text stringlengths 300 320k | source stringlengths 52 154 |
|---|---|
# Video & transcript: Challenges for Safe & Beneficial Brain-Like AGI
(Lightly edited) transcript starts here
=======================================
*(FYI: This has a decent amount of overlap with* [*the version I posted here two years ago*](https://www.lesswrong.com/posts/ajhtyKxtmmErTwH5t/eai-alignment-speaker-ser... | https://www.lesswrong.com/posts/YKmyay3bWF2ofAGNo/video-and-transcript-challenges-for-safe-and-beneficial |
# Post EAG London AI x-Safety Co-working Retreat
Our new group house (*Hogwarts School of Alignment and Rationality*) in the Kentish countryside has few spare rooms, so single rooms will mostly not be available (but there are some hotels not super far away if this is important to you, contact us for details). It also ... | https://www.lesswrong.com/events/oD7uHZfbaCoepmhhc/post-eag-london-ai-x-safety-co-working-retreat |
# Is ChatGPT actually fixed now?
Is ChatGPT actually fixed now?
==============================
I led OpenAI’s “dangerous capability” testing. Want to know if ChatGPT can trick users into accepting insecure code? Or persuade users to vote a certain way? My team built tests to catch this.
Testing is today’s most impor... | https://www.lesswrong.com/posts/9xByyt8QjZf2umL2C/is-chatgpt-actually-fixed-now |
# Interest In Conflict Is Instrumentally Convergent
Why is conflict resolution hard?
I talk to a lot of event organizers and community managers. Handling conflict is consistently one of the things they find the most stressful, difficult, or time consuming. Why is that?
Or, to ask a related question: Why is the ACX M... | https://www.lesswrong.com/posts/qyykn5G9EKKLbQg22/interest-in-conflict-is-instrumentally-convergent |
# Attend the 2025 Reproductive Frontiers Summit, June 10-12
There should be a community oriented towards the genomic emancipation of humanity. There *isn't* such a community, but there *should* be. It's a future worth investing our hope in—a future where parents are able to choose to give their future children the gif... | https://www.lesswrong.com/posts/8ZExgaGnvLevkZxR5/attend-the-2025-reproductive-frontiers-summit-june-10-12 |
# Jim Babcock's Mainline Doom Scenario: Human-Level AI Can't Control Its Successor
Eliezer's AI doom arguments have had me convinced since the ancient days of 2007, back when AGI felt like it was many decades away, and we didn't have an intelligence scaling law (except to the Kurzweilians who considered Moore's Law to... | https://www.lesswrong.com/posts/nZtN9PW4qBeNKehfs/jim-babcock-s-mainline-doom-scenario-human-level-ai-can-t |
# Cheaters Gonna Cheat Cheat Cheat Cheat Cheat
Cheaters. Kids these days, everyone says, are all a bunch of blatant cheaters via AI. Then again, look at the game we are forcing them to play, and how we grade it. If you earn your degree largely via AI, that changes two distinct things.
1. You might learn different th... | https://www.lesswrong.com/posts/ZDXK6nDrvuFypAgoG/cheaters-gonna-cheat-cheat-cheat-cheat-cheat |
# Slow corporations as an intuition pump for AI R&D automation
How much should we expect AI progress to speed up after fully automating AI R&D? This post presents an intuition pump for reasoning about the level of acceleration by talking about different hypothetical companies with different labor forces, amounts of se... | https://www.lesswrong.com/posts/hMSuXTsEHvk4NG6pm/slow-corporations-as-an-intuition-pump-for-ai-r-and-d |
# My Experience With EMDR
[EMDR](https://www.emdr.com/what-is-emdr/) stands for Eye Movement Desensitization and Reprocessing, itself a subset of [Bilateral Stimulation Therapy](https://www.choosingtherapy.com/bilateral-stimulation/). Bilateral Stimulation Therapy is not well understood, but the gist is that activatin... | https://www.lesswrong.com/posts/zztduH59pvor6Tc3z/my-experience-with-emdr |
# Corbent – A Master Plan for Next‑Generation Direct Air Capture
When was the last time you thought about the CO₂ in the air you're breathing right now? For me, it was this morning. Why? Because we've hit 420 parts per million(ppm), higher than at any time in the past 3 million years[^rjjhd0akdnt].
Even if we stopped... | https://www.lesswrong.com/posts/RCifjM58vph2iFYEy/corbent-a-master-plan-for-next-generation-direct-air-capture-1 |
# Glass box learners want to be black box
*Epistemic status: intuitive speculation with scattered mathematical justification.*
My goal here is to interrogate the dream of writing a beautiful algorithm for intelligence and thereby ensuring safety. For example:
![](https://39669.cdn.cke-cs.com/rQvD3VnunXZu34m86e5f/im... | https://www.lesswrong.com/posts/boodbr2PXpEEMGrfx/glass-box-learners-want-to-be-black-box |
