#1Project Glasswing: An Initial Update
Anthropic's Project Glasswing is a collaborative initiative with roughly 50 partners to secure critical software infrastructure before advanced AI models can be weaponized against it. Using Claude Mythos Preview, the project discovered over 10,000 high- or critical-severity vulnerabilities in systemically important software within weeks. The key insight: AI dramatically accelerates vulnerability discovery, but human capacity for verification, disclosure, and patching remains the bottleneck.
