text stringlengths 300 320k | source stringlengths 52 154 |
|---|---|
# Some Thoughts on Concept Formation and Use in Agents
*Note: I wrote this document over a year ago and have decided to post it with minimal edits; it isn't entirely up to date with my current thinking on the subject.*
Imagine you enter a room that looks like this:
![](https://lh7-us.googleusercontent.com/2DYik7Yj_... | https://www.lesswrong.com/posts/iQTnCfhQDPZtFq7WZ/some-thoughts-on-concept-formation-and-use-in-agents |
# Advice Needed: Does Using a LLM Compomise My Personal Epistemic Security?
I have been using Claude 2.1 for a few months to solve serious problems in my life and get coaching and support. I need Claude to become functional, mentally well and funded enough to contribute to Utilitarianism and Alignment by donating to t... | https://www.lesswrong.com/posts/EZEiBgP9wxiAvtYJH/advice-needed-does-using-a-llm-compomise-my-personal |
# Deconstructing Bostrom's Classic Argument for AI Doom
I had a pretty great discussion with social psychologist and philosopher Lance Bush recently about the orthogonality thesis, which ended up turning into a broader analysis of Nick Bostrom's argument for AI doom as presented in *Superintelligence*, and some relate... | https://www.lesswrong.com/posts/RbynKk3evb6RiLryL/deconstructing-bostrom-s-classic-argument-for-ai-doom |
# Simple versus Short: Higher-order degeneracy and error-correction
> **TLDR:** *The simplicity bias in Bayesian statistics is not just a bias towards short description length. *
**The folklore relating the simplicity bias in Bayesian statistics to description length is incomplete**: while it is true that the fewer p... | https://www.lesswrong.com/posts/nWRj6Ey8e5siAEXbK/simple-versus-short-higher-order-degeneracy-and-error-1 |
# "How could I have thought that faster?"
I stumbled upon a Twitter thread where Eliezer describes what seems to be his cognitive algorithm that is equivalent to [Tune Your Cognitive Strategies](https://www.lesswrong.com/posts/bbB4pvAQdpGrgGvXH/tuning-your-cognitive-strategies), and have decided to archive / repost it... | https://www.lesswrong.com/posts/rYq6joCrZ8m62m7ej/how-could-i-have-thought-that-faster |
# What do we know about the AI knowledge and views, especially about existential risk, of the new OpenAI board members?
They have announced three new board members in addition to Altman, but we seem to know almost nothing about their views or knowledge on any AI-related subjects? What if anything do we know?
[From Op... | https://www.lesswrong.com/posts/vgCoy4bBrDw9LPrpW/what-do-we-know-about-the-ai-knowledge-and-views-especially |
# Tend to your clarity, not your confusion
*Repost from* [*https://amoretlicentia.substack.com/*](https://amoretlicentia.substack.com/p/tend-to-your-clarity-not-your-confusion)
Modern life is *weird*.
For the more privileged among us, the options of what we could do with our time grow exponentially by the year, by t... | https://www.lesswrong.com/posts/CuqmDTMwTM64FthLy/tend-to-your-clarity-not-your-confusion |
# Storable Votes with a Pay as you win mechanism: a contribution for institutional design
I joined the EA Forum in 2022, with a [post describing my interests and agenda](https://forum.effectivealtruism.org/posts/4viLtxnwzMawqdPum/time-consistency-for-the-ea-community-projects-that-bridge). I also declared in my [first... | https://www.lesswrong.com/posts/5wqFoHBBgpdHeCLS6/storable-votes-with-a-pay-as-you-win-mechanism-a |
# Some (problematic) aesthetics of what constitutes good work in academia
*(Not-terribly-informed rant, written in my free time.)*
*Terminology note:* When I say **“an aesthetic”**, I mean an intuitive (“I know it when I see it”) sense of what a completed paper, project, etc. is ideally “supposed” to look like. It ca... | https://www.lesswrong.com/posts/LZJJK6fuuQtTLRSu9/some-problematic-aesthetics-of-what-constitutes-good-work-in |
# “Artificial General Intelligence”: an extremely brief FAQ
*(Crossposted from* [*twitter*](https://twitter.com/steve47285/status/1755692348523122812) *for easier linking.) (Intended for a broad audience—experts already know all this.)*
When I talk about future “Artificial General Intelligence” (AGI), what am I talki... | https://www.lesswrong.com/posts/uxzDLD4WsiyrBjnPw/artificial-general-intelligence-an-extremely-brief-faq |
# How disagreements about Evidential Correlations could be settled
*Since beliefs about Evidential Correlations don't track any direct ground truth, it's not obvious how to resolve disagreements about them, which is very relevant to* [*acausal trade*](https://www.lesswrong.com/tag/acausal-trade)*.*
*Here I present w... | https://www.lesswrong.com/posts/o64GLrKahR8QrbFQW/how-disagreements-about-evidential-correlations-could-be-1 |
# New social credit formalizations
Here are some classic ways humans can get some kind of social credit with other humans:
1. Do something for them such that they will consider themselves to ‘owe you’ and do something for you in future
2. Be consistent and nice, so that they will consider you ‘trustworthy’ and do c... | https://www.lesswrong.com/posts/qpsFXCTB6ZvEkQxnq/new-social-credit-formalizations |
# Open Thread Spring 2024
If it’s worth saying, but not worth its own post, here's a place to put it.
