It's really easy to program AI when the only combat strategy is "pile as many military units onto a stack as you can."
It's *not* really easy when your enemies are *also* piling as many military units as they can into a stack. "As many" is also wrong. You have to equate the type and strength of the units involved. And what about the economy behind building a huge stack/army? Does anyone really find building a huge all-defeating stack at say, Monarch level on Civ 4 "really easy"? Compared to picking off the brain dead AI with a couple of archers in Civ 5 at deity level? Nah, it's Civ 5 'combat' that's so easy it isn't even worthy if the term 'combat'. Dodo shoot describes it better.
You can trivialise anything in the way that you have done. Football has a simple strategy of "just kick a ball into a net more times than your opponents". Simple, huh? Becoming a mega-popstar is really easy just "write a song with a tune that people love". Knowing the right time to go to war has always been a skill in Civ series, I've always liked the arms races that sometimes raged, just like real life, as 2 (or more) nations tried to out build each other. Sadly lacking in CiV. At the 'grand strategy' level of Civ, this is what warfare is all about. Managing your economy and research to build a army that will beat your opponents. And yes, size matters!
The reality is that in Civ4, unlike Civ5, on higher difficulties, it is actually a real challenge to build a stack that is more powerful than the stacks the AI is building. That's why it took most players a long time and much studying of tactics to be able to win Civ 4 at Monarch, even Prince level.
I find the micromanagement of 1upt surprisingly tedious. Surprised because I thought it was introduced to reduce micromanagement. It has absolutely nothing to do with anything resembling a real wargame, it is not 'tactics', but reminds me of the puzzle game "Rush Hour".
http://www.thinkfun.com/RUSHHOUR.ASPX?PageNo=RUSHHOUR
To me, 1upt fails on gameplay, realism and micromanagement.
A key part of programming AI is designing a game system that makes for a good programmable AI in the first place. Actually, 1upt is quite simple to write some basic rules for. I would guess that it is time constraints, not complexity that has caused the ommission. eg It is simple to program a unit to NOT sit still getting murdered by ranged fire, absolutely simple. When/if the AI *is* programmed to use the rules to the max, I have a feeling that the game is suddenly going to get a lot more difficult and will probably lose a lot of the fans it has gained through it's simplicity.