Jacob Coxon's AI warning and the control lesson from the OpenAI and Hugging Face incident
Jacob Coxon’s resignation and Evan Hubinger’s AI-risk estimate have revived the safety debate. A documented agent-control failure offers a more immediate lesson.