Fox News AI Newsletter: The AI model that's too dangerous to go public

Anthropic built an AI model called Mythos so effective at finding software vulnerabilities that the company decided it is too dangerous for public release.