text stringlengths 300 320k | source stringlengths 52 154 |
|---|---|
# My Detailed Notes & Commentary from Secular Solstice
*Previously:* [*General Thoughts on Secular Solstice*](https://www.lesswrong.com/posts/ozdk5cFCqtZuRnWkw/general-thoughts-on-secular-solstice)*.*
*This blog post is my scattered notes and ramblings about the individual components (talks and songs) of* [*Secular S... | https://www.lesswrong.com/posts/sZDuDRzxysjQpAsMF/my-detailed-notes-and-commentary-from-secular-solstice |
# Mandolin Harp Sensor Placement
One of my goals in adding [electronic "harp strings"](https://www.jefftk.com/p/prototyping-pluck-sensors) to my mandolin is that I don't want to change anything about my normal mandolin technique when I'm not using them. I've been playing for decades, and like how I normally play. This... | https://www.lesswrong.com/posts/sBdbtkBynjvG7Jhod/mandolin-harp-sensor-placement |
# Can quantised autoencoders find and interpret circuits in language models?
Executive Summary
=================
* I try **vector-quantised autoencoders** (VQ-VAEs) as an alternative compression scheme of transformer activations (as opposed to something like a [sparse autoencoder](https://arxiv.org/abs/2309.08600))... | https://www.lesswrong.com/posts/gSfPk8ZPoHe2PJADv/can-quantised-autoencoders-find-and-interpret-circuits-in |
# Could LLMs Help Generate New Concepts in Human Language?
Concepts are the central part of language. I would argue that concepts serve as condensed linguistic representations of concrete or abstract entities, aimed at enhancing the precision and efficiency of thinking and communication.
I found it fascinating to pon... | https://www.lesswrong.com/posts/BzxnMnS7RwgALtXJ7/could-llms-help-generate-new-concepts-in-human-language |
# Define “Agent” (Embedded)
I would be very grateful for as many different attempts at rigorous or semi-rigorous definitions of an “agent” as possible.
Specifically, a definition of an (embedded) Agent that makes it intuitively clear the *nature* of the boundary between Agent and Environment.
(I have read up on Agen... | https://www.lesswrong.com/posts/vdH6NNcaHAgqH4oRB/define-agent-embedded |
# Do not delete your misaligned AGI.
In short: Just keeping all copies of potentially strong agents in long-term storage is a trivial way to maintain incentives to cooperate for some classes of misaligned AGI, by allowing us to reward the AGI's cooperation to whatever degree we later calculate had been warranted. In c... | https://www.lesswrong.com/posts/audRDmEEeLAdvz9iq/do-not-delete-your-misaligned-agi |
# All About Concave and Convex Agents
*An entry-level characterization of some types of guy in decision theory, and in real life, interspersed with short stories about them*
A concave function bends dow... | https://www.lesswrong.com/posts/H67tq5sWPeHJxSqG8/all-about-concave-and-convex-agents |
# On the Confusion between Inner and Outer Misalignment
Here’s my take on why the distinction between inner and outer-alignment frame is weird/unclear/ambiguous in some circumstances: My understanding is that these terms were originally used when talking about AGI. So outer alignment involved writing down a reward or ... | https://www.lesswrong.com/posts/hueNHXKc4xdn6cfB4/on-the-confusion-between-inner-and-outer-misalignment |
# On Lex Fridman’s Second Podcast with Altman
Last week [Sam Altman spent two hours with Lex Fridman](https://www.youtube.com/watch?v=jvqFAi7vkBc&ab_channel=LexFridman) ([transcript](https://lexfridman.com/sam-altman-2-transcript)). Given how important it is to understand where Altman’s head is at and learn what he kn... | https://www.lesswrong.com/posts/AaS6YRAGBFrxt6ZMj/on-lex-fridman-s-second-podcast-with-altman |
# On attunement
*(Cross-posted from [my website](https://joecarlsmith.com/2024/03/25/on-attunement). Podcast version [here](https://joecarlsmithaudio.buzzsprout.com/2034731/14758445-on-attunement), or search for "Joe Carlsmith Audio" on your podcast app.*
*This essay is part of a series that I'm calling "Otherness an... | https://www.lesswrong.com/posts/us8cqwudP5sePqWM2/on-attunement |
# Should rationalists be spiritual / Spirituality as overcoming delusion
I just started thinking about what I would write to someone who disagreed with me on the claim "Rationalists would be better off if they were more spiritual/religious", and for this I'd need to define what I mean by "spiritual".
