Inside Claude Mythos: Why Anthropic held back its most advanced AI

Anthropic's advanced AI, Claude Mythos, capable of discovering software flaws, was kept from public release due to misuse concerns. Instead, it powered Project Glasswing, a collaborative effort with tech giants to secure critical software, identifying thousands of vulnerabilities. While not fully public, Mythos now offers controlled access for specialized cybersecurity needs, marking a shift towards integrating powerful AI in defense.