Author: Eugene Nalimov
Date: 16:35:23 01/05/99
Go up one level in this thread
On January 05, 1999 at 19:17:11, Christopher R. Dorr wrote: >On January 05, 1999 at 18:56:51, Fernando Villegas wrote: > >>Dear Chris: >>We does not need to speculate about what could happen IF Bill gates decides to >>make a run in the chess field. We already know what happens when you put >>toguether a very great amount of resources: critical mass appears and great >>jumps are the rsult. Manhattan proyect is an example: many phisicyst already >>knew about nuclear reactions et all, but it was needed a huge amount of money, >>personnel etc to make it work. That's the reason Germany was not capable, least >>Japan. > >Well, one reason was that German had lost much of the talent necessary because >of the emigration of Jewish scientists before the war. But this example, I feel, >is not a good model for chess programming. While building an atomic bomb had >many components (delivery, stabilization of fissionable material, production of >that material, implosion of the critical mass, etc.) that could be worked on >independently (under the direction of one overall leader), chess programming has >few such components (opening book, interface, database come to mind) that affect >the strength of the chess engine. I can't really say 'OK georger...you do the >piece square tables...Mike you handle hash table management....Lisa, you get the >evaluation function...Ed, do the learning functions". I've written chess >programs before (albeit very bad ones), and know how interdependent various >portions of them are with each other. > > >>In fact, even if Bill does not do anything about this, it is already happening, >>as in any industry once it has reached certain level of development. I think >>that Ed and Christophe venture is a sample of that. I am sure they have already >>discovered great ideas and new grounds for progress due to his colaboration. In >>any science or technology, when you put people to work toguether and gives them >>money enough, results begin to flow in mass. To think that the esential thing is >>personal creativity of this or that genius is somewhat naive. Truly genuses are >>badly needed when no organization and many tools and resoruces are present, but >>once organization exist, the accumulative work of high intelligences get more >>things that anything an isolated genius can get. >>besides, do not believe we have already reache a top level in chess programming. > >I agree that collaboration is very important. But which do you think is going to >produce the better outcome: two world-class leaders in the field (with vast >amounts of prior experience and education in the field between them) or one team >leader, and 50 fantastic programmers, who know *nothing* about chess >programming? Why do you think that optimal group will be "1 leader + 50 fantastic programmers"? I'd suggest something like "1 leader + 2 grandmasters + 5-7 fantasic programmers + 5 very good testers + 1 specialist on particular CPU architecture + 1-2 administrators + 2 fast computers per developer + lot of *very fast* test computers". And I'm reasonable sure that such a team will produce much better results than team that consist of just 2 leaders. Eugene >>They are very primitive in the fact that, although they get results, the do that >>trought a kind of accumulated practical wisdom in the same sense alchemist got >>things trought many years of practice. A great research team with money to spare >>could get a fundamental jump in terms of creating a real AI chess engine instead >>of what we have now, just a machine that run a list of specific routines that >>works fine toguether in the most unscientific way, just adding weights and >>testing the mix with thousands of games. Alchemy, again. >>Anf , of course, don be deceived by the apparent trivial meaning of "just" going >>from 2600 to 2700 or even 2650; it is a great jump and you as master know it. If >>a great reasearch team can do it, IT will be a wonderfull jump. >>fernando > >But I really don't know where the limit is. You say we're not near it yet, >perhaps we aren't yet. But is it reasonable to assume that with current >hardware, we can't create a 4000 rated program? I think that's pretty clear. >What about 3000? Or 2850? There *is* some limit using currently known >techniques, and I happen to think that we may be fairly near it. > >What about new techniques, you say? Great! But I don't think a team of 50 >inexperienced programmers is even a *fraction* as likely to discover new and >radical ideas (that work) as would be an Ed or a Bob Hyatt. > >I just don't believe that that a Microsoft Chess 2000 product would be that >superior to a Rebel 2000 or a Fritz 7 product; throwing money at a problem >doesn't always solve it. > >Chris Dorr
This page took 0.01 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.