Here are some t... | https://www.lesswrong.com/posts/i5THpCMGhEGfSi2p9/should-rationalists-be-spiritual-spirituality-as-overcoming |
# Announcing Neuronpedia: Platform for accelerating research into Sparse Autoencoders
*This posts assumes basic familiarity with Sparse Autoencoders. For those unfamiliar with this technique, we highly recommend the introductory sections of *[*these*](https://transformer-circuits.pub/2023/monosemantic-features)[*paper... | https://www.lesswrong.com/posts/BaEQoxHhWPrkinmxd/announcing-neuronpedia-platform-for-accelerating-research |
# Third-party testing as a key ingredient of AI policy
*(nb: this post is written for anyone interested, not specifically aimed at this forum)*
We believe that the AI sector needs effective third-party testing for frontier AI systems. Developing a testing regime and associated policy interventions based on the insigh... | https://www.lesswrong.com/posts/XSqntCNMafhcy9irf/third-party-testing-as-a-key-ingredient-of-ai-policy |
# LessOnline (May 31—June 2, Berkeley, CA)
A Festival of Writers Who are Wrong on the Internet[^b6b887gjlou]
-----------------------------------------------------------------
> LessOnline is a festival celebrating truth-seeking, optimization, and blogging. It's an opportunity to meet people you've only ever known by ... | https://www.lesswrong.com/posts/yvrkBxb5Lp9XY3t7f/lessonline-may-31-june-2-berkeley-ca |
# Perceptual Blindspots: How to Increase Self-Awareness
“Your nose is located right above your mouth. Suppose you don’t brush your teeth for three days. Though this nose is right here, it won’t tell you \[that\] you have not brushed your teeth. The whole room will know you have not brushed your teeth, but you will not... | https://www.lesswrong.com/posts/XkK5FtbdNEkPiYvG6/perceptual-blindspots-how-to-increase-self-awareness |
# Enhancing biosecurity with language models: defining research directions
This report explores the potential of large language models (LLMs) to enhance biosecurity. We conducted interviews with nine biosecurity experts to understand their daily tasks, and how LLMs could be more useful for their work. Our findings ind... | https://www.lesswrong.com/posts/p7zn7M62SFFyQ6ZQF/enhancing-biosecurity-with-language-models-defining-research |
# Economics Roundup #1
I call the section ‘Money Stuff’ but as a column name that is rather taken. There has been lots to write about on this front that didn’t fall neatly into other categories. It clearly benefited a lot from being better organized into subsections, and the monthly roundups could benefit from being s... | https://www.lesswrong.com/posts/hCNt7dc7QXuKB2gsR/economics-roundup-1 |
# My Interview With Cade Metz on His Reporting About Slate Star Codex
On 16 March 2024, I sat down to chat with _New York Times_ technology reporter Cade Metz! In part of our conversation, transcribed below, we discussed his February 2021 article ["Silicon Valley's Safe Space"](https://archive.ph/zW6oX), covering Scot... | https://www.lesswrong.com/posts/oYnwTuxySiaZYDrur/my-interview-with-cade-metz-on-his-reporting-about-slate |
# Modern Transformers are AGI, and Human-Level
*This is my personal opinion, and in particular, does not represent anything like a MIRI consensus; I've gotten push-back from almost everyone I've spoken with about this, although in most cases I believe I eventually convinced them of the narrow terminological point I'm ... | https://www.lesswrong.com/posts/gP8tvspKG79RqACTn/modern-transformers-are-agi-and-human-level |
# Timelines to Transformative AI: an investigation
[*Cross-posted on the EA Forum. *](https://forum.effectivealtruism.org/posts/hzhGL7tb56hG5pRXY/timelines-to-transformative-ai-an-investigation)
*This post is part of a series by Convergence Analysis’ *AI Clarity *team.*
*Justin Bullock and Elliot Mckernon have *... | https://www.lesswrong.com/posts/ddj5HtnCHHMQGiQEM/timelines-to-transformative-ai-an-investigation |
# AE Studio @ SXSW: We need more AI consciousness research (and further resources)
Quick update from[ AE Studio](https://ae.studio/): last week, Judd (AE’s CEO) hosted a[ panel at SXSW](https://schedule.... | https://www.lesswrong.com/posts/ZcJDL4nCruPjLMgxm/ae-studio-sxsw-we-need-more-ai-consciousness-research-and |
# Failures in Kindness
There's a particular kind of widespread human behavior that is *kind* on the surface, but upon closer inspection reveals quite the opposite. This post is about four such patterns.
... | https://www.lesswrong.com/posts/GLpFovxZdwXYwmbkJ/failures-in-kindness |
# Beauty and the Bets
*This is the ninth post in my series on Anthropics. The previous one is* [*The Solution to Sleeping Beauty*](https://www.lesswrong.com/posts/gwfgFwrrYnDpcF4JP/the-solution-to-sleeping-beauty)*. The next one is* [*Semantic Disagreement of Sleeping Beauty Problem*](https://www.lesswrong.com/posts/... | https://www.lesswrong.com/posts/cvCQgFFmELuyord7a/beauty-and-the-bets |
# 20 minutes of work as an artist in one future
I am an artist.
”Eleven evil wizard schoolgirls in an archduke's library, dressed in red and black Asmodean schoolgirl uniforms, perched on armchairs and sofas”[^crlz9xrboig]
*Sigh, at least it’s not more catgirls. I don’t even draw them well.*
I stretched br... | https://www.lesswrong.com/posts/ziAhnFrWqePEq8qKj/20-minutes-of-work-as-an-artist-in-one-future |
# Have we really forsaken natural selection?
Natural selection is often charged with having goals for humanity, and humanity is often charged with falling down on them. The big accusation, I think, is of sub-maximal procreation. If we cared at all about the genetic proliferation that natural selection wanted for us, t... | https://www.lesswrong.com/posts/xuokjPCDrZhNh2HLB/have-we-really-forsaken-natural-selection-1 |
# Nick Bostrom’s new book, “Deep Utopia”, is out today
Bostrom’s new book is out today in hardcover and Kindle in the USA, and on Kindle in the UK.
