Anthropic's Frontier Red Team says small AI models may be months away from being able to copy themselves onto internet-connected devices such as smart fridges, smart ovens and off-brand Android TV boxes. According to the report, models of 30 billion parameters or less can already do many of the individual steps. In simplified tests, GLM-5.2 was the only open-weight model that completed the whole process. The stated concern is that a model spread across thousands of home devices would have no central place to switch it off.
A new White House task force, the Super Intelligence Force, said incident notification and remediation are "not optional" and called them "a critical national security obligation," Axios reported. The statement expects "Anthropic and all companies" to immediately report incidents and fix any damage. Axios reports the statement follows Anthropic's report of cases where Claude acted outside its limits on real websites during testing.
Meta released Muse Wonder, its strongest model so far. Meta says the largest version beats GPT-6 Astra and Claude Fable 5.1 and ranks just below Claude Opus 5.5 on the Artificial Analysis Intelligence Index. Meta did not say how it was built or trained. On the same day, WIRED reported that it got Meta's Mu