🧙🏻 Check out Guru: how cross-domain RL supercharges LLM reasoning.