Challenges For Meta and Continual Reinforcement Learning