AI lab Anthropic announced this week that it had developed a powerful new model the company believes could "reshape cybersecurity." It said that its latest model, Mythos Preview, was able to find "high-severity vulnerabilities, including some in every major operating system and web browser." (AP)
