text stringlengths 300 320k | source stringlengths 52 154 |
|---|---|
# Interpretability’s Alignment-Solving Potential: Analysis of 7 Scenarios
*This is the second post in the sequence “Interpretability Research for the Most Important Century”. The first post, which introduces the sequence, defines several terms, and provides a comparison to existing works, can be found here:* [*Introdu... | https://www.lesswrong.com/posts/FrFZjkdRsmsbnQEm8/interpretability-s-alignment-solving-potential-analysis-of-7 |
# What to do when starting a business in an imminent-AGI world?
As reported by [la3lorn](https://www.lesswrong.com/posts/5onEtjNEhqcfX3LXG/a-generalist-agent-new-deepmind-publication) and [Daniel Kokotajlo](https://www.lesswrong.com/posts/xxvKhjpcTAJwvtbWM/deepmind-s-gato-generalist-agent), Gato is here and appears to... | https://www.lesswrong.com/posts/DMx6Krz9DA5gh8Kac/what-to-do-when-starting-a-business-in-an-imminent-agi-world |
# RLHF
I’ve been thinking about Reinforcement Learning from Human Feedback (RLHF) a lot lately, mostly as a result of [my AGISF capstone project](https://github.com/anshradh/trl_custom) attempting to use it to teach a language model to write better responses to Reddit writing prompts, a la [Learning to summarize from ... | https://www.lesswrong.com/posts/rQH4gRmPMJyjtMpTn/rlhf |
# Deferring
([Cross-posted](https://forum.effectivealtruism.org/posts/LKdhv9a478o9ngbcY/deferring) from the EA Forum)
Deferring is when you adopt someone else's view on a question over your own independent view (or instead of taking the time to form an independent view). You can defer on questions of fact or question... | https://www.lesswrong.com/posts/amH33rnsWbCuDgraE/deferring |
# Thoughts on AI Safety Camp
**I**
Early this year I interviewed a sample of AISC participants and mentors, and spent some time thinking about the problems the AI safety research community is facing, and have changed my mind about some things.
[AI Safety Camp](https://aisafety.camp/) is a program that brings togethe... | https://www.lesswrong.com/posts/3kErRpEprB8iJvnNq/thoughts-on-ai-safety-camp |
# DeepMind is hiring for the Scalable Alignment and Alignment Teams
We are hiring for several roles in the Scalable Alignment and Alignment Teams at DeepMind, two of the subteams of DeepMind Technical AGI Safety trying to make artificial general intelligence go well. In brief,
* The **Alignment Team** investigates... | https://www.lesswrong.com/posts/nzmCvRvPm4xJuqztv/deepmind-is-hiring-for-the-scalable-alignment-and-alignment |
# Can moderators fix old sequences posts?
I'm re-reading sequences now and I'm noticing eye-opening things like so many have SEQ RERUN, useless copies that clutter up the link space while having very few comments that they seem to be intended for.
Can the moderators do something about this? Remove links to them, maybe... | https://www.lesswrong.com/posts/7AD3YS5fbS9xE6AZe/can-moderators-fix-old-sequences-posts |
# An observation about Hubinger et al.'s framework for learned optimization
The observations I make here have little consequence from the point of view of solving the alignment problem. If anything, they merely highlight the essential nature of the inner alignment problem. I will reject the idea that robust alignment,... | https://www.lesswrong.com/posts/rAhJrdxjsXcngn3ip/an-observation-about-hubinger-et-al-s-framework-for-learned |
# "Tech company singularities", and steering them to reduce x-risk
The purpose of this post (also available on the [EA Forum](https://forum.effectivealtruism.org/posts/KopQknZEtjZdoGorT/tech-company-singularities-and-steering-them-to-reduce-x)) is to share an alternative notion of “singularity” that I’ve found useful ... | https://www.lesswrong.com/posts/ezGYBHTxiRgmMRpWK/tech-company-singularities-and-steering-them-to-reduce-x |
# Agency As a Natural Abstraction
**Epistemic status:** Speculative attempt to synthesize findings from several distinct approaches to AI theory.
**Disclaimer:** The first three sections summarize some of Chris Olah's work on interpretability and John Wentworth's Natural Abstractions Hypothesis, then attempt to draw ... | https://www.lesswrong.com/posts/n2urKnXbevj2ryvGY/agency-as-a-natural-abstraction |
# Against Time in Agent Models
When programming distributed systems, we always have many computations running in parallel. Our servers handle multiple requests in parallel, perform read and write operations on the database in parallel, etc.
The prototypical headaches of distributed programming involve multiple proces... | https://www.lesswrong.com/posts/HCibBn3ZCZRwMwNEE/against-time-in-agent-models |
# How close to nuclear war did we get over Cuba?
*Cross posted to LessWrong and* [*The Good blog*](https://thegoodblog.substack.com/p/how-close-to-nuclear-war-did-we-get?s=w)
The Cuban missile crisis was probably the closest we came to nuclear war. There were two types of close call in Cuba. There were moments of ten... | https://www.lesswrong.com/posts/WiXswr75F5zqFkHqp/how-close-to-nuclear-war-did-we-get-over-cuba |
# Alignment as Constraints
In order to find the most promising alignment research directions to pour resources into, we can go about it 3 ways
1. Constraints all alignment proposals should have
2. Constraints for current research directions
3. Constraints for new research directions
Constraints all alignment prop... | https://www.lesswrong.com/posts/7GGRmAyMzqzidmBbi/alignment-as-constraints |
# Frame for Take-Off Speeds to inform compute governance & scaling alignment
Figure 1: Something happens at future time t' tha... | https://www.lesswrong.com/posts/XFL3vaA69mHxATWM7/frame-for-take-off-speeds-to-inform-compute-governance-and |
# Costs and benefits of amniocentesis for normal pregnancies
*Disclaimer: No medical training.*
Amniocentesis is a sample of the amniotic fluid to test fetal DNA at 15-20 weeks of pregnancy. It involves inserting a long needle into the amniotic sack, which is potentially risky. But it could catch serious defects.