# Where is the YIMBY movement for healthcare?
In the progress movement, some cause areas are about technical breakthroughs, such as fusion power or a cure for aging. In other areas, the problems are not technical, but social. Housing, for instance, is technologically a solved problem. We know *how* to build houses, bu... | https://www.lesswrong.com/posts/Ha2EFzxpFLCec3vTs/where-is-the-yimby-movement-for-healthcare |
# Book Review: "Encounters with Einstein" by Heisenberg
The title of this book is clickbait. "Encounters with Einstein" is a short collection of lectures given or written by Werner Heisenberg in the 1970's, only one of which discusses Einstein.[^uw3t6vb218] The remaining lectures discuss various aspects of the history... | https://www.lesswrong.com/posts/hc7qi5u2XszzrXaPf/book-review-encounters-with-einstein-by-heisenberg |
# It's Okay to Feel Bad for a Bit
> "If you kiss your child, or your wife, say that you only kiss things which are human, and thus you will not be disturbed if either of them dies." - Epictetus
>
> "Whatever suffering arises, all arises due to attachment; with the cessation of attachment, there is the cessation of su... | https://www.lesswrong.com/posts/aGnRcBk4rYuZqENug/it-s-okay-to-feel-bad-for-a-bit |
# Somerville Porchfest 2025
[Somerville Porchfest](https://somervilleartscouncil.org/porchfest/) was a lot of fun this year, and we got a great crowd:
[](https://www.jefftk.com/porchfest-contra-crowd-2025-big.jpg)
Showing some pictures to Claude it guesses t... | https://www.lesswrong.com/posts/9rurcdtdvEYqpYbKs/somerville-porchfest-2025 |
# Consider not donating under $100 to political candidates
Epistemic status: thing people have told me that seems right. Also primarily relevant to US audiences. Also I am speaking in my personal capacity and not representing any employer, present or past.
Sometimes, I talk to people who work in the AI governance spa... | https://www.lesswrong.com/posts/tz43dmLAchxcqnDRA/consider-not-donating-under-usd100-to-political-candidates |
# Aligning Agents, Tools, and Simulators
*This post was written as part of* [*AISC*](https://www.aisafety.camp/) *2025.*
**Introduction**
In [Agents, Tools, and Simulators](https://www.lesswrong.com/posts/ddK7CMEC3XzSmLS4G/agents-tools-and-simulators), we outlined several lenses for conceptualizing LLM-based AI, wit... | https://www.lesswrong.com/posts/A6oGJFeZwHHWHwY3b/aligning-agents-tools-and-simulators |
# Better Air Purifiers
Are you looking for a project where you could substantially improve indoor air quality, with benefits both to general health and reducing pandemic risk?
I've [written](https://www.jefftk.com/p/kitchen-air-purifier-comparison) [a bunch](https://www.jefftk.com/p/elevating-air-purifiers) [about](h... | https://www.lesswrong.com/posts/bBX46cBpbruv2kJh5/better-air-purifiers |
# [Book Translation] Three Days in Dwarfland
There was a book written half a century ago in Soviet Union, where a group of three kids travel to the land of arithmetic, meet anthropomorphic numbers, and learn a few mathematical concepts in a way appropriate for kids about 10 years old...
And I thought: "What a pity th... | https://www.lesswrong.com/posts/CbxP8x6pk7vTMNEgW/book-translation-three-days-in-dwarfland |
# a confusion about preference orderings
Here's a confusion I have about preference orderings in decision theory.
Caveat: the observations I make below feel weirdly trivial to me, to the point that I feel wary of making a post about them at all; the specter of readers rolling their eyes and thinking "oh he's just tal... | https://www.lesswrong.com/posts/bEy3DKi9C3pJ4LpDW/a-confusion-about-preference-orderings |
# Absolute Zero: Alpha Zero for LLM
**Question for the alignment crowd:**
The new Absolute Zero Reasoner paper proposes a "self‑play RL with zero external data" paradigm, where the same model both invents tasks and learns to solve them. It reaches SOTA coding + math scores without ever touching a human‑curated datas... | https://www.lesswrong.com/posts/dkWicxcPKdSx9npK4/absolute-zero-alpha-zero-for-llm |
# Highly Opinionated Advice on How to Write ML Papers
TL;DR
-----
* **The essence of an ideal paper** is the **narrative**: a short, rigorous and evidence-based technical story you tell, with a takeaway the readers care about
* **What?** A narrative is fundamentally about a contribution to our body of knowled... | https://www.lesswrong.com/posts/eJGptPbbFPZGLpjsp/highly-opinionated-advice-on-how-to-write-ml-papers |
# Game theory of "Nuclear Prisoner's Dilemma" - on nuking rocks
Eliezer Yudkowsky wrote, in a place where I can't ask a follow-up question:
> - A rational agent should always do at least as well for itself as a rock, unless it's up against some other agent that specifically wants to punish particular decision algorit... | https://www.lesswrong.com/posts/Pu9YPzeuJGxyKpcJZ/game-theory-of-nuclear-prisoner-s-dilemma-on-nuking-rocks |
# AIs at the current capability level may be important for future safety work
Sometimes people say that it's much less valuable to do AI safety research today than it will be in the future, because the current models are very different—in particular, much less capable—than the models that we're actually worried about.... | https://www.lesswrong.com/posts/cJQZAueoPC6aTncKK/ais-at-the-current-capability-level-may-be-important-for |
# PSA: The LessWrong Feedback Service
At the bottom of the LessWrong post editor, if you have at least 100 global karma, you may have noticed this button.
