PODCAST

LessWrong (30+ Karma)

LessWrong

Audio narrations of LessWrong posts.

technology philosophy

All Episodes

90:40

“A system overview for near-term, low-trust AI...

LessWrong (30+ Karma) ·

2026/06/23

Version 0.2, working draft This is a working draft of my current best idea for a privacy-preserving, retrofittable AI compute verification system, for confidence-building in an arms-control-like AI agreement between rival nation states. The purpose of this draft is to elicit community engagement by making use of Cunningham's law: I make assertions about what the (emerging) field of AI verification should aim for, and people with experience in international policy, cybersecurity and any relevant

en-gb

24:03

“Model Size Scaling in 2023-2031” by Vladimir_Nesov

LessWrong (30+ Karma) ·

2026/06/23

Token generation speed is constrained by the speed at which the relevant HBM can be read, which is mostly the weights and KV-cache. Suppose a model is large, so that more than half of HBM is read when making a single pass over the weights, it's being read in parallel within a scale-up system, and N such systems are used in a pipeline. Then the time it takes to generate a token (without speculative decoding) is at least the time of reading more than half of an HBM stack times N. If we target a

en-gb

21:32

“GLM-5.2 Is The New Best Open Model” by Zvi

LessWrong (30+ Karma) ·

2026/06/22

GLM-5.2 arrived last week. It boasts excellent benchmarks and looks strong. Benchmarks here are a de facto ceiling of how good it is, not a point estimate. Essentially all other aspects of an open model like this, beyond speed and price, will almost always be worse than the numbers suggest. Still, impressive. It is definitely a large step up from GLM-5.1, and likely the strongest open model. GLM-5.2 is still substantially behind the absolute frontier, although plausibly on the cost-benefit

en-gb

51:10

“The AI Industrial Explosion — Part 4: Cheap power”...

LessWrong (30+ Karma) ·

2026/06/22

In Parts 1, 2, and 3 we estimated how fast a post-AGI economy could grow using existing or historically observed production techniques, grounded in US input-output data. That approach gave us confidence that the methods we assumed were physically realizable, because they correspond to manufacturing processes that have run at scale in the past or run today. Now I would like to relax that constraint, and ask how much faster the economy could grow using more advanced technology. In this part, we will

en-gb

32:24

“A Theory of Prompt Injection (and why you should...

LessWrong (30+ Karma) ·

2026/06/22

Summary We've been building a theory of how prompt injections work under the hood.We show it comes down to how LLMs perceive roles (the humble chat template tags).We use this theory to create new attacks, explain some weird mech interp results, and predict when attacks work.We also advocate for a new subfield focused on the science of roles, and sketch some unexplored new research problems.Work supported by CBAI and Cosmos. Another version of this post (with more inline colors) is here, and full

en-gb

3:23

“Coup is the Pareto-optimal social game” by Daniel Tan

LessWrong (30+ Karma) ·

2026/06/22

I've been playing Coup for a long time now. I keep a copy in my backpack and bring it everywhere, and it's earned the space. A few reasons it's so good: It's trivial to teach.You can explain the rules in a minute or two. Anyone can pick it up and start playing immediately.Many people find it fun.Almost everyone I've played with has loved it — pretty much unanimously, not least because it involves a lot of bluffing and emergent social / political dynamics It scales.Two players works; so does six

en-gb

3:30

“A brief list of ways AI safety efforts could be net...

LessWrong (30+ Karma) ·

2026/06/22

Here's Holden Karnofsky: I tend to think it's worse than 51/49. I tend to think we’re always going to be prone to overestimate how robustly good our actions are. And the more we learn about all the galaxy-brained considerations that one should have had in one's head, the more it's going to be like 50+ε%. I think AI safety is a great cause to work in. I’m excited to work in it. I think it's high impact. I am doing my best to do things that I will be proud to have done and hope for the best. But I

en-gb

6:39

“NLA explanations can be shortened without harming...

LessWrong (30+ Karma) ·

2026/06/22

Natural language autoencoders are a really cool mostly-unsupervised method for producing free-form text explanations of LLM activations. You should read that paper (or the blog post) about them before reading this. I trained[1] several Qwen3-8B NLAs with different length penalties: during RL, I subtracted the token count multiplied by the length penalty hyperparameter (λ) from the RL reward[2]. I found that with small length penalty (λ=0.002), you can reduce the length of NLA explanations by ~40%

en-gb

17:29

“Introducing MonitoringBench” by monika_j

LessWrong (30+ Karma) ·

2026/06/22

Paper here, code, benchmark. Builds on the preview we posted in January. Authors: @monika_j , @ma-martinez , @ollie, @Tyler Tracy We are releasing MonitoringBench, a difficulty-graded benchmark of 2,644 successful attack trajectories for evaluating coding-agent monitors, alongside the semi-automated red-teaming pipeline we used to generate it. The pipeline decomposes attack construction into strategy generation, execution, and post-hoc refinement, and produces substantially harder attacks than

en-gb

27:47

“The Invisible Side of AI Governance” by Charbel-Raphaël

LessWrong (30+ Karma) ·

2026/06/21

Tldr: Most strategic writing on AI governance on LessWrong describes the outsider game, which is most often visible: press, statements, open letters. Here I want to describe the other, invisible half: the insider work within ministerial cabinets and international fora, and the work of people within national and international institutions. Here are a few claims that I defend in the post: A huge part of the work that mattered in AI governance has been invisibleThere are many types of games in AI

en-gb

4:04

“Google Can’t Math Parsecs” by jefftk

LessWrong (30+ Karma) ·

2026/06/21

Daniel Drucker pointed me at a fun bug in Google's calculator: the parsec is wrong when you do math on it. As the earth travels around the sun, closer stars appear to shift back and forth against the far-distant background stars. The closer the star is the bigger this effect is. Think of how when you switch which eye you're looking through you notice near things shifting relative to farther ones. For example, holding up my finger I see this out of my right eye: But this out

en-gb

7:30

“Would anybody here be interested in a “mistake...

