reinforcement-learning

OpenAI's Hide-and-Seek Findings, the Systems Perspective

Yes, the agents cheated, but what does that mean for the system?