text stringlengths 300 320k | source stringlengths 52 154 |
|---|---|
# On oxytocin-sensitive neurons in auditory cortex
*(For the big picture of how I wound up on this topic, see* [*Symbol Grounding and Human Social Instincts*](https://www.lesswrong.com/posts/5F5Tz3u6kJbTNMqsb/intro-to-brain-like-agi-safety-13-symbol-grounding-and-human)*. But I wound up feeling like oxytocin-sensitive... | https://www.lesswrong.com/posts/HA9qiJ8zReBarGAGT/on-oxytocin-sensitive-neurons-in-auditory-cortex |
# No, human brains are not (much) more efficient than computers
Epistemic status: grain of salt (but you can play around with the model yourself). There's lots of uncertainty in how many FLOP/s the brain can perform.
In informal debate, I've regularly heard people say something like, "oh but brains are so much more e... | https://www.lesswrong.com/posts/KsKfvLx7nFBZnWtEu/no-human-brains-are-not-much-more-efficient-than-computers |
# How can we secure more research positions at our universities for x-risk researchers?
Has anyone had experience with this? Is anyone working on this?
Reading [this](https://www.alignmentforum.org/posts/4jFnquoHuoaTqdphu/ai-x-risk-reduction-why-i-chose-academia-over-industry) and [this](https://www.alignmentforu... | https://www.lesswrong.com/posts/PsdnTrwvHp95Nu2B7/how-can-we-secure-more-research-positions-at-our |
# Rejected Early Drafts of Newcomb's Problem
*Discovered inside an opaque box at the University of California and shared by an anonymous source, please enjoy these unpublished variations of* [*physicist William Newcomb's famous thought experiment*](https://www.lesswrong.com/tag/newcomb-s-problem)*.*
Newcomb's Advance... | https://www.lesswrong.com/posts/rxnpuzEw4boJrs3CY/rejected-early-drafts-of-newcomb-s-problem |
# Guitar Pedals on Fiddle
When [Cecilia](https://www.ceciliavacanti.com/) was starting to think about using effects with [Kingfisher](https://www.kingfisherband.com/) I posted [asking for advice](https://www.jefftk.com/p/fiddle-effects-tech) on the right way to hook this up. This is a protocol problem:
* Electric g... | https://www.lesswrong.com/posts/Rsks5tMt4okBXd5gw/guitar-pedals-on-fiddle |
# Framing AI Childhoods
*Generated as part of SERI MATS, under John Wentworth. Thanks to Alex Turner and Garrett Baker for related discussion, and to Justis Mills for draft feedback.*
*All bold claims and ignorant mistakes herein are my own.*
Do you remember when Evan Hubinger became really enamored with ['training ... | https://www.lesswrong.com/posts/saGr6DapTPKFaMhhP/framing-ai-childhoods |
# Progress Report 7: making GPT go hurrdurr instead of brrrrrrr
[First progress post](https://www.lesswrong.com/posts/pvbWz5mEFCS6DFaWD/progress-report-1-interpretability-experiments-and-learning)
Introduction
============
Starting off this project, I was using ROME modification to alter individual weights and get c... | https://www.lesswrong.com/posts/5CGvoviP7t7ftsohh/progress-report-7-making-gpt-go-hurrdurr-instead-of-brrrrrrr |
# First we shape our social graph; then it shapes us
The inside of a womb looks as it did 70,000 years ago, but the world outside has changed. In July 2021, when our daughter was born, the night sky didn’t light up with stars; it was lit up by the warm afterglow of sodium street lamps. Green-clad women carried the bab... | https://www.lesswrong.com/posts/tpewFnmFKnKk5xijh/first-we-shape-our-social-graph-then-it-shapes-us |
# Is training data going to be diluted by AI-generated content?
This is my first post on here so please be lenient if I fail to follow any norms.
An image made by Midjourney AI. The prompt was "the phy... | https://www.lesswrong.com/posts/oXnZovoFSWMgRe88W/is-training-data-going-to-be-diluted-by-ai-generated-content |
# Generators Of Disagreement With AI Alignment
I often find myself disagreeing with most of the things I read about AI alignment. The closest I probably get to accepting a Berkely-rationalism or Bostrom-inspired take on AI is something like Nintil’s [essay on the subject](https://nintil.com/ai-safety). But even that, ... | https://www.lesswrong.com/posts/SmRLWEYDSjEtB9NiP/generators-of-disagreement-with-ai-alignment |
# Galaxy Trucker Needs a New Second Half
I like [Galaxy Trucker](https://boardgamegeek.com/boardgame/31481/galaxy-trucker) quite a bit. It has two very different halves:
1. Everyone grabs tiles and builds a spaceship. Not turn-based, just taking tiles and placing them as fast as you can.
2. We fly our ships th... | https://www.lesswrong.com/posts/GMcjHdwt3oHBZdPNn/galaxy-trucker-needs-a-new-second-half |
# How to Do Research. v1
This post is part of my ["research"](https://pabloernesto.github.io/series/research.html) series.