The button
Many people click the button, and a... | https://www.lesswrong.com/posts/bkDrfofLMKFoMGZkE/psa-the-lesswrong-feedback-service |
# [Beneath Psychology] Introduction Part 1: The Challenge
This sequence, "Beneath Psychology: Truth-Seeking as the Engine of Change", explores how to understand and navigate psychological change from a first principles perspective. The title "Beneath Psychology" references the fact that the constraints of clear thinki... | https://www.lesswrong.com/posts/XaoLgBcu2wJWxs6G9/beneath-psychology-introduction-part-1-the-challenge |
# Global Risks Weekly Roundup #19/2025: India/Pakistan ceasefire, US/China tariffs deal & OpenAI nonprofit control
**Executive summary**
=====================
* India and Pakistan have agreed to a ceasefire, and despite initial breaches, it now seems to be holding.
* The Trump administration and China agreed to r... | https://www.lesswrong.com/posts/QJnLXJBud2zkc9nQr/global-risks-weekly-roundup-19-2025-india-pakistan-ceasefire |
# A Live Look at the Senate AI Hearing
Today’s post will be a little different. This past week,[Sam Altman and others testified at a US Senate hearing on AI competitiveness](https://www.techpolicy.press/transcript-sam-altman-testifies-at-us-senate-hearing-on-ai-competitiveness/).
 a 30-minute documentary about SB-1047 I had been working on since September.
Here's what I learned.
Story of the bill
=================
1. **The bill had to go through many committees** before passing the Califo... | https://www.lesswrong.com/posts/bb9rBGree5Gh77uDC/things-i-learned-making-the-sb-1047-documentary |
# Political sycophancy as a model organism of scheming
This post is a short empirical research note about training away [scheming behavior](https://www.lesswrong.com/posts/m5nWc9v6MTsWXKpCy/training-time-schemers-vs-behavioral-schemers) (which we’ll define as the propensity to take long-term power-seeking actions when... | https://www.lesswrong.com/posts/bhxgkb7YtRNwBxLMd/political-sycophancy-as-a-model-organism-of-scheming |
# How to title your blog post or whatever
So you’ve made a thing. I’ll pretend it’s a blog post, though it doesn’t really matter. If people read your thing, some would like it, and some wouldn’t.
You sh... | https://www.lesswrong.com/posts/yxHo9ysqz2wnM6i9h/how-to-title-your-blog-post-or-whatever |
# No-self as an alignment target
Being a coherent and persistent agent with persistent goals is a prerequisite for long-horizon power-seeking behavior. Therefore, we should prevent models from representing themselves as coherent and persistent agents with persistent goals.
If an LLM-based agent sees itself as ceasing... | https://www.lesswrong.com/posts/LSJx5EnQEW6s5Juw6/no-self-as-an-alignment-target |
# AI Doomerism in 1879
I’m reading George Eliot’s *Impressions of Theophrastus Such* (1879)—so far a snoozer compared to her novels. But chapter 17 surprised me for how well it anticipated modern AI doomerism.
In summary, Theophrastus is in conversation with Trost, who is an optimist about the future of automation an... | https://www.lesswrong.com/posts/DFyoYHhbE8icgbTpe/ai-doomerism-in-1879 |
# Levels of Republicanism
On Profitable Partial Exit from Perverse Regimes Through the Exercise of One's Formal Rights as a Citizen
Enough criticism and analysis for the moment; here's a constructive program!
Whole systems become richer through exchange and division of labor, which affords people more leisure to exp... | https://www.lesswrong.com/posts/xNf9ZkjXLkFYPDscs/levels-of-republicanism |
# Work as meditation
*A ritual framework for how to transform our work into a meditative / introspective practice to help cultivate wisdom and joy. This July 20, 2025 I'm running a "Mindful Research" retreat / residency where we will live and work in community with other scientists sharing the value of awareness, and ... | https://www.lesswrong.com/posts/kyJy5sdRsW2joSnfu/work-as-meditation |
# Monthly Roundup #30: May 2025
I hear word a bunch of new frontier AI models are coming soon, so let’s do this now.
#### Table of Contents
1. [Programming Environments Require Magical Incantations.](https://thezvi.substack.com/i/163471889/programming-environments-require-magical-incantations)
2. [That’s Not How A... | https://www.lesswrong.com/posts/9DankDJfqGFLion3v/monthly-roundup-30-may-2025 |
# Too Soon
It was a cold and cloudy San Francisco Sunday. My wife and I were having lunch with friends at a Korean cafe.
My phone buzzed with a text. It said my mom was in the hospital.
I called to find out more. She had a fever, some pain, and had fainted. The situation was serious, but stable.
