A three-person team has hacked OpenAI after they used Anthropic’s Claude to help exploit a flaw linked to the company’s community forum. The team gained