Introduction
============
Sometimes the need is desperate. You have children to raise, cancer to treat, something important broke and you're going down with it. These kind of situations rarely a... | https://www.lesswrong.com/posts/AFCx9LLGnJFdtaqik/how-to-do-research-v1 |
# 90% of anything should be bad (& the precision-recall tradeoff)
Introduction
============
[Sturgeon’s Law states that 90% of anything is crap.](https://en.wikipedia.org/wiki/Sturgeon%27s_law)
So 90% of sci-fi books are crap,
and 90% of medical research,
and 90% of jazz solos,
and 90% of LessWrong... | https://www.lesswrong.com/posts/nu6BgFnkdQAi7Fzhn/90-of-anything-should-be-bad-and-the-precision-recall |
# Searching for Modularity in Large Language Models
*Produced As Part Of The SERI ML Alignment Theory Scholars Program 2022 Under* [*John Wentworth*](https://www.lesswrong.com/users/johnswentworth)
*See the *[*Google Colab notebook*](https://colab.research.google.com/drive/1yH_foOW7ONKXj5evGiDC6WmuV7qz5SMT?usp=sharin... | https://www.lesswrong.com/posts/rp4CiJtttvwFNHkhL/searching-for-modularity-in-large-language-models |
# Linkpost: Github Copilot productivity experiment
> We recruited 95 professional developers, split them randomly into two groups, and timed how long it took them to write an HTTP server in JavaScript. One group used GitHub Copilot to complete the task, and the other one didn’t. We tried to control as many factors as ... | https://www.lesswrong.com/posts/f2C4CWNmrSKMs6SaK/linkpost-github-copilot-productivity-experiment |
# All AGI safety questions welcome (especially basic ones) [Sept 2022]
**tl;dr: Ask questions about AGI Safety as comments on this post, including ones you might otherwise worry seem dumb!**
Asking beginner-level questions can be intimidating, but everyone starts out not knowing anything. If we want more people in th... | https://www.lesswrong.com/posts/XpeYpKXHvbqhefQi5/all-agi-safety-questions-welcome-especially-basic-ones-sept |
# Solar Blackout Resistance
Many houses have solar panels, but in a power outage almost all of them shut down. Especially with serious blackouts like [Texas had](https://en.wikipedia.org/wiki/2021_Texas_power_crisis) or like Europe may have [this winter](https://en.wikipedia.org/wiki/2022_Russia%E2%80%93European_Union... | https://www.lesswrong.com/posts/fp3hAkLgnbsnzm8d4/solar-blackout-resistance |
# Covid 9/8/22: Booster Boosting
Does it make sense to get the new updated booster? I give my (Not Medical Advice!) definitive take on how I would personally make the decision, and report the opinions of a number of other sources. Unless something changes, this is my take on that subject.
There’s also a variety of sm... | https://www.lesswrong.com/posts/PXyBqrHfx3hNvApdn/covid-9-8-22-booster-boosting |
# What Should AI Owe To Us?
Accountable and Aligned AI Systems via Contractualist AI Alignment
*This is an extended and edited transcript of the talk I recently gave at* [*EAGxSingapore 2022*](https://www.eaglobal.org/events/eagxsingapore-2022/)*. *
*The title has been changed for easier searchability of "Contractua... | https://www.lesswrong.com/posts/Cty2rSMut483QgBQ2/what-should-ai-owe-to-us-accountable-and-aligned-ai-systems |
# Thoughts on AGI consciousness / sentience
Summary
=======
I’ll walk through my take on how consciousness works (basically endorsing “illusionism”) and why I think computer programs can be conscious, and how that relates to moral and ethical obligations towards conscious entities in general and AIs in particular.
U... | https://www.lesswrong.com/posts/32ca3B7rJ93xo9tvb/thoughts-on-agi-consciousness-sentience |
# Postmortem: Trying out for Manifold Markets
Intro
-----
I tried out for [Manifold Markets](https://manifold.markets) in May and June 2022. It didn't work out, welcome to the postmortem.
[Here's](https://www.lesswrong.com/posts/qcYu6fMb6qg4zcftn/create-a-prediction-market-in-two-minutes-on-manifold) an article by t... | https://www.lesswrong.com/posts/EkihsSGujhFeBoz4o/postmortem-trying-out-for-manifold-markets |
# Progress links & tweets, 2022-09-08
**Opportunities**
-----------------
* [Nat Friedman and Daniel Gross are investing $10M in AI-first products, apply here](https://aigrant.org/) (via [@natfriedman](https://twitter.com/natfriedman/status/1564655413785530368))
* [Congressman Jake Auschincloss is hiring a Chief ... | https://www.lesswrong.com/posts/xjJurJZWGT5KwLaC6/progress-links-and-tweets-2022-09-08 |
# [An email with a bunch of links I sent an experienced ML researcher interested in learning about Alignment / x-safety.]