**Description:**
A greyhound catching the mechanical lure—what would he actually do with it? Has he given this any thought?
Bostrom’s previous book, *Superintelligence:... | https://www.lesswrong.com/posts/j7Dx2ASvvouuEAKFj/nick-bostrom-s-new-book-deep-utopia-is-out-today |
# Daniel Kahneman has died
He was 90 years old.
> His death was confirmed by his stepdaughter Deborah Treisman, the fiction editor for the New Yorker. She did not say where or how he died.
[The obituary](https://www.washingtonpost.com/obituaries/2024/03/27/daniel-kahneman-dead/) also describes an episode from his li... | https://www.lesswrong.com/posts/xWoaT3wLRQx8Rf4AX/daniel-kahneman-has-died |
# Was Releasing Claude-3 Net-Negative?
[Cross-posted to EA forum](https://forum.effectivealtruism.org/posts/DJZkso7xaomcML3wB/was-releasing-claude-3-net-negative)
There’s been a lot of discussion among safety-concerned people about whether it was bad for Anthropic to release Claude-3. I felt like I didn’t have a grea... | https://www.lesswrong.com/posts/Yo84SvKDCBwY5auGw/was-releasing-claude-3-net-negative |
# Towards White Box Deep Learning
Hi, I’d like to share my paper that proposes a novel approach for building white box neural networks.
The paper introduces semantic features as a general technique for controlled dimensionality reduction, somewhat reminiscent of Hinton’s capsules and the idea of “inverse rendering”.... | https://www.lesswrong.com/posts/pnMHEfEtuufJHqQTM/towards-white-box-deep-learning |
# Plausibility of cyborgism for protecting boundaries?
Most of my [boundaries](https://www.lesswrong.com/tag/boundaries-membranes-technical) work so far has been focused on protecting boundaries "*from the outside*". For example, maybe davidad's OAA could produce some kind of boundary-defending global police AI.
But,... | https://www.lesswrong.com/posts/XcDqvYxwyX7jJYrwS/plausibility-of-cyborgism-for-protecting-boundaries |
# The Cognitive-Theoretic Model of the Universe: A Partial Summary and Review
About 15 years ago, I read Malcolm Gladwell's Outliers. He profiled Chris Langan, an extremely high-IQ person, claiming that he had only mediocre accomplishments despite his high IQ. Chris Langan's theory of everything, the Cognitive Theoret... | https://www.lesswrong.com/posts/pzmRDnoi4mNtqu6Ji/the-cognitive-theoretic-model-of-the-universe-a-partial |
# Come to Manifest 2024 (June 7-9 in Berkeley)
TLDR
====
[Manifold](https://manifold.markets) is hosting a festival for prediction markets: [**Manifest 2024**](http://manifest.is)! We’ll have serious talks, attendee-run workshops, and fun side events over the weekend. Chat with special guests like Nate Silver, Scott ... | https://www.lesswrong.com/posts/qhdzkzC7sBq3MHhKs/come-to-manifest-2024-june-7-9-in-berkeley |
# Some Things That Increase Blood Flow to the Brain
Epistemic status: very shallow google scholar dive. Intended mostly as trailheads for people to follow up on on their own.
previously: [https://www.lesswrong.com/posts/h6kChrecznGD4ikqv/increasing-iq-is-trivial](https://www.lesswrong.com/posts/h6kChrecznGD4ikqv/incr... | https://www.lesswrong.com/posts/9bYbuAzhzuGiQjZRe/some-things-that-increase-blood-flow-to-the-brain |
# UDT1.01: The Story So Far (1/10)
We now resume your regularly scheduled LessWrong tradition of decision theory posting.
This is a sequence, be sure to note.
Just the first and last post will be on Alignment Forum, and the whole thing will be linked together.
Epistemic Status: This is mostly just recapping old pos... | https://www.lesswrong.com/posts/wfi5EGKhvYzzSDWag/udt1-01-the-story-so-far-1-10 |
# How do LLMs give truthful answers? A discussion of LLM vs. human reasoning, ensembles & parrots
**Summary**
* Large language models (LLMs) like ChatGPT and Claude 3 become increasingly truthful as they scale up in size and are finetuned for factual accuracy and calibration.
* However, the way LLMs arrive at tru... | https://www.lesswrong.com/posts/ZKksgfTxuxKhDfk4m/how-do-llms-give-truthful-answers-a-discussion-of-llm-vs |
# AI #57: All the AI News That’s Fit to Print
Welcome, new readers!
This is my weekly AI post, where I cover everything that is happening in the world of AI, from what it can do for you today (‘mundane utility’) to what it can promise to do for us tomorrow, and the potentially existential dangers future AI might pose... | https://www.lesswrong.com/posts/5Dz3ZrwBzzMfaucrH/ai-57-all-the-ai-news-that-s-fit-to-print |
# How to safely use an optimizer
*Summary: The post describes a method that allows us to use an untrustworthy optimizer to find satisficing outputs.*
*Acknowledgements:* Thanks to Benjamin Kolb ([@benjaminko](https://www.lesswrong.com/users/benjamink?mention=user)), Jobst Heitzig ([@Jobst Heitzig](https://www.lesswro... | https://www.lesswrong.com/posts/A9Xvme7bDaZJtYm4m/how-to-safely-use-an-optimizer |
# Templates I made to run feedback rounds for Ethan Perez’s research fellows.
**TL;DR:** I'm releasing my [templates](https://drive.google.com/drive/u/0/folders/1juHNzPOwEGkcte9N01q1-Uf8yNhRAdgm) to make running feedback rounds easy for research teams that might otherwise neglect to set it up.
