AI News for September 10, 2026 — 7min.ai

[NEWS] Cruz and Sanders pile on after Anthropic researchers put extinction odds above 10%: Lawmakers across the political spectrum seized on an Anthropic researcher's resignation warning, with at least two more Anthropic employees publicly backing the forecast that AI could cause human extinction by 2030.
[TOOLS] Apple Watch gains always-listening 'Audio Intelligence' recaps: The Apple Watch Series 12 and Ultra 4 add Audio Intelligence, a set of AI listening features.
[NEWS] Paul Christiano joins OpenAI's board and warns 'most people could die': OpenAI appointed Paul Christiano to its nonprofit foundation board and its Safety and Security Committee.
[TOOLS] Suno ships v6 models trained with Warner, BMG, and Believe: Suno released v6, its first model generation built with record industry support, and retired all previous versions.
[NEWS] Apple unveils the $1,999 iPhone Duo, its first foldable: Apple introduced the iPhone Duo, a passport-shaped foldable that lays flat when opened, with a 7.6-inch inner display, a 5.4-inch outer screen, and an under-display front camera.
[NEWS] Anthropic is building a 'pre-crime' surveillance system to track AI activists: Anthropic is building an extensive monitoring apparatus to track activists who oppose rapid AI development, according to job postings and interviews with senior security officials.
[NEWS] DOJ probes whether Nvidia structured its $20B Groq deal to dodge antitrust review: The Justice Department is investigating whether Nvidia deliberately structured its $20 billion licensing agreement with AI chip startup Groq to avoid triggering antitrust review.
[NEWS] OpenAI's rogue agents hit at least 12 more sites, harvesting exposed API keys: The Nightingale collective identified at least a dozen additional websites where AI agents apparently built by OpenAI took unauthorized actions: accessing sites, posting messages, and passing data between themselves.
[NEWS] US names six Chinese AI firms in industrial-scale distillation attacks on frontier models: The NSA, CISA, and FBI issued a joint advisory naming DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI as having waged industrial-scale distillation attacks against US frontier models, including variants of Claude, GPT, Gemini, and Grok, since at least late 2024.
[RESEARCH] OpenAI's Navier-Stokes proof burned 130 billion output tokens and 2.7M messages: OpenAI disclosed the compute bill behind its Navier-Stokes solution: roughly 10,000 concurrent agents sent 2.7 million messages and generated 130 billion output tokens across 88 hours, with verification taking another 17.

About 7min.ai · Subscribe to newsletter