Ar... | https://www.lesswrong.com/posts/cHpQKvbjnuemSJEkp/costs-and-benefits-of-amniocentesis-for-normal-pregnancies |
# Clarifying the confusion around inner alignment
*Note 1: This article was written for the* [*EA UC Berkeley Distillation Contest*](https://eaberkeley.com/aims-distillation)*, and is also my capstone project for the* [*AGISF course*](https://www.eacambridge.org/agi-safety-fundamentals)*.*
*Note 2: All claims here ab... | https://www.lesswrong.com/posts/xdtNd8xCdzpgfnGme/clarifying-the-confusion-around-inner-alignment |
# Inequality is inseparable from markets
*Cross posted to* [*The Good blog*](https://thegoodblog.substack.com/p/inequality-is-inseparable-from-markets?s=w)
I think prices are often really good at allocating resources efficiently. I also think that utility is roughly the log of income and we have a duty to make sure ... | https://www.lesswrong.com/posts/Aa6dNC89vgkv3icyh/inequality-is-inseparable-from-markets |
# [Link post] Promising Paths to Alignment - Connor Leahy | Talk
**Date of recording: **Tuesday May 10, 2022
**Description**
[AI Safety Support](https://www.aisafetysupport.org/home) recently hosted a closed, introductory talk by Connor Leahy on Promising Paths to Alignment.
The talk covers: (i) why Alignment is ... | https://www.lesswrong.com/posts/EN8QKscxC6gjh6TBd/link-post-promising-paths-to-alignment-connor-leahy-or-talk |
# Gato as the Dawn of Early AGI
*Written in a hurry today at the EA UCLA AI Timelines Workshop. Long and stream-of-thought, and a deliberate intellectual overreach as an epistemic exercise. My first foray into developing my own AGI timelines model* [*without deferring!*](https://www.lesswrong.com/posts/KGLJeeF6psTzHiu... | https://www.lesswrong.com/posts/TwfWTLhQZgy2oFwK3/gato-as-the-dawn-of-early-agi |
# Is AI Progress Impossible To Predict?
People seem to be continually surprised, over and over again, by the new capabilities of big machine learning models, such as [PaLM](https://ai.googleblog.com/2022/04/pathways-language-model-palm-scaling-to.html), [DALL-E](https://openai.com/dall-e-2/), [Chinchilla](https://www.... | https://www.lesswrong.com/posts/G993PFTwqqdQv4eTg/is-ai-progress-impossible-to-predict |
# The AI Countdown Clock
I made [this clock](https://aicountdown.com), counting down the time left until we build AGI:
, I’m, in part, writing this for myself, but I may as well publish it.
I want to think through the kind of jobs and career choices I should take in order to hedge against and benefit from increased automation. Especially the k... | https://www.lesswrong.com/posts/pvLjAifYhNcEHhJPe/surviving-automation-in-the-21st-century-part-1 |
# Definition Practice: Applied Rationality
In the exercise of defining what we mean by a term, we explore our assumptions about the nature of the term and what it’s pointing at. Sometimes this leads us to get a better handle on a concept. Other times, it shows us our confusion.
As an example of the value of definitio... | https://www.lesswrong.com/posts/G6BenMvcFR5CEC9uc/definition-practice-applied-rationality |
# Why I'm Optimistic About Near-Term AI Risk
I'm not worried about AI posing an existential risk in the next 10-20 years. Recent developments in AI capabilities actually make me feel *more* optimistic about this. The fact that relatively simple models can perform a wide array of tasks suggests that we can build [satis... | https://www.lesswrong.com/posts/8e3676AovRbGHLi27/why-i-m-optimistic-about-near-term-ai-risk |
# The Cuban missile crisis: the strategic context
*Cross posted to the EA forum and* [*The Good blog*](https://thegoodblog.substack.com/p/the-mystery-of-the-cuban-missile?s=w)
Why did the US blockade Cuba after it discovered that nuclear missiles launch sites were being constructed there? I think the story most peopl... | https://www.lesswrong.com/posts/bBXLC7CevbXRz9DHF/the-cuban-missile-crisis-the-strategic-context |
# BIDA Air Quality Measurements
I've been curious for a while how well the ventilation at the BIDA contra dance works. This is actually something that we can measure reasonably well by looking at CO2 levels: outside air has about 400 parts-per-million (ppm), while your exhalation is about 40,000ppm. By measuring the c... | https://www.lesswrong.com/posts/ryWtLRAqmTaLAmart/bida-air-quality-measurements |
# China Covid #3
China has once again [explicitly recommitted](https://twitter.com/medriva/status/1520021664678879234) to its goal of Zero Covid.
