“Anthropic's new AI model, Claude Mythos, reportedly discovered thousands of new software bugs and security vulnerabilities across all major operating systems and web browsers.”