Anthropic has unveiled its new Claude Mythos Preview model, claiming a game-changing leap in cybersecurity. The AI can identify and develop exploits autonomously, making existing defenses obsolete.

While some see this as a critical turning point, others are skeptical. Existing tools already make finding vulnerabilities easier, albeit without the claimed complexity. Nevertheless, industry leaders like Cisco agree it's a significant challenge, with early adopters gaining a crucial edge.

The release to Project Glasswing participants aims to give defenders an advantage over attackers in exploiting these new capabilities. However, critics argue this is just another phase of AI hype, comparing it to learning machine guns when others still use bolt-action rifles.

Regardless, the discussion around Mythos highlights how long it takes for security practices to evolve. Past breaches have led to significant shifts like zero trust and secure by design. Could Mythos be the catalyst for yet another paradigm shift?