I wonder if he could identify those long-term-thinking moves by playing a version of AlphaGo on a smaller board, then gradually increasing the size of the board over successive games once he is able to consistently win or tie at a given board size.
I also wonder if when Alpha Go plays itself, does it always win when it is black or white, does it always tie, or is it a mix?
I was thinking the opposite. AI has already been beating humans on smaller boards, so maybe the only chance we have is to start playing go on bigger and bigger boards.
Huh, interesting. I'm not well-versed in Go at all.
It would be interesting if AlphaGo's maneuvers remain opaque to humans for an extended period of time. Can anyone at this point say with confidence that its strategies will indefinitely remain unknown?
If this plays out anything like it did in chess, in a few years we'll have very strong machine players that we can use to figure out with high confidence the quality of both players' moves.
Obviously the chances for humans to win, increase with the board size.
Given the computational power (=number of cpu/gpu) is fixed and the time for each move to be done is fixed.
Perhaps a future "dr. evil / skynet" A.I. will first try to conquer the microchip production plants to increase its computational power and memory. goodbye taiwan, goodbye south korea, goodbye usa...
There is typically a point advantage of around 7 points when you get the chance to play the first move. In professional matches, white will be given a 6.5 point bonus to compensate for this. So yes I think it's fair to assume that black usually wins. Actually, AlphaGo could probably be used to calculate the perfect number of points, because it has changed over the years. In China they have a 7.5 point bonus for white.
I also wonder if when Alpha Go plays itself, does it always win when it is black or white, does it always tie, or is it a mix?
So much to learn from this.