“Dan Schulman believes Anthropic acted responsibly by not publicly releasing its Mythos AI model due to its ability to expose network vulnerabilities.”