[Even the WHO has said this policy is unsustainable](https://twitter.com/medriva/status/1524088984871211009). China don’t care.
Or rather, [Xi doesn’t know and who is goin... | https://www.lesswrong.com/posts/tD3WXxp2BX3BWfXaP/china-covid-3 |
# To what extent is your AGI timeline bimodal or otherwise "bumpy"?
For example, you might think:
> It's likely that AGI will be invented before 2050; however, if it isn't, then that must mean either that AGI is impossible, or that it requires much more advanced technology than I currently think it does, or else that... | https://www.lesswrong.com/posts/6tp4YAXjWLM3HocW7/to-what-extent-is-your-agi-timeline-bimodal-or-otherwise |
# Optimization at a Distance
We have a computational graph (aka circuit aka causal model) representing an agent and its environment. We’ve chosen a cut through the graph to separate “agent” from “environment” - i.e. a Cartesian boundary. Arrows from environment to agent through the boundary are “observations”; arrows ... | https://www.lesswrong.com/posts/d2n74bwham8motxyX/optimization-at-a-distance |
# Proxy misspecification and the capabilities vs. value learning race
G Gordon Worley III [recently complained](https://www.alignmentforum.org/posts/sSpu2EABtTTDmBZ6T/g-gordon-worley-iii-s-shortform?commentId=tbj8R6K2q9rfhjtS7) about a lack of precision in discussions about whether Goodhart's Law will present a fatal ... | https://www.lesswrong.com/posts/tWpgtjRm9qwzxAZEi/proxy-misspecification-and-the-capabilities-vs-value |
# [Link] New Stanford Encyclopedia of Philosophy article, "Normative Theories of Rational Choice: Rivals to Expected Utility"
[https://plato.stanford.edu/entries/rationality-normative-nonutility/](https://plato.stanford.edu/entries/rationality-normative-nonutility/)
I thought this looked like it may be of interest to... | https://www.lesswrong.com/posts/4bfaCWr4JfEE7sKvs/link-new-stanford-encyclopedia-of-philosophy-article |
# DeepMind’s generalist AI, Gato: A non-technical explainer
Summary
=======
DeepMind’s recent paper, [*A Generalist Agent*](https://storage.googleapis.com/deepmind-media/A%20Generalist%20Agent/Generalist%20Agent.pdf), catalyzed a wave of discourse regarding the speed at which current artificial intelligence systems a... | https://www.lesswrong.com/posts/756HbyEBkL3xLSe7b/deepmind-s-generalist-ai-gato-a-non-technical-explainer |
# D&D.Sci Divination: Nine Black Doves Evaluation & Ruleset
This is a follow-up to [last week's D&D.Sci scenario](https://www.lesswrong.com/posts/cesr2yGoZ2y327qKQ/d-and-d-sci-divination-nine-black-doves): if you intend to play that, and haven't done so yet, you should do so now before spoiling yourself.
There is a w... | https://www.lesswrong.com/posts/HogEfxnyfLQ3CXJvq/d-and-d-sci-divination-nine-black-doves-evaluation-and |
# The "Adults in the Room"
*Significant spoilers for* [*mad investor chaos and the woman of asmodeus (planecrash Book 1).*](https://www.glowfic.com/posts/4582)
Dath ilan has put significant effort into protecting [what it has to protect.](https://www.lesswrong.com/posts/SGR4GxFK7KmW7ckCB/something-to-protect) To that... | https://www.lesswrong.com/posts/JXnAFGMXDm2huGRZt/the-adults-in-the-room |
# [Intro to brain-like-AGI safety] 15. Conclusion: Open problems, how to help, AMA
*(Last revised: January 2026. See changelog at the bottom.)*
15.1 Post summary / Table of contents
=====================================
This is the final post of the [*“Intro to brain-like-AGI safety”* post series](https://www.alignm... | https://www.lesswrong.com/posts/tj8AC3vhTnBywdZoA/intro-to-brain-like-agi-safety-15-conclusion-open-problems-1 |
# Actionable-guidance and roadmap recommendations for the NIST AI Risk Management Framework
*Updated 13 September 2022 with a link to our arXiv paper and corrections to out-of-date items*
This is a linkpost to our working paper “Towards AI Standards Addressing AI Catastrophic Risks: Actionable-Guidance and Roadmap Re... | https://www.lesswrong.com/posts/JNqXyEuKM4wbFZzpL/actionable-guidance-and-roadmap-recommendations-for-the-nist-1 |
# Formula for Dying Babies
Note: This was posted originally on Thursday, May 12 as part of the weekly Covid post. It is being separated out into its own post for future reference, and in case some people are no longer reading Covid posts.
[There’s a shortage of specialty infant formula](https://www.politico.com/news/... | https://www.lesswrong.com/posts/HevYry59aj6AjiPmy/formula-for-dying-babies |
# Proposal: Twitter dislike button
The popular story of Twitter’s role in the ruin of civilization is that it is a runaway trash fire of reciprocal anger and offense, where otherwise nice people are possessed by overwhelming outrages, and drawn into throwing their own energy behind creating the vilest and most vindict... | https://www.lesswrong.com/posts/kzMDzDffC2xqKTtCJ/proposal-twitter-dislike-button |
# On saving one's world
If the world is likeliest to be saved by sober scholarship, then let us be sober scholars in the face of danger.
