Listen

Description

Bloomberg reports a small group of unauthorized users accessed Anthropic's new Mythos AI, a model the company says is powerful enough to enable cyberattacks.