
Anthropic's Claude Mythos and the new rules of cybersecurity in the AI era
Erik Davidsson
Head of AI
AI models can now find and exploit software vulnerabilities faster than any human team. The mean time to exploit a known flaw has collapsed from 63 days in 2018 to -7 in 2025, while patching still takes 39 days on average. This asymmetry favors attackers, and frontier models like Anthropic's Mythos are widening the gap. The good news: the same capability lowers the cost of defending.
