\[\[FYI: I'm just copying this in and removing a few bits; apologies for formatting; I don't intend to post the attachments\]\]
EtA: I (embarrasingly and unfortunately) understated Richard Ngo... | https://www.lesswrong.com/posts/b2Jk3dAmerjyNDzWf/an-email-with-a-bunch-of-links-i-sent-an-experienced-ml |
# Monitoring for deceptive alignment
_This post is a follow-up to “[AI coordination needs clear wins](https://www.alignmentforum.org/posts/vavnqwYbc8jMu3dTY/ai-coordination-needs-clear-wins).” Thanks to Ethan Perez, Kate Woolverton, Richard Ngo, Anne le Roux, Sam McCandlish, Adam Jermyn, and Danny Hernandez for useful... | https://www.lesswrong.com/posts/Km9sHjHTsBdbgwKyi/monitoring-for-deceptive-alignment |
# Most People Start With The Same Few Bad Ideas
*Epistemic status: lots of highly subjective and tentative personal impressions.*
Occasionally people say “hey, alignment research has lots of money behind it now, why not fund basically everyone who wants to try it?”. Often this involves an analogy to venture capital: ... | https://www.lesswrong.com/posts/Afdohjyt6gESu4ANf/most-people-start-with-the-same-few-bad-ideas |
# Samotsvety's AI risk forecasts
*Crossposted to* [*EA Forum*](https://forum.effectivealtruism.org/posts/EG9xDM8YRz4JN4wMN/samotsvety-s-ai-risk-forecasts) *and* [*Foxy Scout*](https://www.foxy-scout.com/samotsvetys-ai-risk-forecasts/)
Introduction
============
In [my review of What We Owe The Future](https://www.goo... | https://www.lesswrong.com/posts/YMsD7GA7eTg2BafQd/samotsvety-s-ai-risk-forecasts |
# Understanding and avoiding value drift
I use [the shard theory of human values](https://www.lesswrong.com/posts/iCfdcxiyr2Kj8m8mT/the-shard-theory-of-human-values) to clarify what value drift is, how it happens, and how it might be avoided by a highly intelligent agent—even if that agent doesn't have any control ove... | https://www.lesswrong.com/posts/jFvFreCeejRKaZv4v/understanding-and-avoiding-value-drift |
# Oversight Leagues: The Training Game as a Feature
*This post is part of my* [*hypothesis subspace*](https://www.lesswrong.com/s/H3xEgE7bPGKvucfQk) *sequence, a living collection of proposals I'm exploring at* [*Refine.*](https://www.lesswrong.com/posts/5uiQkyKdejX3aEHLM/how-to-diversify-conceptual-alignment-the-mode... | https://www.lesswrong.com/posts/i32eyaARtFn6eD9Cg/oversight-leagues-the-training-game-as-a-feature |
# Notes on Resolve
This post examines **resolve** and **decisiveness** (and things in that bailiwick like determination, commitment, dedication, steadfastness, and firmness), as part of a sequence of posts about virtues. I mostly explore what other people have learned about these virtues, rather than sharing my own re... | https://www.lesswrong.com/posts/W4xtftRxp23meoGdB/notes-on-resolve |
# Should you refrain from having children because of the risk posed by artificial intelligence?
Eli Lifland discusses AI risk probabilities [here](https://www.foxy-scout.com/wwotf-review/).
Scott Alexander talks about how everything will change completely in [this post](https://astralcodexten.substack.com/p/slightly-... | https://www.lesswrong.com/posts/fXiqXx8xTssaEdmQQ/should-you-refrain-from-having-children-because-of-the-risk |
# AI alignment with humans... but with which humans?
Note: I'm cross-posting this from EA Forum (where I posted it on Sept 8, 2022), in case anybody on LessWrong or the AI Alignment Forum is interested in commenting; note that there were some very helpful suggested readings in the replies to this: [https://forum.effe... | https://www.lesswrong.com/posts/4HcSZ9pKNEP3iCmSm/ai-alignment-with-humans-but-with-which-humans |
# Interpreting Affordable Housing
In describing affordable housing you'll often see a project described as "60% AMI", which can be confusing:
> *how is 60% of median income affordable?*—[NUMTOT comment](https://www.facebook.com/groups/whatwouldjanejacobsdo/posts/2248919421934333/?comment_id=2249414685218140)
If you ... | https://www.lesswrong.com/posts/L9CeB4huNtYzT4QrT/interpreting-affordable-housing |
# My emotional reaction to the current funding situation
I’m allowed to spend two days a week at Trajan House, a building in Oxford which houses the Center for Effective Altruism (CEA), along with a few EA-related bodies. Two days is what I asked for, and what I received. The rest of the time I spend in the Bodleian L... | https://www.lesswrong.com/posts/gFyJgnu5vAbzELBM8/my-emotional-reaction-to-the-current-funding-situation |
# Swap and Scale
*Produced as part of the SERI MATS program under John Wentworth*
*Thanks to Garrett Barker for pointing out you could think of Swap Symmetry as relabeling and to Matthias G. Mayer for catching my mistake in not multiplying the bias in Scale Symmetry.*
Here are two ways you can change the weights... | https://www.lesswrong.com/posts/G4xCDrfpLpf9JFjKH/swap-and-scale |
# Evaluations project @ ARC is hiring a researcher and a webdev/engineer
The evaluations project at the Alignment Research Center is looking to hire a generalist technical researcher and a webdev-focused engineer. We're a new team at ARC building capability evaluations (and in the future, alignment evaluations) for a... | https://www.lesswrong.com/posts/svhQMdsefdYFDq5YM/evaluations-project-arc-is-hiring-a-researcher-and-a-webdev-1 |
# Web4/Heaven - The Simulation
*[This is a thought experiment / ranty idea. I don’t necessarily agree with everything I say hear but try to think the idea through to its end. Inspired by the hypothetical practice of heaven banning, hence why I call it Heaven.]*
Web4 is the internet that is completely (or at least mos... | https://www.lesswrong.com/posts/gmWiiyjy7syEBXymd/web4-heaven-the-simulation |
# Ought will host a factored cognition “Lab Meeting”
Ought will host a factored cognition “Lab Meeting” on **Friday September 16 from 9:30AM - 10:30AM PT**.
We'll share the progress we've made using language models to decompose reasoning tasks into subtasks that are easier to perform and evaluate. This is part of ou... | https://www.lesswrong.com/posts/6wQyK7RaKcrkBF9nB/ought-will-host-a-factored-cognition-lab-meeting |
# Keeping Time in Epoch Seconds
Unix-like computers keep time in seconds since the [Unix Epoch](https://en.wikipedia.org/wiki/Unix_time)—the 0 second, which is set to midnight, January 1st, 1970, Zulu/UTC. We call a particular second since the Unix Epoch a timestamp.
