Author: Dann Corbit
Date: 10:50:49 01/13/05
Go up one level in this thread
On January 12, 2005 at 22:04:34, Michael Henderson wrote: >>>YES! Advanced Chess is where New Theory will come from... >>>Humans USING Computers! NOT Computers advancing Theory by themselves... >>>The Poor Creatures have NO INTELLIGENCE! They rely on Humans... >>>Don't they? >> >>They rely on humans for now. Eventually, they will program themselves and do it >>much better than us. By the year 2049 a single computer will have the >>intelligence of every human on earth put together. We will be pitifully stupid >>by comparison. >> >>In twenty years, I doubt if humans will have any clue compared to a computer as >>far as chess analysis is concerned. It will be like a 5 year old trying to play >>Kasparov. What can he tell Kasparov that he does not already know? >> >>Obviously nothing at all. > >I know you have to deal with yergin, but I will say this as a programmer: >Computers must rely on code, and code is written by humans. So it's possible to >say that computers will always rely on humans. Code written by programs is a >derivative of a human's code. Hence programs cannot code better than humans, >nor will they be more intelligent than humans. They may be able to do some >tasks faster than us (eg computer chess), but they have to compete with humans >in abstract thought. There are several problems implementing abstract thought >(eg GO) in a computer, unless you can give an example refuting this. I have thought about this idea a lot and I think it is a very sound notion. In short, as computer programs increase in complexity, will they also increase in error probability? If I have a program with one trillion lines of code, will it produce even more mistakes than a human will? It is a fascinating question for which I do not think we can know an answer. I think a partial solution to problems of this nature is to use formally proven components and in parallel. But that is also an imperfect solution. Another method of attack is similar to the RISC verses CISC idea. Make very simple and not so intelligent components that are absurdly fast. Perhaps the ultimate solution to chess is a hardware solution like the Deep Blue approach, on a scale one trillion times bigger. At any rate, we are living in interesting times. Projections of the future are always something a bit commical well into the future, though the science fiction writers have a pretty good record in this regard.
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.