Regarding the actual points you made, they are arguable and counterpoints cannot simply be dismissed with jibes. Yes, stacked units present a formidable challenge to poorly defended cities, no argument there. But there was never anything challenging about countering a stack of doom.
Incorrect. I suspect you never actually played with the so-called "stack of doom". There was, in fact, some challenge involved in "countering" a "stack of doom". A lot of the the primary strategy of a Deity-level player was to avoid and predict early war as much as possible, because war was costly and not preferable to peaceful expansion until the land ran out. It was not always even possible to defend against an enemy stack with the advantages the AI had in the early game, because they started out so far ahead.
As long as the player kept up with the AI in military techs and kept military units up to date, 2-3 units with city walls was usually enough to defend a city or enable it to hang on until reinforcements could arrive. Not really a deep strategy there, and i suppose a stack of doom is tactical in the sense that it was the tactic to city-bust within the confines of the game.
The merits and intricacies of stack-based combat were recently discussed at length in
this thread. I don't feel like retreading the same territory, but suffice it to say, there's a lot of meat and depth to stack-based combat as it was implemented in IV. If you care to read why, the information is a few pages into that thread.
With the current version of the game, i see more complex movements with AI units. They are good at approaching cities, but instead of attacking full-throttle, each individual unit seems to have a bias for its own preservation and in that regard units appear to lack discipline in the face of battle. Barbarians on the other hand seem to be less concerned about their individual worth and do not retreat until a higher threshold of damage has occurred. So on the one hand i like seeing that the AI units aren't all kamikazes, but on the other hand clearly more tuning is needed, like maybe some sort of discipline-like meta-modifier where a value is placed on the importance of any given war (eg, war of expansion an importance of 4, a pillaging/nuisance war a 2, a war to keep another Civ from winning a 6, or whatever numbers work), and that value of 'fighting for the greater good' is then used to modify the self-preservation coding that seems to already be present. In the bigger picture though, i do think the AI can be praised, because it is moving in the right direction, which is to say it is moving away from stacks of doom. Clearly the AI is using 'thought' or else why would they retreat (again compare to barb attack behavior)?.
That's the thing, though. The AI doesn't "think". It has simply been coded to be more timid instead of suicidal. This is actually
worse for it, since it has been given even more significant military advantages in VI as compared to V, in the form of a combat bonus and more starting units. It
should be more aggressive, because it has a large advantage, therefore it can afford to lose some units, it doesn't need to retreat them every time you shoot them with your city. When people complained about the AI being suicidal in V, this wasn't what they wanted, they wanted more coordinated attacks (which have not emerged, as far as I can tell, the AI just wants to try and siege cities but then runs away when it takes damage--I'm sure you can see why this is not a good tactic). There's a
good theory about why the AI behaves the way it does. Far be it from me to critique AI programming, but that's not how I would have approached it. Whatever the reason, the AI has certainly changed a lot from V, for the worse.