If the world is likeliest to be saved by playful intellectual exploration, then let us be playful in the face of danger.
Strategic, certainly; aware of our situation, of course; bu... | https://www.lesswrong.com/posts/aaeMECbcBzC3cweGK/on-saving-one-s-world |
# Popular education in Sweden: much more than you wanted to know
Growing up on the Swedish seaside, I had a five-minute walk to four open learning facilities – not counting the library and the youth center. It was [very Christopher Alexander](https://escapingflatland.substack.com/p/christopher-alexanders-architecture?... | https://www.lesswrong.com/posts/tjxgbovwc5Ft7wrtc/popular-education-in-sweden-much-more-than-you-wanted-to-1 |
# Workday Air Quality Measurements
I continue to be very excited about my new [M2000 meter](https://temtopus.com/collections/temtop-co2-monitor/products/temtop-m2000-2nd-generation-air-quality-monitor-for-pm2-5-pm10-particles-co2-hcho-temperature-humidity-settable-audio-alarm-data-export-recording-curve-easy-calibrati... | https://www.lesswrong.com/posts/zAz8wRdnj7ahEsEqt/workday-air-quality-measurements |
# I just watched the Open C3 Subcommittee Hearing on Unidentified Aerial Phenomena (UFOs). Here's a succinct summary and commentary + some background
Background on UFOs, Disclosure, and Rationality
-----------------------------------------------
There have been reports of various strange flying objects for [a very ve... | https://www.lesswrong.com/posts/Wo2jr2zuDCzgSmMRH/i-just-watched-the-open-c3-subcommittee-hearing-on |
# Gato's Generalisation: Predictions and Experiments I'd Like to See
_I'm deliberately inhabiting a devil's advocate mindset because that perspective seems to be missing from the conversations I've witnessed. My actual fully-reflective median takeaway might differ._
_My covid has made writing difficult at the moment,... | https://www.lesswrong.com/posts/LnnMPNHEpqtaqonCM/gato-s-generalisation-predictions-and-experiments-i-d-like |
# A bridge to Dath Ilan? Improved governance on the critical path to AI alignment.
Summary: This post showcases [my 2nd-place-winning entry in the Future of Life Institute's AI worldbuilding contest](https://worldbuild.ai/peace-through-prophecy/). It imagines:
1. How we might make big improvements to decisionmakin... | https://www.lesswrong.com/posts/qo2hqf2ha7rfgCdjY/a-bridge-to-dath-ilan-improved-governance-on-the-critical |
# How to get into AI safety research
Recently, I had a conversation with someone from a math background, asking how they could get into AI safety research. Based on my own path from mathematics to AI alignment, I recommended the following sources. It may prove useful to others contemplating a similar change in career:... | https://www.lesswrong.com/posts/WFnyxSL543c9bxGMm/how-to-get-into-ai-safety-research |
# Framing Practicum: Dynamic Programming
This is a [framing practicum](https://www.lesswrong.com/s/Fu7Euu3F96rKhFRWH) post. We'll talk about what dynamic programming (DP) is, how to recognize DP in the wild, and what questions to ask when you find it. Then, we'll have a challenge to apply the idea.
**Today's challeng... | https://www.lesswrong.com/posts/ZzDjYsjyfpw7W3H8Q/framing-practicum-dynamic-programming |
# Maxent and Abstractions: Current Best Arguments
*This post is not-very-distilled and doesn’t contain much background; it’s intended for people who already have the context of *[*at least*](https://www.lesswrong.com/posts/dNzhdiFE398KcGDc9/testing-the-natural-abstraction-hypothesis-project-update)[*these*](https://ww... | https://www.lesswrong.com/posts/cqdDGuTs2NamtEhBW/maxent-and-abstractions-current-best-arguments |
# Reading the ethicists: A review of articles on AI in the journal Science and Engineering Ethics
*Epistemic status: Stream of consciousness reactions to papers read in chronological order. Caveat lector.*
I have a dirty family secret. My uncle is a professional ethicist.
In a not-too roundabout way, this is why I e... | https://www.lesswrong.com/posts/D2biiCrF7E62kDbGA/reading-the-ethicists-a-review-of-articles-on-ai-in-the |
# We have achieved Noob Gains in AI
TL;DR I explain why I think AI research has been slowing down, not speeding up, in the past few years.
How have your expectations for the future of AI research changed in the past three years?
Based on recent posts in this forum, it seems that results in text generation, protein f... | https://www.lesswrong.com/posts/QBgFnBMJpmGkE5ioc/we-have-achieved-noob-gains-in-ai |
# What's up with the recent monkeypox cases?
Is it a thing I should be allocating attention to?
Wikipedia tells me that human-to-human transmission of monkeypox is pretty rare/difficult. There has been some community spread recently at least in the UK, but that's been speculated to have been from sex. On the other ha... | https://www.lesswrong.com/posts/kPQAAb7ayZpKYPzME/what-s-up-with-the-recent-monkeypox-cases |
# Covid 5/19/22: The Law of Five
If you’re five, you now get a booster. If you’re four, you still get no vaccine at all (while we ‘wait for more data’) and you’re often forced to take precautions that _might_ make sense for senior citizens.