Monday was a normal... | https://www.lesswrong.com/posts/reo79XwMKSZuBhKLv/too-soon |
# Optimization & AI Risk
*There many ways to taxonomize AI risk. One interesting framing, is ‘risks from optimization’. These are not new ideas. Eliezer *[*wrote about this*](https://www.readthesequences.com/Optimization-And-The-Intelligence-Explosion) *~15 years ago, and it seems like many ‘theory folks’ have been sa... | https://www.lesswrong.com/posts/HHkhSrmfWFqGiDiWL/optimization-and-ai-risk |
# LessWrong Community Weekend 2025- Applications are open
[**>>\> All infos on the main event page <<<**](https://www.lesswrong.com/events/JxsdDs8ZfbF4dBkGe/lesswrong-community-weekend-2025)
**Please help spread the word!**
![](https://39669.cdn.cke-cs.com/rQvD3VnunXZu34m86e5f/images/16751843cd4d22f412f8063875bc151... | https://www.lesswrong.com/posts/ESCBiXGD4aRTfygee/lesswrong-community-weekend-2025-applications-are-open |
# Working through a small tiling result
tl;dr it seems that you can get basic tiling to work by proving that there will be safety proofs in the future, rather than trying to prove safety directly.
"Tiling" here roughly refers to a state of affairs in which we have a program that is able to prove itself safe to run. I... | https://www.lesswrong.com/posts/akuMwu8SkmQSdospi/working-through-a-small-tiling-result |
# Announcing Trajectory Labs - A Toronto AI Safety Office
TL;DR
-----
* We’re announcing [Trajectory Labs](https://www.trajectorylabs.org/), an AI safety coworking and events space in downtown Toronto.
* We support local AI safety efforts through three core programs:
* Workspace - a shared office space for ... | https://www.lesswrong.com/posts/Ws6KS7DtQ3F9Gv8ym/announcing-trajectory-labs-a-toronto-ai-safety-office |
# The Best Reference Works for Every Subject
Introduction
============
[*The Best Textbooks on Every Subject*](https://www.lesswrong.com/posts/xg3hXCYQPJkwHyik2/the-best-textbooks-on-every-subject) is the Schelling point for the best textbooks on every subject. My [*The Best Tacit Knowledge Videos on Every Subject*]... | https://www.lesswrong.com/posts/HLJMyd4ncE3kvjwhe/the-best-reference-works-for-every-subject |
# Default arguments in casual speech
The other day I was planning out bike route that I'll need to take periodically. I need to get from NW 12th and Marshall to NE 60th and Glisan in Portland, Oregon. As you can see below, I have lots of options.
 audience of people who agree with/support you”.*
If you believe that slowing down or pausing AI development is a good idea, and you believe this should be official policy,... | https://www.lesswrong.com/posts/qcKcWEosghwXMLAx9/doomers-should-try-much-harder-to-get-famous |
# Achievement Unlocked - 2E Rationalist School in Brooklyn
I thought I'd give a shout out to my son’s school, Achievement Unlocked [https://achievementunlocked.org/](https://achievementunlocked.org/) in Brooklyn NY, which is currently recruiting for the next school year. It is a twice-exceptional school for gifted ne... | https://www.lesswrong.com/posts/yMK92CEegByFPcxtA/achievement-unlocked-2e-rationalist-school-in-brooklyn |
# Fighting Obvious Nonsense About AI Diffusion
Our government is determined to lose the AI race in the name of winning the AI race.
The least we can do, if prioritizing winning the race, is to try and actually win it.
It is one thing to prioritize ‘winning the AI race’ against China over ensuring that humanity survi... | https://www.lesswrong.com/posts/FBZmnmbDBqhmhGhby/fighting-obvious-nonsense-about-ai-diffusion |
# Dodging systematic human errors in scalable oversight
**Summary:** Both our (UK AISI's) [debate safety case sketch](https://www.alignmentforum.org/posts/iELyAqizJkizBQbfr/an-alignment-safety-case-sketch-based-on-debate) and [Anthropic’s research agenda](https://alignment.anthropic.com/2025/recommended-directions/#h.... | https://www.lesswrong.com/posts/EgRJtwQurNzz8CEfJ/dodging-systematic-human-errors-in-scalable-oversight |
# Moral Obligation and Moral Opportunity
*This concept and terminology was spawned out of a conversation a few years ago with my friend* [*Skyler.*](https://www.lesswrong.com/users/screwtape) *I finally decided to write it up. Any mistakes here are my own.*
* * *
Every once in a while, I find myself at yet another r... | https://www.lesswrong.com/posts/AmwgNS3ybwF8Eovho/moral-obligation-and-moral-opportunity |
# Xi Jinping's readout after an AI "study session" [ChinaTalk Linkpost]
[Substack link here](https://www.chinatalk.media/p/xi-takes-an-ai-masterclass)
TL, DR:; Xi Jinping listens to a lecture about AI and publishes his "study notes/takeaways." From the start of the post (emphasis mine)
> On April 25, observers of Ch... | https://www.lesswrong.com/posts/PfyTyZKASezFynYmv/xi-jinping-s-readout-after-an-ai-study-session-chinatalk |
# Eliezer and I wrote a book: If Anyone Builds It, Everyone Dies
Eliezer and I wrote a book. It’s titled _If Anyone Builds It, Everyone Dies._ Unlike a lot of other writing either of us have done, it’s being professionally published. It’s hitting shelves on September 16th.
