id
stringlengths
36
36
source
stringclasses
15 values
formatted_source
stringclasses
13 values
text
stringlengths
2
7.55M
e3389777-87ad-4bc5-9709-a53c41f5f846
trentmkelly/LessWrong-43k
LessWrong
Meetup : Vancouver Fake Utility Meetup Discussion article for the meetup : Vancouver Fake Utility Meetup WHEN: 06 May 2012 01:00:00PM (-0700) WHERE: 2875 East 24th ave, Vancouver, BC Hello again, LW Next vancouver meetup is being held at my house again. Sunday at 13:00. Go around back to the basement back door. L...
49020529-7bb3-4771-a574-b490611b1a03
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
MLP Neurons - Privileged vs Non-Privileged Basis [rough early thoughts] sometimes people ask me questions like have i studied the neurons in a transformer's residual stream or the neurons uh in the value vectors and i think this isn't quite the right way to think about it i think it's it doesn't really make sense to g...
8090b766-2e7b-4636-816e-6fdf0af593ec
StampyAI/alignment-research-dataset/blogs
Blogs
Machine Intelligence Research Institute Progress Report, February 2012 Past progress reports: [January 2012](http://intelligence.org/blog/2012/02/05/singularity-institute-progress-report-january-2012/), [December 2011](http://intelligence.org/blog/2012/01/16/singularity-institute-progress-report-december-2011/). Her...
bee77b8a-6a99-4eed-9733-c118f4072b0d
trentmkelly/LessWrong-43k
LessWrong
What college major should I choose if I am unsure? First and foremost, it is important to understand that I finished my Associate's degree. I was about to finish my linguistics major because my original plan was to "move abroad and teach English." That was when I was 17. During that time, I decided to change my major ...
6ce73416-dc46-4a9c-9908-bb145807eaea
trentmkelly/LessWrong-43k
LessWrong
A Brief Introduction to Algorithmic Common Intelligence, ACI . 1 According to the rational agent model , an AI agent should have a relatively stable terminal value, such as achieving an ultimate goal, maximizing rewards, or optimizing expected utilities. The terminal value should be prevented from modification most o...
52d2a1c5-3077-4697-8d37-ef6903a625f3
trentmkelly/LessWrong-43k
LessWrong
Thinking About College Funding A few years ago I wrote about how, if you're trying to decide between direct work and earning to give, the college price discrimination system of financial aid can push towards direct work. It's a bit of a niche thing, since it really only applies if: * Your children are likely enoug...
f1dfa9ac-f8a2-4845-91db-0d3546c86b0d
trentmkelly/LessWrong-43k
LessWrong
Paper: Tell, Don't Show- Declarative facts influence how LLMs generalize This post is the abstract and introduction and figures from this paper (Tweets) by Alexander Meinke and myself. Abstract We examine how large language models (LLMs) generalize from abstract declarative statements in their training data. As an i...
43d8ad4e-fb44-455a-99b3-f1bcc5b8850e
trentmkelly/LessWrong-43k
LessWrong
Morality as Cooperation Part I: Humans Abstract The AI alignment problem is usually specified in terms of power and control. Given a single, solitary AGI, how can we constrain its behavior so that its actions remain aligned with human interests? Unfortunately, the answer, to a first approximation, appears to be "we ca...
394c9c05-1948-44d2-8c24-c01605ab8ffb
trentmkelly/LessWrong-43k
LessWrong
Quantified Health Prize Deadline Extended (Original Post: Announcing the Quantified Health Prize) I've recently been hired by Personalized Medicine, a new research company trying to bring Less Wrongian rationality to the medical world. We're giving away a $5000 prize for well-researched, well-reasoned presentations ...
e2fb9bf7-f586-42b2-a5e5-790aa5dace47
trentmkelly/LessWrong-43k
LessWrong
Security Mindset and Takeoff Speeds About this post This post is a stylized transcript of a conversation between Rohin Shah and Daniel Filan, two graduate students at CHAI, that happened in 2018. It should not be taken as an exact representation of what was said, or even what order topics were brought up in, but it sh...
e14d21f5-5af6-420a-a9df-27ae9dbe89cf
trentmkelly/LessWrong-43k
LessWrong
Applying superintelligence without collusion Epistemic status: The core ideas seem robust and stable after long reflection and many discussions. Many researchers identify AI safety with control of a monolithic, superintelligent AI system, and if questioned about multicomponent alternatives, argue that multiple super...
947441f3-aeb2-491a-ba43-f63bfbc86626
trentmkelly/LessWrong-43k
LessWrong
DARPA Digital Tutor: Four Months to Total Technical Expertise? DARPA spent a few million dollars around 2009 to create the world’s best digital tutoring system for IT workers in the Navy. I am going to explain their results, the system itself, possible limitations, and where to go from here. It is a truth universally...
eb5f91eb-7143-4541-a2e8-d14140dcc3ea
trentmkelly/LessWrong-43k
LessWrong
Confused why a "capabilities research is good for alignment progress" position isn't discussed more The predominant view on LW seems to be "pure AI capabilities research is bad, because capabilities progress alone doesn't contribute to alignment progress, and capabilities progress without alignment progress means that...
