PODCAST

LessWrong (30+ Karma)

LessWrong

Audio narrations of LessWrong posts.

technology philosophy

All Episodes

14:28

“The Best Way to Align an LLM: Inner Alignment is Now...

LessWrong (30+ Karma) ·

2025/05/29

This is a link-post for an exciting paper I recently read: Safety Pretraining: Toward the Next Generation of Safe AI by Pratyush Maini, Sachin Goyal, et al. For a couple of years I (and others) have been proposing an approach to alignment: what the authors of this recent paper name "safety pretraining". In a nutshell: that it's best to apply your alignment training as part of the standard pretraining process to produce a base model that is already aligned — simply pretrain it on a lot of examples

en-gb

15:23

“LessWrong Feed [new, now in beta]” by Ruby

LessWrong (30+ Karma) ·

2025/05/28

The modern internet is replete with feeds such as Twitter, Facebook, Insta, TikTok, Substack, etc. They're bad in ways but also good in ways. I've been exploring the idea that LessWrong could have a very good feed. I'm posting this announcement with disjunctive hopes: (a) to find enthusiastic early adopters who will refine this into a great product, or (b) find people who'll lead us to an understanding that we shouldn't launch this or should launch it only if designed a very specific way. You can

en-gb

54:49

“Shift Resources to Advocacy Now (Post 4 of 6 on AI...

LessWrong (30+ Karma) ·

2025/05/28

In my previous post in this series, I estimated that we have 3 researchers for every advocate working on US AI governance, and I argued that this ratio is backwards. When allocating staff, you almost always want to have more people working on the more central activity. I argued that in the case of AI policy, the central activity is advocacy, not research, because the core problem to be solved is fixing the bad private incentives faced by AI developers. As I explained, the problem with these

en-gb

4:38

[Linkpost] “If you’re not sure how to sort a list or...

LessWrong (30+ Karma) ·

2025/05/28

This is a link post. "Getting Things in Order: An Introduction to the R Package seriation":Seriation [or "ordination"), i.e., finding a suitable linear order for a set of objects given data and a loss or merit function, is a basic problem in data analysis. Caused by the problem's combinatorial nature, it is hard to solve for all but very small sets. Nevertheless, both exact solution methods and heuristics are available. In this paper we present the package seriation which provides an

en-gb

19:23

“Beware the Moral Homophone” by ymeskhout

LessWrong (30+ Karma) ·

2025/05/28

Or "Why you should prioritize attacking your allies before anyone else" Homophones are words that look or sound exactly alike, but convey completely different meanings. It's how bear can refer to the large forest animal or the act of carrying a burden, how light can be a weight designation or a source of illumination, or how spring can be a climate season or a compression device. Homophonic confusion is not necessarily a problem, because context usually gives us enough clues to scry the word's

en-gb

5:05

“Briefly analyzing the 10-year moratorium amendment”...

LessWrong (30+ Karma) ·

2025/05/28

This is the result of a half-day research sprint on the recently-introduced amendment to institute a 10-year moratorium on state-level AI regulations in the current budget reconciliation bill, with a focus on figuring out "is it likely to survive the Byrd Rule". It seems quite obviously in violation of the Byrd Rule because it violates at least 2 of the 6 tests: It has no budget effect, i.e. does not change outlays or revenues. (Probably this is not actually true; I think on net we would expect it

en-gb

31:48

“What We Learned from Briefing 70+ Lawmakers on the...

LessWrong (30+ Karma) ·

2025/05/28

Between late 2024 and mid-May 2025, I briefed over 70 cross-party UK parliamentarians. Just over one-third were MPs, a similar share were members of the House of Lords, and just under one-third came from devolved legislatures — the Scottish Parliament, the Senedd, and the Northern Ireland Assembly. I also held eight additional meetings attended exclusively by parliamentary staffers. While I delivered some briefings alone, most were led by two members of our team. I did this as part of my work as a

en-gb

5:47

“Requiem for the hopes of a pre-AI world” by...

LessWrong (30+ Karma) ·

2025/05/27

A few months from now, I turn 55. I've been a transhumanist since my teens in the late 1980s; since I got online in the 1990s, I have participated remotely in the talking shops and virtual salons of Internet transhumanism and, later, rationalism. The upheavals of 21st century politics have provided many distractions, but I have never abandoned the view that it is possible and desirable to reach for something more than the natural human condition. At the very least, one should try to reverse the

en-gb

13:25

“Season Recap of the Village: Agents raise $2,000” by...

LessWrong (30+ Karma) ·

2025/05/27

Four agents woke up with four computers, a view of the world wide web, and a shared chat room full of humans. Like Claude plays Pokemon, you can watch these agents figure out a new and fantastic world for the first time. Except in this case, the world they are figuring out is our world. In this blog post, we’ll cover what we learned from the first 30 days of their adventures raising money for a charity of their choice. We’ll briefly review how the Agent Village came to be, then what the various

en-gb

1:47

“Association taxes are collusion subsidies” by...

LessWrong (30+ Karma) ·

2025/05/27

Under present norms, if Alice associates with Bob, and Bob is considered objectionable in some way, Alice can be blamed for her association, even if there is no sign she was complicit in Bob's sin. An interesting upshot is that as soon as you become visibly involved with someone, you are slightly invested in their social standing—when their social stock price rises and falls, yours also wavers. And if you are automatically bought into every person you notably interact with, this changes your

en-gb

39:40

“Socratic Persuasion: Giving Opinionated Yet...

