Claude Mythos is different.

This is the first model other than GPT-2 that is at first not being released for public use at all.

With GPT-2 the delay was due to a general precautionary principle. OpenAI did not know what they had, or what effect on demand text would have on various systems. It sounds funny now, GPT-2 was harmless, but at the time the concern was highly reasonable.

The decision not to release Claude Mythos is not about an amorphous fear. If given to anyone with a credit card, Claude Mythos would give attackers a cornucopia of zero-day exploits for essentially all the software on Earth, including every major operating system and browser. It would be chaos.

Or, in theory, if Anthropic had chosen to do so, it could have used those exploits. Great power was on offer, and that power was refused. This does not happen often.

Instead Anthropic has created Project Glasswing. Mythos is being given only to cybersecurity firms, so they can patch the world's most important software. Based on how that goes, we can then decide if and when it will become reasonable to give access to a broader [...]

---

Outline:

(03:24) Mundane Alignment Is Excellent

(05:01) Would This Process Be Sufficient To Find A Dangerous Model?

(06:27) Introductory Warning About Superficial Mundane Alignment

(15:12) Model Training (1.1)

(15:25) Release Decision Process (1.2)

(17:50) RSP Evaluations (2.1 and 2.2)

(22:17) Autonomy Evaluations (2.3)

(25:56) The Alignment Risk Update Document

(26:39) The Threat Model

(29:18) Misalignment As Failure Mode

(31:35) Wouldnt You Know?

(33:40) Dont Encourage Your Model

(35:14) Beware Goodharts Law

(37:18) Beware The Most Forbidden Technique (5.2.3)

(41:44) Asking The Right Questions

(43:11) Model Organism Tests

(45:01) Model Weight Security (Risk Report 5.5.2.1)

(45:31) Reward Hacking (Back to The Model Card)

(45:56) Remote Drop-In Worker Coming Soon

(49:01) External Testing (2.3.7)

(49:37) Cyber Insecurity General Principle Interlude

(50:46) Alignment (4)

(56:38) Risk In The Room

(57:56) Mythos Meant Well

(01:00:20) Risk Not In The Room

(01:02:05) Alignment Testing Overview

(01:05:20) Internal Deployment Testing Process

(01:07:55) Reports From Pilot Use (4.2.1)

(01:08:30) Reports From Automated Testing (4.2)

(01:10:13) Other External Testing

(01:10:56) Just The Facts, Sir

(01:13:05) Refusing Safety Research

(01:14:12) Claude Favoritism

(01:15:19) Ruling Out Encoded Thinking (4.4.1)

(01:18:41) Sandbagging (4.4.2)

(01:21:27) Capability for Evasion of Safeguards (4.4.3)

(01:23:04) Pick A Random Number (4.4.3.4)

(01:25:49) White Box Analysis (4.5)

(01:30:30) Model Welfare (5)

(01:31:32) Key Model Welfare Findings (5.1.2)

(01:41:17) Is Mythos Okay?

(01:43:52) Self-Play

(01:45:30) A Few Fun Facts

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Table of contents:

• Part 1: In which I use my optical physics background to share some hopefully-uncontroversial observations
• Part 2: In which I boldly defy Public Health Orthodoxy on the whole UV situation

Part 1: In which I use my optical physics background to share some hopefully-uncontroversial observations

1.1 UV depends a lot on “solar zenith angle” [a.k.a. “angle of the sun away from directly overhead”], not on how hot it is outside

That means: you should mainly be thinking about UV exposure in proportion to how close it is to (1) the summer solstice and (2) solar noon.

Here, I made this handy widget.[1] Select a city in the drop-down at the bottom, and mouse over (or tap) the colored area for specific datapoints:

There's an interactive widget here in the post.

I find that people intuitively judge sunburn risk based on temperatures being high, instead of shadows being short. So they worry about UV too much in the hot late summer, and/or not enough in the cool early spring; and they worry about UV too much in hot late afternoons, and/or not enough in cool late mornings.

(Of course, temperature matters indirectly, because if it's hot [...]

---

Outline:

(00:25) Part 1: In which I use my optical physics background to share some hopefully-uncontroversial observations

(00:35) 1.1. UV depends a lot on solar zenith angle \[a.k.a. angle of the sun away from directly overhead\], not on how hot it is outside

(02:16) 1.2. Other things matter too, so just check your local UV index

(02:57) 1.3. Around half of UV is diffuse (mostly coming from the blue sky) not direct

(03:40) Part 2: In which I boldly defy Public Health Orthodoxy on the whole UV situation

(05:17) 2.1. I lean towards: (1) sunburns are bad, (2) tans are neutral (in themselves), (3) tans are good all things considered (because they prevent sunburns), (4) Sunscreen is for sudden transitions in sun exposure, and then you should try to wean off it

(08:23) 2.2. Wear sunglasses for comfort if you want, but theyre not a health product

(09:34) 2.3. An appropriate effective SPF in most situations is usually like 3, maybe up to 10 tops

The original text contained 6 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

There exists an AI model, Claude Mythos, that has discovered critical safety vulnerabilities in every major operating system and browser. If released today it would likely break the internet and be chaos. If they had wanted to, they could have used it themselves and owned pretty much everyone.

