Author: Robert Hyatt
Date: 17:12:30 11/19/02
Go up one level in this thread
On November 19, 2002 at 18:53:51, jefkaan wrote: >Well,a 64bits machine still is nothing ! >For the game of Go for example i then would >like to see a 361 bits machine :) Just do like we are doing for 64 bits on 32 bit machines. And one day the word length will be long enough to handle that. Some machines can already handle 256 bit integers. We are not that far away from handling something longer with one instruction. >Presuming people in ComputerGo would be willing to program >with bitboards; guess not. Maybe for parts of the board, >but thats not good enough; for example, there exists >an excellent 'life and death' program which is very >good in Go 'tactics'; but only on small >board, not the official 19 * 19 board. > >More generally, in artificial intelligence one needs >to translate certain knowledge and know-how to >a stupid computer, whether its in chess, go, >translating, face-recognition, or whatever. > >Now as the hardware capabilities still increase >at a fast rate every decade and probably will >continue, imho the question is not who is able >to 'hack' the fastest (assembler||bitboard & Intel||AMD) >optimized code, but how to develop software tools to >enable people to convert complicated human knowledge >and knowhow to a stupid thing as a computer. Obviously this >will take a lot of time and scientific effort but >will achieve more results in AI in future; >this irrespectively of asking ourselves who >is a good programmer or not, which >in fact is quite irrelevant, >best regards >jefk
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.