I'm pretty sure the message board that was recently swarmed by OpenAI agents to collude on benchmarks would like to disagree.