I hope they can do some reinforcement learning from Stack Overflow moderator decisions also