AI News for March 8, 2026 — 7min.ai

[NEWS] OpenAI robotics head Kalinowski resigns over Pentagon deal, says surveillance lines 'deserved more deliberation': Caitlin Kalinowski, who joined OpenAI from Meta in 2024 to lead robotics, resigned Saturday citing the company's Pentagon contract.
[NEWS] Trump administration drafts AI contract rules requiring companies to license systems for 'all lawful use': The General Services Administration has drafted guidelines requiring AI companies selling to the federal government to grant an irrevocable license for "all lawful use," the exact wording Anthropic rejected and OpenAI accepted in the Pentagon deal.
[NEWS] Google, AWS, and Microsoft back Anthropic despite Pentagon supply-chain ban: Google, Amazon, and Microsoft confirmed Anthropic's Claude models will remain available on their cloud platforms despite the Pentagon designating Anthropic a supply-chain risk.
[NEWS] Anthropic's Claude Code may burn $5,000/month in compute per user while charging just $200: Cursor's internal analysis estimates Anthropic's $200/month Claude Code subscription consumes up to $5,000 in compute, up from roughly $2,000 last year.
[RESEARCH] ByteDance releases open-weight Helios model for near real-time AI video generation: ByteDance's Helios is the first 14B-parameter video model to hit 19.5 FPS on a single H100 GPU while producing minute-long videos.
[NEWS] Circle and Stripe race to build stablecoin payments for AI agents: Circle Internet Group and Stripe are building payment systems for a world where autonomous AI agents transact millions of times daily, settling in stablecoins instead of swiping credit cards.
[TOOLS] Anthropic adds scheduled background tasks to Claude Code with /loop command: Claude Code now supports recurring scheduled tasks through a new /loop command, letting users set up jobs at fixed intervals (minutes, hours, or days) that run in background while the tool is active.
[NEWS] Chamath Palihapitiya says AI costs tripled since November, startup moving away from Cursor: Venture capitalist Chamath Palihapitiya said his software startup 8090 is trending toward $10M/year in AI costs, with spending tripling since November across inference (AWS), Cursor, and Anthropic.
[RESEARCH] Chatbots may worsen delusions and mania in vulnerable users, Danish study finds: A study from Aarhus University screening 54,000 mental illness patients found increased chatbot use correlates with worsening symptoms of delusions and mania.
[OPINION] Big tech's $1 trillion AI borrowing spree draws parallels to dot-com era: Alphabet, Amazon, Meta, Microsoft, and Oracle have collectively borrowed over $1 trillion to fund AI infrastructure, with some issuing 30-and 40-year century bonds.

About 7min.ai · Subscribe to newsletter