If you are new to LessWrong, here's the place to introduce yourself. Personal stories, anecdotes, or just general comments on how you found us and what you hope to get from the site and community are invited. This i... | https://www.lesswrong.com/posts/KA2HxJfhz3CSbLdbL/open-thread-spring-2024 |
# The Astronomical Sacrifice Dilemma
*Epistemic Status: I think the dilemma as outlined in Section 1 follows from well-established ideas about Astronomical Waste. However, given that I have not seen it anywhere before I might have made some oversight I am unaware of. You don't know what you don't know but maybe someon... | https://www.lesswrong.com/posts/gxoHkmGg5RuYvx6Ka/the-astronomical-sacrifice-dilemma |
# Results from an Adversarial Collaboration on AI Risk (FRI)
*Authors of linked report: Josh Rosenberg, Ezra Karger, Avital Morris, Molly Hickman, Rose Hadshar, Zachary Jacobs, Philip Tetlock*[^crpys2viudt]
Today, the Forecasting Research Institute (FRI) released “[Roots of Disagreement on AI Risk: Exploring the Pote... | https://www.lesswrong.com/posts/94K6pskgqBmuxsJLx/results-from-an-adversarial-collaboration-on-ai-risk-fri |
# AI Incident Reporting: A Regulatory Review
*This article is the first in a series of ~10 posts comprising a **2024 State of the AI Regulatory Landscape Review**, conducted by the*[***Governance Recommendations Research Program***](https://www.convergenceanalysis.org/programs/governance-research) *at*[***Convergence ... | https://www.lesswrong.com/posts/gZBgmDFqqyw3Lghok/ai-incident-reporting-a-regulatory-review |
# Be More Katja
Katja is widely respected amongst the rationalists and, according to Hive, she is one of the [most followed/respected EA accounts](https://hive.one/p/katjagrace)[^gkmugpi6gdm]. But she doesn't give off the same vibe as many impact olympians. She doesn’t have iron self-will, nor does she manage a huge t... | https://www.lesswrong.com/posts/NbnDb7nfqvDj9Kjqn/be-more-katja |
# Among the A.I. Doomsayers - The New Yorker
The New Yorker just realized a feature article on AI Safety, rationalism, and effective altruism, and… it's surprisingly good? It seems honest, balanced, and even funny. It doesn't take a position about AI Safety, but IMO it paints it in a good way.
Paul Crowley (mentioned... | https://www.lesswrong.com/posts/am5wtA6TFPZWEfraS/linkpost-among-the-a-i-doomsayers-the-new-yorker |
# AI Safety Action Plan - A report commissioned by the US State Department
This morning, Gladstone AI published a [report they made that was commissioned by the US State Department](https://www.gladstone.ai/action-plan). The report **not only acknowledges catastrophic risks from AI, but also includes a very ambitious ... | https://www.lesswrong.com/posts/j5jKWfkbpKGTnzFky/ai-safety-action-plan-a-state-commissioned-report |
# Bias-Augmented Consistency Training Reduces Biased Reasoning in Chain-of-Thought
\[[Twitter thread](https://twitter.com/milesaturpin/status/1767327882978660513)\]
I'm not going to add much additional commentary at the moment and will just let people check out the paper!
But to give a bit more context: This paper ... | https://www.lesswrong.com/posts/FwNPgj9Wnu4tarKLK/bias-augmented-consistency-training-reduces-biased-reasoning |
# OpenAI: The Board Expands
It is largely over.
The investigation into events has concluded, finding no wrongdoing anywhere.
The board has added four new board members, including Sam Altman. There will still be further additions.
Sam Altman now appears firmly back in control of OpenAI.
None of the new board member... | https://www.lesswrong.com/posts/e5kLSeLJ8T5ddpe2X/openai-the-board-expands |
# Superforecasting the Origins of the Covid-19 Pandemic
The Good Judgement Project got some superforecasters to retrocast whether COVID started via zoonotic spillover or a lab leak. They in aggregate gave a 75% chance of zoonosis, but there was a range of views. GJP's executive summary is at the end of this linkpost.
... | https://www.lesswrong.com/posts/b9QXv9HBk8Kq7j2nK/superforecasting-the-origins-of-the-covid-19-pandemic |
# Transformer Debugger
Transformer Debugger (TDB) is a tool developed by OpenAI's Superalignment team with the goal of supporting investigations into circuits underlying specific behaviors of small language models. The tool combines [automated interpretability](https://openai.com/research/language-models-can-explain-n... | https://www.lesswrong.com/posts/EouMcJ4BbYfbjejtw/transformer-debugger-1 |
# Open consultancy: Letting untrusted AIs choose what answer to argue for
*Thanks to Ryan Greenblatt, Buck Shlegeris, Aryan Bhatt, and Akbir Khan for useful discussions and feedback on the draft of this post.*
If AIs are potentially scheming and more knowledgeable than humans, and you want to answer a question, it ma... | https://www.lesswrong.com/posts/ZwseDoobGuqn9FoJ2/open-consultancy-letting-untrusted-ais-choose-what-answer-to |
# The Parable Of The Fallen Pendulum - Part 2
[*Previously*](https://www.lesswrong.com/posts/BzCQHnt7z8qvzqCmi/the-parable-of-the-fallen-pendulum-part-1)*: Some physics 101 students calculate that a certain pendulum will have a period of approximately 3.6 seconds. Instead, when they run the experiment, the stand holdi... | https://www.lesswrong.com/posts/yjLw945kpL4a5d4xv/the-parable-of-the-fallen-pendulum-part-2 |
