h/t Lawyer Ex Machina for the California legislature recap:

AI bills that California Gov. Gavin Newson has signed into law so far at the end of California’s legislative session include:

Adam’s Law: SB-1119, named after Adam Raine, regulates chatbot developers who offer their products to users under the age of 18

AI Safety bills: AB-1405 creates a registry of AI auditors and a set of standards “for their independence, transparency, and integrity”; SB-813 requires the state’s Government Operations Agency to develop a framework for evaluating the credentials and expertise of independent verification organizations that may assess AI models for various types of risks

Related: A bill that would regulate the use of AI tools by California attorneys and arbitrators, to address the tide of AI-generated errors included in court filings within and outside of California courts, is still awaiting a signature or veto from the Governor

=> https://www.politico.com/news/2026/09/09/newsom-signs-ai-safety-bills-backed-by-anthropic-openai-01069928

The governor signed one bill from Democratic Assemblymember Rebecca Bauer-Kahan that creates a registry and ethical rules for outside auditors — third-party groups that AI developers could hire to ensure their models comply with state AI laws.

The second bill, authored by former member of Congress and state Sen. Jerry McNerney, tasks the state with establishing criteria for and checking the credentials of so-called “Independent Verification Organizations” and their expertise in assessing the risks posed by AI systems.

Note that the bill that actually required companies get outside audits got vetoed, so. Those clowns in congress did it again.

https://www.patreon.com/MaxRead/posts/ais-sheryl-crow-169107221

Imagine a hypothetical future history of machine learning, perhaps one that focuses on the politics of AI in the 2020s. You may be thinking to yourself, Given the rapid pace of change, it’s hard to say what this book will contain, but there’s one thing I know for certain will not be in this book: Sheryl Crow’s Instagram account.

Well—think again:

“We know and have always known that AI has the capacity to outsmart us and eliminate us in order to continue,” the “All I Wanna Do” singer wrote on Wednesday. “It will have the codes. It will have the capacity to eliminate our energy grids. And, it has the financial data of every single person on the planet.”

In the future, I think, we will remember this as the “Sheryl Crow moment”: The week that, effectively overnight, everyone—which is to say previously uninterested normies, previously skeptical pundits, opportunistic politicians and many other people who had never before heard the words “light cone” and whose browsers had never navigated to LessWrong—suddenly started taking “AI safety” seriously.

E: holy shit that Joe Allan quote. that guy.

From the mirror universe news

https://www.zerohedge.com/news/2026-09-10/campaign-slow-ai-was-ready-viral-warning

ZH’s interpretation is Coxon is some kind of false flag thing to bait Dems into regulating AI

Parker Thayer argued that Coxon’s launch looked like a sophisticated public-relations operation. He traced the earliest amplification to people associated with Encode, the AI Policy Network, and the AI Futures Project.

Brian Chau’s later thread surfaced a direct political-technology link. Newspeak House’s own residency page lists Jacob Coxon among its residents and describes his work as “applying mathematics and data science to inefficient markets.” Newspeak calls itself the London College of Political Technology. Its current calendar advertises Campaign Lab sessions aimed at helping the progressive left “campaign more effectively.” The residency places Coxon inside that political-technology network.

Some of the underlying funding links are public. The Survival and Flourishing Fund’s 2025 grant table lists $2.035 million for the AI Futures Project, $516,000 for Encode AI, and $1.635 million for the AI Policy Institute. The fund attributes those grants to Jaan Tallinn. Anthropic says Tallinn led its Series A financing in 2021.

Encode identifies the Survival and Flourishing Fund as one of several foundations supporting its work. It also says it refuses funding from corporations, foreign governments, and executives at frontier AI companies. Its general counsel, Nathan Calvin, is one of the early amplifiers named in Thayer’s thread.

Some new pins for the conspiracy board I guess. They imagine a “political-technology network” quite the linguistic contortion to not use the word rationalist. No clue what LCPT is but looks like a hacker house type thing?

su3su2u1 review of HPMoR

obi-wan-theres-a-name.png

what ever happened to them anyway? Probably has tenure by now, hopefully?

https://www.theregister.com/ai-and-ml/2026/09/09/ai-models-dont-kill-people-people-kill-people/5295368

At the same time, some of that harm may be balanced on a statistical level by lives saved through AI tech.

