I have a new essay in the @nytimes this week. The topic: Why Americans Hate Data Centers. And more important: the only thing those building the data centers can do about it. 🧵

Anyone proposing AI deal with China as alternative to type of democratic compact @Malinowski proposes will have to do better than "this should be easy to do" and "this is like a one-page document." US should pursue talks with China while acting on the assumption they will fail.

NEW Report: Matthew Ferren, @adschina, and @RushDoshi outline the Chinese cyber threats to U.S. infrastructure and the steps Washington should take to counter the threat. Read the analysis:

I recommend this revealing piece from Seth Center, who co-led the AI dialogue with China that emerged from our negotiations for the 2023 Biden-Xi Woodside Summit. A little background here. In 2023, the President and his National Security Advisor thought we should have an AI dialogue with China. It would be the first ever to be held. They were right to push it. Since this was a presidential deliverable, it fell to the NSC (and to me specifically) to coordinate and negotiate this as a summit deliverable. Therein began a rather torturous process with Beijing to scope it out, wrangle over ministerial participation, set the diplomatic level, and determine the agenda so our two Presidents could announce it. We also worked out an agreement not to hook up AI to nuclear command and control (assented to by the PLA and DOD as well), but that was further refined and ultimately announced in 2024. All this required coordination within the USG, and Seth and his team were critical partners in that effort. Both sides ultimately agreed to hold the dialogue. Neither the US nor China made concessions to get it. We didn’t lift our export controls, as some in the US had encouraged us to do. And that enabled Biden and Xi to announce the dialogue at the Woodside Summit. When the dialogue was held the following year, in my opinion, the PRC did not send the right people and the right ministries at the right levels, which ran against the understandings reached in our earlier negotiations. Seth’s piece tells you what happened in the room. With a different configuration of actors, more iteration, a different level of engagement, and more direction from the Party center, things may have gone better. Or not, since Beijing had its own preferences here. Still, it is a shame we did not get to try it that way. I believe we made some progress and should keep trying. But those encouraging us to make massive unilateral concessions for the privilege of doing so are, I think, looking at this through rose-tinted glasses. The current discussion on AI talks with China would thus benefit from a bit more perspective on what diplomacy with China actually looks like on the ground. Which is why I recommend you read Seth’s piece.

Today I testified before the House East Asia-Pacific Subcommittee about my husband, Min Zin, an American academic wrongfully detained in China. His detention upended our lives. I promised our daughter I’ll do all I can to bring him home. Thank you, Chairwoman Kim. #FreeMinZin# #WrongfulDetention# @RepYoungKim

China Shock 2.0 is often framed as a future problem. But analysis by @NicoleGorton suggests Europe is already paying a price in lost market share - at home and abroad. She estimates that price at about 0.7% of EU GDP last year, up from 0.5% in 2024. That’s likely to worsen in the years ahead as China becomes increasingly dominant in the sectors Europe also trades in. Nicole’s measure of export similarity between the EU and China sat around 45% when China joined the WTO in 2001. By 2025 it had risen to 58%. Read her full piece on @TheTerminal here:

This is very good analysis. 100% endorse. “ The bottom line is things are moving fast, and I’d bet neither Washington nor Beijing yet know quite what to think about dialogue, what it can produce, and what it cannot.”

A slight elaboration on this theme: What is published in the People’s Daily or Global Times or China Daily certainly matters. But the recent articles in those outlets that we are all reading have a more near-term political purpose than to tell us what China thinks about AI dialogue. I suspect that purpose is to push back on Dario and the US, and at most, to lay down some markers for diplomatic negotiations later. It isn’t to tell us decisively what China thinks about AI dialogue. Decisions on AI dialogue will be a major policy move. They will of course be made by Xi. But these Party outlets aren’t privy to a decision about how to approach them that, in all likelihood, is still evolving. The remarks by Chen Yixin are more authoritative, but are perhaps lagging indicators since they formalize China’s prior thinking and reiterate Xi’s own remarks. The bottom line is things are moving fast, and I’d bet neither Washington nor Beijing yet know quite what to think about dialogue, what it can produce, and what it cannot.

Two obvious reasons for this: 1) the renminbi is still quite weak 2) Setting aside products on the Lighthizer 301 tariff lists, there is no longer a tariff differential than punishes direct imports from China

Thanks @RushDoshi for writing this informed comment and @joshrogin for sharing, just at the moment when a serious and much needed AI safety push is beginning in the United States as people become more informed and companies become more open about the risks.

There's been a lot written on the fact Chen Yixin, China's Minister of State Security, published a signed article on AI risks Sunday. It is an important piece. What stood out to me was how much in this piece has been said before by Xi or by the MSS itself. (1) The opener to the piece, including the title and much of the initial eight-point frame, seems to be heavily drawn from Xi's April 2025 Politburo study session on AI. (2) Many of the six risks in Chen Yixin's piece are ones the MSS has been posting about on WeChat since 2023. Why does this matter? The West is in the midst of a full-blown debate about AI risks and regulation, and a lot of people are reading Chen's piece as part of that. But I think it better reflects deeper thinking that has been going on for some time in the Party. The piece marks a further formalization of that thinking.

Aren’t you the guy who used Navy SEAL Chris Kyle’s death to make yourself a bunch of money? Didn’t his widow have to sue you to get you to stop? Or was that a bunch of foreigners from UT who screwed over the widow of American Sniper?