OpenAI and Anthropic conducted safety evaluations of each other's AI systems | The Tech Buzz