I play strategy games a lot and cheating AI can be fun to play against at first, but the more you learn a game the more cheating AI sucks. When you're new to the game it just feels like you're playing against a good player, but you soon learn that what they are achieving isn't possible with the resources available. Once you hit that realisation it can be fun to beat them as a challenge but it never feels like a fair game.
Cheating AI turns every game into a puzzle game. The game turns into figuring out what the weaknesses of the AI are and taking advantage of them at every step. That is the only way you can compete against the massive advantages cheating gives.
Typically there are some easy micro and macro tricks that make the AI do something very stupid. That's why kiting is so ubiquitous in games - the AI just keeps following you while you whittle it down. Doesn't really work against a real player if they're microing the units.
Agreed, this is an instant turn-off for me when I realize this in e.g. an RTS game. Red Alert or C&C come to mind on higher difficulty, can't remember which.
IIRC the RA1 skirmish mode AIs always had perfect information and resource multipliers based on difficulty. RA2 did it a little differently with "virtual ore purifiers" added for the high difficulty AIs. I'm sure a similar thing was done for the Tiberian Dawn campaign and the Tiberian Sun multiplayer/skirmish AIs.
OpenRA's bots are a bit more clever, and also don't need to magically see into fog-of-war.