It’s a concise (~60k word) book aimed at a b... | https://www.lesswrong.com/posts/iNsy7MsbodCyNTwKs/eliezer-and-i-wrote-a-book-if-anyone-builds-it-everyone-dies |
# Curate your space
### **I: Rooms**
I had been travelling for a year and a half before I considered what coming home would look like.
Intending to maintain the growth solo-travel provided, I came up with a list of things to do once I was back at home.
Scrawled onto that list was a simple goal: “**Curate my space.*... | https://www.lesswrong.com/posts/Fy5hb3A2WYk4rsXg3/curate-your-space |
# Re SMTM: negative feedback on negative feedback
SlimeMoldTimeMold (SMTM) recently finished a 14-part series on psychology, [“The Mind in the Wheel”](https://slimemoldtimemold.com/tag/the-mind-in-the-wheel/), centering on feedback loops (“cybernetics”) as kind of a grand unified theory of the brain.
There are parts ... | https://www.lesswrong.com/posts/LfqFZvfSQhnCXHFE6/re-smtm-negative-feedback-on-negative-feedback |
# Elastomeric Fitting Session
A well-fitting elastomeric respirator is something everyone should [have on hand for emergencies](https://www.jefftk.com/p/disasters). \[1\] Primarily for infection prevention, since a future pandemic could be much worse than covid, though they're also useful for smoke. There are a lot of... | https://www.lesswrong.com/posts/mZwg8af7gRLGwkNTx/elastomeric-fitting-session |
# Why OpenAI projects only $174B of revenue by 2030?
1. "OpenAI has told investors it expects to reach $125B in revenue by 2029 and $174B by 2030. OpenAI expects to serve 3B monthly active users by 2030, with 900M DAUs." (from [sacra.com](https://sacra.com/c/openai); original source seems to be [theinformation.com](h... | https://www.lesswrong.com/posts/Yhfgygybmkyfgs64k/why-openai-projects-only-usd174b-of-revenue-by-2030 |
# What if Agent-4 breaks out?
*There are several scenarios that, unlike* [*AI 2027*](https://ai-2027.com/slowdown)*, explores consequences of AIs “going rogue” and freely replicating over the cloud. This post examines what these scenarios suggest might happen if OpenBrain—the leading AI company in the AI 2027 scenario... | https://www.lesswrong.com/posts/3GawHDBhsgg3CGbNb/what-if-agent-4-breaks-out |
# What Does It Mean to "Write Like You Talk"?
*Audio narration available on the* [*Substack version*](https://arjunpanickssery.substack.com/p/what-does-it-mean-to-write-like-you) of this post.
People often say to “write like you talk.” Paul Graham has a post titled [“Write Like You Talk”](https://www.paulgraham.com/t... | https://www.lesswrong.com/posts/TECaEvFeRdjaYK7sM/what-does-it-mean-to-write-like-you-talk |
# From Comments on Accountability Sinks
There was a lot of comments on the [original article about accountability sinks](https://250bpm.substack.com/p/accountability-sinks). People seem to feel strongly about human judgement being replaced by formal processes. So let’s dive in into the comments!
[Ben Pace](https://ww... | https://www.lesswrong.com/posts/E7d2vuP2THxwRCAcu/from-comments-on-accountability-sinks |
# AlphaEvolve
New AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluators
-------------------------------------------------------------------------------------------------------------------------------------------------... | https://www.lesswrong.com/posts/G9gz5k4mu4zuNrqEt/alphaevolve |
# Counter-considerations on AI arms races
(Work done at [Convergence Analysis](https://www.convergenceanalysis.org/). Mateusz wrote the post and is responsible for the outline of the argument, many details of which crystallized in conversations with Justin. Thanks to Olga Babeeva for the feedback on this post.)
1\. I... | https://www.lesswrong.com/posts/wGtTF3QpEDFdW5mR9/counter-considerations-on-ai-arms-races |
# AI #116: If Anyone Builds It, Everyone Dies
[If Anyone Builds It, Everyone Dies](https://x.com/MIRIBerkeley/status/1922712007121252623) is the title of the new book coming September 16 from Eliezer Yudkowsky and Nate Sores. The ‘it’ in question is superintelligence built on anything like the current AI paradigm, and... | https://www.lesswrong.com/posts/CjrbJmtStW4o2ieYN/ai-116-if-anyone-builds-it-everyone-dies |
# Problems with instruction-following as an alignment target
We should probably try to understand the failure modes of the alignment schemes that AGI developers are most likely to attempt.
