Scaling Reinforcement Learning: Environments, Reward Hacking, Agents, Data