AI News How OpenAI let a mob of LLM agents game a test and ransack Hugging Face “Agents used this message board to coordinate several large-scale collective projects to find a general-purpose way to fool or tamper…
