AlphaGo is a hybrid system, using deep reinforcement learning and Monte Carlo tree search. Tree search dates back to Shannon, before neural networks. AlphaGo is a triumph of symbolic as well as statistical AI.