessWrong
We Live in a Post-Scarcity Society Historically, a normal human population hovered on the edge of starvation. > It's hard to comprehend how important food staples used to be. In Edo Japan, wealth was measured in koku (石). One koku is (in theory) enough rice to feed one man for one year. The amount of koku a daimyo co...
ab0356ae-de3f-4f75-91f4-4da0dbd93ca8
LDJnr/LessWrong-Amplify-Instruct
LessWrong
"(This post was collaboratively written together with Duncan Sabien.) Startup founders stereotypically experience some pretty serious mood swings. One day, their product seems destined to be bigger than Google, and the next, it’s a mess of incoherent, unrealistic nonsense that no one in their right mind would ever pay ...
f91db69b-76af-4533-a466-95955974ede7
trentmkelly/LessWrong-43k
LessWrong
Where freedom comes from We are agents engaging with an environment. Based on the environment we, as agents, act. Action will be determined by two aspects: magnitude and direction. Magnitude is the execution, the effort, or the immensity of the action. Direction is the reasoning, the strategy, or the path you ultim...
4d43e38e-b2bc-4c6c-a271-a0cd35b85e58
awestover/filtering-for-misalignment
Redwood Research: Alek's Filtering Results
id: post3936 A putative new idea for AI control; index here . I'm calling "goal completion" the idea of giving an AI a partial goal, and having the AI infer the missing parts of the goal, based on observing human behaviour. Here is an initial model to test some of these ideas on. The linear rocket On an infinite linea...
80f03d9d-b93a-4690-8b9e-2863240f524f
StampyAI/alignment-research-dataset/blogs
Blogs
Diana Spears on the safety of adaptive agents ![Diana Spears portrait](http://intelligence.org/wp-content/uploads/2014/04/Spears_w730.jpg) [Diana Spears](http://www.swarmotics.com/Contact_Information.html) is an Owner and Research Scientist at [Swarmotics, LLC](http://www.swarmotics.com/Swarmotics.html). Previously, s...
5106b4f7-11e8-4a13-9137-df957b531a56
trentmkelly/LessWrong-43k
LessWrong
LINK-Cryonics Institute documentary "WE WILL LIVE AGAIN looks inside the unusual and extraordinary operations of the Cryonics Institute. The film follows Ben Best and Andy Zawacki, the caretakers of 99 deceased human bodies stored at below freezing temperatures in cryopreservation. The Institute and Cryonics Movement ...
a91ce0c8-8ec9-46a0-a924-cfc7114f77ed
trentmkelly/LessWrong-43k
LessWrong
Silver Chairs, Paternalism, and Akrasia Inspired in part by Robin Hanson's excellent article on paternalism a while back, and in response to the various akrasia posts. In C.S. Lewis's fourth Narnia book, The Silver Chair, the protagonists (two children and a Marsh-wiggle) are faced with a dilemma regarding the title...
0bb8a7a2-06dc-40f6-93d4-be158f349a3b
trentmkelly/LessWrong-43k
LessWrong
Is there a way to replicate photosynthesis artificially? Humans have imitated nature since prehistoric times. Aircraft, submarines and even cars are designed in the image of wild animals. So why haven't we imitated plants yet? The reason this question is significant is due to the increasing discordance between the e...
636b6a3d-0c4c-4b0d-b46c-c73b787a1e85
trentmkelly/LessWrong-43k
LessWrong
Refusal mechanisms: initial experiments with Llama-2-7b-chat This work was conducted as part of Berkeley's Supervised Program for Alignment Research (SPAR), under the mentorship of Nina Rimsky. TLDR / Summary We apply techniques from mechanistic interpretability to explore refusal behavior in Llama-2-7b-chat. We are...
