The Hugging Face incident was an unintended side effect of OpenAI's frontier model evaluations.