PODCAST

LessWrong (30+ Karma)

LessWrong

Audio narrations of LessWrong posts.

technology philosophy

All Episodes

6:53

“AI Safety Can’t Afford a Second Cause” by atlasaligned

LessWrong (30+ Karma) ·

2026/07/08

Imagine an astronomer who discovers an asteroid with a 50% chance of hitting Earth in 2035. She goes on TV. She testifies before Congress. She founds the Asteroid Deflection Institute and starts doing fundraising rounds. And then, in between appearances, she prolifically tweets about zoning reform. I think that most people agree that this would be pretty weird. Her zoning takes might even be correct, but why would you do that in the first place? If she has attention left over for zoning, how real

en-gb

10:12

“Superhuman Articulacy as an LLM Safety Target” by...

LessWrong (30+ Karma) ·

2026/07/08

TL;DR: Current LLMs are bad communicators relative to their agentic capabilities. I claim that articulacy is useful (and perhaps necessary) for AI safety and suggest a path for improving articulacy. Briefly: a theory for articulacy Frequently, LLM agents miscommunicate with their human operators, such as when they write documentation or respond to queries about their activity during a coding session. Any given communication failure can be ascribed to either or both of these two factors: Articulacy

en-gb

8:00

“Experiments With Fabel’s Fiction” by Tomás B.

LessWrong (30+ Karma) ·

2026/07/08

In keeping with my tradition and given I will lose access to Fabel in a couple days, I have asked fable to read all my organically written short stories and attempt to write one in my style. It output the following: There's LLM-written text here. Claude Fable writes:en-US-AvaMultilingualNeural__ The Last Complainer I want it on the record that I complained about the job before I took it. This is, if you think about it, the strongest possible qualification, and the Ministry thought about it,

en-gb

35:22

“No Space Like J-Space” by Zvi

LessWrong (30+ Karma) ·

2026/07/07

There is a new very cool Anthropic paper: Verbalizable Representations Form a Global Workspace in Language Models. You can read the blog post verison here. I encourage reading of the whole original blog post or paper, if you have the time. Table of Contents Through A Different Lens. Establishing J-Space As A Global Workspace. Are You Pondering What I’m Pondering? Assistant J. The Power Of Virtuous Thinking. High Praise. Everyone Remains Confused About Consciousness. Further

en-gb

5:44

“Current views on large-scale longtermist...

LessWrong (30+ Karma) ·

2026/07/07

This post summarizes some of my views. It states uncertain things boldly in part to [surface disagreements / help readers point out things that seem wrong or missing]. Note that many things are uncertain and many claims are not justified here. One tl;dr: money is great even on the margin, and it's crucial to invest extremely well rather than merely well. In 2026, philanthropists will donate about 1.6 billion dollars to AI safety nonprofits (being slightly picky about what counts) (~1.2 billion

en-gb

33:25

“Data filtering works a lot worse than you would...

LessWrong (30+ Karma) ·

2026/07/07

This work was largely done during Neel Nanda's MATS 10.0 Exploration Phase. J Rosser and Dohun Lee are co-first authors for this post with equal contribution. Josh Engels and Neel Nanda supervised the project, and provided guidance and feedback throughout. TLDR Models can acquire undesirable traits from during supervised fine-tuning (SFT). A natural thing to try is to identify the data points with these traits and filter them out and retrain.To our surprise, across most of our broad OLMo SFT

en-gb

16:41

“A conceptor by any other name” by Keenan Pepper

LessWrong (30+ Karma) ·

2026/07/07

I just had one of those delightful moments where I have a very specific idea, and then I search for it (Claude Research in this case), and it turns out that people have already been using that exact concept but just calling it by a name I'd never heard of. That name is conceptor, but the basic construction is not new and has gone by many other names. What a conceptor is I'll quote from "Conceptors for Semantic Steering" (Triantafyllopoulos et al.) since I couldn't say it any better myself: Given a

en-gb

37:45

“Some Important Models for Health and Fitness” by benwr

LessWrong (30+ Karma) ·

2026/07/07

This is a synthesis of many facts I've learned over the last few years, mainly about metabolism and exercise,[1] that have helped me become much healthier, and might help you. It's focused mainly on “basic” models that, in my opinion, high school health class ought to cover (though mine emphatically did not). Because of various personality diso... traits, simply doing what people tell me to do basically never works. In order to achieve things that are easy for many people, I tend to need to (a)

en-gb

5:23

“Claude Code as a Claude Coach” by Brendan Long

LessWrong (30+ Karma) ·

2026/07/07

Exercise is hard but it's even harder if you have to use your brain and muscles at the same time. I wish a personal trainer would just teleport into my house whenever I work out, tell me exactly what to do, and then record my progress (and complaints) to improve the program going forward. Apps are too rigid or too complicated; personal trainers are expensive and require scheduling, but using Claude Code as a personal trainer has worked out well for me. I'm still inconsistent, but more consistent

en-gb

15:40

“Bounding eval awareness of ~human-level AI across...

LessWrong (30+ Karma) ·

2026/07/07

In our last post, we argued that measuring evaluation awareness is fundamentally challenging because of the safe-to-dangerous distributional shift: we cannot directly measure the evaluation awareness of a model without deploying it, but we cannot safely deploy it until we know it is not scheming. We expect sufficiently superhuman AI will be eval aware, but this post outlines a tentative solution for bounding the awareness of a ~human-expert-level[1] AI across this safe-to-dangerous shift: Instead

en-gb

26:26

“Desiderata for functional welfare experiments on...