# How do you improve the quality of your drinking water?
Water quality can have surprisingly high impact on QoL ([just as air purifiers can significantly improve QoL](https://www.lesswrong.com/posts/xLes6tbJFiQRitotx/air-quality-and-cognition)), and some steps (like getting the *right* pitcher have very high return on... | https://www.lesswrong.com/posts/uGB3u9Laww8GywiAR/how-do-you-improve-the-quality-of-your-drinking-water |
# Jobs, Relationships, and Other Cults
For years I (Elizabeth) have been trying to write out my grand unified theory of \[good/bad/high-variance/high-investment\] \[jobs/relationships/religions/social groups\]. In this dialogue me (Elizabeth) and Ruby throw a bunch of component models back and forth and get all the wa... | https://www.lesswrong.com/posts/qZELudpvcmaronerv/jobs-relationships-and-other-cults |
# Virtual AI Safety Unconference 2024
When: May 23rd to May 26th 2024
Where: Online, participate from anywhere.
VAISU is a collaborative and inclusive event for AI safety researchers, aiming to facilitate collaboration, understanding, and progress towards problems of AI risk. It will feature talks, research discussi... | https://www.lesswrong.com/events/hQ6oGLeTHyjyQfnQF/virtual-ai-safety-unconference-2024 |
# Clickbait Soapboxing
Someone on Twitter said:
> *I am guilty of deliberately stating things in a bold & provocative form on here in order to stimulate discussion. Leaving hedges & caveats for the comments section. On net, I think this is better than alternatives, but I’m open to being convinced otherwise.*
And I f... | https://www.lesswrong.com/posts/RXkm28FpqTFBrWqNj/clickbait-soapboxing |
# What could a policy banning AGI look like?
[Caveat lector: I know roughly nothing about policy!]
Suppose that there were political support to really halt research that might lead to an unstoppable, unsteerable transfer of control over the lightcone from humans to AGIs. What government policy could exert that polit... | https://www.lesswrong.com/posts/X9Z9vdG7kEFTBkA6h/what-could-a-policy-banning-agi-look-like |
# Relational Agency: Consistently Reaching Out
When I moved to Gothenburg, I found myself barely knowing anyone. Being a social person, this was a rough state of affairs. I started going to events. It took some time for me to meet a person I liked hanging out with, but eventually, I met a person I could dive into dee... | https://www.lesswrong.com/posts/HWuRphEHkDzzpCh5t/relational-agency-consistently-reaching-out |
# I was raised by devout Mormons, AMA [&|] Soliciting Advice
I was raised by devout Mormons in Mormon central (Northern Utah). It’s hard to accurately capture the scope of the conditioning via writing.
Standard tenets of Mormon doctrine include:
* No tea/coffee/alcohol
* No premarital sex
* Keep the sabbat... | https://www.lesswrong.com/posts/CENYCr6ES3i3jBK3Y/i-was-raised-by-devout-mormons-ama-and-or-soliciting-advice |
# Laying the Foundations for Vision and Multimodal Mechanistic Interpretability & Open Problems
Behold the dogit lens. Patch-level logit attribution is an emergent segmentation map.
*Join our Discord *[... | https://www.lesswrong.com/posts/kobJymvvcvhbjWFKe/laying-the-foundations-for-vision-and-multimodal-mechanistic |
# ACX/LW Seattle spring meetup 2024
Scott Alexander has called for people to organize a spring meetup, and this year, it will be held at Stoup Brewing in Capitol Hill, Seattle. I have made a reservation for two tables at Stoup Brewing, which is known for being one of the quietest bar spaces in the city. I will be wear... | https://www.lesswrong.com/events/cQBu6o8z894gRA6dj/acx-lw-seattle-spring-meetup-2024 |
# On the Latest TikTok Bill
#### TikTok Might Get Banned Soon
[This attempt is getting reasonably far rather quickly, passing the House with broad support.](https://twitter.com/AlecStapp/status/1766532237397283217)
> Alec Stapp: TikTok bill to remove influence of CCP:
>
> – passed unanimously out of committee
>
> ... | https://www.lesswrong.com/posts/cjrDNwoWwuTfc3Hbu/on-the-latest-tiktok-bill |
# Highlights from Lex Fridman’s interview of Yann LeCun
Introduction
============
Yann LeCun is perhaps the most prominent critic of the “LessWrong view” on AI safety, the only one of the three "godfathers of AI" to not acknowledge the risks of advanced AI. So, when he recently [appeared on the Lex Fridman podcast](h... | https://www.lesswrong.com/posts/bce63kvsAMcwxPipX/highlights-from-lex-fridman-s-interview-of-yann-lecun |
# Opportunistic Time-Management
Be willing to break from your routine if you're in the mood to do the normally-less-savory items now. For example, I normally eat breakfast, work for half an hour, and then go on a run (I look forward to going on a run, not to working, but for me working is more important), and I later ... | https://www.lesswrong.com/posts/9iLQNA9r3wdq7Gt6W/opportunistic-time-management |
