Someone Made an Infra-Bayesian RL Agent, and it One-Boxes!

Someone made an infra-bayesian RL agent, and it learns to one-box. Arxiv paper, GitHub repo. Discuss