Binance's initial findings trace the 29 October theft to a tampered AI coding tool download in June and a malicious interview repository in September; on 20 October the attackers used an internal AI tool to take over two wallet engineers' systems.
Sen. Bernie Sanders, New York Mayor Zohran Mamdani and former FTC chair Lina Khan attacked the plan after The New York Times reported that schools rejecting AI would lose student Search; Google replied on X that schools are in control.
The research firm raised its estimate after identifying more than 9GW of possible capacity; SpaceX has not published these figures itself, and Elon Musk reposted the report.
The Information says the guidance is not mandatory and covers future contracts, not existing deals; cloud customers are also asked to disclose end users, and the FT says it was passed on verbally by sales teams.
Officials told Indian press the agency is testing whether the model can find and exploit vulnerabilities in a closed environment; neither Anthropic nor CERT-In has published a statement.
OpenAI says Astra produced the enclosure, circuit board and firmware without direct human edits, but still missed manufacturing tolerances, and that hardware-team token use is up 23-fold since February.
Wu said on X that Friday was her last day and Anthropic confirmed the exit to Business Insider; she gave no reason and did not say where she is going.
Gary King and Steven Worthington compared 134 human reviews from their rapidPeer service with Claude Opus reviews; AI won every pair under a GPT-5.5 judge and 87% when three former journal editors judged.
The 10-K risk factors say agents could alter code, send messages or complete transactions without instruction, and that Apple's ability to oversee third-party agents may be limited.
Data Center revenue hit a record $8.5bn, up 97% year on year, on EPYC CPUs and the MI450 GPU launch; the shares rose about 7% after hours.
Its ops and infrastructure lead told an AI Engineer conference that about 2,000 engineers now default to Z.ai's GLM-5.3, served on 56 Nvidia B200 GPUs at about $0.30 per million tokens.
The AI server maker's gross margin was 12.7% and operations used $434m of cash; the shares rose nearly 8% after hours.