Last month’s preannouncement of Anthropic’s Mythos, an unreleased AI model that the company warns is extraordinarily good at finding and exploiting security vulnerabilities in everything from operating systems and web browsers to industrial software and health-care platforms, sent More »