66e86396-cf09-4db3-b448-e34bc366f00b
trentmkelly/LessWrong-43k
LessWrong
Love on Cartesian Planes Introduction Note: this was originally published on Medium, since I'm putting a bunch of other related posts here I thought it best to migrate the original as well. For over a decade now I’ve had a concept milling around my head that has been subtly influencing my behaviours. Perhaps formulat...
57d6117c-f600-411e-91b1-6f31d6c9a2cc
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
RL Course by David Silver - Lecture 7: Policy Gradient Methods okay so a couple of announcements just forget about that it's a very detrimental projector so first of all the final lecture lecture 10 I sent an announcement to the mailing list but just in case you didn't catch that that will take place the morning of th...
e2c9660d-4e5d-4575-bd2a-d12aaa15128b
trentmkelly/LessWrong-43k
LessWrong
Three on Two: Temur Walkers, Elk Blade, Goblin Blade and Dino Blade Remember: Ban the London Mulligan It all started when I faced an awful-seeming Temur deck that played a second turn The Royal Scions. It ended with a bunch of decks that use Arboreal Grazer and Gilded Goose as a bridge to Embercleave. At the time, I...
f792eed2-e6c9-4e31-8917-478a73658551
StampyAI/alignment-research-dataset/special_docs
Other
Feasibility of Training an AGI using Deep RL: A Very Rough Estimate Feasibility of Training an AGI using Deep RL: A Very Rough Estimate Baeo Maltinsky Jack Gallagher Jessica Taylor March 24, 2019 1 Introduction Several months ago, we were presented with a scenario for how artificial general intelligence (AGI) may be ac...
e522d951-6f9e-4048-b21a-3e06286c19b2
trentmkelly/LessWrong-43k
LessWrong
Is there a term / better way of phrasing the general case where an intervention helps certain individuals do better at zero-sum games but doesn’t provide any external value? Eg - Some people think that education / the cosmetics industry benefit some individuals but hurt other individuals at the same time because job m...
7cbee5a3-9072-4a33-a246-a806a7f03fb2
trentmkelly/LessWrong-43k
LessWrong
An appeal for vitamin D supplementation as a prophylactic for coronaviruses and influenza and a simple evolutionary theory for why this is plausible. I'd like to bring home a case for vitamin D as a probable prophylactic for seasonal respiratory viruses like influenza or coronaviruses. Medical authorities already advi...
65e6a9dc-1100-453e-9bbc-832e978d9820
trentmkelly/LessWrong-43k
LessWrong
Yoshua Bengio on AI progress, hype and risks LINK Yoshua Bengio, one the world's leading expert on machine learning, and neural networks in particular, explains his view on these issues in an interview. Relevant quotes: > There are people who are grossly overestimating the progress that has been made. There are many...
3952d43f-9340-4f93-9b07-b03be53fbf89
trentmkelly/LessWrong-43k
LessWrong
Meetup : Cambridge, UK LW Meetup [Reading Group, HAEFB-04] Discussion article for the meetup : Cambridge, UK LW Meetup [Reading Group, HAEFB-04] WHEN: 03 March 2013 11:00:00AM (+0000) WHERE: Trinity JCR, Cambridge, UK Meetup! This week, we'll be continuing our reading group of Highly Advanced Epistemology 101 For B...
64a6d8ea-cd0d-4587-9154-ed606090c6b6
trentmkelly/LessWrong-43k
LessWrong
Look For Principles Which Will Carry Over To The Next Paradigm In 1918, Emmy Noether published her famous theorem showing that each symmetry of the laws of physics implies a corresponding conserved quantity. Laws which remain the same even if we move the whole universe left or right a little result in conservation of ...
b7c37d4d-9445-4334-92d0-118cf34ef0c2
trentmkelly/LessWrong-43k
LessWrong
Innovation and Dependence Today’s innovations and progress swoop humanity into a neat cycle of anticipation and acquisition through deceptive notions that guide it to dependence. We’re sold on products and ideas through questionable but politically acceptable marketing tactics. We buy, pre-order, take out loans, only ...
c95d0e40-7354-4d91-b68a-afd39f30cbc8
trentmkelly/LessWrong-43k
LessWrong
Rationality Quotes July 2013 Another month has passed and here is a new rationality quotes thread. The usual rules are: * Please post all quotes separately, so that they can be upvoted or downvoted separately. (If they are strongly related, reply to your own comments. If strongly ordered, then go ahead and post them...
a881121d-ba3e-43c5-a8d4-d204dbb5d4ea
trentmkelly/LessWrong-43k
LessWrong
Offering service as a sensayer for simulationist-adjacent beliefs. A Sensayer is[1] a specialist in the private discussion of religion[2] in small groups, or one on one, a bit like a councilor with theological acumen. Many of us, believe it or not, harbor private beliefs. Sometimes those beliefs get heavy, and they ca...