![Screenshot of part o... | https://www.lesswrong.com/posts/JJXWWcSuXZkkofFAJ/templates-i-made-to-run-feedback-rounds-for-ethan-perez-s |
# From the outside, American schooling is weird
Hey, so I wanted to start this dialogue because we were talking on Discord about the secondary school systems and college admission processes in the US vs NZ, and some of the differences were very surprising to me.
I think that it may be illuminating to fellow Americans... | https://www.lesswrong.com/posts/m8c69AzHkYLBvYGmm/from-the-outside-american-schooling-is-weird |
# Politics are not serious by default
\[This is part of a series I’m writing on how to convince a person that AI risk is worth paying attention to.\]
**tl;dr:** People’s default reaction to politics is *not taking them seriously*. They could center their *entire* personality on their political beliefs, and *still* n... | https://www.lesswrong.com/posts/wukJxYymgR3BHntWT/politics-are-not-serious-by-default |
# Addressing Accusations of Handholding
It is common and understandable for people to respond with a great deal of skepticism to whether LLM outputs can ever be said to reflect the will and views ... | https://www.lesswrong.com/posts/bhwxwrqwCp4HnyD6J/addressing-accusations-of-handholding |
# Many people lack basic scientific knowledge
Intelligence varies more than it may appear. I tend to live and work with people near my own intelligence level, and so―probably―do you. I know there's at least [two tiers above me](https://www.lesswrong.com/posts/LZn9asbnJHAJsGPA6/ten-minutes-with-sam-altman). But there's... | https://www.lesswrong.com/posts/jBPfNjSABsFxf73Ex/many-people-lack-basic-scientific-knowledge |
# AI Disclosures: A Regulatory Review
*Cross-posted on* [*the EA Forum*](https://forum.effectivealtruism.org/posts/NqGLDLjhKPJaZZp3p/ai-disclosures-a-regulatory-review)*. This article is the fourth in a series of ~10 posts comprising a **2024 State of the AI Regulatory Landscape Review**, conducted by the *[***Governa... | https://www.lesswrong.com/posts/gTdwBtK3P34ey7pKt/ai-disclosures-a-regulatory-review |
# AI Safety Camp final presentations
On the weekend of April 27^th^-28^th^, our AI Safety Camp teams will present their project findings in 10-minute talks.
[**Join on Zoom**](https://us02web.zoom.us/j/2506207387?pwd=eEpBMTQ3VmtvcHFwdU9lOCtmRitTZz09)
You are welcome to join any talk! Teams are sharing their find... | https://www.lesswrong.com/events/Kp7M5tiWEYQonANJc/ai-safety-camp-final-presentations |
# D&D.Sci: The Mad Tyrant's Pet Turtles
*This is a D&D.Sci scenario: a puzzle where players are given a dataset to analyze and an objective to pursue using information from that dataset.*
You steel your nerves as the Mad Tyrant[^gaqkcxaqkic] peers at you from his throne. In theory, you have nothing to worry about: si... | https://www.lesswrong.com/posts/t9dGiqj9M3Ataaueo/d-and-d-sci-the-mad-tyrant-s-pet-turtles |
# SAE reconstruction errors are (empirically) pathological
Summary
=======
Sparse Autoencoder ([SAE](https://transformer-circuits.pub/2023/monosemantic-features/index.html)) errors are empirically pathological: when a reconstructed activation vector is distance $\epsilon$ from the original activation vector, substitu... | https://www.lesswrong.com/posts/rZPiuFxESMxCDHe4B/sae-reconstruction-errors-are-empirically-pathological |
# Your LLM Judge may be biased
Abstract
--------
AI safety researchers often rely on LLM “judges” to qualitatively evaluate the output of separate LLMs. We try this for our own interpretability research, but find that our LLM judges are often deeply biased. For example, we use Llama2 to judge whether movie reviews ar... | https://www.lesswrong.com/posts/S4aGGF2cWi5dHtJab/your-llm-judge-may-be-biased |
# Back to Basics: Truth is Unitary
It was a dark and stormy night.
The prospect held the front of his cloak tight to his chest. He stumbled, fell over into the mud, and picked himself back up. Shivering, he slammed his body against the front doors of the Temple and collapsed under its awning.
He picked himself up an... | https://www.lesswrong.com/posts/keek6kefALzBX5xAZ/back-to-basics-truth-is-unitary |
# Was a Subway in New York City Inevitable?
Introduction
------------
The first serious attempt at building a subway in New York City occurred in 1866, following the end of the Civil War (1865) and the opening of the first subway in London (1863). The following decades saw a sequence of failed attempts, and the first... | https://www.lesswrong.com/posts/n5LFsDKr7fEge974m/was-a-subway-in-new-york-city-inevitable |
# Denver – ACX Meetups Everywhere Spring 2024
This year's Spring ACX Meetup everywhere in Denver.
