The Trouble with Computers and their application to solving chess

The Trouble with Computers and their application to solving chess

Icouldabeenachampion
Icouldabeenachampion
Mar 20, 2018, 9:29 PM |
0

Will computers ever solve chess?

Here’s the thing about computers. A computer, if sophisticated enough – which is only ever related to speed of calculation, speed of search and its storage capacity, can solve any problem that has a finite number of possibilities given enough time. It is envisaged that there are some 10^120, so a database of the said number is required and the time and speed to search the database. This is not difficult in the true sense of the word, it is just whether the amount of resources will ever be available or should be made available to dedicate to such a pointless endeavour.

 We hear of quantum computers but this is only a computer using a different mathematical number base that can process millions of  times faster than the latest crop of supercomputers. In fact despite all the bollacks you hear on the internet no such computer exists that can be usefully used. The issue of course is that the current crop of computers can not make intuitive decisions and neither will quantum computers. These so called intuitive decisions lauded by the science world are only the application of very smart searches on a finite number of possible solutions and a finite number of stored cross references.

Here is a simple test, one of only many that proves a point.

 Can you tell the difference between a dog and a cat ? of course you can, kids of four can do it without training or intensive programming; kids of four have not been fed a picture of every dog and cat breed in existence, they, like you, just know. This very simple task for a child which will yield almost perfect results (although the Chinese seem intent on breeding dogs that look like nothing else on Earth) is beyond any computer. A computer faced with the same task would require a database containing images, and digital recognition points like those used in facial recognition, of every breed of dog and cat in the world to do the same task. I do not say that this will always be impossible, but until the algorithm for recognition is developed and input it will never happen, and it will still be an algorithm using a finite number of solution cases. We do not know how a child performs this trick so how can it be imitated or programmatically designed.

Judging from the talk on this forum, I suspect that I am preaching to the converted, however there are plenty of dimwits out there  and I can already hear the howls coming from the aficionados, Tesla owners, and watchers of too many Star Wars videos. Screams of “AI” in a thousand voices led by the lovies (and that space litterer Musk) of Silicon valley. When they talk of AI it is a synthetic AI and not what is implied by Intelligence at all. It is a supercomputer containing enough stored information and cross references to be able to find the solution to its current problem, everything that comes out is the sophisticated result of a sophistic programmers input.

So to answer the basic question. I don’t know, but I expect that humankind will have destroyed itself long before the necessary resources to do so ever become available.