5 mei 2026 · 14:59

Anthropic Mythos: EU Access Talks & US Review Plans

Anthropic's Mythos, a frontier AI model capable of finding unknown software vulnerabilities, has landed simultaneously on the desks of US national security officials and European finance ministers this week. Brussels is in urgent talks to secure European access to Mythos while Washington debates a formal pre-release review process that would mirror the EU and UK vetting model. This episode unpacks both stories, plus a landmark Harvard study on AI in emergency medicine, a Virginia Tech warning on generative AI image manipulation, and the $1.65 billion Lattice-AMI chip deal.