A researcher who quit Anthropic says AI will dominate daily news next year and should be treated as a weapon rather than a product — a warning aimed at governments, investors, and anyone with a job that software can do.

• Jacob Coxin resigned from Anthropic after three years of pre-training research there and at OpenAI, walking away from unvested stock, and said both companies are racing to self-improving superintelligence.
• His resignation post drew 175 million views and helped push AI onto the agenda at a White House meeting and in talks with Xi.
• The core claim: next year's models will be good enough to automate large chunks of work, including the jobs of AI researchers themselves, which makes progress accelerate.
• He argues the technology is powerful enough that it should be nationalized, and finds it strange that Trump is leaving the keys with Sam Altman and Dario Amodei.
• Private labs are already making calls that belong to governments — who ends up unemployed, whether a hostile state can steal a model, whether someone builds a bioweapon.

Outlook: Expect louder pressure for government control over the big AI labs as model capabilities jump again next year.