PODCAST

LessWrong (30+ Karma)

LessWrong

Audio narrations of LessWrong posts.

technology philosophy

All Episodes

12:57

“Is 90% of code at Anthropic being written by AIs?”...

LessWrong (30+ Karma) ·

2025/10/22

In March 2025, Dario Amodei (CEO of Anthropic) said that he expects AI to be writing 90% of the code in 3 to 6 months and that AI might be writing essentially all of the code in 12 months.[1] Did this prediction end up being true? We'll charitably interpret the prediction as being about Anthropic itself. Recently, various articles have claimed that Dario said that AIs are writing 90% of the code at Anthropic (7 months after the prediction of 3-6 months was made). This is a reasonable summary of

en-gb

30:11

“Stratified Utopia” by Cleo Nardo

LessWrong (30+ Karma) ·

2025/10/22

Summary: "Stratified utopia" is a possible outcome where mundane values get proximal resources (near Earth in space and time) and exotic values get distal resources (distant galaxies and far futures). I discuss whether this outcome is likely or desirable. 1. Introduction 1.1. Happy Coincidence I hold mundane values, such as partying on the weekend, the admiration of my peers, not making a fool of myself, finishing this essay, raising children, etc. I also have more exotic values, such as maximizing

en-gb

2:58

“Remarks on Bayesian studies from 1963” by dynomight

LessWrong (30+ Karma) ·

2025/10/22

In 1963, Mosteller and Wallace published Inference in an Authorship Problem, which used Bayesian statistics to try to infer who wrote some of the disputed Federalist Papers. (Answer: Madison) Anyway, at the end they have a list of "Remarks on Bayesian studies" which is astonishing to read 62 years later: Study of variation of results with different priors is recommended. Bracketing the prior is often easy. When founded in data, the choice of the prior has a status like that of the data

en-gb

2:40

[Linkpost] “21st Century Civilization curriculum” by...

LessWrong (30+ Karma) ·

2025/10/21

This is a link post. I’ve just released a curriculum on foundational questions in modern politics, which I drew up in collaboration with Samo Burja. I’ve copied the introductory text and the section headings below; you can find the full curriculum at www.21civ.com. Sign up here by 27 October to join the first cohort of discussion groups (which will meet weekly to discuss each of the 11 sections of the curriculum). This curriculum is about Western civilization, and how it enables citizens of the

en-gb

39:20

“Symbiogenesis vs. Convergent Consequentialism” by...

LessWrong (30+ Karma) ·

2025/10/21

(Cross-posted from SayIt archive.) Background for conversation: After an exchange in the comments of Audrey's LW post where plex suggested various readings and got a sense that there were some differences in models worth exploring, plex suggested a call. Some commentors were keen to read the transcript, and plex thinks in order for any human, other sentient, current AI, or currently existing structure to not be destroyed in the near future, we need to either coordinate to pivot from the rush

en-gb

16:48

“EU explained in 10 minutes” by Martin Sustrik

LessWrong (30+ Karma) ·

2025/10/21

If you want to understand a country, you should pick a similar country that you are already familiar with, research the differences between the two and there you go, you are now an expert. But this approach doesn’t quite work for the European Union. You might start, for instance, by comparing it to the United States, assuming that EU member countries are roughly equivalent to U.S. states. But that analogy quickly breaks down. The deeper you dig, the more confused you become. You try with other

en-gb

15:12

“Can you find the steganographically hidden message?”...

LessWrong (30+ Karma) ·

2025/10/20

tl;dr: I share a curated set of transcripts of models successfully executing message passing steganography from our recent paper. I then give a few thoughts on how I think about risks from this kind of steganography. Background I recently was a co-first author on a paper (LW link) where we evaluated the steganographic capabilities of frontier models. Model monitoring has emerged as a leading strategy for detecting instances of misuse or misalignment, and steganography is one plausible way that

en-gb

0:41

[Linkpost] “How Stuart Buck funded the replication...

LessWrong (30+ Karma) ·

2025/10/20

This is a link post. Stuart Buck has perhaps the largest shapely value of any one individual in creating the replication crisis first in psychology, and then in many fields. This is his personal account of how and why he made the choices he did. ---First published: October 20th, 2025 Source: https://www.lesswrong.com/posts/6it9vp8sWsYEcHSiv/how-stuart-buck-funded-the-replication-crisis Linkpost

en-gb

50:29

“Consider donating to Alex Bores, author of the RAISE...

LessWrong (30+ Karma) ·

2025/10/20

Written by Eric Neyman, in my personal capacity. The views expressed here are my own. Thanks to Zach Stein-Perlman, Jesse Richardson, and many others for comments. Over the last several years, I’ve written a bunch of posts about politics and political donations. In this post, I’ll tell you about one of the best donation opportunities that I’ve ever encountered: donating to Alex Bores, who announced his campaign for Congress today. If you’re potentially interested in donating to Bores, my suggestion

en-gb

27:27

“Considerations around career costs of political...

LessWrong (30+ Karma) ·

2025/10/20

I’m close to a single-issue voter/donor. I tend to like politicians who show strong support for AI safety, because I think it's an incredibly important and neglected problem. So when I make political donations, it's not as salient to me which party the candidate is part of, if they've gone out of their way to support AI safety and have some integrity.[1] I think many people who focus on AI safety feel similarly. But working in government also seems important. I want the government to have the tools

en-gb

12:47

“Scenes, cliques and teams - a high level ontology of...

