Cooperation Through Reinforcement Learning 8 February 2025 Playing against another AI produces unimaginably useful results.