
Anthropic Unveils AI Cybersecurity Model It Deems Too Dangerous for Public Release
Anthropic announced Project Glasswing, using its unreleased AI model Claude Mythos Preview to find vulnerabilities for major tech companies while restricting public access.












