PODCAST

LessWrong (30+ Karma)

LessWrong

Audio narrations of LessWrong posts.

technology philosophy

All Episodes

4:42

“Supervillain Monologues Are Unrealistic” by Algon

LessWrong (30+ Karma) ·

2025/11/01

Supervillain monologues are strange. Not because a supervillain telling someone their evil plan is weird. In fact, that's what we should actually expect. No, the weird bit is that people love a good monologue. Wait, what? OK, so why should we expect supervillains to tell us their master plan? Because a supervillain is just a cartoonish version of a high agency person doing something the author thinks is bad. So our real life equivalent of a supervillain, or rather the generalization of one, is just

en-gb

6:17

“Anthropic’s Pilot Sabotage Risk Report” by dmz

LessWrong (30+ Karma) ·

2025/11/01

As practice for potential future Responsible Scaling Policy obligations, we're releasing a report on misalignment risk posed by our deployed models as of Summer 2025. We conclude that there is very low, but not fully negligible, risk of misaligned autonomous actions that substantially contribute to later catastrophic outcomes. We also release two reviews of this report: an internal review and an independent review by METR. Our Responsible Scaling Policy has so far come into force primarily to

en-gb

37:36

“OpenAI Moves To Complete Potentially The Largest...

LessWrong (30+ Karma) ·

2025/10/31

OpenAI is now set to become a Public Benefit Corporation, with its investors entitled to uncapped profit shares. Its nonprofit foundation will retain some measure of control and a 26% financial stake, in sharp contrast to its previous stronger control and much, much larger effective financial stake. The value transfer is in the hundreds of billions, thus potentially the largest theft in human history. I say potentially largest because I realized one could argue that the events surrounding the

en-gb

9:12

“Resampling Conserves Redundancy & Mediation...

LessWrong (30+ Karma) ·

2025/10/31

Audio note: this article contains 86 uses of latex notation, so the narration may be difficult to follow. There's a link to the original text in the episode description. Around two months ago, John and I published Resampling Conserves Redundancy (Approximately). Fortunately, about two weeks ago, Jeremy Gillen and Alfred Harwood showed us that we were wrong. This proof achieves, using the Jensen-Shannon divergence ("JS"), what the previous one failed to show using KL divergence

en-gb

30:26

“Steering Evaluation-Aware Models to Act Like They...

LessWrong (30+ Karma) ·

2025/10/30

📄Paper, 🖥️Code, 🤖Evaluation Aware Model Organism TL, DR:; We train an evaluation-aware LLM.Specifically, we train a model organism that writes Python type hints in evaluation but not in deployment. Additionally, it recognizes that a certain evaluation cue always means that it is being tested. The evaluation cue is intended to simulate hard-to-remove signs of an eval being artificial, e.g. being LLM-written. Our training process was designed to mimic how evaluation awareness could arise in

en-gb

2:04

[Linkpost] “AISLE discovered three new OpenSSL...

LessWrong (30+ Karma) ·

2025/10/30

This is a link post. The company post is linked; it seems like an update on where we are with automated cybersec. So far in 2025, only four security vulnerabilities received CVE identifiers in OpenSSL, the cryptographic library that secures the majority of internet traffic. AISLE's autonomous system discovered three of them. (CVE-2025-9230, CVE-2025-9231, and CVE-2025-9232) Some quick thoughts: OpenSSL is one of the most human-security-audited pieces of open-source code ever, so discovering 3 new

en-gb

35:58

“Sonnet 4.5’s eval gaming seriously undermines...

LessWrong (30+ Karma) ·

2025/10/30

According to the Sonnet 4.5 system card, Sonnet 4.5 is much more likely than Sonnet 4 to mention in its chain-of-thought that it thinks it is being evaluated; this seems to meaningfully cause it to appear to behave better in alignment evaluations. So, Sonnet 4.5's behavioral improvements in these evaluations may partly be driven by growing tendency to notice and game evaluations rather than genuine alignment. This is an early example of a phenomenon that is going to get increasingly problematic: as

en-gb

8:04

“ImpossibleBench: Measuring Reward Hacking in LLM...

LessWrong (30+ Karma) ·

2025/10/30

This is a post about our recent work ImpossibleBench: Measuring LLMs' Propensity of Exploiting Test Cases (with Aditi Raghunathan, Nicholas Carlini) where we derive impossible benchmarks from existing benchmarks to measure reward hacking.Figure 1: Overview of the ImpossibleBench framework. We start with tasks from established coding benchmarks and create impossible variants by mutating test cases to conflict with natural language specifications. The resulting cheating rate serves as a direct

en-gb

3:01

[Linkpost] “Emergent Introspective Awareness in Large...

LessWrong (30+ Karma) ·

2025/10/30

This is a link post. New Anthropic research (tweet, blog post, paper): We investigate whether large language models can introspect on their internal states. It is difficult to answer this question through conversation alone, as genuine introspection cannot be distinguished from confabulations. Here, we address this challenge by injecting representations of known concepts into a model's activations, and measuring the influence of these manipulations on the model's self-reported states. We find that

en-gb

8:00

“An Opinionated Guide to Privacy Despite...

