What AlphaGo Can Teach Us About How People Learn