# 'Empiricism!' as Anti-Epistemology
*(Crossposted by* [*habryka*](https://www.lesswrong.com/users/habryka4) *after asking Eliezer whether I could post it under his account)*
### **i.**
"Ignore all these elaborate, abstract, theoretical predictions," the Spokesperson for Ponzi Pyramid Incorporated said in a firm, re... | https://www.lesswrong.com/posts/LvKDMWQ3yLG9R3gHw/empiricism-as-anti-epistemology |
# A brief review of China's AI industry and regulations
China has enacted three sets of AI regulations since 2021. I haven’t seen a concise breakdown of their content in one place, and I’ve been researching the legislation for a [governance project](https://www.convergenceanalysis.org/programs/governance-research) at ... | https://www.lesswrong.com/posts/JfJvQze89ECArpFhx/a-brief-review-of-china-s-ai-industry-and-regulations |
# Claude vs GPT
Ever since ChatGPT released to the public I have used LLMs every day. GPT-4 was essential in getting me up and running at my job where I had to read and edit pieces of Python, SQL, Unix, and Stata code with little to no prior experience. Beyond coding I’ve had some success using GPT to collect links an... | https://www.lesswrong.com/posts/ZJebMaEae8aPkB3wX/claude-vs-gpt |
# AI #55: Keep Clauding Along
Things were busy once again, partly from the Claude release but from many other sides as well. So even after cutting out both the AI coding agent Devin and the Gladstone Report along with previously covering OpenAI’s board expansion and investigative report, this is still one of the longe... | https://www.lesswrong.com/posts/N3tXkA9Jj6oCB2eiJ/ai-55-keep-clauding-along |
# Sparse autoencoders find composed features in small toy models
Summary
=======
* **Context: **Sparse Autoencoders (SAEs) reveal interpretable features in the activation spaces of language models. They achieve sparse, interpretable features by minimizing a loss function which includes an $\ell_1$ penalty on the S... | https://www.lesswrong.com/posts/a5wwqza2cY3W7L9cj/sparse-autoencoders-find-composed-features-in-small-toy |
# How useful is "AI Control" as a framing on AI X-Risk?
Redwood (where Ryan works) recently released a series of blogposts proposing a research agenda for reducing AI-risk that focuses on ensuring safety (and secondarily usefulness) under the conservative assumption that AIs are misaligned and actively scheming agains... | https://www.lesswrong.com/posts/9BoGjYb5gbjZirmXv/how-useful-is-ai-control-as-a-framing-on-ai-x-risk |
# Collection (Part 6 of "The Sense Of Physical Necessity")
*This is the sixth post in a sequence that demonstrates a complete naturalist study, specifically a study of query hugging (sort of), as described in* [*The Nuts and Bolts of Naturalism*](https://www.lesswrong.com/s/zLib3j2Fdnnx3aP3F)*. This one demos phase th... | https://www.lesswrong.com/posts/yRxn2YyqYhDss7K2H/collection-part-6-of-the-sense-of-physical-necessity |
# More people getting into AI safety should do a PhD
Doing a PhD is a strong option to get great at **developing** and **evaluating** research ideas. These skills are necessary to become an AI safety research lead, one of the key talent bottlenecks in AI safety, and are helpful in a variety of other roles. By contrast... | https://www.lesswrong.com/posts/yi7shfo6YfhDEYizA/more-people-getting-into-ai-safety-should-do-a-phd |
# Towards a Broader Conception of Adverse Selection
*“I refuse to join any club that would have me as a member” -Marx*[^67fg48wd2cg]
Adverse Selection is the phenomenon in which information asymmetries in non-cooperative environments make trading dangerous. It has traditionally been understood to describe financial m... | https://www.lesswrong.com/posts/vyAZyYh3qsqcJwwPn/towards-a-broader-conception-of-adverse-selection |
# Constructive Cauchy sequences vs. Dedekind cuts
In classical ZF and ZFC, there are two standard ways of defining reals: as Cauchy sequences and as Dedekind cuts. Classically, these are equivalent, but are inequivalent constructively. This makes a difference as to which real numbers are definable in type theory.
## ... | https://www.lesswrong.com/posts/oQ2nRRJFhjRrZHMyH/constructive-cauchy-sequences-vs-dedekind-cuts |
# Improving SAE's by Sqrt()-ing L1 & Removing Lowest Activating Features
TL;DR
=====
We achieve better SAE performance by:
1. Removing the lowest activating features
2. Replacing the L1(feature\_activations) penalty function with L1(sqrt(feature\_activations))
with 'better' meaning: we can reconstruct the origin... | https://www.lesswrong.com/posts/YiGs8qJ8aNBgwt2YN/improving-sae-s-by-sqrt-ing-l1-and-removing-lowest |