Keeping time this way is pretty handy, but looking... | https://www.lesswrong.com/posts/Cq7g4KZuWz5uNbsvt/keeping-time-in-epoch-seconds |
# Path dependence in ML inductive biases
In this post, we define path dependence as the sensitivity of a model's behavior to the details of the training process and training dynamics.[^j9an6vasz5d] High path-dependence indicates that small changes to the training process can cause significant changes to how the final... | https://www.lesswrong.com/posts/bxkWd6WdkPqGmdHEk/path-dependence-in-ml-inductive-biases |
# Quintin's alignment papers roundup - week 1
Introduction
------------
I've decided to start a weekly roundup of papers that seem relevant to alignment, focusing on papers or approaches that might be new to safety researchers. Unlike the Alignment Newsletter, I'll be spending relatively little effort on summarizing ... | https://www.lesswrong.com/posts/7cHgjJR2H5e4w4rxT/quintin-s-alignment-papers-roundup-week-1 |
# Put Dirty Dishes in the Dishwasher
> *"It's not worth using the dishwasher: by the time you've gotten your dishes clean enough to put in, you might as well just finish washing them yourself."*
People have a wide range of ways they deal with dishes before putting them in the dishwasher:
1. Rinse well. The dishes s... | https://www.lesswrong.com/posts/mHpXFFXdf37mqTKbN/put-dirty-dishes-in-the-dishwasher |
# A Pin and a Balloon: Anthropic Fragility Increases Chances of Runaway Global Warming
*Alexey Turchin,* [*alexeiturchin@gmail.com*](mailto:alexeiturchin@gmail.com)
*Anonymous coauthor 1 *
*Anonymous coauthor 2*
**TL;DR:** Because of survival bias we underestimate the location of climate tipping points and thus ou... | https://www.lesswrong.com/posts/rmwAuWXYTo24E5nnX/a-pin-and-a-balloon-anthropic-fragility-increases-chances-of |
# AI Risk Intro 1: Advanced AI Might Be Very Bad
*This is a submission for the* [*AI Safety Public Materials contest.*](https://www.lesswrong.com/posts/gWM8cgZgZ9GQAYTqF/usd20k-in-bounties-for-ai-safety-public-materials#How_to_submit)
**Introduction**
================
If human civilisation is destroyed this century,... | https://www.lesswrong.com/posts/bJgEMfiD48fEJJxjm/ai-risk-intro-1-advanced-ai-might-be-very-bad |
# Making a New Table Leaf
We got our dining room table used, and it didn't come with any leaves. I made one out of plywood when we first got it, and it's been pretty helpful over the past ~7y:
[](https://www.jefftk.com/old-table-leaf-big.jpg)
With the leaf we can comfort... | https://www.lesswrong.com/posts/ttBa7yuvg4FyRd4ox/making-a-new-table-leaf |
# Briefly thinking through some analogs of debate
\[This post is an almost direct summary of a conversation with John Wentworth.\]
Note: This post assumes that the reader knows what [AI Safety via Debate](https://www.alignmentforum.org/tag/debate-ai-safety-technique-1) is. I don't spend any time introducing it or exp... | https://www.lesswrong.com/posts/pLDd8dJFq2iWuCD9h/briefly-thinking-through-some-analogs-of-debate |
# You Don't Have To Click The Links
[There](https://www.lesswrong.com/posts/46qnWRSR7L2eyNbMA/the-lens-that-sees-its-flaws) [are](https://www.lesswrong.com/posts/3nZMgRTfFEfHp34Gb/the-meditation-on-curiosity) [a](https://www.lesswrong.com/posts/nYkMLFpx77Rz3uo9c/belief-as-attire) [lot](https://www.lesswrong.com/posts/... | https://www.lesswrong.com/posts/JZuqyfGYPDjPB9Lne/you-don-t-have-to-click-the-links |
# Argument against 20% GDP growth from AI within 10 years [Linkpost]
Mohammed Bavarian, a research scientist at OpenAI, tweeted [this thread](https://twitter.com/mobav0/status/1566821953230364672) arguing that he could see "the overall US GDP growth rising from recent avg 2-3% to 20+% in 10 years." Feel free to check ... | https://www.lesswrong.com/posts/tMriCctwgpdFNWzwF/argument-against-20-gdp-growth-from-ai-within-10-years |
# Freeloading?
Selections from the category of "provider expects people in general will make it worth their while, and you're behaving in a way where you won't", roughly ordered from my impression of least to most acceptable:
* Taking free newspapers from distribution boxes to use in craft projects or for heating y... | https://www.lesswrong.com/posts/WrXGGAfFJ52esLgtf/freeloading |
# I’ve written a Fantasy Novel to Promote Effective Altruism
Back in February I got a grant through the extended set of ACX grants to write this. I now have a draft that is sufficiently polished that I’m comfortable showing to the community, though I suspect there will be a lot of things that I will want to change bas... | https://www.lesswrong.com/posts/aYuJdRKTvvgRd6HWr/i-ve-written-a-fantasy-novel-to-promote-effective-altruism |
# Alignment via prosocial brain algorithms
In this post, I want to briefly propose a semi-novel direction for alignment research that I'm excited about. Though some of these ideas are not brand new—they purposefully bear resemblance to recent (highly promising) work in [SHARD theory](https://www.lesswrong.com/posts/xq... | https://www.lesswrong.com/posts/yKgai84JhFmCkWQ8R/alignment-via-prosocial-brain-algorithms |
# Why do People Think Intelligence Will be "Easy"?
In many discussions I've hard with people around AI takeoff, I've frequently encountered a belief that intelligence is going to be "easy".
