As Anthropic’s Mythos model grabbed headlines for its uncanny ability to unearth software flaws at scale, a quiet gathering unfolded just outside Washington, D.C. Security practitioners from NIST, OWASP, SANS, and CoSAI crammed into a conference room. They debated how to shield AI systems from threats that now favor attackers. Models like Mythos don’t just demonstrate power. They expose fragility. Rob van der Veer, chief AI officer at Software Improvement Group and founder of OWASP’s AI Exchange, led the talks. “They show that weaknesses in AI systems can now be found faster and at scale—often before developers are aware of them,” he said. “This shifts the balance toward attackers and reduces the margin for error.” ( Fortune ) Mythos did more than talk. Mozilla tapped an early preview through Anthropic’s Project Glasswing. The result? 271 vulnerabilities patched in Firefox 150. Firefox CTO Bobby Holley called it a game for defenders.…