Researchers reportedly use Anthropic's Claude to hack OpenAI

Anthropic says Claude is helping build its own successor, leading 26% of the company’s AI research and development under human supervision.