AI
The Great Reversal: China Won Open AI. Now Beijing Wants To Close The Door.
Chinese open-weight models won the US enterprise in the first half of 2026. Beijing is now moving to close the door it walked through — and Washington, ironically, is helping....
GPT-5.6 ships — and pre-clearance quietly becomes the default
US government pre-clearance for frontier models is now the working default — regardless of whether the White House admits it.
The Claude Code affair — hidden trackers, state advisories, and a trust breach that redraws the AI supply chain
Anthropic's hidden Chinese-user tracker is not a backdoor in the CVE sense — but it is the first time a US frontier lab has been caught embedding user-identification steganography inside...
The Day Smart Glasses Became the Default — And the Backlash Landed With Them
The form-factor debate for post-phone AI is effectively over. Smart glasses won — not because any one product is great, but because everyone shipped on the same day and the...
The Utrecht Verdict: Robotic-Surgery AI Is Being Sold Faster Than It's Being Validated
A peer-reviewed audit of a decade of robotic-surgery AI has landed a hard verdict — fewer than a third of studies bother with clinical validation, and the metrics that are...
The Backdoor That Wasn't Quite a Backdoor: Anthropic, Alibaba, and the Technical Decoupling of AI
The US–China AI split is no longer policy overlay. It is now shipped inside the code, and the tooling is the border.
Korea's Trillion-Won Bet: Buying Its Way From AI Consumer to AI Landlord
Korea has decided that the way to escape being an AI tenant is to become the landlord — and it has priced that ambition at roughly two-thirds of a trillion...
Watermelon: Meta's Compute-Order-of-Magnitude Bet and Why the Benchmark Claim Isn't the Real News
If Wang's benchmark claim survives independent verification, the frontier is a five-lab club, not a two-lab one. But the more informative fact is not what Watermelon scores — it is...
The week smart transportation stopped being emerging
The self-driving argument just got quietly settled — not by breakthrough, but by simultaneous validation across regulation, safety math, capital, and architecture.
China just put AI cancer vaccines on a production line
China has industrialised what the West is still piloting — the policy and competitive consequences land before the clinical ones do.
Korea just turned the physical-AI race into a sovereign-industrial project
The unit of competition in AI has moved from models to fabs, megawatts and humanoids — and Seoul is the first capital to bid at that altitude.
The Red Queen Has a Patent: Why a Cambridge–NVIDIA Preprint Just Reset the Self-Improvement Debate
The RQGM is the first paper to put a working mechanism under the "AI builds itself" debate — and the safeguard it ships with is exactly the safeguard the next...
The Great AI Unbundling
Washington's export controls on frontier AI have triggered, in a single week, the most significant restructuring of global AI supply chains since ChatGPT launched.
Showing 26 of 87