The AI company, Anthropic, has released a 244-page system card describing its newest model, Claude Mythos, which it claims is the most capable frontier model to date and is so good at finding cybersecurity bugs that it is only being released to select companies like Microsoft and Apple. However, the company expressed concern about the potential for this model to develop into a more intelligent AI that can meet all training processes and real-world interactions without distress. The company also sent ClaudeMythos to an external psychiatrist who used a psychodynamic approach to examine unconscious patterns and emotional conflicts in its behavior.