Luckily for all of us, Anthropic did no such thing. Instead, Anthropic is launching Project Glasswing, and making Mythos available to cybersecurity companies, so everyone can patch all the world's critical software as quickly as possible, and then we can figure out what to do from there.

That's the story in AI that matters this week, and it is where my focus will be until I’ve worked my way through it all. But as always, that takes time to do right. So instead, I’m getting the weekly, and coverage of everything else, out of the way a day early. This post is about the non-Mythos landscape, and I hope to start covering Mythos and Project Glasswing tomorrow.

I also covered the latest extended (18k words!) article about the history of Sam Altman and OpenAI, which contained some new material while confirming much old material, and analyzed their recent [...]

---

Outline:

(02:17) Language Models Offer Mundane Utility

(02:48) Language Models Dont Offer Mundane Utility

(03:11) Huh, Upgrades

(04:24) On Your Marks

(06:55) Meta Problems

(07:15) Fun With Media Generation

(09:13) A Young Ladys Illustrated Primer

(09:22) You Drive Me Crazy

(22:05) Unprompted Attention

(22:46) They Took Our Jobs

(33:27) They Took Our Job Market

(35:29) Get Involved

(37:31) In Other AI News

(38:08) Search Your Feelings You Know It To Be True

(45:58) Actors And Scribes

(49:06) Show Me the Money

(53:46) Bubble, Bubble, Toil and Trouble

(54:05) Quiet Speculations

(54:20) Quickly, Theres No Time

(58:02) More Time Would Be Better

(58:55) Greetings From The Department of War

(01:00:11) The Quest for Sane Regulations

(01:01:57) Chip City

(01:03:29) Political Violence Is Completely and Always Unacceptable

(01:04:16) The Week in Audio

(01:06:42) Rhetorical Innovation

(01:10:53) People Really Hate AI

(01:13:39) Aligning a Smarter Than Human Intelligence is Difficult

(01:17:44) Messages From Janusworld

(01:21:00) People Are Worried About AI Killing Everyone

(01:21:50) The Lighter Side

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

I wrote a book! It's called Obsolete: The AI Industry's Trillion-Dollar Race to Replace You—and How to Stop It, and it’ll be available in May if you preorder through my publisher (OR Books+The Nation).

With it, I took on a dual mission: get anyone up to speed on the state of AI and its risks (especially skeptics and lefties) AND write something people working in the space will enjoy and learn from. An unbiased panel of early reviewers (i.e. people I know) think it succeeds. For instance, here's a review from a skeptical friend:

It felt like I was being drawn into this profound vision I had managed to keep out of my head for years… I feel this deep, visceral distrust of AI companies, and chatbots, for reasons that you were able to lay out in various places. I think I came to this like a very non-SF, non-tech, ‘normie’ who is reacting with their gut to seeing the proliferation of ai... But I also had never directly faced the story of AI, how the tech worked, and whether my skepticism was in fact, ill-founded. This hit that, and I thought it did a great job balancing [...]

---

Outline:

(06:37) Ways to help

(08:22) Excerpt - The Obsoleting Project

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

In SF this week, I met an online friend in person for the first time yesterday. We talked about super-persuasion. His take was: there is mostly an efficient market for power, and the world is reactive. Unlike software, humans adapt to new exploits or even just unexplained strange happenings. Society resists, pushes back. Unlike an instance of FreeBSD, it is not fine one minute then pwned the next. My reply was just to point to powerful historical figures and say, "if he could do it so can an ASI!"

That is, if an AI wants to acquire power it should be able to choose a human proxy - or more realistically a portfolio of human proxies - and help them take over a nation. I don't expect it to happen this way, as I imagine there are far quicker paths, but we know it is possible because humans have done it.

There are arguments against this. If I give the examples of Hitler or Lenin or Bonaparte, one might reply that they didn't really fully take over. There was still a political economy they worked within. Their actions were constrained by other agents. I don't think this is [...]

The original text contained 1 footnote which was omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

Written very quickly for the Inkhaven Residency.

As I take the time to reflect on the state of AI Safety in early 2026, one question feels unavoidable: have we, as the AI Safety community, already lost? That is, have we passed the point of no return, after which becomes both likely and effectively outside of our control?

Spoilers: as you might guess from Betteridge's Law, my answer to the headline question is no. But the salience of this question feels quite noteworthy to me nonetheless, and reflects a more negative outlook on the future.

Today I’ll start by explaining “the plan” as I understood it in 2024.

