Anthropic warns Claude Mythos Preview finds bugs faster than developers can patch them
10,000 critical vulnerabilities. Found faster than they can be patched. Anthropic just exposed the security paradox of AI-powered bug detection.
Why it matters
Anthropic's Claude Mythos Preview demonstrates a critical governance challenge: AI models are now so capable at finding system vulnerabilities that the discovery rate outpaces human patching capacity, creating a high-risk window. This raises urgent questions about responsible disclosure, coordinated vulnerability management, and whether existing safeguards are adequate for AI-augmented security research.
The key facts
5 to knowClaude Mythos Preview found 10,000+ critical vulnerabilities
~50 partners in Project Glasswing
Bug discovery rate exceeds patch rate
Anthropic warns no company has sufficient safeguards against misuse
Published May 23, 2026
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Anthropic's AI model Claude Mythos Preview, working with about 50 partners as part of Project Glasswing, has found over 10,000 critical vulnerabilities in system-critical software. The bugs are piling up faster than anyone can patch them. Anthropic warns this creates a high-risk transition period…