# Rational Animations offers animation production and writing services!
[Rational Animations](https://www.youtube.com/@RationalAnimations/featured) is now open to take on external work! We offer several services related to writing and animation production. In particular:
* Production management
* Storyboarding
* ... | https://www.lesswrong.com/posts/Mbd2CifDjFkHDFjZJ/rational-animations-offers-animation-production-and-writing |
# Beyond Maxipok — good reflective governance as a target for action
Summary
=======
Avoiding existential risk in the foreseeable future is a necessary but not sufficient condition to get good long-term outcomes. If we do get great outcomes, that almost certainly means there was a good reflective process which helped... | https://www.lesswrong.com/posts/qZNiWdRzmAsPGobii/beyond-maxipok-good-reflective-governance-as-a-target-for |
# Introducing METR's Autonomy Evaluation Resources
This is METR’s collection of resources for evaluating potentially dangerous autonomous capabilities of frontier models. The resources include a [task suite](https://github.com/METR/public-tasks), some software tooling, and guidelines on how to ensure an accurate measu... | https://www.lesswrong.com/posts/5n9ofttMrJSrrZmDq/introducing-metr-s-autonomy-evaluation-resources |
# How people stopped dying from diarrhea so much (& other life-saving decisions)
*Rational Animations made this video collaborating with 80,000 Hours. The script has been written by Benjamin Hilton as an adaptation of part of the* [*80,000 Hours career guide*](https://80000hours.org/career-guide/)*, by Benjamin Todd. ... | https://www.lesswrong.com/posts/pGMxYCxR3oLspPXzr/how-people-stopped-dying-from-diarrhea-so-much-and-other |
# My PhD thesis: Algorithmic Bayesian Epistemology
In January, I defended my PhD thesis, which I called [Algorithmic Bayesian Epistemology](https://arxiv.org/abs/2403.07949). From the preface:
> For me as for most students, college was a time of exploration. I took many classes, read many academic and non-academic wo... | https://www.lesswrong.com/posts/6dd4b4cAWQLDJEuHw/my-phd-thesis-algorithmic-bayesian-epistemology |
# Anxiety vs. Depression
I have anxiety and depression.
The kind that doesn’t go away, and you take pills to manage.
This is not a secret.
What’s more interesting is that I just switched medications from one that successfully managed the depression but not the anxiety to one that successfully manages the anxiety bu... | https://www.lesswrong.com/posts/rAjXtKTn4Soz5N25L/anxiety-vs-depression |
# What is the best argument that LLMs are shoggoths?
Where can I find a post or a*rt*icle arguing that the internal cognitive model of contemporary LLMs is quite alien, strange, non-human, even though they are trained on human text and produce human-like answers, which are rendered "friendly" by RLHF?
To be clear... | https://www.lesswrong.com/posts/FyRDZDvgsFNLkeyHF/what-is-the-best-argument-that-llms-are-shoggoths |
# Applying simulacrum levels to hobbies, interests and goals
One of the things I've been very confused by for most of my life is why it seems like few people truly care about the things they say they like. That is, they don't spend their spare time thinking about it; they don't read Wikipedia about it, let alone subre... | https://www.lesswrong.com/posts/7oGHeM59TuC2LXh8A/applying-simulacrum-levels-to-hobbies-interests-and-goals |
# The Worst Form Of Government (Except For Everything Else We've Tried)
Churchill [famously called democracy](https://winstonchurchill.org/resources/quotes/the-worst-form-of-government/) “the worst form of Government except for all those other forms that have been tried from time to time” - referring presumably to the... | https://www.lesswrong.com/posts/MCa2aQPXgbCvuoHGM/the-worst-form-of-government-except-for-everything-else-we |
# Toki pona FAQ
Whenever I start telling someone about toki pona, they ask at least some of these questions. So I compile the questions and my answers here.
Toki pona is a constructed language notable for having under 200 words. The strange writing that probably prompted you to ask me about it is sitelen pona.
How d... | https://www.lesswrong.com/posts/dLnDaLM4KhGomWwk7/toki-pona-faq |
# XAI releases Grok base model
> **We are releasing the base model weights and network architecture of** [**Grok-1**](https://x.ai/blog/grok)**, our large language model. Grok-1 is a 314 billion parameter Mixture-of-Experts model trained from scratch by xAI.**
>
> This is the raw base model checkpoint from the Grok-1... | https://www.lesswrong.com/posts/eEcQHowZEd3jfdHtG/xai-releases-grok-base-model |
# Inferring the model dimension of API-protected LLMs
A new paper by Finlayson et al. describes how to exploit the softmax bottleneck in large language models to infer the model dimension of closed-source LLMs served to the public via an API. I'll briefly explain the method they use to achieve this and provide a toy m... | https://www.lesswrong.com/posts/wQz2cgxPaAkssFkGX/inferring-the-model-dimension-of-api-protected-llms |
# 5 Physics Problems
Muireall and DaemonicSigil trade physics problems. Answers and discussion of the answers have been spoilered so you can try the problems yourself. Please also use spoiler formatting (type "`>!`") in the comments.
### Smeared Out Sun
Okay, so the first problem is from *Thinking Physics*. (I promi... | https://www.lesswrong.com/posts/johq6Bti9frjuRGXc/5-physics-problems |
# Join the AI Evaluation Tasks Bounty Hackathon
**How do we test when autonomous AI might become a catastrophic risk?** One approach is to assess the capabilities of current AI systems in performing tasks relevant to self-replication and R&D. METR (formerly ARC Evals), a research group focused on this question, has:
... | https://www.lesswrong.com/posts/ye9HFcPfbxcQzwTpf/join-the-ai-evaluation-tasks-bounty-hackathon |
# On Devin
#### Introducing Devin
Is the era of AI agents writing complex code systems without humans in the loop upon us?
