Anthropic's Mythos AI model sparks fears of turbocharged hacking
Anthropic announced its Mythos AI model on April 7 and refused to release it publicly, citing its unprecedented ability to find and exploit software vulnerabilities . The company is instead limiting a