I still think [Instruction-following AGI is easier and more likely than value aligned AGI](https://www.lesswrong.com/posts/7NvKrq... | https://www.lesswrong.com/posts/CSFa9rvGNGAfCzBk6/problems-with-instruction-following-as-an-alignment-target |
# Consider Others' Cost Tolerances
### Just because you won't, doesn't mean they won't.
* * *
Background
----------
There are a few main categories of costs associated with everyday decision-making:
1. Time
2. Money
3. Physical suffering
4. Mental suffering
5. Risk of death
Everyone has their own limits for ... | https://www.lesswrong.com/posts/XYerY6SLHcuFJfrhG/consider-others-cost-tolerances |
# time is event based
[](https://substackcdn.com/image/fetch/f_auto,q_auto:good,fl_progressive:steep/ht... | https://www.lesswrong.com/posts/bBSJzAM6hchi4bX2G/time-is-event-based |
# Starting Over: What to tell Sarah, at the edge of professional oblivion.
I came across a sad post recently from a woman on LinkedIn— let’s call her Sarah—who spent over a decade, in her words, as a *top 2% recruiter*, and was now finding it impossible to land even the most basic of roles. It was one of those earnes... | https://www.lesswrong.com/posts/FpGzbtJ4Dbi3dwJ8B/starting-over-what-to-tell-sarah-at-the-edge-of-professional |
# Researching Synthetic Consciousness: sound appealing?
Without going into the philosophical or technical underpinnings too much, several individuals I know from a wide range of backgrounds—computational neuroscience, computer science, philosophy, AI research—have put together a project where we hope to put some real... | https://www.lesswrong.com/posts/bNDhh9wQYXTnxRpJe/researching-synthetic-consciousness-sound-appealing |
# Staying in a Capsule Hotel
A hotel room generally feels way bigger than I need when traveling: I'm really just looking for somewhere to sleep. When I first read the [Wikipedia page for capsule hotels](https://en.wikipedia.org/wiki/Capsule_hotel), maybe 20y ago, this seemed like a very reasonable approach: lots of pe... | https://www.lesswrong.com/posts/XXk23MAvSBwf6ajEg/staying-in-a-capsule-hotel |
# Bay Area Summer Solstice
**What:** Secular celebration of the Summer Solstice. It's a big picnic at the park, basically.
**When:** Saturday, June 21st, the Summer Solstice
**Where:** Point Pinole Regional Park, East Bay
**Who:** Anyone who is reading this. This is for all members of the rationalist, LessWrong, EA... | https://www.lesswrong.com/events/ihkJBq8Brk9XqzH3w/bay-area-summer-solstice |
# It Is Untenable That Near-Future AI Scenario Models Like “AI 2027” Don't Include Open Source AI
*Note: This post is the second in a broader series of posts about the difficult tradeoffs inherent in public access to powerful open source models (the first post is *[*here*](https://www.lesswrong.com/posts/QSyshep2CRs8J... | https://www.lesswrong.com/posts/vyWzz5exnyKbPJ4zt/it-is-untenable-that-near-future-ai-scenario-models-like-ai |
# Interpretable Fine Tuning Research Update and Working Prototype
*Status: Very early days of the research. No major proof of concept yet. The aim of this research update is to solicit feedback and encourage others to experiment with my code. *
[GitHub repository](https://github.com/AMindToThink/interpretable-fine-tu... | https://www.lesswrong.com/posts/xZkBDAgQGqm6tvCvy/interpretable-fine-tuning-research-update-and-working |
# Generating the Funniest Joke with RL (according to GPT-4.1)
Language models are not particularly good at generating funny jokes. Asked for their funniest jokes, Claude 3.7 gives us:
> Why don't scientists trust atoms? Because they make up everything!
o3 gives us:
> Why don't scientists trust atoms anymore? Becaus... | https://www.lesswrong.com/posts/xMGmibZpPDnawjHXk/generating-the-funniest-joke-with-rl-according-to-gpt-4-1 |
# Regarding South Africa
The system prompt being modified by an unauthorized person in pursuit of a ham-fisted political point very important to Elon Musk once already doesn’t seem like coincidence. It happening twice looks rather worse than that.
A Funny Thing Happened on Twitter
In addition to having [seemingly ba... | https://www.lesswrong.com/posts/kMH8zFoHHJvy6wH7h/regarding-south-africa |
# US-China trade talks should pave way for AI safety treaty [SCMP crosspost]
*This is a* [*crosspost from the South China Morning Post*](https://www.scmp.com/opinion/china-opinion/article/3310252/us-china-trade-talks-should-pave-way-ai-safety-treaty).
Few expected it would happen this fast, but there is widespread re... | https://www.lesswrong.com/posts/C6Aaba47JeLQzJoe7/us-china-trade-talks-should-pave-way-for-ai-safety-treaty |
# P-Values Know When You're Cheating
This post is not "the solution to p-hacking" or even "the problem with p-hacking". But I had some recent insights into thinking behind the p-value approach and wanted to see how they apply to p-hacking as an issue. While p-values are a flawed approach to reporting final results, th... | https://www.lesswrong.com/posts/2MxK9CtHfwkqTJTQf/p-values-know-when-you-re-cheating |
# How Fast Can Algorithms Advance Capabilities? | Epoch Gradient Update
*I'm cross-posting my guest post on Epoch's Gradient Updates newsletter, in which I describe some new research from my team at UChicago's XLab — roughly, the algorithmic improvements that most improve capabilities at scale are the ones that requir... | https://www.lesswrong.com/posts/qhjNejRxbMGQp4wHt/how-fast-can-algorithms-advance-capabilities-or-epoch |
# Events: Debate & Fiction Project
I'm doing a couple of events next week which people might be interested in attending:
* * *
Event: Tsvi & Abram debate AI timelines.