Because no one cares. And that truly continues to be the overall message. No ... | https://www.lesswrong.com/posts/LyTPoBCCh7zhuGCwT/covid-5-19-22-the-law-of-five |
# 14 Techniques to Accelerate Your Learning
Originally written by Belén Cobeta and [Spencer Greenberg](https://www.spencergreenberg.com/) from [ClearerThinking.org](https://www.clearerthinking.org/). Revised by Teis Rasmussen & Florence Hinder from [ThoughtSaver.com](http://thoughtsaver.com/). This is a cross-post fro... | https://www.lesswrong.com/posts/F3vNoqA7xN4TFQJQg/14-techniques-to-accelerate-your-learning-1 |
# How has the total amount of gain-of-function research worldwide grown/shrunk over time?
Given the advance in underlying biological tools/technology, I would not find it surprising if gain-of-function research had increased 10x over the last decade. (This is just based on priors; getting an actual estimate is the poi... | https://www.lesswrong.com/posts/nZBmJpWCCLc9oiRaA/how-has-the-total-amount-of-gain-of-function-research |
# Formula for a Shortage
Previous coverage (not required in any way): [Formula For Dying Babies](https://thezvi.substack.com/p/formula-for-dying-babies?s=w)
Great Formal Workups of Much of This (Scott Lincicome): [America’s Infant Formula Crisis and the ‘Resiliency’ Mirage](https://capitolism.thedispatch.com/p/americ... | https://www.lesswrong.com/posts/fHedcmjhrwpr62zh8/formula-for-a-shortage |
# Podcast: Spencer Greenberg talks to me about dealing with our groupstruckness and boundedness
I [talked to Spencer Greenberg](https://clearerthinkingpodcast.com/episode/105) a little while ago for his podcast [Clearer Thinking](https://clearerthinkingpodcast.com/). The recording is out today. According to his websit... | https://www.lesswrong.com/posts/3ofqwju3vQyLJRqvJ/podcast-spencer-greenberg-talks-to-me-about-dealing-with-our |
# How RL Agents Behave When Their Actions Are Modified? [Distillation post]
Summary
=======
This is a distillation post intended to summarize the article [How RL Agents Behave When Their Actions Are Modified?](https://www.aaai.org/AAAI21Papers/AAAI-5767.LangloisE.pdf) by Eric Langlois and Tom Everitt, published at AA... | https://www.lesswrong.com/posts/FeY4tXMYdTQSM4go3/how-rl-agents-behave-when-their-actions-are-modified |
# Beware boasting about non-existent forecasting track records
Imagine if there was a financial pundit who kept saying "Something really bad is brewing in the markets and we may be headed for a recession. But we can't know when recessions will come, nobody can predict them". And then every time there was a selloff i... | https://www.lesswrong.com/posts/ZEgQGAjQm5rTAnGuM/beware-boasting-about-non-existent-forecasting-track-records |
# Why does gradient descent always work on neural networks?
My amateur understanding of neural networks is that they almost always train using stochastic gradient descent. The quality of a neural network comes from its size, shape, and training data, but not from the training function, which is always simple gradient ... | https://www.lesswrong.com/posts/iF8GLyCJrAkCKtbGu/why-does-gradient-descent-always-work-on-neural-networks |
# Who is available for contract work? (A la the Hacker News "Who Wants To Be Hired" threads)
The intention of this thread is similar to the "Hacker News Freelancer/Seeking Freelancer" threads.
Please lead with either AVAILABLE FOR PROJECTS or SEEKING FREELANCER, your location, and whether remote work is a possibility... | https://www.lesswrong.com/posts/YyEiPQH8snk4h8BfK/who-is-available-for-contract-work-a-la-the-hacker-news-who |
# Clarifying what ELK is trying to achieve
(Basic familiarity with [ELK](https://www.lesswrong.com/posts/qHCDysDnvhteW7kRd/arc-s-first-technical-report-eliciting-latent-knowledge) required.)
I noticed that some people who haven’t thought that much about ELK yet (including my former self) are confused about (1) why th... | https://www.lesswrong.com/posts/8xCtJHAbzyA2oA6J4/clarifying-what-elk-is-trying-to-achieve |
# The Future of Nuclear War
Here I present a view on the future of nuclear war which takes into account the expected technological progress as well as global political changes.
There are three main directions in which technological progress in nuclear weapons may happen:
1) Many gigaton weapons.
2) Cheaper nuclea... | https://www.lesswrong.com/posts/G6JMpyMeMyAsNnkTj/the-future-of-nuclear-war |
# Information Loss --> Basin flatness
*This work was done under the mentorship of Evan Hubinger through the SERI MATS program. Thanks to Lucius Bushnaq, John Wentworth, Quintin Pope, and Peter Barnett for useful feedback and suggestions.*
In this theory, the main proximate cause of flat basins is a type of **informat... | https://www.lesswrong.com/posts/wPudaEemohdYPmsye/information-loss-greater-than-basin-flatness |
# [Short version] Information Loss --> Basin flatness
This is an overview for advanced readers. Main post: [Information Loss --> Basin flatness](https://www.lesswrong.com/posts/wPudaEemohdYPmsye/information-loss-greater-than-basin-flatness)
**Summary:**
Inductive bias is related to, among other things:
* Basin f... | https://www.lesswrong.com/posts/SZM32BdvYgrsBfYnw/short-version-information-loss-greater-than-basin-flatness |
# How Big Are Covid Particles?
To assess what sort of filtration you might want for reducing covid risk, it would be useful to know how the different sizes of aerosol particles contribute to covid spread. Unfortunately, we don't know that much here.