LessWrong (30+ Karma) ·

2025/10/29

I've created a highly specific and actionable privacy guide, sorted by importance and venturing several layers deep into the privacy iceberg. I start with the basics (password manager) but also cover the obscure (dodging the millions of Bluetooth tracking beacons which extend from stores to traffic lights; anti-stingray settings; flashing GrapheneOS on a Pixel). I feel strongly motivated by current events, but the guide also contains a large amount of timeless technical content. Here's a

en-gb

17:39

“The End of OpenAI’s Nonprofit Era” by garrison

LessWrong (30+ Karma) ·

2025/10/29

Key regulators have agreed to let the company kill its profit caps and restructure as a for-profit — with some strings attached This is the full text of a post first published on Obsolete, a Substack that I write about the intersection of capitalism, geopolitics, and artificial intelligence. I’m a freelance journalist and the author of a forthcoming book called Obsolete: Power, Profit, and the Race to Build Machine Superintelligence. Consider subscribing to stay up to date with my work. This

en-gb

13:15

“Please Do Not Sell B30A Chips to China” by Zvi

LessWrong (30+ Karma) ·

2025/10/29

The Chinese and Americans are currently negotiating a trade deal. There are plenty of ways to generate a win-win deal, and early signs of this are promising on many fronts. Since this will be discussed for real tomorrow as per reports, I will offer my thoughts on this one more time. The biggest mistake America could make would be to effectively give up Taiwan, which would be catastrophic on many levels including that Taiwan contains TSMC. I am assuming we are not so foolish as to seriously

en-gb

22:12

“AI Craziness Mitigation Efforts” by Zvi

LessWrong (30+ Karma) ·

2025/10/29

AI chatbots in general, and OpenAI and ChatGPT and especially GPT-4o the absurd sycophant in particular, have long had a problem with issues around mental health. I covered various related issues last month. This post is an opportunity to collect links to previous coverage in the first section, and go into the weeds on some new events in the later sections. A lot of you should likely skip most of the in-the-weeds discussions. What Are The Problems There are a few distinct phenomena we have

en-gb

15:43

“When Will AI Transform the Economy?” by Andre.Infante

LessWrong (30+ Karma) ·

2025/10/29

Substack version here: https://andreinfante.substack.com/p/when-will-ai-transform-the-economy A caricature of a common Twitter argument: ”Hey it seems like language models can do a lot of humanlike cognitive tasks, look at this impressive example! That seems like something that would transform the economy.” “Well, smart guy, these things have been out for a couple of years now. Why isn’t the economy transformed yet?” This does strike me as a little bit mysterious. The models can do impressive

en-gb

14:05

“Some data from LeelaPieceOdds” by Jeremy Gillen

LessWrong (30+ Karma) ·

2025/10/29

I've been curious about how good LeelaPieceOdds is, so I downloaded a bunch of data and graphed it. For context, Leela is a chess bot and this version of it has been trained to play with a handicap.This is BBNN odds, meaning Leela starts without bishops and knights. I first heard about LeelaQueenOdds from simplegeometry's comment, and I've been playing against it occasionally since then. LeelaPieceOdds is the version with more varied handicap options, which is more interesting to me because

en-gb

3:59

“Workshop on Post-AGI Economics, Culture, and...

LessWrong (30+ Karma) ·

2025/10/29

This is an announcement and call for applications to the Workshop on Post-AGI Economics, Culture, and Governance taking place in San Diego on Wednesday, December 3, overlapping with the first day of NeurIPS 2025. This workshop aims to bring together a diverse range of expertise to deepen our collective understanding of how the world might evolve after the development of AGI, and what we can do about it now. The draft program features: Anton Korinek on the Economics of Transformative AI Alex

en-gb

1:56

“Introducing the Epoch Capabilities Index (ECI)” by...

LessWrong (30+ Karma) ·

2025/10/29

We at Epoch AI have recently released a new composite AI capability index called the Epoch Capabilities Index (ECI), based on nearly 40 underlying benchmarks. Some key features... Saturation-proof: ECI "stitches" benchmarks together, to enable comparisons even as individual benchmarks become saturated. Global comparisons: Models can be compared, even if they were never evaluated on the same benchmarks. Difficulty-based task weighting: ECI uses a simple statistical model (similar to those used in

en-gb

15:56

“Mottes and Baileys in AI discourse” by Raemon

LessWrong (30+ Karma) ·

2025/10/29

This post kinda necessarily needs to touch multiple political topics at once. Please, everyone, be careful. If it looks like you haven't read the LessWrong Political Prerequisites, I'm more likely than usual to delete your comments. I think some people are (rightly) worried about a few flavors of Motte and Baily-ing with the IABIED discourse, and more recently, with the Superintelligence Statement. With IABIED: "Sure, the motte is 'If any company or group, anywhere on the planet, builds an

en-gb

9:04

“LLM robots can’t pass butter (and they are having an...

LessWrong (30+ Karma) ·

2025/10/29

TLDR: Andon Labs, evaluates AI in the real world to measure capabilities and to see what can go wrong. For example, we previously made LLMs operate vending machines, and now we're testing if they can control robots at offices. There are two parts to this test: We deploy LLM-controlled robots in our office and track how well they perform at being helpful. We systematically test the robots on tasks in our office. We benchmark different LLMs against each other. You can read our paper "Butter-Bench"

en-gb

21:28

“The Memetics of AI Successionism” by Jan_Kulveit

LessWrong (30+ Karma) ·

2025/10/28

TL;DR: AI progress and the recognition of associated risks are painful to think about. This cognitive dissonance acts as fertile ground in the memetic landscape, a high-energy state that will be exploited by novel ideologies. We can anticipate cultural evolution will find viable successionist ideologies: memeplexes that resolve this tension by framing the replacement of humanity by AI not as a catastrophe, but as some combination of desirable, heroic, or inevitable outcome. This post mostly

en-gb

2,933 results

LessWrong (30+ Karma)

All Episodes

Similar Podcasts

Lead With AI

Tamara Nall

Linux in the Ham Shack

Black Sparrow Media

Data Engineering Podcast

Tobias Macey

Eye On A.I.

Craig S. Smith

Reflections

Livewire Labs