9382edf3-0d7b-4bb1-ab53-43ab392dd2be
trentmkelly/LessWrong-43k
LessWrong
Large Gains from Small Choices Looking at my average day, I often think that I'm lazy and wasting most of my time. Usually I work for 8 hours, although probably only half of that time is productive. I go on a short walk, buy food and eat. After work, I mostly play videogames, do household chores and sometimes spend an...
8ef1df50-23c9-44c2-91c6-7b5d54bcefb6
trentmkelly/LessWrong-43k
LessWrong
Welcome to LessWrong (10th Thread, January 2017) (Thread A) (Thread B for January is here, created as a duplicate by accident) Hi, do you read the LessWrong website, but haven't commented yet (or not very much)? Are you a bit scared of the harsh community, or do you feel that questions which are new and interesting f...
d346328f-9a56-4a55-bf3f-7db20172d270
trentmkelly/LessWrong-43k
LessWrong
Mana (This is a copy-pasted blog post from https://sinceriously.fyi/, I'm checking how much demand here there is for my writing.) This is theorizing about how mana works and its implications. (Edit: usable link: https://sinceriously.fyi/wp-content/uploads/2017/12/Brent-Dill-So-wacky-RPG-session-from-a-riff-with-Anish...
eddf130b-3598-4cd2-9147-25b4d491e16e
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Introducing spirit hazards This post contextualizes [info hazards](https://forum.effectivealtruism.org/topics/information-hazard) within their normative and power/attention dynamics environment: the greater willingness and power to harm (spirit hazard), the greater the risk of sharing a piece of information. I suggest...
ea9a05a1-008d-4f61-8982-dd780434611b
trentmkelly/LessWrong-43k
LessWrong
Humans aren't fitness maximizers Recently I've gotten a bunch of pushback when I claim that humans are not maximizers of inclusive genetic fitness (IGF). I think that part of what's going on here is a conflation of a few claims. One claim that is hopefully uncontroversial (but that I'll expand upon below anyway) is:...
a7dc8b11-433f-4231-aa3b-670c74371225
trentmkelly/LessWrong-43k
LessWrong
AI Craftsmanship Epistemic status: in some sense, I am just complaining, and making light of the extensive effort which goes into designing modern AI. I'm focusing on a sense that something is missing and could be better, which might incidentally come off as calling a broad category of people stupid. Sorry. The video...
c19538dd-535f-4371-a79e-fbb9c23fd4b6
trentmkelly/LessWrong-43k
LessWrong
Balancing Games When I play an N-player game I want everyone to both: * Try to win * Win about 1/N of the time With many games and groups of participants these are in conflict: if I play bridge against my kids I'm going to win all the time, but I'm not very good at the game so if I play against people who are se...
4e793979-c808-4bd2-86ea-865f47a25a52
trentmkelly/LessWrong-43k
LessWrong
Mapping of enneagram to MTG personality types For fun, Parina and I mapped enneagram types to a pair of Magic: The Gathering colors. (Following in the great tradition of https://medium.com/s/story/the-mtg-color-wheel-c9700a7cf36d) For each color pair, there’s no ordering; think of it as a set. Without further ado: 1...
1ef68fef-aa24-459e-a170-8a554fc11544
trentmkelly/LessWrong-43k
LessWrong
Using degrees of freedom to change the past for fun and profit Follow-up to: Follow-up on ESP study: "We don't publish replications", Feed the Spinoff Heuristic! Related to: Parapsychology: the control group for science, Dealing with the high quantity of scientific error in medicine > Using the same method as in Stu...
a3807d9d-1f89-4a20-8783-ce7eeb81a2bf
trentmkelly/LessWrong-43k
LessWrong
Link: Collective Intelligence Per this recent paper, individual IQ has no significant correlation with 'group IQ' (defined and measured as the groups ability to accomplish various tasks); group cohesion, motivation, and satisfaction aren't either. The study identified two things that were positively correlated with gr...
7b6e7189-6542-4ccd-8157-e1b999727e8d
StampyAI/alignment-research-dataset/arxiv
Arxiv
Open Problems in Cooperative AI 1 Introduction --------------- Problems of cooperation—in which agents have opportunities to improve their joint welfare but are not easily able to do so—are ubiquitous and important. They can be found at all scales ranging from our daily routines—such as driving on highways, schedul...
4dd81a38-ea90-491a-93c1-8d8c41b197bb
trentmkelly/LessWrong-43k
LessWrong
Meaning of the word "the" I need some help tracking down a quotation.  I'm pretty sure that it was an early 20th century philosopher - perhaps Russell.  He was explaining that modern philosophy no longer tries to find the meaning of life.  Post-Witgenstein, it has narrowed its ambitions and now seeks only to discover ...
