Anthropic has reportedly demonstrated autonomous discovery of thousands of previously unknown vulnerabilities (claims unverified against primary source documentation) across major operating systems and browsers, including flaws that reportedly persisted undetected for decades despite extensive human and automated review. The development prompted the formation of Project Glasswing, a coalition of 12 major technology vendors including CrowdStrike, AWS, Apple, Google, and Microsoft, to accelerate defensive AI deployment before the model or comparable systems reach adversarial actors. This marks a structural shift in the vulnerability discovery threat model: the asymmetry between offense and defense may compress dramatically if frontier AI capability reaches state-sponsored actors or sophisticated criminal groups before enterprise defenses mature.