Anthropic’s Mythos AI model tests limits of global cyber defences

Scritto il 18/04/2026
da
New system has sparked fears it could turbocharge hacking and expose weaknesses faster than they can be fixed