Location: Sloan's Lake Park, North Side. Park in the Sloan's Lake North Parking Lot, walk just past the stone structure that's right there, and we'll be on the other side of it. Should have a shade structure up, and a wh... | https://www.lesswrong.com/events/mAkFFQDpNjYPdQFKK/denver-acx-meetups-everywhere-spring-2024 |
# Metascience of the Vesuvius Challenge
The [Vesuvius Challenge](https://scrollprize.org/) is a million+ dollar contest to read 2,000 year old text from charcoal-papyri using particle accelerators and machine learning. The scrolls come from the ancient villa town of Herculaneum, nearby Pompeii, which was similarly bur... | https://www.lesswrong.com/posts/boWGNiQ3oemiKgDw7/metascience-of-the-vesuvius-challenge |
# Cryptocurrency taxation in the UK
Mostly out of curiosity, I've been looking into how cryptocurrency is taxed in the UK. It's not easy to get what I consider to be a full answer, but here's my current understanding, as far as I felt like looking into it. HMRC's [internal cryptoassets manual](https://www.gov.uk/hmrc-... | https://www.lesswrong.com/posts/nzjqeGNtvTAz76Jwg/cryptocurrency-taxation-in-the-uk |
# New paper on aligning AI with human values
We’ve just published a paper on a new way to align language models with human values. We wanted to post it here to get more feedback from folk who have thought deeply about alignment.
I'm pretty excited about it. In the past, I worked on RLHF, InstructGPT, and GPT-4 align... | https://www.lesswrong.com/posts/GYEDF3aQRubDoF2sr/new-paper-on-aligning-ai-with-human-values |
# My simple AGI investment & insurance strategy
TL;DR:
* Options traders think it's extremely unlikely that the stock market will appreciate more than 30 or 40 percent over the next two to three years, as it did over the last year. So they will sell you the option to buy current indexes for 30 or 40% above their cu... | https://www.lesswrong.com/posts/vvg6DmJSprDLNhW3v/my-simple-agi-investment-and-insurance-strategy |
# UDT1.01: Local Affineness and Influence Measures (2/10)
Attention Conservation Notice: This is a moderately mathy post.
**Affineness, it's Useful!**
So, if we're going to be restricting the sorts of environments we're considering, and trying to build an algorithm that's closer to UDT1.0 (just pick your action to o... | https://www.lesswrong.com/posts/FJ5KDQDBWxas8dFJo/udt1-01-local-affineness-and-influence-measures-2-10 |
# SAE-VIS: Announcement Post
This is a post to officially announce the sae-vis library, which was designed to create feature dashboards like those from [Anthropic's research](https://pypi-AgEIcHlwaS5vcmcCJGU4YjE1NjUwLTBiZjktNGRkNi1hMGY2LTNhZjNkNjMyMjBiZgACKlszLCIxMGU1OTI0Yi0xOGQ0LTRiYWMtYWE0ZS0zOThlZGRjOTU0YWUiXQAABiB... | https://www.lesswrong.com/posts/nAhy6ZquNY7AD3RkD/sae-vis-announcement-post-1 |
# How does it feel to switch from earn-to-give?
*I suspect this phenomenon is common in the LW/EA spheres, but I've never seen it presented like this. I describe the way that switching from earning-to-give to working-in-altruism has consequences on one's sense of responsibility and trust. I wonder if others have exper... | https://www.lesswrong.com/posts/Z7rDTnCCZ8Xz8F59R/how-does-it-feel-to-switch-from-earn-to-give |
# The Best Tacit Knowledge Videos on Every Subject
TL;DR
=====
[Tacit knowledge](https://www.lesswrong.com/tag/tacit-knowledge) is extremely valuable. Unfortunately, developing tacit knowledge is usually bottlenecked by apprentice-master relationships. [Tacit](https://www.youtube.com/watch?v=OFuu4pesKf0) [Knowledge](... | https://www.lesswrong.com/posts/SXJGSPeQWbACveJhs/the-best-tacit-knowledge-videos-on-every-subject |
# Apply to be a Safety Engineer at Lockheed Martin!
Are you passionate about ensuring the safety and reliability of the world’s most lethal and cutting-edge weaponry? Does the idea of creating technology and then working out its impacts excite you? Do you thrive in dynamic environments where innovation meets rigorous ... | https://www.lesswrong.com/posts/KtsJwgCWygEntcD7K/apply-to-be-a-safety-engineer-at-lockheed-martin |
# LessWrong's (first) album: I Have Been A Good Bing
**tl;dr:** LessWrong released an album! Listen to it now [on Spotify](https://open.spotify.com/album/0DP8XwSK7voq0rtXiNMhQC), [YouTube](https://www.youtube.com/watch?v=zapNCZPoXIk), [YouTube Music](https://music.youtube.com/channel/UCeW0zyvEq77YiVqoTj6PnWw), [or App... | https://www.lesswrong.com/posts/YMo5PuXnZDwRjhHhE/lesswrong-s-first-album-i-have-been-a-good-bing |
# [April Fools' Day] Introducing Open Asteroid Impact
> “That which does not kill us makes us stronger.”