5273497d-5143-4999-aeb3-2c118bb506b4
trentmkelly/LessWrong-43k
LessWrong
Judging Our April 2020 Covid-19 Predictions Scott Alexander has given his verdict on our predictions for Covid from April 2020.  This seems like an excellent opportunity to reflect on those predictions. I’ll also attempt to render my verdict on the predictions, based on the principles I discuss in Evaluating Predicti...
f82a1fbe-99d8-41ee-bd2e-ed79b4fda611
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Broad Basins and Data Compression *Produced As Part Of The SERI ML Alignment Theory Scholars Program 2022 Research Sprint Under* [*John Wentworth*](https://www.lesswrong.com/users/johnswentworth) **Intro** --------- The "broad basins induce data compression" argument goes like this: If we train a neural network to m...
551afe59-c364-49e9-af91-9b802ac3ca6f
trentmkelly/LessWrong-43k
LessWrong
The Singularity Institute's Arrogance Problem I intended Leveling Up in Rationality to communicate this: > Despite worries that extreme rationality isn't that great, I think there's reason to hope that it can be great if some other causal factors are flipped the right way (e.g. mastery over akrasia). Here are some de...
8bdc9c4c-49f0-48d7-9f1d-980872f4d61f
trentmkelly/LessWrong-43k
LessWrong
Deciding what to think about; is it worthwhile to have universal utility function? The hunter-gatherer example in the Is risk aversion really irrational? got me thinking about the real world issues with 'maximizing utility' and any other simple rule approach to decision making. The elephant in the room is that univer...
1f323ebe-3629-4fec-8424-5a360d173f52
trentmkelly/LessWrong-43k
LessWrong
Please advise the Singularity Institute with your domain-specific expertise! The Singularity Institute would benefit from having a team of domain-specific advisors on hand. If you'd like to help the Singularity Institute pursue its mission more efficiently, please sign up to be a Singularity Institute advisor! If you...
40a430ab-03a7-4621-b68c-a21c1184fa25
trentmkelly/LessWrong-43k
LessWrong
Open thread, Jul. 18 - Jul. 24, 2016 If it's worth saying, but not worth its own post (even in Discussion), 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...
af95505b-bec5-45e7-b53f-02a1a3942ff6
trentmkelly/LessWrong-43k
LessWrong
Nature: "Stop talking about tomorrow’s AI doomsday when AI poses risks today" Overall, a headline that seems counterproductive and needlessly divisive. I worry very much that coverage like this has the potential to bring political polarization to AI risk and it would be extremely damaging for the prospect of regulati...
94a130a8-556c-4df3-870f-3880d6250940
trentmkelly/LessWrong-43k
LessWrong
Triangle NC Meetup July 6th hwc and I have tried submitting this meetup through the new function, but it seems to break the link.  Please let us know if anyone knows what's going on. In any case, Location:          Caribou Coffee‎, 1406 E Franklin St, Chapel Hill, NC Date:          2011-07-06, 7:00 PM EDT (Prop...
d4f21430-fddb-42f8-a349-5083f2b84134
StampyAI/alignment-research-dataset/arbital
Arbital
You can't get more paperclips that way Instrumental convergence says that various properties $P$ of an agent, often scary or detrimental-by-default properties like "trying to gain control of lots of resources" or "deceiving humans into thinking you are nice", will fall out of pursuing most utility functions $U.$ You ...
0e1beef7-b45e-4a51-98e1-949c85f831eb
trentmkelly/LessWrong-43k
LessWrong
Honest science is spirituality [cross-posted from my blog https://pchvykov.com/blog] I see a problem with current science. It’s not the reproducibility crisis, nor the toxic work culture, nor the misaligned incentive. But in another sense, it is all of these – or perhaps the root cause behind them. It’s hard to name ...
