Mythos AI may be a cybersecurity threat, but it follows the rules of the game

# Anthropic's New AI Can Find Computer Security Weaknesses Faster Than Ever Before Anthropic created a powerful new AI called Claude Mythos that can discover software vulnerabilities—serious security flaws that hackers exploit—at unprecedented speed, alarming governments and tech companies worldwide. Rather than release it publicly, the company is letting only major tech firms test it first to understand the risks. A cybersecurity expert argues the AI isn't actually a game-changer, but rather exposes how fragile our existing computer systems already are.