Re: computer-go: Applying Moore's Law to Computer Go

[Sam Sloan <sloan@xxxxxxxxxxxxx>]

I disagree.

It was predictable that the strength of chess playing programs would
increase as the speed of computers increased because by the mid-1980s we
had solved the basic problems of how to write a good chess playing program.
I competed in the World Computer Chess Championship in Cologne, Germany in
1986 and the programs we had then work the same as now. The programs now
are faster, more efficient and contain fewer bugs but they are basically
the same as in 1986. The main reason they are stronger now is the CPUs run
faster.

In the case of go, nobody has come up with the idea yet of how to write a
strong playing program. We do not even have the basic algorithms yet, so
how can we make them work better or more efficiently?

Until somebody finally comes up with THE BIG IDEA, I do not believe there
will be any significant improvement in the strength of go playing programs.

I am in Japan now. I attended the 10th Annual World Amateur Go Pairs
Championship in Tokyo on Sunday and met for the first time many of the
people I have featured on my web site.

Sam Sloan

http://www.samsloan.com/umezawa.htm

At 00:48 99/11/14 -0600, Clay ChipSmith wrote:
>Applying Moore's Law to Computer Go raises guarded hope.
>
>(See http://webopedia.internet.com/TERM/M/Moores_Law.html for a definition
>of Moore's Law).
>
>Computer Go strength grew 1 rank each 20 months
>during the last decade.  I reckon this trend will continue
>in part for the foreseeable future.  Particularly, let us
>speculate growth to continue 1 rank each 18 to 30 months
>for another 2 decades.  For example, in 2020 Computer
>Go strength could reach 4k* to 1d* (IGS scale).
>
>The profound scaleability of cmos is the foundation
>of Moore's Law.  Computer Go had both hardware and
>software dependency for strength growth.  Moore's
>Law may cover the hardware side of Computer Go
>strength dependency.  A combination of programming
>efficiency and ever increasing colaboration may cover
>the software side of Computer Go strength dependency.
>
>Best Wishes, Clay ChipSmith ><>
>
>
>
>