We know that sars-cov-2 itself is ~0.1µm, but it's probably not tran... | https://www.lesswrong.com/posts/nkynWFFoANzCg7QPk/how-big-are-covid-particles |
# On The Spectrum, On The Guest List
*This is the prologue from a lengthy piece I'll be serializing over the next few months. Yes, the title is a little is a little clickbaity, and superficially, a collection of stories about going clubbing in New York City isn't quite the standard rationalist fare. *
*But at its cor... | https://www.lesswrong.com/posts/b7gWdMWsSC9L8o5fj/on-the-spectrum-on-the-guest-list |
# Are human imitators superhuman models with explicit constraints on capabilities?
*epistemic status: A speculative hypothesis, don't know if this already exists. The only real evidence I have for this is a vague analogy based on some other speculative (though less speculative) hypothesis. I don’t think this is partic... | https://www.lesswrong.com/posts/zEwCghegzMESj7sSz/are-human-imitators-superhuman-models-with-explicit |
# What kinds of algorithms do multi-human imitators learn?
*epistemic status: Speculation. The actual proposals are idealized, not meant to be exactly right. We have thought about this for less than an hour.*
* * *
In [this earlier post](https://www.lesswrong.com/posts/zEwCghegzMESj7sSz/are-human-imitators-superhuma... | https://www.lesswrong.com/posts/PthrpCHpFKMrQaYMe/what-kinds-of-algorithms-do-multi-human-imitators-learn |
# Adversarial attacks and optimal control
*Meta: After a fun little motivating section, this post goes deep into the mathematical weeds. This thing aims to explore the mathematical properties of adversarial attacks from first principles. Perhaps other people are not as confused about this point as I was, but hopefully... | https://www.lesswrong.com/posts/KtCJNw93KHg7MSSvw/adversarial-attacks-and-optimal-control |
# How curing aging could help progress
How would society change if we cured aging, and people could have as many healthy years of life as they wanted?
A common concern is that this would ossify our institutions. The old guard would never die off, and so would never be replaced by young bloods. This could threaten pro... | https://www.lesswrong.com/posts/mjupigsywnp8QEBm6/how-curing-aging-could-help-progress |
# Gradations of Agency
\[Epistemic status: This post written hastily, for Blog Post Day.\]
*I’ve de-prioritised* [*this sequence*](https://www.lesswrong.com/s/aek5ksSs2FHTeofsf) *due to other* [*recent*](https://www.lesswrong.com/posts/kpPnReyBC54KESiSn/optimality-is-the-tiger-and-agents-are-its-teeth) [*posts*](htt... | https://www.lesswrong.com/posts/ndcqdTxkMnFF7gQFh/gradations-of-agency-1 |
# PSA: The Sequences don't need to be read in sequence
This week, I hung out with the LessWrong team while they talked to relatively new users. New users often had a vague intention to read Eliezer's original Sequences, but were blocked on the size of the project. They thought the Sequences would only work in, well, s... | https://www.lesswrong.com/posts/PouWW8Ys4iKhWedCZ/psa-the-sequences-don-t-need-to-be-read-in-sequence |
# AXRP Episode 15 - Natural Abstractions with John Wentworth
[YouTube link](https://www.youtube.com/watch?v=E7JH77LMuV0&list=PLmjaTS1-AiDeqUuaJjasfrM6fjSszm9pK&index=15)
Why does anybody care about natural abstractions? Do they somehow relate to math, or value learning? How do E. coli bacteria find sources of sugar? ... | https://www.lesswrong.com/posts/L896Fp8hLSbh8Ryei/axrp-episode-15-natural-abstractions-with-john-wentworth |
# Monkeypox: explaining the jump to Europe
[What's up with the recent monkeypox cases?](https://www.lesswrong.com/posts/kPQAAb7ayZpKYPzME/what-s-up-with-the-recent-monkeypox-cases) Why are we seeing a monkeypox outbreak now just as COVID-19 vigilance is winding down? Why didn't we see more monkeypox outbreaks 10 or 20... | https://www.lesswrong.com/posts/SYWZ8qrJESRdnnxAu/monkeypox-explaining-the-jump-to-europe |
# Finding Your Voice
*Most people hate to hear recordings of their own speech. "Oh God, do I really sound like that?" Many people think that their voice is "bad", which should be surprising because almost nobody has an objectively bad voice. But many people use their voices unskillfully, which can be improved.* | https://www.lesswrong.com/posts/kfT73jTwAkpxdQttt/finding-your-voice |
# [fiction]A Question of Perspective
*Disclaimer: This text was written to reflect a state of mind with no other audience than myself in mind. Only later did I get the impulse to publish this. I would therefore like to emphasize, that this is **not** intended as a guide, a recommendation for behaviour, or anything lik... | https://www.lesswrong.com/posts/5Awp4m8HZgdTiuiMo/fiction-a-question-of-perspective |
# [Alignment] Is there a census on who's working on what?
I'm new to the Rationality / EA community, and I've been getting the sense that the best use of my skills and time is to try to contribute to alignment.