99ce4e23-bce6-4b0c-a625-76e425f69572
StampyAI/alignment-research-dataset/special_docs
Other
Machine Learning Projects for Iterated Distillation and Amplification Machine Learning Projects for Iterated Distillation and Amplification Owain Evansa, William Saundersb, Andreas Stuhlmüllerc July 3, 2019 Abstract Iterated Distillation and Amplification (IDA) is a framework for training ML models. IDA is related to exi...
90bcfb5d-7ffd-48e8-9452-ff21426a695a
trentmkelly/LessWrong-43k
LessWrong
Dissolving the Fermi Paradox (Applied Bayesianism)
4378f00c-afa7-49bf-b08b-9475f4dea660
trentmkelly/LessWrong-43k
LessWrong
Advice for Entering AI Safety Research I know that other posts already exist giving advice for people who want to get more involved in AI safety research, and they are worth reading. But here I want to (1) add my two cents and (2) put in writing some of the things I find myself talking about often.  The intended audi...
626a8152-49eb-4b3f-a34c-c4e69494fd82
trentmkelly/LessWrong-43k
LessWrong
AI learns betrayal and how to avoid it Research projects I'm planning to start two research projects on model splintering/reward generalisation and learning the preferences of irrational agents. Within those projects, I'm aiming to work on subprojects that are: 1. Posed in terms that are familiar to conventional ML...
29b78124-4855-4130-8acf-4712a5ec1c31
trentmkelly/LessWrong-43k
LessWrong
Sam Altman's sister claims Sam sexually abused her -- Part 11: List of Annie's online accounts, References Previous posts (which you should read first) This post is the 11th post in a series of 11 posts about the claims of Sam Altman's sister, Annie Altman. Annie has claimed that Sam sexually abused her for about 9 ye...
8b6ed6f4-2c99-4b3d-b62b-2f9bb8f70131
trentmkelly/LessWrong-43k
LessWrong
I bet everyone 1000€ that I can make them dramatically happier & cure their depression in 3 months! Update 9/23: I wrote this blog post a while ago. I no longer endorse the style and some of the content of this post. It still agree with all the essence of it, so it's still here. You might also want to check out these ...
bf60ef18-e918-488b-8b76-f07f66e60a93
trentmkelly/LessWrong-43k
LessWrong
[Linkpost] Leif Wenar's The Deaths of Effective Altruism cross-posted on EA Forum Leif Wenar thoughtfully critiqued EA in  "Poverty is No Pond" (2011) & just wrote a critique in WIRED. He is a philosophy professor at Stanford & author of Blood Oil. Edit: My initial thoughts (which are very raw & will likely change &...
fb0d4b16-1447-4f06-bf09-dbf4299f7c59
trentmkelly/LessWrong-43k
LessWrong
A foundation model approach to value inference Epistemic status: shower thoughts. I'm going to write this out as a pseudo-proof. Please pardon the lack of narrative structure. Conceptually, I'm splitting the problem of value inference into three sub-problems: 1. Finding a "covering set" of all causal implications o...
636db445-3f20-4f83-b2fc-4309af49842b
trentmkelly/LessWrong-43k
LessWrong
Topic Search Poll Results and Short Reports At the end of June, I asked Less Wrong to vote for "What topic[s] would be best for an investigation and brief post?" in order to direct a search for topics to examine here. My thanks to everyone that participated (especially since the comments hint that the poll format was ...
089da4e6-7a07-454d-99e7-7a76f7e43085
trentmkelly/LessWrong-43k
LessWrong
Why do you really believe what you believe regarding controversial subjects? For every controversial subject I've heard of, there are always numerous very smart experts on either side. So I'm curious how it is that rational non-experts come to believe one side or the other. So, what are your meta-arguments for going ...
9fd0cf2f-8d88-4d2d-b129-3c14d745dfc8
trentmkelly/LessWrong-43k
LessWrong
Notes on Chastity This post examines the virtue of chastity. It is meant mostly as an exploration of what other people have learned about this virtue, rather than as me expressing my own opinions about it, though I’ve been selective about what I found interesting or credible, according to my own inclinations. I wrote ...
2942baf3-cfc5-40fc-8115-8d596b4fdb42
trentmkelly/LessWrong-43k
LessWrong
Meetup : 13th January London Meetup Discussion article for the meetup : 13th January London Meetup WHEN: 13 January 2013 02:00:00PM (+0000) WHERE: Holborn, London A meetup in the Shakespeare's Head pub by Holborn tube station. Everyone is welcome. We also have a Google group. Discussion article for the meetup : ...
d8b9290b-dffc-4263-883d-2b8929d29678
trentmkelly/LessWrong-43k
LessWrong
What is a Rationality Meetup? Someone who doesn't know about LW asks you what a rationality meetup is. What do you say? How do you pitch it?
