From alchemy to electricity: how our community can create breakthroughs in reinforcement learning