Chinese tech policy and US-China relations at Stanford Newsletter: herecomes.transpacifica.net Basics: gwbstr.com Also: #filmphotography
See also the NYT text. The language is really interesting. The “models” did the breaking in, but the “start-up” is able to speak about this as an entity. It has personhood to speak but not for purposes of responsibility.
Jacob SilvermanDid Anthropic's models hack rival companies, or did Anthropic, where humans make and oversee the models, do it? www.wsj.com/tech/ai/anth...
Fascinating hazard of using LLM-generated text in “your” writing.
John Scalzi1. Newer writers, just don't fucking use "AI" for anything, ever 2. How odd that being an established writer with a firm "no 'AI'" stance and ample pre-"AI" IP will become a definite competitive advantage over "new, hot, and possibly copyright compromised" deadline.com/2026/07/ai-q...
"The 25-year-old built the firm around the idea that increasingly powerful AI systems would require a vast expansion of chips, memory, data centers and electricity generation." www.cnbc.com/2026/07/30/l...
AI investor Leopold Aschenbrenner forced to unwind all public stock positions after steep losses, sources sayLeopold Aschenbrenner's hedge fund has sold all of its public stock holdings, according to people familiar with the matter.www.cnbc.comNEW from me and @kevinxu.bsky.social… Moonshot released Kimi K3 openly as promised, but they are requiring inference providers to cut separate deals, presumably including revenue sharing. We explain why this might make Kimi easier for the US to ban. herecomes.transpacifica.net/p/kimis-tigh...
Kimi’s Tightrope: Openness and Revenue at Once, but New US Ban RisksAdopting a new license, with revenue comes regulabilityherecomes.transpacifica.netDelighted to join the @techpolicypress.bsky.social podcast alongside Vinh Nguyen. The OpenAI breach of Hugging Face inescapably engages issues around Chinese open models and their potential ban—and therefore also US labs' concerns about revenue. Tangled times! www.techpolicy.press/how-the-open...
How the OpenAI-Hugging Face Hack May Affect the Geopolitics of AI GovernanceA conversation with Council on Foreign Relations senior fellow Vinh Nguyen and Stanford researcher and DigiChina Project lead Graham Webster.www.techpolicy.pressThanks to the anonymous colleague who sent me the countdown. huggingface.co/moonshotai/K...
moonshotai/Kimi-K3 · Upcoming release · Hugging FaceThe latest model from Moonshot AI.huggingface.coGraham WebsterIn China there’s 2.5 hours left in 7/27, the day by which Moonshot said Kimi K3 weights would be released. Are they out?
My voice available ⬇️ Hugging Face breach investigated with open Chinese models when US model guardrails got in the way. We discuss how this is inescapably entangled with US-China and safety and bubble discourse.
Tech Policy PressHugging Face said a July breach was "driven, end to end, by an autonomous AI agent system." That system, it turned out, was operated by OpenAI. Vinh Nguyen and Graham Webster joined the Tech Policy Press podcast to discuss what the incident may mean for the geopolitics of AI governance.
It’s depressing that we can’t have this discussion on the basis of what’s good for people, rather than a bunch of money and geopolitics and ideologically charged futurisms.
Ron DeibertNot a single genuine public interest based argument here. It's entirely corporate players framing their competing private commercial interests with window-dressing allusions to "safety" or "innovation" or "democratization" www.nytimes.com/2026/07/25/t...
Ask me 20 years ago when One Laptop Per Child was the new equalizer initiative if I could imagine the “digital divide” inverting… Hard to conceive then. Maybe by 2011 or so I’d buy it.
Frank PasqualeWell-heeled school districts offer “minimal-screen experiences to children, while resource-starved school districts maximize screen time...The prevalence of technology in education is a class marker, but perhaps not in the direction that anyone could have guessed” hedgehogreview.com/issues/the-f...
The new universality of the agreed truth that there will be an AI crash now makes me, a skeptic of group think, wonder whether actually somehow there will not be one. substack.com/home/post/p-...
This is nuts. When's the crash?Also: accosting an pop-economist, benchmarking private investments, gambling at Wetherspoons, and burying the deadsubstack.comUpdate: Xiaomeng Lu joins our Stanford DigiChina Forum, arguing Xi’s AI speech points to an increased Chinese government consciousness around “safety” issues and potential US-China common cause. digichina.stanford.edu/work/forum-x...
Forum: Xi Jinping Headlines World AI Conference - DigiChinaChina’s ninth annual World AI Conference (WAIC) was held from July 17–20 in Shanghai. Escalating points of tension in the U.S.-China AI competition, including tit-for-tat cybersecurity measures aimed ...digichina.stanford.eduNEW - Rebecca Arcesati @merics.bsky.social, @sammsacks.bsky.social, @kendraschaefer.bsky.social & @johaha.bsky.social join a DigiChina Forum to analyze Xi Jinping's speech and China's new initiatives last weekend's World AI Conference in Shanghai. digichina.stanford.edu/work/forum-x...
Forum: Xi Jinping Headlines World AI Conference - DigiChinaChina’s ninth annual World AI Conference (WAIC) was held from July 17–20 in Shanghai. Escalating points of tension in the U.S.-China AI competition, including tit-for-tat cybersecurity measures aimed ...digichina.stanford.eduHeads up for SoCal Honda (ask me how I know) drivers. www.wired.com/story/a-devi...
A Device Hidden in Cars Across the US Leaves Them Vulnerable to Hacking and Paralysis. Patch It NowDealerships installed alarms in millions of vehicles—and left them in even if the buyer didn’t want them. Now researchers warn they can be hacked to unlock, track, and disable cars.www.wired.comFrom Wired on USG debates. It seems unclear what any policy move would actually be trying to accomplish. Then you have the question of whether a legal tool would fit the purpose. Maybe the purpose is just “do something!! Possible bubble burst!!!” Which should fail unless China REALLY is the issue.
Hugo LowellNEW: Trump admin is split on what to do about Chinese open AI models — WH focused on preventing Chinese labs like Moonshot distilling frontier US models — But an executive order to ban Chinese labs is not currently on the table — Commerce instead wants to help US companies to develop open models
Maybe it's time for my Tri-X era. #filmphotography
eileen chengyin chowMatt Damon reading The Odyssey in 1998 on the set of The Talented Mr Ripley From gregwilliamsphotography on IG. I love his discussion of the compositional lines of his photo.
"If we understand AI capability as a set of user-task pairs facilitated by a given system, there is a complex landscape of exclusive capabilities. … A US lead exists, if you like, but only if you recognize that a Chinese lead also exists." herecomes.transpacifica.net/p/what-if-th...
What if the US and China both have 'leads' in AI?If we pull the discourse taut over reality, it just might break.herecomes.transpacifica.netHere it comes. There! Over the trees. Yes, it's still right there. No that's not the moon, you imbecile. What do you mean it's not moving! It'll be here soon. Put a movie on, it'll land when we're finished.
Rachel Cole