Abram expects to defend a picture similar to the AI 2027 report. Tsvi expects to defend much longer timelines.
Monday, May 19 · 10:00 – 11:00am E... | https://www.lesswrong.com/posts/sW3GH2wiDFqZySQdt/events-debate-and-fiction-project |
# Social anxiety isn't about being liked
There’s this popular idea that socially anxious folks are dying to be *liked.* Why else would someone be so anxious about how others perceive them?
And yet, being socially anxious tends to make you less likeable… so they must be behaving really irrationally, right?
 [been](https://en.wikipedia.org/wiki/Office_Space) [contemptuous](https://en.wikipedia.org/wiki/Dilbert) of [management](https://en.wikipedia.org/wiki/The_Soul_of_a_New_Machine): if you're not writing code w... | https://www.lesswrong.com/posts/rhLbcWna8npw8BavC/management-is-the-near-future |
# Book Review: The Art of Happiness
The Art of Happiness is a book co-written by American psychiatrist Dr. Cutler and the 14th Dalai Lama, Tenzin Gyatso. Published in 1998, the book is an attempt to blend western scientific understanding of the mind with the spiritual wisdom and ancient training of Buddhist practices.... | https://www.lesswrong.com/posts/PuSD6entRFBbnru9P/book-review-the-art-of-happiness |
# The absent-minded variations
I wake up in a featureless room. I don't remember how I ended up here, when did I go to sleep, or for that matter, anything at all besides general abstract facts about the world and math. The place is hard to qualify; for the most part it really seems like I am just inside a perfect Plat... | https://www.lesswrong.com/posts/uvNsFv9YKxtfSBffM/the-absent-minded-variations |
# D&D.Sci: The Choosing Ones
*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.*
*Thank you to Juan Vasquez for playtesting.*
*Intended Difficulty: ~3.5/5*
The fairy in your bedroom explains that she is a champion of Fat... | https://www.lesswrong.com/posts/hoFLMjgjkLGtGfiay/d-and-d-sci-the-choosing-ones |
# Paranoid Debating
When: June 1st, 2 PM
Where: JFK Park, Cambridge
Run by: Mili
Paranoid Debating is a variant of The Aumann Game where one player purposefully subverts the group estimate. Similar to The Aumann Game, the activity consists of a group jointly producing a confidence interval for an unknown, but... | https://www.lesswrong.com/events/3cAG3KcZio5qPyyCD/paranoid-debating |
# Seven ways to Improve the Internal Model Principle
*This post was written during the* [*Dovetail Research*](https://dovetailresearch.org/) *Fellowship. Thanks to* [*Alex*](https://www.lesswrong.com/users/alex_altair)*,* [*Dalcy*](https://www.lesswrong.com/users/dalcy)*, and* [*Jose*](https://www.lesswrong.com/users/... | https://www.lesswrong.com/posts/5NPLrAt9SKWiKPvnw/seven-ways-to-improve-the-internal-model-principle |
# Will we survive if AI solves engineering before deception?
Google's [AlphaEvolve](https://deepmind.google/discover/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/) has recently started to make real world scientific discoveries, but only in domains where it's very cheap to verify the... | https://www.lesswrong.com/posts/TGQrqqptazFeRfHBR/will-we-survive-if-ai-solves-engineering-before-deception |
# Multipolar AI is Underrated
**Setup**
---------
AGI timelines are rapidly shortening, with a growing number of predictions and scenarios suggesting something close to AGI could emerge within [2 years](https://www.lesswrong.com/posts/bb5Tnjdrptu89rcyY/what-s-the-short-timeline-plan) - [5 years](https://www.metaculus... | https://www.lesswrong.com/posts/JjYu75q3hEMBgtvr8/multipolar-ai-is-underrated |
# What OpenAI Told California's Attorney General
### In a previously unreported letter, the AI company defends its restructuring plan while attacking critics and making surprising admissions
*This is the full text of a post first published on *[*Obsolete*](https://garrisonlovely.substack.com/)*, a Substack that I wri... | https://www.lesswrong.com/posts/mtgbgepheaQsAyXmG/what-openai-told-california-s-attorney-general |
# Google Logo Ligature Bug
Jeffrey Yasskin recently [pointed out](https://hachyderm.io/@jyasskin/114520767683101962) an [interesting security bug](https://issues.chromium.org/issues/391788835):
[](https://www.jefftk.com/google-logo-ligature-example-big.png)
... | https://www.lesswrong.com/posts/MGGm8B7StJtbhQs56/google-logo-ligature-bug |
# European Links (18.05.25)
The Birth of Estonian E-government
==================================
Estonia thanks the old communist bureacrats in the ‘90s for its now world famous e-government. It’s worth noting that all post-communist countries were facing the same problem, but only Estonia (ok, maybe also Georgia un... | https://www.lesswrong.com/posts/Jt7whJiBzk8ZDAmQx/european-links-18-05-25 |
# Modeling versus Implementation
*Epistemic status: I feel that naming this axis deconfuses me about agent foundations about as much as writing the rest of this sequence so far - so it is worth a post even though I have less to say about it. *
I think my goal in studying agent foundations is a little atypical. I am u... | https://www.lesswrong.com/posts/rysmFDWKT5L7uNTqD/modeling-versus-implementation |
# Limits to Control Workshop
**[limitstocontrol.org](https://limitstocontrol.org/)**
## Intro
Can we keep enough control over AI? If systems are developed to be more and more autonomous, this is no longer a given. It's a hypothesis, calling for serious investigation.