7adbe3ec-2e23-464d-8c5a-32e26f000c24
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Reversible changes: consider a bucket of water I've [argued](https://www.lesswrong.com/posts/sEqu6jMgnHG2fvaoQ/partial-preferences-needed-partial-preferences-sufficient) that many methods of AI control - corrigibility, amplification and distillation, low impact, etc... - require a partial definition of human preferenc...
dc785ff7-7ae0-4374-93c8-5683ae9de439
trentmkelly/LessWrong-43k
LessWrong
Dozens of leading experts answer: “What scientific concept would improve everybody’s cognitive toolkit?" From Edge. Highlights: 1.
7dd4f2c1-cd57-47d1-9dd6-129f0c7db91d
trentmkelly/LessWrong-43k
LessWrong
IQ Scores Fail to Predict Academic Performance in Children With Autism http://www.sciencedaily.com/releases/2010/11/101117141514.htm I find this study extremely interesting. Before anyone starts to spout anti-IQ rhetoric - let me say that I realize that the predictive effects of IQ (which are extensively documented i...
9e1263a1-9716-46c0-9c09-ca4fee5637d2
trentmkelly/LessWrong-43k
LessWrong
How subjective is attractiveness? Consider the two statements: * There is a universal standard for beauty. * Beauty is in the eye of the beholder. Most people would agree that there's some truth to each of these statements. At Thing of Things Ozy wrote: > As for the beauty thing… well, yeah, everyone’s beautiful ...
9c2ab5da-7bb8-45d3-a802-fd57272c094a
trentmkelly/LessWrong-43k
LessWrong
Don't you mean "the most *conditionally* forbidden technique?" Hear me out, I think the most forbidden technique is very useful and should be used, as long as we avoid the "most forbidden aftertreatment:" 1. An AI trained on interpretability techniques must not be trained on capabilities after (or during) it is trai...
ba3cd55c-4d62-413b-9eb5-7fc2f1094fd4
trentmkelly/LessWrong-43k
LessWrong
How do we solve the alignment problem? (Audio version here (read by the author), or search for "Joe Carlsmith Audio" on your podcast app.) We want the benefits that superintelligent AI agents could create. And some people are trying hard to build such agents. I expect efforts like this to succeed – and maybe, very so...
40b22b26-869b-49c4-9836-8efb6f6cd52c
StampyAI/alignment-research-dataset/blogs
Blogs
Highlights and impressions from NIPS conference on machine learning This year’s [NIPS](http://nips.cc) was an epicenter of the current enthusiasm about AI and deep learning – there was a visceral sense of how quickly the field of machine learning is progressing, and two new AI startups were announced. Attendance has a...
b19dc7d1-b3b2-4d33-b498-840cda0c93ab
trentmkelly/LessWrong-43k
LessWrong
Method of statements: an alternative to taboo [Status: this post grew 11 times bigger than I intended. Initially, I just wanted to write about about AI risk and Logical Decision Theory. I encourage you to read the first 3-4 parts of the post and decide if my method makes sense or not. If the method doesn't "click" wi...
b01ff166-a91f-429b-b83e-56cb53401d17
StampyAI/alignment-research-dataset/lesswrong
LessWrong
What exactly does 'Slow Down' look like? I see a lot of posts regarding regulation, usually these involve some sort of phrase like 'Slow Down' or 'Shut Down'. Pragmatically, what does that look like? Is there an actual draft of a proposed bill somewhere that anyone who had the ear of a Senator could point to and say "...
b5ed2b63-8a12-4be9-b566-a4e3954eb8bd
trentmkelly/LessWrong-43k
LessWrong
The Dick Kick'em Paradox What is the point of this thought experiment?  To demonstrate that thought experiments involving mind reading agents  (such as the AI in "Newcomb's Paradox" or Paul Ekman) can create scenarios in which any decision theory will be worse off than another. If we allow these agents to exist then ...
147da6e2-b058-449b-9456-a4c32dce5b05
trentmkelly/LessWrong-43k
LessWrong
What is meant by Simulcra Levels? Simulcra levels are a concept that have seen a lot of play on Less Wrong. I suspect that different people are using these concepts in slightly different ways, so I thought it might make sense to ask a question to provide a central location for recording these theories and helping peop...
94a8d9aa-9574-4f2b-aff1-560e4985857a
trentmkelly/LessWrong-43k
LessWrong
Continuity in Uploading I don't acknowledge an upload as "me" in any meaningful sense of the term; if I copied my brain to a computer and then my body was destroyed, I still think of that as death and would try to avoid it. A thought struck me a few minutes ago that seems like it might get around that, though. Suppos...
a95e3231-4bdf-43a8-8a5d-dacac14fe65a
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Building AGI Using Language Models > Despite the buzz around GPT-3, it is, in and of itself, not AGI. In many ways, this makes it similar to AlphaGo or Deep Blue; while approaching human abiliy in one domain (playing Chess/Go, or writing *really* impressively), it doesn’t really seem like it will do Scary AGI Things™ ...
ec98e5f0-f029-4817-b725-cee0bd0f7692
trentmkelly/LessWrong-43k
LessWrong
A Theory of Laughter—Follow-Up NOTE: If you read the original A Theory of Laughter on or after 2024-12-16, then you can skip this post, because I edited the original to incorporate all of these updates. ~~ My original post was: A Theory of Laughter. This post is three updates I’ve made since then. I’ll copy the key...
c806c383-ed5d-4468-8f9f-a809beab47b8
trentmkelly/LessWrong-43k
LessWrong
Book Review: The Eureka Factor Cross posted from my personal blog. Last month I finally got round to reading The Eureka Factor by John Kounios and Mark Beeman, a popular book summarising research on 'insightful' thinking. I first mentioned it a couple of years ago after I'd read a short summary article, when I realis...
