Thank you too WeatherLover, I've read the article.
AlphaGo Zero also discovered new knowledge, developing unconventional strategies and creative new moves that echoed and surpassed the novel techniques it played in the games against Lee Sedol and Ke Jie."
What exactly is this new knowledge other than extremely fast calculation and trial and error of every possible move that could've be done? If the data is build up by playing against itself, it has to know the rules before it can decide what will be a better move and what isn't. By being able to play a zillion games in no time and ongoing trial and error it seems to have reached perfection. That's cool, having the potential to scan and evaluate every move that could've been made by anyone anytime anywhere. Of course, the outcome will surprise us. We are individuals, not knowing what all other players know. Humans also have emotions and sometimes do things by intuition, not knowing where and when the same situation occured in the entire history of this game. This "cold" approach and the relentless way the algorhythm is programmed gives it a unique helicopter view that is unparalleled in exhausting all the meta-data.
Impressive? Yes, very! But I remain sceptical because it still didn't go beyond what is has been programmed for... play against yourself, remember and store the data, improve by trial and error and become better and better and better up to the point of perfection. No human being will be able to have access to so much data at once by using memory, but we can build machines that do it for us because we program them to do that specific task, they have no free will.