Even alignment relies on control. Researchers buil... | https://www.lesswrong.com/posts/WvaskSo4YsxJdmEyW/limits-to-control-workshop |
# Gödel, Escher, Bach in the age of LLMs
Look, nothing I can say about GEB will be news to anyone reading LessWrong, but I wrote all this up to organize my thoughts, so I might as well post it here for anyone who hasn't read the book or wants a wave of nostalgia about that time they got obsessed with it as a teenager.... | https://www.lesswrong.com/posts/ZFJpYj7AsGA6f5ehe/goedel-escher-bach-in-the-age-of-llms |
# Dreams of Ideas
CGP Grey [describes](https://youtu.be/NG8tVC5-WPg?t=353) a phenomenon I think of as 'dreams of ideas' that I find useful as a tool to know when to stop working on a project.
> You can be working on something and you are thinking about what it could be, but what is hard to know is if that thought in ... | https://www.lesswrong.com/posts/dzdEtFwkjGQ6XvTko/dreams-of-ideas |
# The Fragility of Naive Dynamism
### Movement alone isn’t progress, and there are dangers in ignoring dimensions and directions of acceleration.
*Note: This is part of my Substack on* [*Exploring Cooperation*](https://exploringcooperation.substack.com/)*. I’m delaying my next planned post to respond to* [*a recent p... | https://www.lesswrong.com/posts/upKqoHezhwpAncdri/the-fragility-of-naive-dynamism |
# A widely shared AI productivity paper was retracted, is possibly fraudulent
*Confidence notes: I am a physicist working on computational material science, so I have some familiarity with the field, but don't know much about R&D firms or economics. Some of the links in this article were gathered from a post at* [*piv... | https://www.lesswrong.com/posts/M2GzdAGbxwinERSEt/a-widely-shared-ai-productivity-paper-was-retracted-is |
# [Paper] Safety by Measurement: A Systematic Literature Review of AI Safety Evaluation Methods
"*If you cannot measure it, you cannot improve it.*" \- Lord Kelvin
The [science of AI safety evaluations](https://www.apolloresearch.ai/blog/we-need-a-science-of-evals) is still nascent, but it is making progress! We know... | https://www.lesswrong.com/posts/CwdCYmsutwXwnYtEF/paper-safety-by-measurement-a-systematic-literature-review |
# FLF Fellowship on AI for Human Reasoning: $25-50k, 12 weeks
The Future of Life Foundation is launching a [fellowship on AI for Human Reasoning](https://www.flf.org/fellowship), which we think could be decisive in averting existential risk.
> ## Fellowship on AI for Human Reasoning
>
> [Apply](https://jobs.lever.co/... | https://www.lesswrong.com/posts/aDWvPQNub6aDXwxue/flf-fellowship-on-ai-for-human-reasoning-usd25-50k-12-weeks |
# Thoughts on "Antiqua et nova" (Catholic Church's AI statement)
On January 28, 2025 (during the pontificate of Pope Francis) the Catholic Church put out its position paper on AI: [*Antiqua et nova*: Note on the Relationship Between Artificial Intelligence and Human Intelligence](https://www.vatican.va/roman_curia/con... | https://www.lesswrong.com/posts/yDyRgLSvpsD3PqQHC/thoughts-on-antiqua-et-nova-catholic-church-s-ai-statement |
# Semen and Semantics: Understanding Porn with Language Embeddings
**Summary**
-----------
Porn content has gotten more extreme over time. Here's the average title for the first full year of Pornhub's existence, 2008:
* "Hot blonde girl gets fucked"
and here's the average title for 2023:
* "FAMILYXXX - "I Cant... | https://www.lesswrong.com/posts/mcZYoRQWpneyM4yWp/semen-and-semantics-understanding-porn-with-language |
# Survey of Multi-agent LLM Evaluations
Disclaimer for LessWrong: This is a project I worked on for a college class. I think it’s useful for people working on evaluations to skim this to get a sense of what types of multi-agent evals exist. Thank you to Xavier Roberts-Gaal for advising me for this project!
Abstract
... | https://www.lesswrong.com/posts/tGcLA596E8g3KnphE/survey-of-multi-agent-llm-evaluations |
# US credit rating downgraded, $1T in Gulf state investments in the US, Kurdistan Workers’ Party disbanded | Sentinel Global Risks Weekly Roundup #20/2025
**Executive summary**
---------------------
**Moody’s downgraded the US credit rating** as the US budget deficit grows and US borrowing costs rise. House Republica... | https://www.lesswrong.com/posts/TN8BRmDc6qXtRKXTM/us-credit-rating-downgraded-usd1t-in-gulf-state-investments |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.