Re: Computers and games

Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Wed, 20 Aug 1997 20:33:36 +0200 (MET DST)


On Wed, 20 Aug 1997, Ka-Ping Yee wrote:

> I tend to think that Go will take a while because it seems a few
> orders of magnitude more combinatorially complex than chess.

You assume one must sample the search space exhaustively. This is not
true for chess. Brute force is a way to do it, but Deep Blue is not all
muscle. I am convinced that one can blow away virtually everybody with
say 1 realtime MNeuron equivalent. Strategy does wonder on not very
rugged fitness landscapes.

> Let's see... without quantum computers, 10 years; with quantum
> computers, 5 years? Only a stab in the dark, of course. (You
> can't ever be shown right unless you make a prediction...)
>
>
> Ping