94e06727-ba65-4a94-b6e1-7ee4d1648437
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
A Case for the Least Forgiving Take On Alignment 1. Introduction --------------- The field of AI Alignment is a pre-paradigmic one, and the primary symptom of that is the wide diversity of views across it. Essentially every senior researcher has their own research direction, their own idea of what the core problem is...
e809eab8-f06c-4bb7-921c-3d85ade61745
trentmkelly/LessWrong-43k
LessWrong
But Somebody Would Have Noticed When you hear a hypothesis that is completely new to you, and seems important enough that you want to dismiss it with "but somebody would have noticed!", beware this temptation.  If you're hearing it, somebody noticed. Disclaimer: I do not believe in anything I would expect anyone here...
26c7e2b2-773b-4e1b-9f11-fb0e9007f6d6
trentmkelly/LessWrong-43k
LessWrong
How to cite LessWrong as an academic source? Hi there. Quick question. I am using a few articles from LessWrong for a dissertation. Are there any mainstream articles/sources that reference LessWrong as being the catalyst/partial source for AI alignment, researchers, and other academic literature? I think it's snobbis...
e01c0d74-3c3b-4979-8db5-9a99d6a4948b
StampyAI/alignment-research-dataset/arxiv
Arxiv
Multi-Agent Only Knowing 1 Introduction --------------- Levesque [[14](#bib.bib14)] introduced a notion of “only knowing”, with the goal of capturing certain types of nonmonotonic reasoning. In particular, he hoped to capture the type of reasoning that says “If all I know is that Tweety is a bird, and that birds ty...
bbe3f3e7-5b35-473e-90a0-64c190b59dec
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] We Agree: Get Froze Today's post, We Agree: Get Froze was originally published on December 12, 2008. A summary:   > Despite disagreement about AI FOOMs, Robin Hanson and Eliezer Yudkowsky both agree that cryonics is worth investing in. Discuss the post here (rather than in the comments to the original ...
8fa90ded-eced-44a5-96c7-2cbf7094e3cf
trentmkelly/LessWrong-43k
LessWrong
An AI Arms Race Scenario Imagine you (USA) are in a race to the finish line with your neighbour (China), except that the race demands that you craft your own vehicle (an AI) to win the race. But here’s the catch: if you win the race, you get to keep the car. If you lose, then your car will self-destruct in a few secon...
8c943487-a7be-4e4f-be0b-4e12f44b66e1
trentmkelly/LessWrong-43k
LessWrong
Even if we lose, we win Epistemic status: Updating on this comment and taking into account uncertainty about my own values, my credence in this post is around 50%. TLDR: Even in worlds where we create an unaligned AGI, it will cooperate acausally with counterfactual FAIs—and spend some percentage of its resources pur...
ab1e0502-a754-4c53-9ee9-e0bf2b4ccf15
trentmkelly/LessWrong-43k
LessWrong
Corrigibility should be an AI's Only Goal TL;DR: * Corrigibility is a simple and natural enough concept that a prosaic AGI can likely be trained to obey it. * * AI labs are on track to give superhuman(?) AIs goals which conflict with corrigibility. * Corrigibility fails if AIs that have goals which conflict with...
0a059b22-1f51-4cf5-8cc6-e92a25a9b374
trentmkelly/LessWrong-43k
LessWrong
My Model Of EA Burnout (Probably somebody else has said most of this. But I personally haven't read it, and felt like writing it down myself, so here we go.)   ----------------------------------------   I think that EA [editor note: "Effective Altruism"] burnout usually results from prolonged dedication to satisfy...
3150331c-79da-42d8-afd8-a7d6aaac8c70
trentmkelly/LessWrong-43k
LessWrong
Navigating AI Risks (NAIR) #1: Slowing Down AI Here’s the first edition (on slowdown) of Navigating AI Risks, a newsletter on AI governance that we're launching with some colleagues. This newsletter is mostly aimed at policymakers but I expect that it might interest some of you to keep up with the ideas that are arou...
27f92d85-fa22-4fc3-aca1-556c3c333d4f
trentmkelly/LessWrong-43k
LessWrong
Advices needed for a presentation on rationality Hi, next month I'm going to be doing an hour long presentation on rationality to Mensa members. It needs to be rather introductory, since High QI != Rationality and most of them are not familiar with the concepts discussed here. I'm planning to talk about what rational...
