Anthropic’s Claude Mythos Preview Uncovers 10,000+ 0-Days in Project Glasswing

Anthropic has revealed the staggering initial results of Project Glasswing, a collaborative cybersecurity initiative designed to secure critical infrastructure using advanced AI before malicious actors can exploit it. In its first month, the project leveraged the unreleased Claude Mythos Preview model to autonomously discover over 10,000 high- and critical-severity zero-day vulnerabilities across the world’s most […] The post Anthropic’s Claude Mythos Preview Uncovers 10,000+ 0-Days in Project Glasswing appeared first on Cyber Security News.