Anthropic's AI Model Identifies Thousands of System Vulnerabilities

Anthropic has introduced Claude Mythos Preview, an artificial intelligence model developed to detect security flaws. This AI successfully identified thousands of high-severity and zero-day vulnerabilities across various systems. Notably, the model uncovered long-standing issues, including a 27-year-old OpenBSD bug and a 16-year-old FFmpeg flaw, showcasing its capability in cybersecurity analysis.