This week’s post makes the case for a political deal: frontier safety for federal preemption. Soon, mainstream politics and big money will make AI policy a worse place. Accelerationists and safetyists share an interest in setting sound foundations – while they still can.

One Tuesday night, as my wife and I sat down for dinner, a sheriff’s deputy knocked on the door to serve me a subpoena from OpenAI. I held back on talking about it because I didn't want to distract from SB 53, but Newsom just signed the bill so... here's what happened: 🧵

On Friday evening, my father-in-law Pastor Ezra Jin was arrested by Chinese government officials, along with more than 20 other pastors and church workers across 7 cities in the most extensive attack on a Chinese church in 4 decades. I have been quiet here mostly because helping

Excellent new thread from Page (former OAI employee and a much better attorney than I) discussing previously unknown details about the subpoenas. He’s new to twitter and I would love if folks helped share.

This relatively short essay by Jack Clark is a good indicator of the attitude of many people inside the AI labs, and what they think is happening right now in AI development. You do not have to believe him, of course, but it is worth noting: importai.substack.com/p/import-ai-43…

Technological Optimism and Appropriate Fear - an essay where I grapple with how I feel about the continued steady march towards powerful AI systems. The world will bend around AI akin to how a black hole pulls and bends everything around itself.

1/ I think this thread by Chucheng Feng captures the logic and motive for Beijing's approach to the latest spike in US-China trade war escalation. Both sides are playing a high-stakes game. It's not entirely clear that both sides are playing the same game, though. (short 🧵)

Had a great time discussing China's AI+ initiative and the big-picture context of economic sea change with Grace Shao! Check it at her must-read substack AI Proem, YouTube, or Spotify aiproem.substack.com/p/ai-plus-unde… More links....

Yes, China clearly prepared rare earths measures well in advance and yes they are way out of proportion w/the US 50 percent rule, but nothing in latest reporting confirms CN would have issued now w/o the US move. These facts are just as consistent with contingency plan. Demetri

Who has Xi purged from the CCP Central Committee? Attendance at next week's Fourth Plenum will reveal the extent of his recent "purge surge" If all rumors are true, only 168/205 members (82%) and 159/171 alternates (93%) will be there And only 17/44 of PLA members (39%)! 1/

The moment friction of tech diffusion clicked for me: Winter 2017 I'm riding the Metra into Chicago, reading 2013 study saying 47% jobs are at risk in near future. Then ticket taker taps me on the shoulder, asks for my paper ticket, punches it with his hole puncher, and moves on.

One of China’s top AI evaluation teams just released results from a new eval. Some interesting takeaways, like a model pretending to have used a tool when it didn’t. Curious to hear takes on this (utility, novelty, sophistication, etc) from other eval folks.

Last month, China released its most sophisticated document yet on how it views AI risks and what mitigations it can implement to address them. Very interesting bits on open-source risks, labor market impacts, and CBRN. In our latest for Carnegie Endowment, Matt Sheehan and I break

my new essay “The China Tech Canon” is out in Asterisk's 12th issue, themed “Book.” link in reply. in this piece, i trace the intellectual map of china’s tech founders by asking a simple question: what books have shaped their worldview? what emerged is that the “china

If you said: “We should have real-time incident reporting for large-scale frontier AI cyber incidents.” A lot of people in DC would say: “That sounds ea/doomer-coded.” And yet incident reporting for large-scale, non-AI cyber incidents is the standard practice of all major

Mechanize hope a historical trend toward greater redistribution persists through full automation. But full automation changes the relative leverage: it removes the implicit threat that workers lay down their work, pick up arms, etc. That difference might well break the trend!

Interesting study from Tsinghua on how gender + seating assignments affect test scores on Gaokao. TLDR: when you seat women together, their scores go up, while there isn't a corresponding affect on male students based on seating. sciencedirect.com/science/articl…

If you remember this video, then I've got a treat for you: I just talked to its creator, a 29-year-old actor in Hangzhou named Mu Tianran 穆天然, about how he captures AI slop's essence so well, and why Sora 2 made him worry about his job. wired.com/story/made-in-…