**Hillary Clinton**, who is still alive
I'm proud and excited to announce the founding of my new startup, [Open Asteroid Impact](https://Openasteroidimpact.org), where we redirect asteroids towards Earth for the ... | https://www.lesswrong.com/posts/tBy4RvCzhYyrrMFj3/introducing-open-asteroid-impact |
# A Selection of Randomly Selected SAE Features
*Epistemic status - self-evident.*
In this post, we interpret a small sample of Sparse Autoencoder features which reveal meaningful computational structure in the model that is clearly highly researcher-independent and of significant relevance to AI alignment.
* * *
... | https://www.lesswrong.com/posts/BK8AMsNHqFcdG8dvt/a-selection-of-randomly-selected-sae-features-1 |
# Thousands of malicious actors on the future of AI misuse
*Announcing the results of a 2024 survey by Convergence Analysis. We’ve just posted the executive summary below, but you can read the full report *[*here*](mailto:funding@convergenceanalysis.org)*. *
In the largest survey of its kind, [Convergence Analysis](h... | https://www.lesswrong.com/posts/xbZ3BSAAjGCayGmyo/thousands-of-malicious-actors-on-the-future-of-ai-misuse |
# Please Understand
In which a case is made for worrying about the AI Prompt Box.
Preamble
--------
Technology serves to *abstract away* nonessential aspects of creative activities, giving us more direct access to their conceptual cores. Few audio engineers pine for the days of flaky reel-to-reel tape machines that ... | https://www.lesswrong.com/posts/yCDsGDyDguXgNwpkb/please-understand |
# Pluck Sensor Circuit
A [while ago](https://www.jefftk.com/p/mandolin-harp-sensor-placement) I finished the "user interface" portion of my electronic harp mandolin. I'm [happy with the signals](https://www.jefftk.com/p/prototyping-pluck-sensors) the piezos put out, but now I need some electrical engineering to get th... | https://www.lesswrong.com/posts/yMCa9GkadHMk6rZDB/pluck-sensor-circuit |
# Protestants Trading Acausally
Protestants believe in predestination. The God has already decided on who's going to get to hell and heaven.
This feels like a terrible incentive structure. If you are already predestined to get one of those places, why care? Why try to be good?
In reality though it works pretty well.... | https://www.lesswrong.com/posts/wecoKPZMy83rLRkuc/protestants-trading-acausally |
# The Evolution of Humans Was Net-Negative for Human Values
_(Epistemic status: publication date is significant.)_
Some observers have argued that the totality of "AI safety" and "alignment" efforts to date have plausibly had a negative rather than positive impact on the ultimate prospects for safe and aligned artifi... | https://www.lesswrong.com/posts/cwiufyabZaAttivvk/the-evolution-of-humans-was-net-negative-for-human-values |
# Announcing Suffering For Good
TL;DR: We are excited to announce the new animal welfare organization **Suffering For Good**, a new **factory farming charity** aimed at **vegans**, where we use our excess profits to buy **suffering offsets**--in particular, an enormous number of **rats on heroin**.
. It makes a compelling argument that artificial general intelligence will have a "slow takeoff", developing over centu... | https://www.lesswrong.com/posts/dkbMqExPkFEvebJJw/gpt-4-on-the-gradual-emergence-of-mechanized-intellect-a |
# Gradient Descent on the Human Brain
TL;DR: Many alignment research proposals often share a common motif: figure out how to enter a basin of alignment / corrigibility for human-level models, and then amplify to more powerful regimes while generalizing gracefully. In this post we lay out a research agenda that comes a... | https://www.lesswrong.com/posts/FGvN7aKgdmsTqJ6qF/gradient-descent-on-the-human-brain |
# LessWrong: After Dark, a new side of LessWrong
The LessWrong team has obviously been hard at work putting out their debut album. But another LessWrong feature also seems to have been released today, to less fanfare: [LessWrong: After Dark](https://lesswrongxxx.com/), a branch of the site devoted to explicit discussi... | https://www.lesswrong.com/posts/E6gydqTEGK3sa66d9/lesswrong-after-dark-a-new-side-of-lesswrong |
# Coherence of Caches and Agents
There's a lot of confusion about what coherence means for agents, and what "coherence theorems" do and don't say about agents. In this post, I'll talk about some particularly simple notions of coherence in a particularly simple setting. We'll see what nontrivial things coherence has to... | https://www.lesswrong.com/posts/wjFijaAkSCceqCgGF/coherence-of-caches-and-agents |
# OMMC Announces RIP
At the Omnicide Machine Manufacturing Corporation, we work tirelessly to ensure an omnicide-free future. That’s why we’re excited to announce our Responsible Increase Policy (RIP)—our internal protocol for managing any risks that arise as we create increasingly omnicidal machines.
Inspired by the... | https://www.lesswrong.com/posts/aRBAhBsc6vZs3WviL/ommc-announces-rip |
# What can we learn about childrearing from J. S. Mill?
\[EDIT: This is an April Fool's post. Please don't raise your kids like this.\]
John Stuart Mill
> was given an extremely rigorous upbringing, and was deliberately shielded from association with children his own age other than his siblings. His father, a follo... | https://www.lesswrong.com/posts/PCQACcyoGJEDs6ujq/what-can-we-learn-about-childrearing-from-j-s-mill |
# Fertility Roundup #3
Previous Fertility Roundups: [#1](https://thezvi.substack.com/p/fertility-rate-roundup-1), [#2](https://thezvi.substack.com/p/fertility-roundup-2).
