The X algorithm should be RL'd with its reward being the Kardashev scale