LessWrong (30+ Karma) ·

2025/05/27

The full post is long, but you can 80/20 the value with the 700 word summary! Over half the post is eight optional case studies. Thanks to Jemima Jones, Claude 4 Opus and Gemini 2.5 Pro for help copy-editing and drafting TL;DR: I recommend giving advice by asking questions to walk someone through key steps in my argument — often I’m missing key info, which comes up quickly as an unexpected answer, while if I’m right I’m more persuasive, and can still express my case. This is useful in a wide range

en-gb

1:11

[Linkpost] “Formalizing Embeddedness Failures in...

LessWrong (30+ Karma) ·

2025/05/27

This is a link post. AIXI is a dualistic agent that can't work as an embedded agent... right? I couldn't find a solid formal proof of this claim, so I investigated it myself (with Marcus Hutter). It turns out there are some surprising positive and negative results to be derived as easy corollaries of the paper "Universal Prediction of Selected Bits." Interestingly, further technical advances in algorithmic information theory could substantially strengthen our results - I would welcome collaborations

en-gb

8:07

“New website analyzing AI companies’ model evals” by...

LessWrong (30+ Karma) ·

2025/05/26

I'm making a website on AI companies' model evals for dangerous capabilities: AI Safety Claims Analysis. This is approximately the only analysis of companies' model evals, as far as I know. This site is in beta; I expect to add lots more content and improve the design in June. I'll add content on evals, but I also tentatively plan to expand from evals to evals and safeguards and safety cases (especially now that a company has said its safeguards are load-bearing for safety!). Some cherry-picked bad

en-gb

0:58

“New scorecard evaluating AI companies on safety” by...

LessWrong (30+ Karma) ·

2025/05/26

The new scorecard is on my website, AI Lab Watch. This replaces my old scorecard. I redid the content from scratch; it's now up-to-date and higher-quality. I'm also happy with the scorecard's structure: you can click on rows, columns, and cells and zoom in to various things. Check it out! Thanks to Lightcone for designing the site. While it is a scorecard, I don't feel great about the numbers; I mostly see it as a collection of information. --- First published: May 26th, 2025

en-gb

20:59

“Alignment Proposal: Adversarially Robust...

LessWrong (30+ Karma) ·

2025/05/26

Epistemic Status: Over years of reading alignment plans and studying agent foundations, this is my first serious attempt to formulate an alignment research program that I (Cole Wyeth) have not been able to find any critical flaws in. It is far from a complete solution, but I think it is a meaningful decomposition of the problem into modular pieces that can be addressed by technical means - that is, it seems to solve many of the philosophical barriers to AI alignment. I have attempted to make the

en-gb

1:31

[Linkpost] “Priming effects are fake, but framing...

LessWrong (30+ Karma) ·

2025/05/25

This is a link post. A few decades ago, it was pretty common to mush together priming effects and framing effects and see them as two closely connected parts of a single Bigger Truth about the human mind. Of course, everyone understood that the effects themselves were a bit different, but one common view was that they were providing evidence for the same larger picture. That larger picture said: People's judgments are radically unstable, easily pushed around by subtle and almost unnoticeable

en-gb

17:58

“Meditations on Doge” by Martin Sustrik

LessWrong (30+ Karma) ·

2025/05/25

Lessons from shutting down institutions in Eastern Europe. This is a cross post from: https://250bpm.substack.com/p/meditations-on-doge Imagine living in the former Soviet republic of Georgia in early 2000's: All marshrutka [mini taxi bus] drivers had to have a medical exam every day to make sure they were not drunk and did not have high blood pressure. If a driver did not display his health certificate, he risked losing his license. By the time Shevarnadze was in power there were hundreds,

en-gb

118:44

“Claude 4 You: Safety and Alignment” by Zvi

LessWrong (30+ Karma) ·

2025/05/25

Unlike everyone else, Anthropic actually Does (Some of) the Research. That means they report all the insane behaviors you can potentially get their models to do, what causes those behaviors, how they addressed this and what we can learn. It is a treasure trove. And then they react reasonably, in this case imposing their ASL-3 safeguards on Opus 4. That's right, Opus. We are so back.Yes, there are some rather troubling behaviors that Opus can do if given the proper provocations. If you tell it to

en-gb

7:48

“It’s hard to make scheming evals look realistic” by...

LessWrong (30+ Karma) ·

2025/05/25

Abstract Claude 3.7 Sonnet easily detects when it's being evaluated for scheming. Surface‑level edits to evaluation scenarios, such as lengthening the prompts, or making conflict of objectives less salient, do improve realism of evaluation scenarios for LLMs, yet these improvements remain modest. The findings confirm that truly disguising an evaluation context demands removal of deep stylistic and structural cues rather than superficial prompt adjustments. For future LLMs the situation is likely to

en-gb

3:26

“That’s Not How Epigenetic Modifications Work” by...

LessWrong (30+ Karma) ·

2025/05/24

Ask an epigenetics researcher what they study, and the standard story you'll hear goes something like this... "Sometimes a little methyl group (i.e. -CH3) gets stuck on the side of a strand of DNA. Turns out these guys are pretty important! They're copied over when cells replicate, so they stick around long-term, and they can activate or repress (usually repress) nearby genes on the DNA strand. In particular, different types of cells all have the same DNA code, but something has to be different in

en-gb

4,489 results

LessWrong (30+ Karma)

All Episodes

Similar Podcasts

Talking Tech w/ Wayne

Wayne Hunter

The Everything Feed - All Packet Pushers Pods

Packet Pushers

The Good Tech Companies

HackerNoon

Engadget News + Next

Engadget

Cybersecurity Tech Brief By HackerNoon

HackerNoon