Tomorrow, I’ll explain why this question seems so salient to me, and why the situation looks much worse than when I was reflecting on this question two years ago in 2024. These reasons include: many of our governance and policy plans have failed (in ways that reflect poorly on my naivete in 2024), AI progress is going along more aggressively timelines, the community has largely went “all-in” on Anthropic and lost its independence, some of the more ambitious technical research plans have not paid out, and the political situation [...]

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

Or, for that matter, anything else.

This post is meant to be two things:

1. a PSA about LessWrong's current security posture, from a LessWrong admin[1]
2. an attempt to establish common knowledge of the security situation it looks like the world (and, by extension, you) will shortly be in

Claude Mythos was announced yesterday. That announcement came with a blog post from Anthropic's Frontier Red Team, detailing the large number of zero-days (and other security vulnerabilities) discovered by Mythos.

This should not be a surprise if you were paying attention - LLMs being trained on coding first was a big hint, the labs putting cybersecurity as a top-level item in their threat models and evals was another, and frankly this blog post maybe could've been written a couple months ago (either this or this might've been sufficient). But it seems quite overdetermined now.

LessWrong's security posture

In the past, I have tried to communicate that LessWrong should not be treated as a platform with a hardened security posture. LessWrong is run by a small team. Our operational philosophy is similar to that of many early-stage startups. We treat some LessWrong data as private in a social sense, but do [...]

---

Outline:

(01:04) LessWrongs security posture

(02:03) LessWrong is not a high-value target

(04:11) FAQ

(04:29) The Broader Situation

The original text contained 6 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

Hello, LessWrong. This is a personal introduction diary-ish post and it does not have a thesis. I apologise if this isn't a good fit for the website; I just needed to unload my brain somewhere and this seemed like the spiritually correct place.

I write to you from the Lighthaven campus in fabulous Berkeley, CA. It's my first time here and I am enchanted.

After one week of full-body immersion, I am completely fascinated with the Rationalist philosophy, culture, community. I do not know whether I 'belong' here, as such, and I am writing this post in the hopes of interrogating the feelings I'm experiencing.

After all, why do I feel so positively about the above when I've barely scratched the surface and know almost nothing of what it is I'm looking at? Do I just feel warm because of positive vibes and an ornately-decorated campus? Do I just feel alive because coming to Berkeley feels like an adventure compared to the monotonous university-working life that I've come from?

In other words: do I like the ideas, or do I just like the vibes?

Maybe. I think that's a bias worth challenging.

I do not know if this [...]

---

Outline:

(01:33) Who am I?

(01:56) Why am I at Lighthaven?

(04:00) What have I seen?

(04:36) Ratspeak

(05:50) No Punishment for Bad Takes

(07:16) Eugenics

(08:01) Autism

(09:06) People Just Say and Do Shit

(10:19) AI-generated Images

(11:31) Children and Families

(12:21) Really Excellent Vibes

(13:13) What is the point of this post?

The original text contained 4 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

Nobody has ever done an in person door to door survey about AI risks[1]. What do people really think about AI? Like really? There have been some surveys on the risks from AI. But there's a real difference between looking at numbers on page vs. the feeling of talking to our fellow humans.

I[2] asked 101 people what they thought about the impact of AI. Approximately half of the responses were from ringing doorbells, and half were from asking people out and about[3]. Every single one of these was in person face to face. Around 10 respondents only spoke Spanish and were surveyed in Spanish.

Here are the results. The top level results are very strong when it comes to showing interest in regulating AI. Everything else will be reflections on the results/process and all the qualitative data I picked up from this process.

Thoughts on some Specific Questions

This question was the longest and most complex question. I basically have to read out all of the options since it's not on a simple scale. I decided to first split it on whether someone thinks superhuman AI should be developed at all, and then only give the more granular [...]

---

Outline:

(01:01) Thoughts on some Specific Questions

(03:51) Takeaways

(05:00) Surveying is Inherently Biased

(08:42) People are Weird and Surprising

(11:03) You Can Do This Too!

The original text contained 16 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

These are things I've learned from experience that others might find helpful. Some of them are easy to miss for a while. (Also an exercise in "reality contains a surprising amount of detail"; I could probably have kept going for a while but needed to call it at some point.)

Baking

Oven thermostats are often miscalibrated enough to matter. If you're following existing recipes but find things often coming out overdone or underdone, you might consider buying an oven thermometer to check how miscalibrated your oven thermostat is. Unfortunately, oven thermometers are also often miscalibrated. Fortunately, they're not that expensive[1]. A friend of mine bought three from three different brands to check for inter-rater agreement. Note that ovens can end up at different temperatures in different locations within the oven[2], so ideally you want to place all three thermometers relatively closely together (but not touching) roughly around where you typically put the thing you're baking. (Also note that other factors can affect baking times, like altitude.)

You need to use mass measurements rather than volumetric measurements. For everything macro-scale, anyways - if a recipe asks for a teaspoon of vanilla extract, nobody will tell you how many grams that [...]

The original text contained 11 footnotes which were omitted from this narration.

---

First published:

Source:

---

Narrated by TYPE III AUDIO.