LessWrong (30+ Karma) ·

2025/10/20

Ontological status: Yes, this is ontology Groups of people are one of the most important things. If I were to list all the things and rank them by importance, groups of people would be near the top. Love, truth and freedom and other such things might score higher from some angles, but these things are usually found in groups of humans anyway. And groups are complex. That's why I started using this categorisation of groups recently. I have found it very helpful in conversations about community and

en-gb

6:49

“Frontier LLM Race/Sex Exchange Rates” by Arjun...

LessWrong (30+ Karma) ·

2025/10/20

This is a cross-post (with permission) of Arctotherium's post yesterday "LLM Exchange Rates, Updated." It uses a similar methodology to the CAIS "Utility Engineering" paper, which showed e.g. "that GPT-4o values the lives of Nigerians at roughly 20x the lives of Americans, with the rank order being Nigerians > Pakistanis > Indians > Brazilians > Chinese > Japanese > Italians > French > Germans > Britons > Americans." Highlights from the linked post (emphasis is from

en-gb

8:46

“Humanity Learned Almost Nothing From COVID-19” by...

LessWrong (30+ Karma) ·

2025/10/19

Summary: Looking over humanity's response to the COVID-19 pandemic, almostsix years later, reveals that we've forgotten to fulfill our intent at preparing for the next pandemic. I rant. content warning: A single carefully placed slur. If we want to create a world free of pandemics and other biological catastrophes, the time to act is now. —US White House, “ FACT SHEET: The Biden Administration's Historic Investment in Pandemic Preparedness and Biodefense in the FY 2023 President's Budget ”,

en-gb

47:16

“The Dark Arts of Tokenization or: How I learned to...

LessWrong (30+ Karma) ·

2025/10/19

Audio note: this article contains 225 uses of latex notation, so the narration may be difficult to follow. There's a link to the original text in the episode description. Introduction There are <span>_208_</span> ways to output the text ▁LessWrong[1] with the Llama 3 tokenizer, but even if you were to work with Llama 3 for thousands of hours, you would be unlikely to see any but one. An example that generalizes quite widely: if you prompt Llama 3.2 3B Base with the text You're

en-gb

7:19

“Give Me Your Data: The Rationalist Mind Meld” by...

LessWrong (30+ Karma) ·

2025/10/19

I don’t want your rationality. I can supply my own, thank you very much. I want your data. If you spot a logical error in my thinking, then please point it out. But short of that, among mostly-rational people, I think most disagreements come down to a difference of intuitions, which are rooted in a difference in the data people have been exposed to, and instead of presenting a logical counter-argument, you’re better off doing a “Rationalist mind meld” where you share your data. I find myself making

en-gb

7:27

“Meditation is dangerous” by Algon

LessWrong (30+ Karma) ·

2025/10/18

Here's a story I've heard a couple of times. A youngish person is looking for some solutions to their depression, chronic pain, ennui or some other cognitive flaw. They're open to new experiences and see a meditator gushing about how amazing meditation is for joy, removing suffering, clearing one's mind, improving focus etc. They invite the young person to a meditation retreat. The young person starts making decent progress. Then they have a psychotic break and their life is ruined for years, at

en-gb

4:36

“I’m an EA who benefitted from rationality” by juliawise

LessWrong (30+ Karma) ·

2025/10/17

This is my personal take, not an organizational one. Originally written May 2025, revived for the EA Forum's Draft Amnesty Week. Cross-posted from the EA Forum but cross-posting didn't work the first time. When I discovered the beginnings of EA, it felt like coming home. I had finally found a group of people who cared as much as I did about saving lives. When I discovered rationality not long after, it felt new. The intensity of LessWrong was both offputting and fascinating to me, and I read a

en-gb

11:07

“Finding Features in Neural Networks with the...

LessWrong (30+ Karma) ·

2025/10/17

Audio note: this article contains 63 uses of latex notation, so the narration may be difficult to follow. There's a link to the original text in the episode description. Summary Kernel regression with the empirical neural tangent kernel (eNTK) gives a closed-form approximation to the function learned by a neural network in parts of the model space. We provide evidence that the eNTK can be used to find features in toy models for interpretability. We show that in Toy Models of Superposition and a

en-gb

20:08

“Reducing risk from scheming by studying trained-in...

LessWrong (30+ Karma) ·

2025/10/17

In a previous post, I discussed mitigating risks from scheming by studying examples of actual scheming AIs.[1] In this post, I'll discuss an alternative approach: directly training (or instructing) an AI to behave how we think a naturally scheming AI might behave (at least in some ways). Then, we can study the resulting models. For instance, we could verify that a detection method discovers that the model is scheming or that a removal method removes the inserted behavior. The Sleeper Agents paper

en-gb

11:39

“Rogue internal deployments via external APIs” by...

LessWrong (30+ Karma) ·

2025/10/16

Once AI companies build powerful AIs, they may: Give internal AIs access to sensitive internal privileges (e.g. access to the internal infra that touches model weights, help with training of the next generation of AIs, …); Have mitigations specific to those internal deployments enforced via monitoring of the internal LLM API; Have an external API deployment without those mitigations. In this situation, an internal AI may build an agent scaffold that makes calls to the external API and uses the

en-gb

2,881 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