The Information reports that the three largest labs have discussed a joint organization for AI safety standards. Which means the standards their own models would be measured against.
At the BRICS summit in New Delhi over the weekend, Xi Jinping said China would take the lead on AI cooperation among developing countries. The Minister of State Security weighed in on Sunday.
Anthropic's business lead put the ballpark at $70 to $120 per user per month in an interview with WealthManagement. Wealth.com will feed Claude trust and will summaries.
SoftBank dropped 10% in Japan, HPE 8%. Palo Alto and CrowdStrike rose double digits - both sit in on early tests of unreleased Anthropic and OpenAI models.
Mustafa Suleyman says the document has been in the works for about five months. It's meant to govern development starting in 2027 - and Microsoft wants feedback first.
Trump called the warnings a hoax and a scam on Monday. His own AI adviser pushes back on Amodei for an entirely different reason: product liability.
OpenAI has filed confidentially, but it is not going public this year. The interesting part is Altman's reason: not the market, but the safety debate.
Bloated skill descriptions, mandatory reading before every edit, strict approval rules: what kept older models on track now slows Astra down. Almost all of it maps onto Claude Code.