To quantify what I mean by "easy", many people seem to expect that the *marginal* returns on cognitive investment (whether ... | https://www.lesswrong.com/posts/nHp9dmeKrj6uQe2BF/why-do-people-think-intelligence-will-be-easy |
# [Linkpost] A survey on over 300 works about interpretability in deep networks
[**Toward Transparent AI: A Survey on Interpreting the Inner Structures of Deep Neural Networks**](https://arxiv.org/abs/2207.13243)
**Tilman Räuker* (**[traeuker@gmail.com](mailto:traeuker@gmail.com))
**Anson Ho*** ([anson@epochai.org]... | https://www.lesswrong.com/posts/WWcPFBZqpwA5kzE5y/linkpost-a-survey-on-over-300-works-about-interpretability |
# Time is not the bottleneck (on making progress thinking about difficult things)
... at least, not for me. The bottleneck is something like [mental energy](https://en.wikipedia.org/wiki/Mental_energy). I can only make progress on a (subjectively) difficult topic for a handful of hours (at best) each day before hittin... | https://www.lesswrong.com/posts/jc4DAHDcK2hgbnjvi/time-is-not-the-bottleneck-on-making-progress-thinking-about |
# EA & LW Forums Weekly Summary (5 - 11 Sep 22')
*Supported by Rethink Priorities*
Cross-posted on the EA forum [here](https://forum.effectivealtruism.org/posts/FWcJzBdfNF3mCKP47/ea-and-lw-forums-weekly-summary-5-11-sep-22).
This is part of a weekly series - you can see the full collection [here.](https://www.lesswr... | https://www.lesswrong.com/posts/4tyMgGthbfHCCMhYr/ea-and-lw-forums-weekly-summary-5-11-sep-22 |
# Which LessWrong content would you like recorded into audio/podcast form?
For the last few months, LessWrong has been working with Solenoid_Entity to record each [curated post](https://www.lesswrong.com/recommendations) into audio form. Recordings to date can be [found here](https://www.buzzsprout.com/2037297/) and o... | https://www.lesswrong.com/posts/GsX2cgcBvLEkGyBW6/which-lesswrong-content-would-you-like-recorded-into-audio |
# Deep Q-Networks Explained
*Note: This is a long post. The post is structured in such a way that not everyone needs to read everything - Sections 1 and 2 are skippable background information, and Sections 4 and 5 go into technical detail that not everybody wants or needs to know. Section 3 on its own is sufficient to... | https://www.lesswrong.com/posts/kyvCNgx9oAwJCuevo/deep-q-networks-explained |
# An investigation into when agents may be incentivized to manipulate our beliefs.
Contemporary research on both AI capabilities and alignment often features more intricate interaction patterns between humans and the AI-in-training than the standard approach of defining an reward signal once and running an optimizatio... | https://www.lesswrong.com/posts/Z9Rgek2FstkJ3L7ri/an-investigation-into-when-agents-may-be-incentivized-to |
# New tool for exploring EA Forum, LessWrong and Alignment Forum - Tree of Tags
Explore the hidden treasures of the forums
==========================================
With this tool you can zoom in on yo... | https://www.lesswrong.com/posts/cecqH7PvsNkrxFvwe/new-tool-for-exploring-ea-forum-lesswrong-and-alignment-1 |
# Simple proofs of the age of the universe (or other things)
I saw some time ago Carl Sagan asked schoolchildren for proofs that the earth is round. I can no longer find it, but they gave him sound, simple answers. One was watching a ship seem to sink below the horizon as it grows distant. One was photos of Earth from... | https://www.lesswrong.com/posts/KeTYY5N28JRHArsyv/simple-proofs-of-the-age-of-the-universe-or-other-things |
# Trying to find the underlying structure of computational systems
*This post was written in the* [*SERI MATS*](https://www.serimats.org/) *program under* [*John Wentworth*](https://www.lesswrong.com/users/johnswentworth)*.*
In our quest to find steering mechanisms in computational systems, we first have to find the ... | https://www.lesswrong.com/posts/od6zKB5swBzGL3LqE/trying-to-find-the-underlying-structure-of-computational |
# VA Rationalists - Norfolk Meetup
Hi!
Come on out to Norfolk's weekly social meetup, all are welcome :)
Suggested reading: [Twelve Virtues of Rationality](https://www.lesswrong.com/posts/7ZqGiPHTpiDMwqMN2/twelve-virtues-of-rationality)
We meet on Wednesdays from 17:00-19:30 EDT.