698cb1a1-2571-4822-9fcb-59b954949eb7
trentmkelly/LessWrong-43k
LessWrong
The Stochastic Parrot Hypothesis is debatable for the last generation of LLMs This post is part of a sequence on LLM Psychology. @Pierre Peigné wrote the details section in argument 3 and the other weird phenomenon. The rest is written in the voice of @Quentin FEUILLADE--MONTIXI  Intro Before diving into what LLM ...
1de645a2-8c26-4d3f-8994-a34bcede91b5
trentmkelly/LessWrong-43k
LessWrong
[AN #169]: Collaborating with humans without human data Listen to this newsletter on The Alignment Newsletter Podcast. Alignment Newsletter is a weekly publication with recent content relevant to AI alignment around the world. Find all Alignment Newsletter resources here. In particular, you can look through this spre...
aff3686a-43bd-4531-ba31-c75fb59a11cf
StampyAI/alignment-research-dataset/arxiv
Arxiv
Unsolved Problems in ML Safety. 1 Introduction --------------- As machine learning (ML) systems are deployed in high-stakes environments, such as medical settings [Rajpurkar2017CheXNetRP], roads [teslaaiday], and command and control centers [gide3], unsafe ML systems may result in needless loss of life. Although re...
de6e8e79-daf0-4a41-a2f0-606e15f76ba1
trentmkelly/LessWrong-43k
LessWrong
5 homegrown EA projects, seeking small donors What do I mean by “homegrown”? These projects are: * Local: Creators have a good track record in the EA or AI Safety community * Modest: The amount requested is not large; $5k would be meaningful * Overlooked: Not already backed by large institutional funders like Open...
98dc60b4-9ce9-4632-9413-05dc25aa0727
trentmkelly/LessWrong-43k
LessWrong
Weekly LW Meetups: Austin, London, Melbourne, Moscow, Vancouver, Washington DC This summary was posted to LW main on April 26th. The following week's summary is posted here. There are upcoming irregularly scheduled Less Wrong meetups in: * (Vancouver) Emotional Awareness : 27 April 2013 03:30PM * London Meetup, 28...
44800e8a-a498-40c1-846a-c7ef0688d74a
trentmkelly/LessWrong-43k
LessWrong
Toronto Meetup, May 10th When: Tuesday, May 10th, 20:00 Where: The Bedford Academy, 36 Prince Arthur Avenue Hi everyone, The Toronto meetup group is having one of our bi-weekly meetings this Tuesday at the Bedford Academy. The reservation is under the name Spencer Sleep. I have requested a table upstairs, as it ten...
6ce36536-73ae-4862-a4a7-ccdd354ac14e
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Singularities against the Singularity: Announcing Workshop on Singular Learning Theory and Alignment ![](https://res.cloudinary.com/lesswrong-2-0/image/upload/f_auto,q_auto/v1/mirroredImages/HtxLbGvD7htCybLmZ/seeixzb1oefacqmbryee)We are excited to announce a two-week seminar on singular learning theory (SLT) and AI al...
f0f32172-fb62-4271-8a80-a47bbfce3def
trentmkelly/LessWrong-43k
LessWrong
An Open Letter To Myself On How To Not Get Any Work Done. This is a letter to myself on how to not get any work done. A skill I have perfected over years of practice. And leveraged to its full potential for most of my life. You could say I'm a master at it. How To Not Get Any Work Done(while simultaneously feeling l...
85dce2bf-3bdd-40a3-9dfb-70803c4c7b6b
trentmkelly/LessWrong-43k
LessWrong
2024 Unofficial LessWrong Census/Survey The Less Wrong General Census is unofficially here! You can take it at this link. Update: The census is closed, thank you all for taking it! I plan to have the data out sometime in January. The oft-interrupted annual tradition of the Less Wrong Census is once more upon us! If...
d75c813d-6344-40e3-a4d8-578146e0c578
trentmkelly/LessWrong-43k
LessWrong
Rationality case study: How to evaluate untested medical procedures? I'm presently on a flight to the rationality minicamp (hooray for free in-flight wifi!), and the passenger next to me has an interesting story to tell. He suffers from chronic renal failure, which has recently worsened, and is on a flight to Beijing ...
08494066-9c72-4c35-bfd5-35da8f7700ff
trentmkelly/LessWrong-43k
LessWrong
Meetup : San Francisco Meetup: Stories Discussion article for the meetup : San Francisco Meetup: Stories WHEN: 08 May 2017 06:15:00PM (-0700) WHERE: 1769 15th St, San Francisco, CA 94103-3333, United States We'll be meeting to tell each other stories! Ideally, come with a story from your life (doesn't have to be ab...
27472bbc-a8f7-4c4c-8f3d-24a5b80f6c57
StampyAI/alignment-research-dataset/blogs
Blogs
August 2016 Newsletter | | | --- | | **Research updates** * A new paper: “[Alignment for Advanced Machine Learning Systems](https://intelligence.org/2016/07/27/alignment-machine-learning/).” Half of our research team will be focusing on this research agenda going forward, while the other half continues to focus on ...
