From: Joonas Kekoni <mlist <at> AURORA <dot> TKY <dot> HUT <dot> FI>
Date: 18 may 1997
Subject: Re: Real chess computing

> I don't know that this is necessarily
> the major breakthrough in Artificial Intelligence which it is being touted as?

I dont think it is. Chess programs have been beter than normal people for
long long time. I is big event in chess history though.

But is computer chess really artifical intelligence at all? It is counting
values for limited number of moves depending on algorithms and databases.

It is more clever than newtonian iteration? The goal is to give as
accurate solution to pregiven problem as possible. It is much more complex
and complicate that, but that is not what i am talking about.

There are much more cleverer ai techniques around. When limited to games
self learning programs and programs playing games by _rules only_.

> Is there any chance that this machine can or will be programmed to play shogi?
> I would imagine that the problem of playing shogi (larger board, promotion,
> drops) might give Deep Blue a slight headache? After that, we can try it with
> Go?

Well It is only a parallel processor. Many fast cpu's hooked together with
fast nework. Since chess is easilly splittable, even fast network is not
needed. Hey when shall we play against Gasparov using the "internet chess
machine"? Similarry than the dess challence.

What _REALLY_ maters is not the computer but the program.
And the program is written for chess, becouse that what the big (western)
public knows. It is also easiest to program.

> Should the NSR think about organising a challenge to Deep Blue or am I just
> being silly?

That would be the next challence, but does it give enough advertisement to
IBM that it would cost. That i$ the que$tion.