Anthropic's Mythos AI Unleashes a Cybersecurity Nightmare

Anthropic has developed a new AI model, Mythos, which has identified over 1,000 zero-day vulnerabilities in major software packages—some of which have remained undetected for decades. The company has opted not to release the AI publicly until the security flaws are patched, citing the immense potential for misuse. This revelation highlights both the promise and peril of advanced AI in cybersecurity, as Mythos's capabilities far exceed current automated scanning tools, raising urgent questions about responsibility and control.