A new AI model automates the search for cybersecurity bugs and system vulnerabilities. This capability accelerates both defensive patching and offensive exploitation. While Claude Mythos and Project Glasswing highlight these risks, the tool's dual-use nature creates a volatile environment. Security practitioners must now defend against machine-speed vulnerability discovery that bypasses traditional human analysis.