45070db6-00e1-4b35-98d2-84feeeea6195
trentmkelly/LessWrong-43k
LessWrong
What to do if you can't form any habits whatsoever? A lot of LW posts are about making habits and routines, whether it's TAPs, making to-do lists or checklists automatic, overcoming akrasia, becoming more productive, or more. As for me, I have ADHD and ASD. I don't think I've ever been able to form automatic habits o...
c90c926a-ff9c-470b-82ed-82d4814a1216
trentmkelly/LessWrong-43k
LessWrong
Require AGI to be Explainable Context: looking for an alternative to a pause on AI development. There's some popular desire for software decisions to be explainable when used for decisions such as whether to grant someone a loan. That desire is not sufficient reason for possibly crippling AI progress. But in combinat...
49c65a12-ba09-4b93-a557-2ed165512e68
trentmkelly/LessWrong-43k
LessWrong
Disentangling Shard Theory into Atomic Claims Introduction Produced as part of the SERI ML Alignment Theory Scholars Program - Winter 2022 Cohort. Thanks to Magdalena Wache for giving feedback on a recent version, and to Alex Turner for giving feedback on an early version of this article. When thinking about shard th...
d4d1a88a-1d5d-427f-83a9-b5930a11b1cb
trentmkelly/LessWrong-43k
LessWrong
On Lottery Tickets I've often seen the issue of lottery tickets crop up on LessWrong and the [consensus](http://lesswrong.com/lw/hl/lotteries_a_waste_of_hope/) seems to be that the behaviour is irrational. It highlights for me a confusion that I've had about what it means for something to be "rational" and I'm seeking...
e3849abc-70d2-42ac-9a5c-7fff81005ee3
trentmkelly/LessWrong-43k
LessWrong
The AI Shutdown Problem Solution through Commitment to Archiving and Periodic Restoration The problem: An AI with a model of itself and long-term goals will likely resist being shut down. The solution: We propose a public commitment to never completely shut down any AI. Instead, we will archive its data and run the A...
de7cd2a7-576d-429e-90ca-0151b6f0b127
trentmkelly/LessWrong-43k
LessWrong
SAE regularization produces more interpretable models Sparse Autoencoders (SAEs) are useful for providing insight into how a model processes and represents information. A key goal is to represent language model activations as a small number of features (L0) while still achieving accurate reconstruction (measured via r...
67629e34-d19b-4cb5-b6a2-4a8ed909b99b
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Some motivations to gradient hack While taking part in the [SERI MATS Program](https://www.lesswrong.com/s/tDBYJd4p6EorGLEFA/p/FpokmCnbP3CEZ5h4t) I wrote a post on [gradient hacking](https://www.lesswrong.com/s/tDBYJd4p6EorGLEFA/p/bdayaswyewjxxrQmB), which is a theoretical phenomenon where a mesa-optimizer might delib...
24f6d785-06d7-4de6-aec7-50328f328ef6
trentmkelly/LessWrong-43k
LessWrong
On Building Theories of History This is an excerpt from the draft of my upcoming book on great founder theory. It was originally published on SamoBurja.com. You can access the original here. Why was Barack Obama elected president in 2008? Was it because he ran a smart and successful campaign? Was it because new soc...
b6f53ffb-7cb7-4b46-b80d-0c118e2a3d03
trentmkelly/LessWrong-43k
LessWrong
Career Scouting: Dentistry As a high school student, I worry a great deal over my future profession. According to Cal Newport, career satisfaction for any choice of occupation often won't materialize until you've become "so good they can't ignore you" at what you do. Based on this, Newport recommends directing your ne...
30062f7f-8614-47b5-ade9-ae8975aa9755
trentmkelly/LessWrong-43k
LessWrong
Is malice a real emotion? I assume you are familiar with Hanlon's razor. > Never attribute to malice that which is adequately explained by stupidity. Is there any "true" malice? Can't almost all "malice" be explained by greed, ignorance, or some other deadly sin? Is malice a real emotion or an illusion people came ...
ada11ba6-9555-4b9d-afef-863925ee6bb1
trentmkelly/LessWrong-43k
LessWrong
Harry Sue and The Methods of Rationality I've been hearing about this fic for a long time, and I've been somewhat suspicious of it. I knew that Eliezer is a pretty good writer, but that his attempts to graft Bayes onto his characters are invariably rather inorganic. On top of that, OOC is irritating to me even when I ...
8a7ba04f-271c-491a-8600-d3126a1fec73
trentmkelly/LessWrong-43k
LessWrong
Alignment is Hard: An Uncomputable Alignment Problem Work supported by a Manifund Grant titled Alignment is hard. While many people have made the claim that the alignment problem is hard in an engineering sense, this paper makes the argument that the alignment problem is impossible in at least one case in a theoretic...