While I'm not going to say the AI doesn't need work, again, I think you're vastly simplifying the AI routines required for unit preservation (and stockpiling).
For example, City-States are excellent (perhaps too keen) at maintaining giant armies. But they only ever have a single City to focus on. Barbarians are decent enough too (if reluctant to siege Cities) but have
no City to care about.
Enforcing a minimum army size would impact AI build orders, economic priorities and their ability to gun for both Districts and Wonders. Like a player, they're meant to be able to adapt, and not be forced to some baseline of military spending.
You also don't know the existing routines because your multi-aggressor logic could already be implemented. After all, AIs can declare war at the same time, and can declare war independently of any player agency or action. However, it could be being superceded by Agenda weighting which leads to heavy-handed denouncements leading straight into war, and so on.
You don't consider any of the related AI routines when suggesting the logic that you say isn't complicated. Well, if you exclude them all, then of course it won't be.
While my response was more of ironic amusement concerning Civ V's AI, I will endeavor to share my opinion towards your comment, Gorbles.
Your quite right in that VP will not work for Civ 6, and that it was a bad suggestion ( however ironically amusing that may have been). The main reason is that VP will not work for 6, as it was specifically created for Civ V AI. My best guess is that the talented modders of VP may be working on a similar mod for 6, providing they get all the tools necessary for them to create the mod.
However, with that said, I am curious as to your rather single minded viewpoint concerning Civ 6 ( and I mean no disrespect here), especially considering there have been well articulated counter-arguments concerning this topic by some of the vets on these boards. You will occasionally agree with some points made, but overall, you seem unwilling to entertain the idea that other posters conjectures may have validity concerning the OP's opening thesis. Why?
The AI is trash, despite the occasional posts stating that the AI managed to surprise a player ( on Diety diff, IIRC).
The UI is horrible, even after the patch.
TT's, even on a high end rig, are ridiculously long.
Units are still OP'd, and UP'd depending on type.
Information that a player needs is still hidden, incomplete, or takes forever to view when hovering.
This is just a few issues from a rather lengthy list, that should have been found before launch from a triple AAA company.
Look, I do my best to try to understand the viewpoints of posters on these boards. But your rather narrow focus ( again, no offense intended) on this particular topic is puzzling.
Because people keep talking about it? Generally how threads go, really.
If I don't argue with a point it's because either I have no information on it, or I don't disagree with it. Again, generally how all posters work. You'll notice a weight towards criticism over praise, and that just isn't because of the state of the product; it's how all discourse generally evolves over time (especially on a particular topic). People don't argue over things they agree on, by definition. People argue over things they disagree on, which includes the validity of the
other person's disagreement.
I've responded to plenty of other well-articulated posts, several in this thread alone. That said, the status of someone as a "veteran" means very little here, as even peoples' pedigrees as software developers are seen as "making an argument from authority" (which is a misapplication of the fallacy because it only applies in a discussion with equal grounding).
To your list specifically, in order:
1. The AI is arguable. It suffers against higher-skilled players more than it does against less-skilled players, and CFC has a specific weighting towards a higher-skilled playerbase (on average) compared to the average Steam player. Further to this, simply saying "the AI is bad" doesn't help anyone or anything. Which is why I've been trying to point out what is constructive r.e. AI improvements and what isn't ("lol CiV AI works better").
2. The UI definitely needs a lot of work, even from the perspective of giving each area equal polish. Some widgets are noticeably more placeholder and out-of-theme from the Age of Exploration than others.
3. Performance is always arguable. I see comparable times to Beyond Earth (though I haven't exactly labbed it), which vastly improved on CiV (and said improvements were backported at least in part to CiV I think, resulting in improved turn performance there, later in BNW's lifecycle).
4. Unit balance is a delicate thing and I think we need more playtime before going "stuff is OP". I think unit / upgrade pacing needs work before we can say ranged is better than melee, or anti-cavalry is better than dedicated melee, etc. I've seen complaints that ranged is too useful counterargued by pointing out their frailty in melee combat. Mounted melee units definitely have an advantage, and I honestly think people just need to get used to the new movement rules. You can call them worse, but the sooner people accept they're here to stay, the sooner their own playstyles will evolve to suit.
5. UI and UX. The information is there, just buried. See point 2.