LessWrong (30+ Karma) ·

2026/07/06

TLDR LLMsappear to havefunctional welfare:coherent sets ofbehaviourthat track how well things are going relative to their goals.Improving model functional welfare matters for safety (low welfare may amplify misalignment) and for moral reasons (models may be or become moral patients).Naive interventions can fail in non-obvious ways. We argue any successful intervention must: A)shift multiple welfare-constituting channels together and coherently.B)avoidcorruptingthe model's ability to register

en-gb

50:56

“A Review of Anthropic’s Global Workspace Paper” by...

LessWrong (30+ Karma) ·

2026/07/06

The below is a public review Anthropic asked me to write for their new global workspace paper. I recommend at least skimming their paper first. TLDR: I think this is a fantastic paper - it presents compelling evidence for some kind of "cognitive space" in models, that is used as a "working memory" for intermediate variables during a forward pass, shows that J-Lens is a useful technique for accessing this space. I believe these key claims.I believe J-Lens will be a useful (but limited) tool in

en-gb

9:45

“SFF is very suboptimal” by Zach Stein-Perlman

LessWrong (30+ Karma) ·

2026/07/06

I recently served as a recommender in SFF's annual funding round (grants will be decided and announced in September). I'm deeply grateful for SFF's funders, and I hope more AI safety donors appear in the future. Unfortunately, the SFF experience is bad for both applicants and recommenders, it's slow, and it lacks some other desiderata. (The actual funding decisions are moderately good by my lights — probably over half of the value of grants that a better system with the same budget would produce,

en-gb

23:28

“Visioning: Concretely Imagining What You Want” by...

LessWrong (30+ Karma) ·

2026/07/06

When John told me (Gretta) his practice of “visioning,” I was skeptical at first. I gave it a try, a little bit out of spite, to show him I was capable of it. It blew me away. Here's what I told him about it a few days later, edited for brevity and clarity. Something internal really felt different. Awake, alive, hungry, energetic. Selfish. A little angry. Do you know the 1984 Apple Macintosh ad? Ignore the specific words and politics and computer-adness of it all, and just feel the energy of the

en-gb

33:04

“A global workspace in language models” by wesg

LessWrong (30+ Karma) ·

2026/07/06

[This is the blog post for our new paper Verbalizable Representations Form a Global Workspace in Language Models Readers might also be interested in: the Public commentary, Github and Neuronpedia] As you read this sentence, circuits in your brain are adjusting your posture, controlling your breathing, and transforming lines and curves on the screen into recognizable words. Most of this processing is invisible to you. But some of what takes place in your brain you do have access to—an image that

en-gb

7:19

“Sub-agent delegation chaining” by David Rein

LessWrong (30+ Karma) ·

2026/07/06

Epistemic status: pretty confident in the validity of the core proposal, not that confident in specific implementation details TL;DR: we should cryptographically verify that sub-agent instances/sessions are downstream of human instructions Frontier AI labs have started using LLM-based monitoring systems to check for misbehavior from their internal AI agents, which often run unwatched by humans for hours or days. One central reason to be concerned about rogue internal deployments, where AIs subvert

en-gb

31:09

“Tie training can make DPO/RLHF-trained AIs...

LessWrong (30+ Karma) ·

2026/07/06

This post covers our recent ICML paper: Spurious Correlation Learning in Preference Optimization: Mechanisms, Consequences, and Mitigation via Tie Training. TL;DR Our theorems and experiments suggest that DPO and RLHF have an unwelcomeconsequence: they make AIs care about everyfeature of actions that correlates with true value on the training distribution.[1] That's true even if the training set contains no misspecified preference data.And it's true even in the infinite-data limit.So AIs trained

en-gb

16:11

“Claude’s malicious compliance and normalization of...

LessWrong (30+ Karma) ·

2026/07/06

It was January 27, 1986, the night before the Space Shuttle Challenger was scheduled for launch. The goal was to have a shuttle that could land back on Earth and be reused for future missions; its first-planned priorities were satellite deployment, comet observation, and science education. The last of these would involve students from around the world watching live broadcasts from Christa McAuliffe, the first civilian teacher in space. (The following has been lightly edited:) The forecast for the

en-gb

35:04

“We need 3rd party Training-Run Assessments” by Alex...

LessWrong (30+ Karma) ·

2026/07/05

Training-run assessments conducted by a 3rd party should become a standard part of frontier AI safety. By a Training-Run Assessment, or TRA, I mean an in-depth analysis of the post-training pipeline and dynamics leading up to a frontier model release. A TRA can look at intermediate checkpoints, training rollouts, RL environments, reward signals, SFT datasets, and the process by which the developer responded to warning signs.[1] In this post I will argue that: Final-checkpoint evaluations will be

en-gb

6:13

“Harry Potter and the Rules of Quidditch” by Tomás B.

LessWrong (30+ Karma) ·

2026/07/05

Ron's face pulled into a scowl. "If you don't like Quidditch, you don't have to make fun of it!" "If you can't criticise, you can't optimise. I'm suggesting how to improve the game. And it's very simple. Get rid of the Snitch." "They won't change the game just 'cause you say so!" "I am the Boy-Who-Lived, you know. People will listen to me. And maybe if I can persuade them to change the game at Hogwarts, the innovation will spread." A look of absolute horror was spreading over Ron's face. "But, but

en-gb

4,451 results

LessWrong (30+ Karma)

All Episodes

Similar Podcasts

Reflections

Livewire Labs

Netz-Barrierefrei - Digitale Barrierefreiheit verstehen

netz-barrierefrei.de

Life and Technology with Charlie Brown

Tapt Media

LessWrong (30+ Karma)

LessWrong

Hacker Newsroom

pod pub