Cheers,
Willa | https://www.lesswrong.com/events/hnfobwnWjWskFa3Pg/va-rationalists-norfolk-meetup |
# Simple 5x5 Go
I've been [playing 5x5 Go](https://www.jefftk.com/p/5x5-go-board) with a lot of people new to the game, and have been using some [simplified](https://www.jefftk.com/p/simplifying-board-games) rules for faster teaching:
1. You can play on any empty intersection, unless it would make the board [look li... | https://www.lesswrong.com/posts/vwxs3HREyZnLsRPWB/simple-5x5-go |
# Dan Luu on Futurist Predictions
*Epistemic status: perspective derived from following Dan Luu's output for the last 5 years or so. Trying to vaguely gesture at a few things at once. Please ask questions if you find something confusing.*
Dan Luu has written a interesting post analysing the track record of futurist... | https://www.lesswrong.com/posts/2nouzDJnGyM3d97Ne/dan-luu-on-futurist-predictions |
# Git Re-Basin: Merging Models modulo Permutation Symmetries [Linkpost]
See [Arxiv](https://arxiv.org/pdf/2209.04836.pdf) and [Twitter](https://twitter.com/SamuelAinsworth/status/1569719494645526529). From the abstract:
> We argue that neural network loss landscapes contain (nearly) a single basin, after accounting f... | https://www.lesswrong.com/posts/4J8Cucvb5k7HHnkrL/git-re-basin-merging-models-modulo-permutation-symmetries |
# Some ideas for epistles to the AI ethicists
Some papers, or ideas for papers, that I'd loved to see published in ethics journals like *Minds and Machines* or *Ethics and Information Technology*.[^fsk6e604vrk] I'm probably going to submit one of these to a [2023 AI ethics conference](https://ethics.iit.edu/node/904?m... | https://www.lesswrong.com/posts/23cMcXb2zFbfJKz3n/some-ideas-for-epistles-to-the-ai-ethicists |
# Why Do People Think Humans Are Stupid?
Introduction
============
On more than one occasion, I've seen the following comparisons used to describe how a superintelligence might relate to/perceive humans:
* Humans to ants
* Humans to earthworms
* And similar
More generally, people seem to believe that humans a... | https://www.lesswrong.com/posts/yrekdsZfLsgfaFjFp/why-do-people-think-humans-are-stupid |
# The Defender’s Advantage of Interpretability
**TL;DR:** In this post, I want to argue why Interpretability & Transparency tools have a defender’s advantage if they are used correctly, i.e. they improve alignment much more than new capabilities and therefore mitigate risks of dual use. I draw parallels from biosecuri... | https://www.lesswrong.com/posts/6ReBeYwsDeNgv6Dr5/the-defender-s-advantage-of-interpretability |
# What would happen if we abolished the FDA tomorrow?
When I wax lyrical about how terrible the FDA is and how we should abolish it, most people react with "but you surely can't mean we should have no drugs regulation at all can you?".
I usually respond with explaining some alternative schemes which might work well, ... | https://www.lesswrong.com/posts/SrvwzkPwzauBErXh3/what-would-happen-if-we-abolished-the-fda-tomorrow |
# Transhumanism, genetic engineering, and the biological basis of intelligence.
I recently had a [fascinating conversation with Dr. Steven Hsu](https://www.youtube.com/watch?v=Ks8lLlogbOY), theoretical physicist and serial entrepreneur, about computational genomics and genetic engineering.
There were a number of real... | https://www.lesswrong.com/posts/zazA44CaZFE7rb5zg/transhumanism-genetic-engineering-and-the-biological-basis |
# Precise P(doom) isn't very important for prioritization or strategy
People spend time trying to determine the probability that AI will become an existential risk, sometimes referred to as P(doom).
One point that I think gets missed in this discussion is that a precise estimate of P(doom) isn't that important for ... | https://www.lesswrong.com/posts/c7fDt27pBdDDrEaZo/precise-p-doom-isn-t-very-important-for-prioritization-or |
# Guesstimate Algorithm for Medical Research
This document is aimed at subcontractors doing medical research for me. I am sharing it in the hope it is more broadly useful, but have made no attempts to make it more widely accessible.
Intro
=====
[Guesstimate](https://getguesstimate.com/) is a tool I have found quite... | https://www.lesswrong.com/posts/d9r68yE7bG2tA64w8/guesstimate-algorithm-for-medical-research |
# ACT-1: Transformer for Actions
> ACT-1 can take a high-level user request and execute it. The user simply types a command into the text box and ACT-1 does the rest. In this example, this requires repeatedly taking actions and observations over a long time horizon to fulfill a single goal.
...
> While we’re excited... | https://www.lesswrong.com/posts/pmhcZiv32FqSgckEE/act-1-transformer-for-actions |
# When does technical work to reduce AGI conflict make a difference?: Introduction
*This is a pared-down version of a longer draft report. We went with a more concise version to get it out faster, so it ended up being more of an overview of definitions and concepts, and is thin on concrete examples and details. Hopefu... | https://www.lesswrong.com/posts/oNQGoySbpmnH632bG/when-does-technical-work-to-reduce-agi-conflict-make-a |
# When would AGIs engage in conflict?
Here we will look at two of the claims introduced in the previous post: AGIs might not avoid conflict that is costly by their lights ([*Capabilities aren’t Sufficient*](https://www.lesswrong.com/posts/oNQGoySbpmnH632bG/when-does-technical-work-to-reduce-agi-conflict-make-a#Capabil... | https://www.lesswrong.com/posts/cLDcKgvM6KxBhqhGq/when-would-agis-engage-in-conflict |