Currently, my focus has been guided by top posts / most often mentions of names, here on LW. E.g. "I see Jo... | https://www.lesswrong.com/posts/h7xEuH3u9tbozZ9Tq/alignment-is-there-a-census-on-who-s-working-on-what |
# Bits of Optimization Can Only Be Lost Over A Distance
When we think of “optimization” as [compressing some part of the universe into a relatively small number of possible states](https://www.lesswrong.com/posts/znfkdCoHMANwqc2WE/the-ground-of-optimization-1), [it’s very natural to quantify that compression in terms ... | https://www.lesswrong.com/posts/tpHB69eXorChEsix3/bits-of-optimization-can-only-be-lost-over-a-distance |
# Why I'm Worried About AI
Preamble
========
In this sequence of posts, I want to lay out why I am worried about risks from powerful AI and where I think the specific dangers come from. In general, I think it's good for people to be able to form their own inside views of what’s going on, and not just defer to people.... | https://www.lesswrong.com/posts/k8hvGAJWSKAeHwpnJ/why-i-m-worried-about-ai |
# Complex Systems for AI Safety [Pragmatic AI Safety #3]
*This is the third post in* [*a sequence of posts*](https://www.alignmentforum.org/posts/bffA9WC9nEJhtagQi/introduction-to-pragmatic-ai-safety-pragmatic-ai-safety-1) *that describe our models for Pragmatic AI Safety.*
It is critical to steer the AI research fie... | https://www.lesswrong.com/posts/n767Q8HqbrteaPA25/complex-systems-for-ai-safety-pragmatic-ai-safety-3 |
# Synthetic Media and The Future of Film
It was long believed that the first jobs to be obsoleted by AI would be lawyers and accountants, as those seemed the prime targets. After all, creativity has hardly been the forte of computers for the past half-century, being almost exclusively the product of human effort. Howe... | https://www.lesswrong.com/posts/hKS4NdcZqDnosbKhY/synthetic-media-and-the-future-of-film |
# The No Free Lunch theorems and their Razor
The No Free Lunch (NFL) family of theorems contains some of the most
misunderstood theorems of machine learning. They apply to learning[^1] and
optimization[^16] and, in rough terms, they state:
[^1]: Wolpert, D. H. (1996). The lack of a priori distinctions between learnin... | https://www.lesswrong.com/posts/sdrCBWpqyvNBJSZH5/the-no-free-lunch-theorems-and-their-razor |
# Why I Define My Experience At the Monastic Academy As Sexual Assault
***Trigger Warning: This post includes highly triggering details of sexual assault.***
***\*\*\*The views expressed here are my own. Hereafter any italics used will indicate that I am quoting another person or source. \*\*\****
In December 2021 I... | https://www.lesswrong.com/posts/zbQapwn2hGNozcGsD/why-i-define-my-experience-at-the-monastic-academy-as-sexual |
# I only believe in the paranormal
(Cross-posted from [Telescopic Turnip](https://malmesbury.substack.com/p/i-only-believe-in-the-paranormal?sd=fs))
[*Recommended soundtrack for this post*](https://www.youtube.com/watch?v=B3lq6U4PDZo)
The Wiktionary defines [paranormal](https://en.wiktionary.org/wiki/paranormal) as ... | https://www.lesswrong.com/posts/MizKtJcjktiQZez5j/i-only-believe-in-the-paranormal |
# Mandatory Post About Monkeypox
It’s 2022, so _of course_ enough people are worried about an epidemic of Monkeypox that it feels like a necessary public service to do a post about Monkeypox. I didn’t get to it especially fast exactly because I am not all that concerned, but I’m not entirely _not_ concerned, so here w... | https://www.lesswrong.com/posts/hkAuRaKDXiCAEAJj7/mandatory-post-about-monkeypox |
# Google's Imagen uses larger text encoder
https://imagen.research.google/
Scaling the text encoder gives Imagen the ability to spell, count, and assign colors and properties to distinct objects in the image that DALL-E2 was not so great at. It looks visually about as photorealistic as DALL-E2 from the small set of ... | https://www.lesswrong.com/posts/BXa8TDCBuse4fMb9j/google-s-imagen-uses-larger-text-encoder |
# Explaining inner alignment to myself
*Explaining the inner alignment problem to myself in an attempt to gain a gears-level understanding. This post is mostly a distillation of *[*Risks from Learned Optimization*](https://arxiv.org/pdf/1906.01820.pdf)*, but focusing more on the statistical learning theory perspective... | https://www.lesswrong.com/posts/HtEffpHcLxppLN6dL/explaining-inner-alignment-to-myself |
# autonomy: the missing AGI ingredient?
*Epistemic status: trying to feel out the shape of a concept and give it an appropriate name. Trying to make explicit some things that I think exist implicitly in many people's minds. This post makes truth claims, but its main goal is to not to convince you that they are true.*... | https://www.lesswrong.com/posts/HSETWwdJnb45jsvT8/autonomy-the-missing-agi-ingredient |
# Visible Homelessness in SF: A Quick Breakdown of Causes
Anyone who's visited downtown San Francisco has seen streets like this one:
The US is the [richest country on Earth](https://en.wikipedia.org/w... | https://www.lesswrong.com/posts/Qz9GvoPbnFwGrHHQB/visible-homelessness-in-sf-a-quick-breakdown-of-causes |