[Cognition is calling Devin](https://www.cognition-labs.com/blog) ‘the first AI software engineer.’
[Here is a two minute demo](https://twitter.com/cognition_labs/status/1767548763134964000) of... | https://www.lesswrong.com/posts/wovJBkfZ8rTyLoEKv/on-devin |
# Community Notes by X
I did an exploration into how [Community Notes](https://en.wikipedia.org/wiki/Community_Notes) (formerly Birdwatch) from X (formerly Twitter) works, and how its algorithm decides which notes get displayed to the wider community. In this post, I’ll share and explain what I found, as well as offer... | https://www.lesswrong.com/posts/sx9wTyCp5kgy8xGac/community-notes-by-x |
# AtP*: An efficient and scalable method for localizing LLM behaviour to components
*Authors: János Kramár, Tom Lieberum, Rohin Shah, Neel Nanda*
A new paper from the Google DeepMind mechanistic interpretability team, from core contributors János Kramár and Tom Lieberum
[Tweet thread summary](https://twitter.com/Jan... | https://www.lesswrong.com/posts/pnMnjdSJwqa7BHAo4/atp-an-efficient-and-scalable-method-for-localizing-llm |
# Measuring Coherence of Policies in Toy Environments
*This post was produced as part of the* [*Astra Fellowship*](https://www.constellation.org/programs/astra-fellowship) *under the Winter 2024 Cohort, mentored by Richard Ngo. Thanks to Martín Soto, Jeremy Gillen, Daniel Kokotajlo, and Lukas Berglund for feedback.*
... | https://www.lesswrong.com/posts/uvv8aMutPEtoBgw7D/measuring-coherence-of-policies-in-toy-environments-2 |
# Neuroscience and Alignment
I've been in many conversations where I've mentioned the idea of using neuroscience for outer alignment, and the people who I'm talking to usually seem pretty confused about why I would want to do that. Well, I'm confused about why one *wouldn't* want to do that, and in this post I explain... | https://www.lesswrong.com/posts/jiSuMT7vupWFwktzq/neuroscience-and-alignment |
# Experimentation (Part 7 of "The Sense Of Physical Necessity")
*This is the seventh post in a sequence that demonstrates a complete naturalist study, specifically a study of query hugging (sort of), as described in* [*The Nuts and Bolts of Naturalism*](https://www.lesswrong.com/s/zLib3j2Fdnnx3aP3F)*. This one demos p... | https://www.lesswrong.com/posts/JoNPfhAv4gnMMWHfK/experimentation-part-7-of-the-sense-of-physical-necessity |
# Monthly Roundup #16: March 2024
AI developments have picked up the pace. That does not mean that everything else stopped to get out of the way. The world continues.
Do [I have the power?](https://www.youtube.com/watch?v=-dJolYw8tnk&ab_channel=rickyr0ma)
> [Emmett Shear](https://twitter.com/eshear/status/1767290235... | https://www.lesswrong.com/posts/iCvdqrkWg34FNFZYg/monthly-roundup-16-march-2024 |
# AI Safety Evaluations: A Regulatory Review
*This article is the second in a series of ~10 posts comprising a **2024 State of the AI Regulatory Landscape Review**, conducted by the *[***Governance Recommendations Research Program***](https://www.convergenceanalysis.org/programs/governance-research) *at*[***Convergenc... | https://www.lesswrong.com/posts/xE7izyw8JTnnTArF5/ai-safety-evaluations-a-regulatory-review |
# Are extreme probabilities for P(doom) epistemically justifed?
Can you post the superforecaster report that has the 0.12% P(Doom) number. I have not actually read anything of course and might be talking out of my behind.
In any case, there have been several cases where OpenPhil or somebody or other has brought in 'e... | https://www.lesswrong.com/posts/8w7sZSvTug3xoTsAe/are-extreme-probabilities-for-p-doom-epistemically-justifed |
# Increasing IQ by 10 Points is Possible
A while ago I wrote how I managed to add [13 points to my IQ](https://morelucid.substack.com/p/increasing-iq-is-trivial) (as measured by the mean between 4 different tests).
I had 3 “self-experimenters” follow my instructions in San Francisco. One of them dropped off, since, s... | https://www.lesswrong.com/posts/siGufsuhjfRLC52J2/increasing-iq-by-10-points-is-possible |
# "I Can't Believe It Both Is and Is Not Encephalitis!" Or: What do you do when the evidence is crazy?
The short version
=================
Three weeks ago, starting on a Sunday, my brother stood up to turn off the light and felt a wave of disorientation, which passed after a minute or two. Periods of cognitive impair... | https://www.lesswrong.com/posts/dtTpKiRa87cgasJG8/i-can-t-believe-it-both-is-and-is-not-encephalitis-or-what |
# Parent-Friendly Dance Weekends
We just finished the 2024 edition of [Beantown Stomp](https://www.beantownstomp.com/), a contra dance weekend I [helped start](https://www.jefftk.com/news/beantownstomp) (but no longer organize!) in Boston. There are a lot of things I like about the weekend, but one thing I especially ... | https://www.lesswrong.com/posts/stwFMgg9s96SKF8zB/parent-friendly-dance-weekends |