# When is intent alignment sufficient or necessary to reduce AGI conflict?
In this post, we look at conditions under which [*Intent Alignment isn't Sufficient*](https://www.lesswrong.com/posts/oNQGoySbpmnH632bG/when-does-technical-work-to-reduce-agi-conflict-make-a#Intent_Alignment_isn_t_Sufficient_) or [*Intent Ali... | https://www.lesswrong.com/posts/fMJhfNZXFzCNpCL8v/when-is-intent-alignment-sufficient-or-necessary-to-reduce |
# Seeing the Schema
For the past month or two, I've been regularly playing a mental rotation game, available [here](https://0xf00ff00f.github.io/rotator/). In the game, the task is to find which two of t... | https://www.lesswrong.com/posts/rrCzJipDgcDmwaDnW/seeing-the-schema |
# Effective altruism in the garden of ends
This is [a crosspost from the EA Forum](https://forum.effectivealtruism.org/posts/AjxqsDmhGiW9g8ju6/effective-altruism-in-the-garden-of-ends) that I was encourage to post here as well. I imagine it may have even more experience here, since the most "hardcore" EAs that I know ... | https://www.lesswrong.com/posts/YDHRa5cmKQCLGrCWj/effective-altruism-in-the-garden-of-ends |
# Coordinate-Free Interpretability Theory
Some interpretability work assigns meaning to activations of individual neurons or small groups of neurons. Some interpretability work assigns meaning to directions in activation-space. These are two different ontologies through which to view a net’s internals. Probably neithe... | https://www.lesswrong.com/posts/sxhfSBej6gdAwcn7X/coordinate-free-interpretability-theory |
# Forecasting thread: How does AI risk level vary based on timelines?
*Crossposted to* [*EA Forum*](https://forum.effectivealtruism.org/posts/nYgw4FNpHf9bmJGEi/forecasting-thread-how-does-ai-risk-level-vary-based-on)
While there have been many previous surveys asking about the chance of existential catastrophe from A... | https://www.lesswrong.com/posts/h7Sx4DBL4JZnbTpes/forecasting-thread-how-does-ai-risk-level-vary-based-on |
# Sequencing Intro II: Adapters
A couple weeks ago I wrote a short [sequencing intro](https://www.jefftk.com/p/sequencing-intro). Here's a bit more, starting with a minor puzzle. I'm working with California wastewater sequencing data ( [Rothman et al 2021](https://journals.asm.org/doi/full/10.1128/AEM.01448-21)) and I... | https://www.lesswrong.com/posts/BBRXq3is35jjmZz3W/sequencing-intro-ii-adapters |
# General advice for transitioning into Theoretical AI Safety
**Update 2025: This post is completely out of date, don't follow its advice.**
*Cross-posted from the* [*EA Forum*](https://forum.effectivealtruism.org/posts/cEpYmdDMeZrsKJa8x/general-advice-for-transitioning-into-theoretical-ai-safety)*.*
*During the pas... | https://www.lesswrong.com/posts/mEDAqbdvg6ivy7eRp/general-advice-for-transitioning-into-theoretical-ai-safety |
# Capability and Agency as Cornerstones of AI risk — My current model
tldr: In this post I am explaining my current model of AI risk after failing at my goal of understanding everything completely, without relying on arguments by authority. I hope this post is valuable in explaining the topic and also by giving an ex... | https://www.lesswrong.com/posts/Ezhu43CRahQSdsWug/capability-and-agency-as-cornerstones-of-ai-risk-my-current |
# What's the longest a sentient observer could survive in the Dark Era?
Periodically I look at [this graph](https://en.wikipedia.org/wiki/Future_of_an_expanding_universe), and I'm like, "holy shit, that is *so much Nothing.* Wut. Like, they converted it to log-scale so it would fit into my brain and it's still *so muc... | https://www.lesswrong.com/posts/YHDu7EmuysTEuWWym/what-s-the-longest-a-sentient-observer-could-survive-in-the |
# FDT defects in a realistic Twin Prisoners' Dilemma
*Thanks to Caspar Oesterheld for discussion and helpful comments, as well as Tristan Cook, James Faville, Daniel Kokotajlo and Lukas Finnveden.*
Summary
-------
[Updateless decision theory](https://www.lesswrong.com/tag/updateless-decision-theory) (UDT)/[functiona... | https://www.lesswrong.com/posts/QpqKBYzPKdZpByZS3/fdt-defects-in-a-realistic-twin-prisoners-dilemma |
# Why deceptive alignment matters for AGI safety
**Update:** I changed the title from "Why AGI safety researchers should focus mostly on deceptive alignment" to "Why deceptive alignment matters for AGI safety". I think my original message was too strong and I'm actually much more uncertain about failure modes from AGI... | https://www.lesswrong.com/posts/oBFMbhQMt9HkmfF6d/why-deceptive-alignment-matters-for-agi-safety |
# Are Human Brains Universal?
[\[Previously\]](https://www.lesswrong.com/posts/yrekdsZfLsgfaFjFp/why-do-people-think-humans-are-stupid)
Introduction
============
After reading and updating on the answers to [my previous question](https://www.lesswrong.com/posts/yrekdsZfLsgfaFjFp/why-do-people-think-humans-are-stupid... | https://www.lesswrong.com/posts/f4mbXjhQ2yaMrgLBG/are-human-brains-universal |
# Covid 9/15/22: Permanent Normal
I may or may not have had Covid about two weeks ago.
At the time, I knew I had come down with _something._ I largely lost my voice and spent about a week carefully conserving what remained. I spent over a week coughing a decent amount. I was having some trouble breathing when any pre... | https://www.lesswrong.com/posts/kWbuF3oKS29yAnfkY/covid-9-15-22-permanent-normal |
# Rational Animations' Script Writing Contest
[Cross-posted to the EA Forum.](https://forum.effectivealtruism.org/posts/p8aMnG67pzYWxFj5r/rational-animations-script-writing-contest)
**Edit:** The deadline has been removed! So you can continue to send scripts.
Introduction
------------
I'm announcing a script wr... | https://www.lesswrong.com/posts/RH8nGG5vnuXc4eKu5/rational-animations-script-writing-contest |
