AI Policy
A Sanctioned Chipmaker Kept Buying Nvidia Anyway
Inspur Group was blacklisted by the US in March 2023. Weeks later its US arm renamed itself Aivres and kept selling Nvidia Blackwell servers into 2026.
AI Policy
AI Policy
OpenAI's Chief Scientist Says Alignment Isn't Solved
OpenAI's own chief scientist says no lab, including his, has solved AI alignment - and calls for mandatory third-party safety audits, days after OpenAI's most aligned model yet shipped.
AI Policy
Newspapers Sued Over AI Training. The Number Was Filed by Microsoft.
The Seattle Times and Newsday joined the AI copyright fight this week. Microsoft's own court filing, drawn from the plaintiffs' sample, found reproduction under one percent.
AI Policy
OpenAI Wants Its Own Incident Clock. Brussels Started One First.
OpenAI proposed a voluntary misalignment-reporting framework on September 5, the same week it denied its lawyers slowed a rogue-agent disclosure. The EU already runs a clock of its own.
AI Policy
Brussels Just Tested Its New Power on 30 AI Companies
The European Commission's AI Office sent its first-ever Article 91 requests for information to more than 30 general-purpose AI providers on September 1, 2026, probing safety practices and copyright compliance, with fines up to EUR 15 million for misleading answers.
AI Policy
A New US Bill Wants Every AI Agent Named and Logged
The bipartisan Stop Rogue AI Act, introduced September 3, 2026, directs NIST to build a machine-readable inventory of every AI agent, tamper-proof action logs, and continuous verification standards. It targets agents, not models, a governance object the EU AI Act does not yet cover.
AI Policy
US Bill Would Jail AI Developers for 20 Years
Sanders and Casar's Ban Artificial Superintelligence Act proposes 20-year prison terms and corporate dissolution, far harsher than the EU AI Act's EUR 35 million fine cap.
AI Policy
NYC Bars AI Chatbots for Pre-K to 8th Grade
NYC's one-year ban on ChatGPT and AI chatbots covers 600,000 students through grade 8, the strictest US schools policy yet, months ahead of any EU classroom AI rule.
AI Policy
A Security Badge Told The Pentagon Nothing About What Grok Would Generate
The Pentagon deployed Grok for Government and ChatGPT Mil to 3 million personnel on August 31, 2026, both accredited at Impact Level 5. That accreditation certifies data-hosting security, not what a model generates - and Grok's adult-content generation problem was never evaluated at all.
AI Policy
Your AI Vendor Just Got a Letter From Brussels
The EU AI Office issued its first formal Requests for Information under the AI Act on August 29, 2026, 27 days after gaining enforcement power. Two separate tracks target model safety and training-data transparency.
AI Policy
A 45-Year-Old Bug Delayed an AI Model's Release
Z.ai's GLM-5.3 found 2,436 vulnerabilities, including one from 1981, during post-training alone. The open-weight release it promised for August 28 has not shipped.
AI Policy
Judge Voids Pentagon's Anthropic Risk Label
Judge Rita Lin vacated the Pentagon's Anthropic risk label on August 27, ruling it retaliation, not a security finding - and set a precedent.
Page 4 / 12
The Operator Brief
One considered note on infrastructure, governance, and measurement, most mornings. No theory.