[citation needed]

Here is a roundup of the “october tech movie” superfecta

https://www.hollywoodreporter.com/movies/movie-news/musk-artificial-social-reckoning-holmes-documentary-1236693959/

is there a kalshi market for which one tanks the hardest? Or will we get crossover sequels for musk vs altman a la MCU?

New (to me) lore from https://www.theargumentmag.com/p/god-is-back-on-the-timeline

This all kicked off last year when religious Substacker Ethan Muse wrote a heavily sourced essay arguing that the Miracle of the Sun at Fatima — a 1917 incident where large crowds of people reported seeing the sun dance and emit radiant colors — was a genuine miracle, unexplainable by any other method other than divine intervention. This inspired rationalist blogger Scott Alexander to write a more skeptical post titled “The Fatima Sun Miracle: Much More Than You Wanted To Know,” which led to responses and counter-responses.

The discourse continued into 2026 when other bloggers began jumping into the fray and Alexander began investigating similar sun miracles in Buddhism. It seems to have wrapped up with Alexander’s recent pilgrimage to a modern-day holy site where people still claim to see sun miracles to this day.

lol pilgramage what. is he converting?

Not compared to the supreme court, no. E: your recollection kinda proves the point imho.

Another oped from StudyHacks Cal Newport:

Opinion: How Scared Should We Be of A.I. Right Now? | NYT

The time has come to push back. Once we understand the key ideas of Rationalism, we can begin to filter them out. Be very wary, for example, when A.I. leaders or online commentators begin talking in sweeping generalities about the future of the technology, such as stating with confidence that many jobs will soon be automated or that their language models will soon become sentient beings intent on escaping their creators’ control. (No, OpenAI’s autonomous agents conducting hacks doesn’t mean that A.I. is rising up against its human overlords.)

The correct reply is to promptly turn the conversation back to the possibilities, harms and viability of the products that actually exist. To extend an analogy I’ve made before, if Ford automobiles were spontaneously catching on fire, we wouldn’t tolerate the chief executive’s changing the subject to discuss our general lack of preparedness for a future with flying cars. We’d demand that the company address the acute problem happening right now.

(Hey remember when there was a guy that actually blew the whistle on exploding cars, ran for president? “We” tolerated exploding cars and people dying for years, until someone actually stepped up and forced the issue. How’s the country been treating that guy?)

It’s not surprising that so many Americans feel negatively about A.I. at the moment, but these sentiments aren’t inevitable. Our current crop of L.L.M.-powered tools are exciting — and worrisome, and promising, and frustrating, all at the same time. With proper direction, they’ll hopefully make our lives better. But they don’t have to be terrifying. We thought Silicon Valley was trying to warn us of an unavoidable danger, but maybe they were just being weird.

wonder if such a moderated, watered down position in a large publication ends up sending more eyeballs to Yudkowsky than it does to innoculate the general public, could end up a net negative.

Did you ever ask him about his day job? https://en.wikipedia.org/wiki/Neal_Stephenson#Outside_of_writing is pretty colorful and the sword kickstarter especially was not a good look.

I feel like if a you want to train something off of my software or my other intellectual property, we can come up with a commercial licensing agreement for that. Or you can obey the terms of the publicly posted software license. Instead they are arguing that I am not allowed to negotiate and that big companies get to do whatever they want with code from small companies.

They already have an army of lawyers, so why do they need a get out of jail free card too? I wrote the thing, so what do I get? Fucked over is what I get.

I dug into the link at the bottom of the rfp, it’s definitely a lot more crypto / ethereum leaning than ai, none of the names rang any bells to me as I’m less acquainted with that group. https://www.paradigm.xyz/frontiers-2026 and https://www.paradigm.xyz/writing but they do have shocking amounts of js on their pages, hung the tab, probably trying to mine something.

in which le gwern and sword guy neal sponsor a writing contest

https://www.gpuworld.org/

Premise

Imagine that AI frontier progress stops as of 1 September 2026: AI becomes faster and cheaper, but it never becomes superhuman or improves considerably across the board.

So the Singularity never happens—but GPUs keep getting made. By 2040, there may be the equivalent of 8 billion GPUs globally and everyone has access to a frontier LLM.

What happens in this ‘business as usual’ future?