# How should DeepMind's Chinchilla revise our AI forecasts?
*Acknowledgements: I wrote this report as part of a six-hour paid work-trial with* [*Epoch AI*](https://epochai.org/)*.*
*Epistemic status: My dataset analysis is a bit simplistic but the inference I draw from it seems likely. The implications for TAI timel... | https://www.lesswrong.com/posts/kixewxJfuZ23DQDfF/how-should-deepmind-s-chinchilla-revise-our-ai-forecasts |
# Understanding Conjecture: Notes from Connor Leahy interview
I recently listened to Michaël Trazzi interview Connor Leahy (co-founder & CEO of [Conjecture](https://www.conjecture.dev/), a new AI alignment organization) on a podcast called The Inside View. *Youtube video* [*here*](https://www.youtube.com/watch?v=Oz4G9... | https://www.lesswrong.com/posts/KFLdfuw35qkgjzWer/understanding-conjecture-notes-from-connor-leahy-interview |
# AstralCodexTen and Rationality
Meetup Organisers’ Retreat —
Europe, Middle East, and Africa 2023
**TL;DR: Retreat for EMEA meetup organisers, Summer 2023, apply** [**here**](https://forms.gle/sxttpRqpN71kNT437)
Application deadline: 31ˢᵗ Oct 2022
**Why a retreat?**
We think we all benefit from having an interna... | https://www.lesswrong.com/posts/u4DytKwQiFpxjwHjG/astralcodexten-and-rationality-meetup-organisers-retreat |
# Affordable Housing Investment Fund
Lots of people are looking for investments they can feel good about, which makes me think people might be interested in putting a portion of their savings toward helping resolve the housing crisis if it offered a reasonable return. The idea is that you construct 100% [affordable ho... | https://www.lesswrong.com/posts/amFYJ3NgZBSnBboH9/affordable-housing-investment-fund |
# Guidelines for Mad Entrepreneurs
*Significant planecrash spoilers up through* [*Book 5, crisis of faith.*](https://www.glowfic.com/posts/5977)
*Generated at the suggestion of John Wentworth.*
*Epistemic status: I personally have no experience with project management; here I'm just presenting the ideas on the topic... | https://www.lesswrong.com/posts/EosbLKPrJ9xahGvbT/guidelines-for-mad-entrepreneurs |
# I wrote a fantasy novel to promote EA: More Chapters
Note to Royal Road admin, yes, I plan to publish there. Not plagiarism.
I’m posting now several more chapters here, and of course the [google doc](https://docs.google.com/document/d/1ZppL3mlO6M98TLQk2IAdL3nM__Pmjxirt59WXGzbIMM/edit?usp=sharing) link where you... | https://www.lesswrong.com/posts/qqmhmdszdhMnRrzmQ/i-wrote-a-fantasy-novel-to-promote-ea-more-chapters |
# Representational Tethers: Tying AI Latents To Human Ones
*This post is part of my* [*hypothesis subspace*](https://www.lesswrong.com/s/H3xEgE7bPGKvucfQk) *sequence, a living collection of proposals I'm exploring at* [*Refine.*](https://www.lesswrong.com/posts/5uiQkyKdejX3aEHLM/how-to-diversify-conceptual-alignment-t... | https://www.lesswrong.com/posts/h7BA7TQTo3dxvYrek/representational-tethers-tying-ai-latents-to-human-ones |
# Levels of goals and alignment
*This post was written as part of* [*Refine*](https://www.lesswrong.com/posts/5uiQkyKdejX3aEHLM/how-to-diversify-conceptual-alignment-the-model-behind)*. Thanks to Adam Shimi, Lucas Teixeira, Linda Linsefors, and Jonathan Low for helpful feedback and comments. *
*Epistemic status: high... | https://www.lesswrong.com/posts/rzkCTPnkydQxfkZsX/levels-of-goals-and-alignment |
# Katja Grace on Slowing Down AI, AI Expert Surveys And Estimating AI Risk
Katja runs AI Impacts and recently published [What do ML researchers think about AI in 2022](https://aiimpacts.org/what-do-ml-researchers-think-about-ai-in-2022/), a new survey of AI Experts. We start this episode by discussing what Katja is cu... | https://www.lesswrong.com/posts/CZKzr82AzCZZ7r7qN/katja-grace-on-slowing-down-ai-ai-expert-surveys-and |
# Why are we sure that AI will "want" something?
I have no doubt that AI will some day soon surpass humans in all aspects of reasoning, that is pretty obvious. It is also clear to me that will surpass humans in the ability to do something, should it "want" to do it. And if requested to do something drastic, it can acc... | https://www.lesswrong.com/posts/pZk5b7bC9fxnR9hvR/why-are-we-sure-that-ai-will-want-something |
# Refine Blogpost Day #3: The shortforms I did write
Today was Refine's third blogpost day. [I find it hard to write blogposts](https://www.lesswrong.com/posts/EeTq9vbzMT4Zb4oWo/all-the-posts-i-will-never-write). The voice of the inner critic is too loud. The voice of Adam Shimi too is powerful. I'm caught between a r... | https://www.lesswrong.com/posts/iA9S8fLCgbjFF6fw4/refine-blogpost-day-3-the-shortforms-i-did-write |
# Towards a philosophy of safety
We live in a dangerous world. Many hazards come from nature: fire, flood, storm, famine, disease. Technological and industrial progress has made us safer from these dangers. But technology also creates its own hazards: industrial accidents, car crashes, toxic chemicals, radiation. And ... | https://www.lesswrong.com/posts/MefzHmmcZvia5vrJH/towards-a-philosophy-of-safety |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.