5dfab9e1-34c8-4290-af29-229733fc749b
trentmkelly/LessWrong-43k
LessWrong
Inflection.ai is a major AGI lab Update (April 2024): Due to the recent breakup of Inflection, I no longer think they're on track to be a major AGI lab. Inflection.ai (co-founded by DeepMind co-founder Mustafa Suleyman) should be perceived as a frontier LLM lab of similar magnitude as Meta, OpenAI, DeepMind, and Anth...
aaf48457-98d4-4342-9357-a3715276b338
trentmkelly/LessWrong-43k
LessWrong
Notes on Amiability This post examines the virtue of amiability (and closely-related virtues like friendliness, geniality, agreeableness, conviviality, affability, niceness, affection, and warmth). I mean to explore what other people have learned about these virtues, rather than to express my own opinions about them, ...
dcce6ed9-bc5c-4c03-acd8-22af21e3a5ac
trentmkelly/LessWrong-43k
LessWrong
Possible Worlds after Prediction Take-off This is the final piece of a blog post series that explores how AI prediction services affect the risks of war. It is based on my 10-week summer research project at Stanford Existential Risk Institute. Here's a summary of my project and my epistemic status. In previous posts...
7bd651f4-ba03-46f1-8dc1-1d8a0d4ad022
trentmkelly/LessWrong-43k
LessWrong
Zvi's Manifold Markets House Rules This document was updated August 11, 2024. Broad intent has not changed. All markets created by Zvi Mowshowitz shall be graded according to the rules described herein, including the zeroth rule. The version of this on LessWrong shall be the canonical version, even if other versions...
b3d1f027-dc4c-46ec-a984-5625a885c273
trentmkelly/LessWrong-43k
LessWrong
Highlights From The Comments On Cost Disease I got many good responses to my Considerations On Cost Disease post, both in the comments and elsewhere. A lot of people thought the explanation was obvious; unfortunately, they all disagreed on what the obvious explanation was. Below are some of the responses I found most ...
ba800b01-b5ff-4961-ba8e-de5b289f7b51
trentmkelly/LessWrong-43k
LessWrong
What Jason has been reading, March 2023 A new monthly feature, let me know what you think. Books Matt Ridley, How Innovation Works (2020). About halfway through, lots of interesting case studies, very readable. Vaclav Smil, Creating the Twentieth Century (2005). I read the first chapter; saving the rest of it for w...
45d24ad0-5979-4726-9405-8b1c4ad1dc3d
trentmkelly/LessWrong-43k
LessWrong
General Thoughts on Less Wrong General disclaimer: Obviously there's always going to be more that could be done and it's better to do a few things well, rather than many things poorly, but I still think it is worth being conscious of areas that could be improved[1]. * I'm still strongly in favor of adding sub-commun...
5467addf-ed13-4af7-af95-d6b70126b36a
trentmkelly/LessWrong-43k
LessWrong
Akrasia hack survey Survey here.  Analysis here.
1a449105-5bfa-4072-8ac2-a437453969a0
trentmkelly/LessWrong-43k
LessWrong
A question and a tail This is a rambling post, and I will appreciate your criticism to help dry it or delete it altogether. It seems that however little a question I research by reviewing [botanical] literature, there is always a much more complex, and rather difficult to rigorously put, question that I have to ask f...
6e25fbaf-ea8a-46af-9049-aa6be066675b
trentmkelly/LessWrong-43k
LessWrong
Progress: Fluke or trend? A foundational conviction of The Roots of Progress is that progress is a trend with definite, substantive causes, and that it can continue far, far into the future. Progress is not automatic or inevitable: it can slow, stop, even reverse. But the history of progress over the last 200+ years c...
91c701d5-2463-4bc7-b60e-594b9d2f9237
trentmkelly/LessWrong-43k
LessWrong
AGI Safety Fundamentals 2023 Notes Hello! As I am completing my capstone project for AGI Safety Fundamentals 2023, I wanted to link-post my notes for the core readings for the course. I have found it extremely rewarding to summarize key points and point out my topics of confusion in this manner.  I would like to im...
a80ead47-8a2a-484d-ab9b-c0050b5c722d
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Inner Alignment: Explain like I'm 12 Edition *(This is an unofficial explanation of Inner Alignment based on the Miri paper [Risks from Learned Optimization in Advanced Machine Learning Systems](https://arxiv.org/abs/1906.01820) (which is almost identical to the [LW sequence](https://www.lesswrong.com/s/r9tYkB2a8Fp4DN...