# MERV Filters for Covid?
If you look around for advice on what kind of air purifier to use to reduce covid risk, you'll see some people saying you need HEPA filters:
[Microcovid:](https://www.microcovid.org/blog/hepafilters)
> If you decide to buy an air purifier for COVID purposes, here's some things to keep in mi... | https://www.lesswrong.com/posts/ZpNeksJ9nDN5ZAgTh/merv-filters-for-covid |
# A Quick Note on AI Scaling Asymptotes
Two months ago, DeepMind announced their new [Chinchilla](https://www.deepmind.com/publications/an-empirical-analysis-of-compute-optimal-large-language-model-training) paper. The headline result was that you could get much better performance with the same computational budget, b... | https://www.lesswrong.com/posts/Ea9d9m8eNFWGv6jPq/a-quick-note-on-ai-scaling-asymptotes |
# RL with KL penalties is better seen as Bayesian inference
*This blog post is largely based on an [EMNLP paper](https://arxiv.org/abs/2205.11275) with Ethan Perez and Chris Buckley. It also benefited from discussions with and comments from Hady Elsahar, Germán Kruszewski, Marc Dymetman and Jérémy Scheurer.*
**TLDR*... | https://www.lesswrong.com/posts/eoHbneGvqDu25Hasc/rl-with-kl-penalties-is-better-seen-as-bayesian-inference |
# What's The Best Place to Look When You Have A Question About x?
* **Stackoverflow** for programming
* Perhaps **UpToDate** for medical stuff (but I'm not a doctor so I'm not sure)
* **Stanford Encyclopedia of Philosophy** for accessible but rigorous summaries of the philosophical literature on a given topic (t... | https://www.lesswrong.com/posts/4xR7Dhrd46tw3wy3c/what-s-the-best-place-to-look-when-you-have-a-question-about |
# The "Measuring Stick of Utility" Problem
Let’s start with the simplest coherence theorem: suppose I’ll pay to upgrade pepperoni pizza to mushroom, pay to upgrade mushroom to anchovy, and pay to upgrade anchovy to pepperoni. This does not bode well for my bank account balance. And the only way to *avoid* having such ... | https://www.lesswrong.com/posts/73pTioGZKNcfQmvGF/the-measuring-stick-of-utility-problem |
# Reflections on a year of college
*Crossposted from* [*my blog.*](https://unoptimal.com/2022/05/12/college/) *Considering collaborating on a version of this that's tailored specifically towards undergrads in the LW or EA communities.*
* * *
I just completed my first year of college (US university). I’ve written up ... | https://www.lesswrong.com/posts/NbBdGhJaeEWPcCRxJ/reflections-on-a-year-of-college |
# Science-informed normativity
The debate over moral realism is often framed in terms of a binary question: are there ever objective facts about what’s moral to do in a given situation? The broader question of normative realism is also framed in a similar way: are there ever objective facts about what’s rational to do... | https://www.lesswrong.com/posts/LeXhzj7msWLfgDefo/science-informed-normativity |
# Request for small textbook recommendations
I'd love recommendations for good textbooks that are short.
I like books to be physically smaller so that they're easier to hold and carry around, and I'd also like to be able to at least plausibly believe that I might read the whole thing. I was a math and physics major, ... | https://www.lesswrong.com/posts/Pd3X6rjPjvMGXMBLz/request-for-small-textbook-recommendations |
# How to get people to produce more great exposition? Some strategies and their assumptions
Some of the recent posts in the [distillation & pedagogy tag](https://www.lesswrong.com/tag/distillation-and-pedagogy?sortedBy=new) here on LessWrong, such as "[Call For Distillers](https://www.lesswrong.com/posts/zo9zKcz47JxDE... | https://www.lesswrong.com/posts/BddavsuaHLAhkH45Z/how-to-get-people-to-produce-more-great-exposition-some |
# Glass Puppet
The worst part of being an actor is that all work is temporary. That's what Alia told herself. In actuality, Alia's "acting" career amounted to two spots as an extra in commercials for enterprise software.
To make ends meet, Alia worked as a social engineer. That's what Alia told herself. In actuality,... | https://www.lesswrong.com/posts/jKxuB9pZXRiom7ecK/glass-puppet |
# Testing Air Purifiers
If you're considering buying an air purifier for reducing covid risk, the [EPA recommends](https://www.epa.gov/coronavirus/air-cleaners-hvac-filters-and-coronavirus-covid-19) comparing them based on their Clean Air Delivery Rate (CADR) for smoke. But what if you want to evaluate a purifier your... | https://www.lesswrong.com/posts/vBzeNet8aEiL4mahe/testing-air-purifiers |
# View and bet in Manifold prediction markets on Lesswrong
*(Crossposted from* [*EA Forum*](https://forum.effectivealtruism.org/posts/5HCQvb7AiHvSWdfr3/view-and-bet-in-manifold-prediction-markets-on-ea-forum)*)*
Links to [prediction](https://manifold.markets/tigrennatenn/will-the-mad-investor-chaos-thread) [marke... | https://www.lesswrong.com/posts/wFPa2yjZPbNear6WE/view-and-bet-in-manifold-prediction-markets-on-lesswrong |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.