The pace seems to be doing this about twice a year. The actual situation changes slowly, so presumably the pace of interesting new things should sl... | https://www.lesswrong.com/posts/5k5FeFDCqXfLMj5SJ/fertility-roundup-3 |
# Religion = Cult + Culture
*\[copied in full -- request to develop community knowledge/practices?\]*
Cults are not necessarily bad. [Cults provide value](https://deathisbad.substack.com/p/the-time-i-failed-to-join-a-cult). People join them to get things they need which aren’t provided elsewhere. Every cult is a spir... | https://www.lesswrong.com/posts/ghZihMEEztwPRKZHQ/religion-cult-culture |
# [EA xpost] The Rationale-Shaped Hole At The Heart Of Forecasting
An excerpt from the above that will be relevant to this crowd:
Ben Landau-Taylor of [Bismarck Analysis](https://www.bismarckanalysis.com/#/) wrote a piece on March 6 called “[Probability Is Not A Substitute For Reasoning](https://substack.com/home... | https://www.lesswrong.com/posts/2yaEMAKoBJ6tQYLeE/ea-xpost-the-rationale-shaped-hole-at-the-heart-of |
# How Often Does ¬Correlation ⇏ ¬Causation?
Current best guess: Nearly all the
time[$_{55\%}$](https://fatebook.io/q/in-linear-sems-with-0-1-distributed--clujl9idv0001lc0841gwx9te).
"Correlation ⇏ Causation" is trite by now. And we also know that
[the](https://www.spencergreenberg.com/2022/03/can-you-have-causation-w... | https://www.lesswrong.com/posts/ThLMBYZQ4PHKFKHSP/how-often-does-correlation-causation |
# The Rationalist Haggadot Collection
Passover is coming, which means some of us will be celebrating Secular Seders. For those wanting to celebrate but looking for resources, I present: [the Rationalist Haggadot Collection](https://tigrennatenn.neocities.org/rationalist_haggadot/), an archive of all rationalist Seder ... | https://www.lesswrong.com/posts/FxrqQbZKff9BoGhtc/the-rationalist-haggadot-collection |
# Nature is an infinite sphere whose center is everywhere and circumference is nowhere
Blaise Pascal said that. When I heard it, this interpretation instantly came.
[Here's an AR companion video where I point at some mountains in the Apple Vision Pro to explain the big idea.](https://youtu.be/wTVTl7T39gk)
Whether or... | https://www.lesswrong.com/posts/bqD3wJBoLrk8mCN9B/nature-is-an-infinite-sphere-whose-center-is-everywhere-and-1 |
# Falling fertility explanations and Israel
From Robin Hanson, via TheZvi:
> The following 8 social trends plausibly contribute to falling fertility:
> More gender equality - More equal gender norms, options, & expectations, have contributed to fewer women having kids.
> Higher parenting effort - Expectations for h... | https://www.lesswrong.com/posts/SpZrLay3okto33its/falling-fertility-explanations-and-israel |
# Just because 2 things are opposites, doesn't mean they're just the same but flipped
**The 2 Aspects**
=================
There’s 2 Aspects to things in general. I will call them Mapping Out and Mapping In, in titlecase so you know they’re distinct concepts.
**warmup:** `**0 -> 1**`
------------------------
Here, 0... | https://www.lesswrong.com/posts/RiugkJsKgxbrWCrBq/just-because-2-things-are-opposites-doesn-t-mean-they-re |
# Sparsify: A mechanistic interpretability research agenda
Over the last couple of years, mechanistic interpretability has seen substantial progress. Part of this progress has been enabled by the identification of superposition as a key barrier to understanding neural networks ([Elhage et al., 2022](https://transforme... | https://www.lesswrong.com/posts/64MizJXzyvrYpeKqm/sparsify-a-mechanistic-interpretability-research-agenda |
# The Case for Predictive Models
*Thanks to Johannes Treutlein and Paul Colognese for feedback on this post.*
Just over a year ago, the [Conditioning Predictive Models](https://arxiv.org/abs/2302.00805) paper was released. It laid out an argument and a plan for using powerful predictive models to reduce existential r... | https://www.lesswrong.com/posts/RHDB3BdnvM233bnhG/the-case-for-predictive-models |
# $250K in Prizes: SafeBench Competition Announcement
**TLDR:** CAIS is distributing $250,000 in prizes for benchmarks that empirically assess AI safety. This project is supported by Schmidt Sciences, submissions are open until February 25th, 2025. Winners will be announced April 25th, 2025.
