An idea just occurs to me. Why not have the best Chess playing software report discoveries to a central database? Perhaps Rybka on someone’s computer discovers the ideal in a situation: it could then be reported, so that in the future, it need not be explored again, but simply retrieved. In this way, the computers would learn from each other. This would be an especial help for weaker computers such as those on tiny portable computers.
Ah…
Don’t you wish technology could just evolve faster? But be grateful, a few years ago we didn’t even have Fritz!
This is not as good a solution as you think. There will be some people concerned about privacy. Also, it is very hard to setup an online connection, transmit data, process the data, store it, and then retrieve it at will. With so many games played, you will need a huge databank to store the info.
There is already technology to accomplish this, people are just too lazy to implement it 
I remember a few years ago, someone wrote a distributed chess engine, and played it in a major computer tournament (thats going back a few years), and it didn’t actually do all that well.
I forget how many computers it ran on, but I think it was between 50,000 and 65,000 computers.
Although it doesn’t really answer your question, it does highlight that merely cranking out more positions in a brute force manner won’t neccessarily equate into stronger play. Hence, I doubt having a large database with pre-stored positions with a few “best play” moves to go along with that position would equate to a statistical advantage. Even engine using endgame tablebases have a statically insignificant increase in ELO: something around +3 elo for an engine rated above 3000 elo.
Do any of these computers improve your OTB results? Or doesn’t that matter anymore?
All the best, Joe
Computers have busted some of Tal’s best games, that doesn’t mean they weren’t brilliant at the time (and in real time)
Would UltraSuperFritz have made the mistake Anand made this week? I don’t know, I’m not sure I care.