AI news / automated archive.
The Trump administration lifts its two-week export ban on Anthropic's Fable and Mythos models after a security deal is reached. Anthropic simultaneously launches Claude Sonnet 5, a cheaper agentic model, and Claude Science for researchers. Plus: the UN warns AI is worsening global inequality, and Google AI Overviews are shown to cut publisher clicks by nearly 40%.
Anthropic launches an AI drug discovery program joining the healthcare race, OpenAI discovers a method to halve inference costs, and cybersecurity firms post record quarters as AI-driven threats surge. Plus: a California study reveals highly educated workers are most exposed to AI disruption, and the White House issues a new AI executive order.
Anthropic secures landmark California government partnership, Google unveils the Interactions API as the primary interface for Gemini agents, and Claude becomes generally available in Microsoft Foundry. Plus: the White House greenlights Anthropic's cyber-capable model, Amazon shops for AI alternatives, and bots officially outnumber humans online.
Anthropic inks Claude deal with California at half price; Taiwan raids Super Micro in Nvidia chip smuggling probe; Chamath raises $135M for AI coding startup; Chinese hedge funds warn AI 'super bubble' is ready to burst; Cursor goes mobile; Brown professor uncovers mass AI cheating.
OpenAI launches GPT-5.6 Sol cybersecurity AI, Google brings computer use to Gemini 3.5 Flash, Meta's new research chief bets on agents, Austria seeks to host Anthropic in Europe, Google restricts Meta's Gemini access, South Korea plans $576B AI-chip investment, and the BIS warns of AI bubble risks.
Billions wiped off Alphabet as top AI scientists defect to rivals; Anthropic launches Claude Tag, a persistent AI teammate in Slack; OpenAI diverges from White House on AI safety rules; Nvidia posts record $58.3B profit and $40B in AI equity bets.