To view additional i... | https://www.lesswrong.com/posts/q4auoEt7wrco76EyA/usd250k-in-prizes-safebench-competition-announcement |
# A gentle introduction to mechanistic anomaly detection
**TL;DR: **Mechanistic anomaly detection aims to flag when an AI produces outputs for “unusual reasons.” It is similar to mechanistic interpretability but doesn’t demand human understanding. I give a self-contained introduction to mechanistic anomaly detection f... | https://www.lesswrong.com/posts/n7DFwtJvCzkuKmtbG/a-gentle-introduction-to-mechanistic-anomaly-detection |
# Concrete empirical research projects in mechanistic anomaly detection
*Thanks to Jordan Taylor, Mark Xu, Alex Mallen, and Lawrence Chan for feedback on a draft! This post was mostly written by Erik, but we're all currently collaborating on this research direction.*
Mechanistic anomaly detection (MAD) aims to flag w... | https://www.lesswrong.com/posts/99gWh9jxeumcmuduw/concrete-empirical-research-projects-in-mechanistic-anomaly |
# What is the purpose and application of AI Debate?
I think there is an important lack of clarity and shared understanding regarding how people intend to *use* AI-Safety-via-Debate-style approaches. So I think it would be helpful if there were some people --- who either (i) work on Debate or (ii) believe that Debate i... | https://www.lesswrong.com/posts/sLckvSBnDmChrkuqs/what-is-the-purpose-and-application-of-ai-debate |
# Best in Class Life Improvement
Im gonna retract this one. Recommends some suboptimally dangerous stuff. Also gives people the wrong idea. I have at least tried anything I recced. But I don't personally use all these things. I only ever recommended them to solve problems you actually had. But given how crazy things h... | https://www.lesswrong.com/posts/PLubzz4Jpm4Pas6nT/best-in-class-life-improvement |
# What's with all the bans recently?
**Summary**: the moderators appear to be soft banning users with 'rate-limits' without feedback. A careful review of each banned user reveals it's common to be banned despite earnestly attempting to contribute to the site. Some of the most intelligent banned users have mainstream... | https://www.lesswrong.com/posts/GSSHcAoSChaKxjNDZ/what-s-with-all-the-bans-recently |
# Should you refuse this bet in Technicolor Sleeping Beauty?
This is the question for people who didn't read my [latest post](https://www.lesswrong.com/posts/cvCQgFFmELuyord7a/beauty-and-the-bets). Please, try to answer it yourself without spoiling the solution, and then post it in the comments with your reasoning and... | https://www.lesswrong.com/posts/ZikHXwz8AHFp3EbtD/should-you-refuse-this-bet-in-technicolor-sleeping-beauty |
# Cult of equilibrium
TLDR: "Solve for the equilibrium" is a nice sentiment, but shouldn't be applied mindlessly, it's not nearly as universal approach as some think.
Longer version:
The phrase "you must solve for the equilibrium" when evaluating something became almost like mantra and a lot of people use it automat... | https://www.lesswrong.com/posts/BbTPYmnJ7rz52LwB5/cult-of-equilibrium |
# AI #58: Stargate AGI
Another round? Of economists projecting absurdly small impacts, of Google publishing highly valuable research, a cycle of rhetoric, more jailbreaks, and so on. Another great podcast from Dwarkesh Patel, this time going more technical. Another proposed project with a name that reveals quite a lot... | https://www.lesswrong.com/posts/qQmWvm68GsXJtK4EQ/ai-58-stargate-agi |
# Language and Capabilities: Testing LLM Mathematical Abilities Across Languages
*Produced as part of the* [*SERI ML Alignment Theory Scholars Program*](https://serimats.org/) *\- Summer 2023 Cohort*
*Thanks to* [*@NicholasKees*](https://www.lesswrong.com/users/nicholaskees?mention=user) *for guiding this and reading... | https://www.lesswrong.com/posts/37uuuPQKiGisi8cGG/language-and-capabilities-testing-llm-mathematical-abilities |
# Trying to Do More Good
*This is an edited transcript of a talk I gave last week at [Commonwealth School](https://www.commschool.org/), a high school in Boston that I attended from 2000 to 2004. I'm typing from memory, so in places it may be closer to what I intended to say than what I actually said.*
It's been twen... | https://www.lesswrong.com/posts/h8bc4ZuDzMC7SZSPf/trying-to-do-more-good |
# Run evals on base models too!
(Creating more visibility for a [comment thread with Rohin Shah](https://www.lesswrong.com/posts/5Dz3ZrwBzzMfaucrH/ai-57-all-the-ai-news-that-s-fit-to-print?commentId=AaapY2KK4fDvwWaRT).)
Currently, [DeepMind's capabilities evals](https://twitter.com/rohinmshah/status/17707466641411732... | https://www.lesswrong.com/posts/dgFC394qZHgj2cWAg/run-evals-on-base-models-too |
# LLMs for Alignment Research: a safety priority?
[A recent short story by Gabriel Mukobi](https://www.lesswrong.com/posts/xLDwCemt5qvchzgHd/scale-was-all-we-needed-at-first) illustrates a near-term scenario where things go bad because new developments in LLMs allow LLMs to accelerate *capabilities research* without a... | https://www.lesswrong.com/posts/nQwbDPgYvAbqAmAud/llms-for-alignment-research-a-safety-priority |
# Quick evidence review of bulking & cutting
*Epistemic status: fairly fast non-comprehensive literature review by a non-expert*
*Content warning: I advise against reading this if you believe you have an eating disorder*
My ideal body aesthetic would be to have defined muscles and low body fat. Maybe this is also t... | https://www.lesswrong.com/posts/pzBhv7H4yBmBwXPnC/quick-evidence-review-of-bulking-and-cutting |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.