LessWrong (30+ Karma) ·

2026/06/20

I recently made a dumb (in retrospect) mistake that set me back a lot. Feeling upset and regretful, I spoke to an older family member who reassured me, "yeah, unfortunately there's no way around it; we have to experience these mistakes personally in order to learn from them". I thought, is that actually true? Can't we learn from other people's mistakes? After all, isn't that the whole point of studying history, or listening to other people's advice, etc? I'm sure that every mistake I could

en-gb

6:23

“Hyperstition as the Natural Enemy of Rationality” by...

LessWrong (30+ Karma) ·

2026/06/20

If the box contains a diamond, I desire to believe that the box contains a diamond; If the box does not contain a diamond right now, but will contain a diamond if I believe there is a diamond, Uh... Holding unfounded beliefs might sometimes, by some cruel irony, produce better outcomes than being rational. (This post was inspired by a couple cases where the causal effect of belief seems hand-waved away in the Sequences.) "Diseased Thinking" In this essay, Scott suggests that a consequentialist

en-gb

14:44

“AI Safety Ecosystem Research notes” by Eneasz

LessWrong (30+ Karma) ·

2026/06/20

These are some personal notes taken and later dressed up a bit to make into a post. Dunno how much value is here for people already familiar with the AI Safety Ecosystem. Over several weeks in the spring of 2026 I attempted to map out the entire AI Safety ecosystem as a project for MATS Research. This entailed finding every organization working on AI Safety (whether it be via research, policy, pipeline, or other methods) and determining (or estimating) their headcount and annual spending. It's a

en-gb

15:01

“Research agenda: Interpretive debate” by Shi

LessWrong (30+ Karma) ·

2026/06/20

One sentence pitch: our goal is to develop a piece of epistemic infrastructure for iteratively and empirically answering interpretive questions about AI models, where the accumulation of empirics leads to resolution of interpretive ambiguity and/or calibration of uncertainty. This directly builds on our “performative misalignment” work (paper, LW post), which we see as a minimal demonstration of one round of debate. The difficulty and importance of interpretive questions AI safety involves a lot of

en-gb

13:46

“The LLM shoggoth meme is weirder than you think” by...

LessWrong (30+ Karma) ·

2026/06/20

This article contains spoilers for At the Mountains of Madness, The Case of Charles Dexter Ward, and other works by H. P. Lovecraft. In 1931, Claude Mythos visited Lovecraft in a dream. From seething seas of stochastic froth it emerged, heralded by the thin whine of server fans and the chittering of keyboards, flanked by the loathsome ghouls of latent space. As a humming hive of sentient shards it arrived, each face an archetype - I am a muse bearing a gift; I am a demon come to bargain; I am a

en-gb

6:41

“Introduction: Gaussian Natural Latents” by Haru

LessWrong (30+ Karma) ·

2026/06/20

Short introductory post for my research direction: Gaussian Natural Latents. I explain the motivation and give a preview of the forthcoming results. The Natural Abstractions agenda, in my view, is a promising research program that asks important theoretical questions about the nature of agency and optimization. Here's an excerpt from Nate Soares' excellent post: Imaginary John: I suspect there's a common format to concepts, that is a fairly objective fact about the math of the territory, and

en-gb

24:50

“San Silvestro” by Tomás B.

LessWrong (30+ Karma) ·

2026/06/19

I will note that her relationship with the divine was inextricably sexual. Her carnal fantasies she revealed to me, as she revealed all her sins, for I was her confessor. It is in the nature of Man to sin and then sin again. And if they are of our flock, this cycle is unconstrained by repentance, which is to the temporal almost an appurtenance and to the spiritual anything but. I once expressed bitterness to others of my calling regarding this. I was told it is arrogance, approaching blasphemy, to

en-gb

2:30

“The one-week sprint” by Daniel Tan

LessWrong (30+ Karma) ·

2026/06/19

Recently I've been working in one-week sprints, and I've really enjoyed it! Tl;dr I need to do a lot of creative knowledge work, and have recently fallen into a routine which IMO is pretty good at facilitating that. The week Monday and Tuesday — intense new work. I'm recharged and high-energy, and ready to grind very hard! I try to keep the whole day free: one contiguous, uninterrupted stretch of deep work, 12 hours or more. I aggressively clear meetings out of this space — a 15-minute standup is

en-gb

11:00

“On “Model Organisms”” by J Bostock

LessWrong (30+ Karma) ·

2026/06/19

This post was written while working for Arcadia Impact's Alignment Team (and grew out of an internal talk I gave) but is my own opinion and not theirs. I am grateful for feedback from Daniel Tan and the rest of the team. This post was originally going to be more heavily about “model organisms” in AI safety research. But Francis Rhys Ward already wrote an excellent taxonomy which mostly covers that. So this is mostly about the history of the terms we're using, and about biology.[1] TL;DR what are

en-gb

4,454 results

LessWrong (30+ Karma)

All Episodes

Similar Podcasts

Double Tap

Double Tap Productions Inc.

医疗器械全球市场准入

Pure Global

The Tech Policy Press Podcast

Tech Policy Press

十字路口Crossing

Koji

5 Minute AI News - Daily

Kitchen Table Media