It's easier not to think of it as cheating with all the emotional baggage that that entails. If you just regard it as being the way the world works in the game then there is no problem coming to terms with it. I think everyone accepts that the AI will not be perfect because that is too difficult to achieve within a reasonable budget for such a complex game and we all hope and expect that it will nevertheless be getting better with each new version.
The most important point is that the "cheats" should not be too obvious in that they should not show as blatantly unrealistic behaviour and thus spoil the illusion of the game. So if, say, the AI builds things five percent faster or has a five percent weighting in combat that would not hurt the game (it is just a way of having a stronger opponent) but if units suddenly pop up on your doorstep when you know they were not there before, that would be rather like adding magic (or matter transmitters) to the game, but only for the AI.
As an intersting corollary, whenever we exploit our knowledge of how the AI works rather than making logical decisions based on the game model then we are cheating just as much. For example, fortifying a phalanx on a mountain because we know the AI will expend resources attacking it for no sound reason until it is destroyed, or (in the original game) changing production to a wonder to get the bonus from caravans and then changing it back.
For my part I am happy to take a middle line. If Firaxis make a decent effort with the AI (and they will) and the "cheats" are not too obvious, then I won't poke around overmuch in the inner workings and I will enjoy the game for what it is intended to be. Willing suspension of disbelief helps just as much with this as it does when reading fiction.
Referring back to the opening post, the computer's ability to make rapid arithmetic calculations is no cleverer than an animal's ability to breath, walk etc. The human brain does far more work far quicker than a computer does and if it had an arithmetic circuit built in, then it would calculate just as quickly (in fact that may explain how some people are apparently able to do complex arithmetic almost instantaneously).
The advantage chess computing has is the enormous resource that has been thrown at the problem for the last forty years.
By the way, Go is a million miles from having infinite possibilities. I doubt if it even has as many as chess. One of the reasons that so much AI research has been devoted to chess is that although it is a closed system (and therefore relatively easy to define theoretically) it is thought to have sufficient complexity to act much like an open system (which is impossible to model completely) and therefore is fertile ground for developing algorithms of use in far more complex systems like economic behaviour and weather.
I've got a feeling I may have said most of this before. Hope I haven't turned it upside down this time
------------------
"Ridicule can do much...but one thing is not given to it, to put a stop permanently to the incursion of new and powerful ideas"
-Aaron Nimzovitch
