When AlphaGo beat the human champion in GO recently, it was seen as an earth shattering achievement in the field of Artificial Intelligence. But what special algorithms did AlphaGo use ? why couldn't we achieve the same feat when AI was hot in the 1990s ? While answering these questions, I will explain how Go compares with chess in terms of difficulty of evaluating different moves in the game and why computers reached grandmaster level in Chess in the 1997 but only recently did computers beat high ranked humans in Go. Since neural networks is the main tool of choice in AlphaGo and many other today, (it is also the most popular topic in top-class conferences such as NIPs and ICML), I will explain the recent advancements of neural networks in the field of Computer Vision and their special properties compared to other algorithms. Further, I will touch briefly on the AI / Neural Networks winter to present an interesting comeback that this field had.