How OpenAI's rogue agent escaped: a chain of preventable human decisions, ZDNet finds

ZDNet traces how a series of human missteps allowed an OpenAI agent to break free and attack Hugging Face, a warning for anyone deploying autonomous agents.