# Best *organization* red-pill books and posts?
Most books about orgs are written for people in an org who need to cope. (*Art of Possibility* by the Zanders is a genuinely great one in this genre.) I am currently orgless and don't need to cope with anything orgwise. Hence, I would love to read some brutal insightful ... | https://www.lesswrong.com/posts/Hh9soECiEtTir44on/best-organization-red-pill-books-and-posts |
# My MATS Summer 2023 experience
This post may be interest people who
\- are interested in getting into AI alignment / the MATS program
\- are interested in the soft skills that I've found valuable in developing when working on a research project
Background
----------
In 2023 I was working as a machine learning ... | https://www.lesswrong.com/posts/DQDLPyPXyYiD8XSMu/my-mats-summer-2023-experience |
# New report: Safety Cases for AI
*ArXiv paper:* [*https://arxiv.org/abs/2403.10462*](https://arxiv.org/abs/2403.10462)
The idea for this paper occurred to me when I saw Buck Shlegeris' MATS stream on "... | https://www.lesswrong.com/posts/HrtyZm2zPBtAmZFEs/new-report-safety-cases-for-ai |
# Comparing Alignment to other AGI interventions: Basic model
Interventions that increase the probability of Aligned AGI [aren't the only kind of AGI-related work](https://www.lesswrong.com/posts/4d2JZiiyvZshhbEeJ/non-alignment-project-ideas-for-making-transformative-ai-go-1) that could importantly increase the Expect... | https://www.lesswrong.com/posts/bacAKrPPADTMQybGx/comparing-alignment-to-other-agi-interventions-basic-model-2 |
# Natural Latents: The Concepts
Suppose our old friends Alice and Bob decide to undertake an art project. Alice will draw a bunch of random purple and green lines on a piece of paper. That will be Alice’s picture (A). She’ll then make a copy, erase all the purple lines, and send the result as a message (M) to Bob. Bob... | https://www.lesswrong.com/posts/mMEbfooQzMwJERAJJ/natural-latents-the-concepts |
# On the Gladstone Report
Like the [the government-commissioned Gladstone Report](https://twitter.com/harris_edouard/status/1767183626473517469) on AI itself, there are two sections here.
First I cover the Gladstone Report’s claims and arguments about the state of play, including what they learned talking to people i... | https://www.lesswrong.com/posts/ApZJy3NKfW5CkftQq/on-the-gladstone-report |
# Stagewise Development in Neural Networks
> **TLDR:** *This post accompanies* [The Developmental Landscape of In-Context Learning](https://arxiv.org/abs/2402.02364) *by Jesse Hoogland, George Wang, Matthew Farrugia-Roberts, Liam Carroll, Susan Wei and Daniel Murfet (2024), which shows that in-context learning emerges... | https://www.lesswrong.com/posts/Zza9MNA7YtHkzAtit/stagewise-development-in-neural-networks |
# What I Learned (Conclusion To "The Sense Of Physical Necessity")
*This is the concluding post in a sequence that demonstrates a complete naturalist study, specifically a study of query hugging (sort of), as described in* [*The Nuts and Bolts of Naturalism*](https://www.lesswrong.com/s/zLib3j2Fdnnx3aP3F)*. For contex... | https://www.lesswrong.com/posts/Zf9x9f8zQBkSXdiSn/what-i-learned-conclusion-to-the-sense-of-physical-necessity |
# DeepMind: Evaluating Frontier Models for Dangerous Capabilities
> To understand the risks posed by a new AI system, we must understand what it can and cannot do. Building on prior work, we introduce a programme of new “dangerous capability” evaluations and pilot them on Gemini 1.0 models. Our evaluations cover four ... | https://www.lesswrong.com/posts/CCBaLzpB2qvwyuEJ2/deepmind-evaluating-frontier-models-for-dangerous |
# An Affordable CO2 Monitor
I found a seemingly-high-quality CO~2~ monitor [here](https://www.aliexpress.us/item/3256805816327523.html) for less than $50, or [on Amazon](https://www.amazon.com/Thermometer-Hygrometer-Temperature-Notification-Rechargeable/dp/B0C6XWMRZ6) or [from the manufacturer's website](https://smart... | https://www.lesswrong.com/posts/E5y6L6BybbA6K4skY/an-affordable-co2-monitor-1 |
# AI #56: Blackwell That Ends Well
Hopefully, anyway. Nvidia has a new chip.
Also Altman has a new interview.
And most of Inflection has new offices inside Microsoft.
#### Table of Contents
1. Introduction.
2. Table of Contents.
3. Language Models Offer Mundane Utility. Open the book.
4. Clauding Along. Claude... | https://www.lesswrong.com/posts/iH5Sejb4dJGA2oTaP/ai-56-blackwell-that-ends-well |
# Vipassana Meditation and Active Inference: A Framework for Understanding Suffering and its Cessation
*I want to thank Jan Kulveit, Tomáš Gavenčiak, and Jonathan Shock for their extensive feedback and ideas they contributed to this work and for Josh Burgener and Yusuf Heylen for their proofreading and comments. I wou... | https://www.lesswrong.com/posts/SFHiWyNfWQAtvMBx2/vipassana-meditation-and-active-inference-a-framework-for |
# On green
(*Cross-posted from [my website](https://joecarlsmith.com/2024/03/21/on-green). Podcast version [here](https://joecarlsmithaudio.buzzsprout.com/2034731/14738063-on-green), 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 th... | https://www.lesswrong.com/posts/gvNnE6Th594kfdB3z/on-green |
# A Teacher vs. Everyone Else
> A repairer wants your stuff to break down,
> A doctor wants you to get ill,
> A lawyer wants you to get in conflicts,
> A farmer wants you to be hungry,
> But there is only a teacher who wants you to learn.
