Deep reinforcement learning is interesting and has plenty of potential. But highlighting AlphaGo as an example of reinforcement learning is like undermining the concepts of reinforcement learning.