Anthropic's Mythos Found Millions of Security Vulnerabilities | Artificial Intelligence: AI News, ChatGPT, OpenAI, LLM, Anthropic, Claude, Google AI | Podwise