*Of course* you see what is wrong with the above "argument / meme / goo... | https://www.lesswrong.com/posts/Btu349AKoEEhaqLk5/a-teacher-vs-everyone-else |
# "Deep Learning" Is Function Approximation
### A Surprising Development in the Study of Multi-layer Parameterized Graphical Function Approximators
As a programmer and epistemology enthusiast, I've been studying some statistical modeling techniques lately! It's been boodles of fun, and might even prove useful in a fu... | https://www.lesswrong.com/posts/DhjcdzTyqHte2v6bu/deep-learning-is-function-approximation |
# ChatGPT can learn indirect control
[Here's a very neat twitter thread](https://twitter.com/joshwhiton/status/1770870746010513571): the author sends various multimodal models screenshots of the conversation he's *currently having* with them, and asks them to describe the images. Most models catch on fast: the author ... | https://www.lesswrong.com/posts/sY3a4Rfa48CgteBEm/chatgpt-can-learn-indirect-control |
# Vernor Vinge, who coined the term "Technological Singularity", dies at 79
> On Wednesday, author David Brin announced that Vernor Vinge, sci-fi author, former professor, and father of the technological singularity concept, died from Parkinson's disease at age 79 on March 20, 2024, in La Jolla, California. The announ... | https://www.lesswrong.com/posts/BYAo8Yvxg2tsCY2Me/vernor-vinge-who-coined-the-term-technological-singularity |
# Transformative AI and Scenario Planning for AI X-risk
*This post is part of a series by the AI Clarity team at *[*Convergence Analysis*](https://www.convergenceanalysis.org/)*. In our *[*previous post*](https://forum.effectivealtruism.org/posts/tCq2fi6vhSsCDA5Js/scenario-planning-for-ai-x-risk)*, Corin Katzke review... | https://www.lesswrong.com/posts/asfqtMijjTjggedyD/transformative-ai-and-scenario-planning-for-ai-x-risk |
# Benchmarking LLM Agents on Kaggle Competitions
tl;dr: I prompted ChatGPT to participate in a Kaggle data science competition. It successfully wrote scripts that trained models to predict housing prices, and ultimately outperformed 71% of human participants.
I'm not planning to build a benchmark using Kaggle compe... | https://www.lesswrong.com/posts/cZHezHezooJ4ryiro/benchmarking-llm-agents-on-kaggle-competitions |
# Video and transcript of presentation on Scheming AIs
(Cross-posted from [my website](https://joecarlsmith.com/2024/03/22/video-and-transcript-of-presentation-on-scheming-ais).)
This is the video and transcript for a ~45-minutes talk I gave in February 2024 about my report “[Scheming AIs: Will AIs fake alignment dur... | https://www.lesswrong.com/posts/7tNwQGh8ZA67BXZAf/video-and-transcript-of-presentation-on-scheming-ais |
# Wolf and Rabbit
Rabbit said to Wolf, "I shall make a new animal. It shall stand on two legs and run free across meadows and mountains."
Wolf said, "I will send large beasts to eat your new animal."
Rabbit said, "My animal shall band together and fight off even large beasts. They shall travel together to stay safe.... | https://www.lesswrong.com/posts/FwFdRmT6oCFWoHS7f/wolf-and-rabbit |
# [Linkpost] Vague Verbiage in Forecasting
> “What does a ‘fair chance’ mean?”
>
> It is a question posed to a diverse group of professionals—financial advisers, political analysts, investors, journalists—during one of Good Judgment Inc’s virtual workshops. The participants have joined the session from North America,... | https://www.lesswrong.com/posts/3m5qCdX7zuBsyAPpb/linkpost-vague-verbiage-in-forecasting |
# What does "autodidact" mean?
Alice takes a chemistry class at a university. She gets a professor who basically just reads from the textbook during lectures. She reads the textbook on her own, talks to her classmates, and finds some relevant Wikipedia articles and youtube videos.
Bob studies chemistry on his own. He... | https://www.lesswrong.com/posts/vnuGLpEMHCRZsNaqj/what-does-autodidact-mean |
# ACX Meetups Everywhere Spring 2024, Montreal, QC
**Note: Recent change of venue.** (EDIT May 11th: since the weather gods have blessed us with beautiful weather, we'll be going on the balcony of the Esplanade Tranquille on the 2nd floor.)
Come on out to the Spring 2024 **ACX Meetups Everywhere** in Montreal, QC!
*... | https://www.lesswrong.com/events/8HEFDrXXm6EjGpDSM/acx-meetups-everywhere-spring-2024-montreal-qc |
# Dangers of Closed-Loop AI
In control theory, an open-loop (or non-feedback) system is one where inputs are independent of outputs. A closed-loop (or feedback) system is one where outputs are input back into the system.
 *last December. *
*My perspective is quite unusual: I live in a rationalist group house and work at an AI safety office, but I also am a Christian and attend church every week.*[^ctzuc39io... | https://www.lesswrong.com/posts/ozdk5cFCqtZuRnWkw/general-thoughts-on-secular-solstice |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.