I may as well through my name in the hat, I could use the cash. but, who the hell gave them $100k for this? I guess BTC is back up.

https://www.thetechbubble.info/p/the-future-is-not-tech-forward

whole thing is good but long, but then THAT GUY pops up again. Humans First has apparently already came and went, that other article left out this fun bit of lore:

Politics comes for Humans First too. In late March, there’s a high-profile spat online where pro-AI White House allies like Marc Andreessen and David Sacks retweet a piece critical of Humans First. Shortly after, the organization splits along partisan lines. The conservative wing keeps the Humans First name and makes the right-wing political activist Amy Kremer, who helped organize the January 6 “Save America” rally in DC, its chair; Allen leaves the group. When I ask Allen about the changeup, he is curiously coy (but characteristically body-horrible): The group was “undergoing mitosis,” he says. “I am now a lone organelle floating in the residual cytoplasm.”

It’s ok little buddy you still have steve bannon. /s

Being on YT is a big driver here, and Lex is smart enough to work the algo. Have John Carmack on, get views when people search for Quake. Have Elon on, get subs when people search for Tesla. Have Mr Beast on and get some kids in the funnel. He’s very good at piercing through filter bubbles to get extra reach and pull numbers, unlike an Alex Jones. (EDIT: For this, the bland persona is a feature not a bug)

I’d ballpark that he made $20k off this DHH interview and picked up some rails devs along the way.

Whenever new coworkers bring up Lex, I always feel like it’s a test, like are they oblivious, or are they secret closet magas trying to test the water. Slate Star has a similar thing. And YT is bad enough but now I have to see it pop up on LinkedIn too.

The thing that bothers me the most is that while this happens in Cambridge but over at Data Colada, Ariely is just like “whoopsie” and gets to keep his job. https://datacolada.org/138 - personally I consider fabricating your entire study to be worse than sloppy quotes, but it is insane how these two people are treated so differently.

I also think it’s really lazy that BBC couldn’t even mention Claudine Gay (AND Martin Luther King) since it was the exact same thing. Different country I guess.

https://www.nakedcapitalism.com/2026/09/ai-off-balance-sheet-obligations-hit-a-subprime-like-reset-wall-in-2027-and-2028.html by Yves / naked capitalism

Critically, it is these “frontier labs” that contract for, lease and build data centers. The very little I have seen about this process has caused me concern. For instance, I read of Meta contracting for a data center for what sounded like 20 years but was 4 back-to-back five year leases. Recall that the once AAA rated General Electric has taken a big fall in investor grace. How strong will the now-mighty-seeming Meta be, particularly if AI bets come a cropper?

Groundbreaker has identified a key element of risky AI leverage, which are obligations that are not accounted for as debt at the AI sponsor. Those of you who are corporate finance literate will know this is a cardinal sin. When I was a wee young thing at Goldman, where we went through hard copies of financial statements to extract key data which we entered into green accountants’ ledger paper, one of the aims was to determine total debt to assets. One of the footnote items we would root out was operating lease payments. For instance, big companies often use leased mainframes. We would capitalize them and treat them as a long-term liability.

=> https://www.groundbrkr.com/p/the-teaser-period-why-the-ai-boom

The take-or-pay compute contract - the instrument at the center of the AI build-out - has a structural feature that almost no one prices: its payments do not begin at signing. They begin at delivery. A lab signs a multi-year capacity commitment today, but the payments do not start until the data center is energized, the capacity is accepted, and the contractual ramp schedule commences - an interval set not by finance, but by construction: siting, powering, and filling a gigawatt-scale campus takes 24-to-36 months from signature - mirroring the two-to-three-year teaser of a subprime ARM.

Going to take me a while to digest this one

I looked her up, she was at MetaMed/Palantir/Longevity Research Inst, so definitely in the thick of things and in that world. That lying problem piece hits a lot harder post-sbf.

over on linked in

as a statistician, I’m so tired of hearing they’re “updating their priors.” No you didn’t. gwern at least posts his R code, but you are just an idiot regurgitating something you half heard, probably on Lex.

(00:39:39) Let me put it. Omarchy is a beautiful, modern, and opinionated Linux system. It’s an alternative operating system to macOS and Windows, and it’s freaking amazing.

No it isn’t. It’s your glorified dotfiles. too bad Lex can’t tell the difference. I feel bad that Lex is exposing literally a million people to your crap. /s

these guys.

Maybe this is just me being slow to put things together, but why exactly is that startup for AI cheating overlay for zoom job interviews now putting out workplace / startup humor shorts? They seem to have paid a host of clippers too. Ling Long is everywhere. Did the AI thing not work out for cluely?

eg https://www.youtube.com/watch?v=vmLOORRGW4U