Hmm, this might be useful for my project to use reinforcement learning on-chain! I'm toward the end of learning most of what I need, so it's data time.