Civ V has one undeniable and crushing flaw: You do nothing while playing it.
For over two hours per game. Probably much more than that.
so true!
it's one of the reasons i do not enjoy civ4 as much as i could have. during long turn times, my mind starts to wander off and i lose focus. when i lose focus i make stupid mistakes and get all pissed about 'em.
as i have heard, this issue is even more serious in civ5. therefore buying/playing civ5 is out of the question. as one gets older, he/she tends to value time more
There's no longer the RNG element that was used to stand in for real diplomacy in previous Civs. Most Civs will start out neutral and then turn to friendly. It will be things you do in the game that determines the course of their actions.
The most common reason AI goes to war in the early game is over land.
The most common reason AI goes to war in the late game is over threats.
no RNG? seriously?

as to reasons to DoW, what about "you [player] are pursing the same victory condition as i [AI] am" (yep, on turn 14)?
The AI doesn't really usually "need" what it asks for, it doesn't use it well once it receives the favor, and it's not coded to generally ask for things that support its overall strategy. In other words, it's not a human. It never was, it never will be. It's an AI code.
won't be surprised if AI's "needs" are guided by the wise RNG God
don't agree about the last statement though. allow me to introduce
Artificial Intelligence: A Modern Approach (3rd Edition)
incorporating at least 10% of knowledge of this book in civ AI will insanely increase it's "intelligence". trust me
I'd argue that it is. You'd be hard pushed to find a game in the 90s that has Civ 5s capabilities, unpolished as they are.
try Call to Power (release March 1999)
However, as a professional I know from experience that a comment such as "chunky code" gets the point across without alienating anyone.
To detail however, each turn requires the AI to make building decisions, city management choices, unit movements, winning scenario decisions, exploration, worker maintainance, diplomatic relations with other civs, assesments of your relative positon and perceived strategy, defensive and offensive military build\action choices, graphics, scoring, barbarian movements....the list can go on and on and on and is done for each civilisation in turn, so on a standard map it's running that 7 times....all of that = some fairly chunky code to run inside a 30 second window.
u kidding bro?
some suggestions right off the top of my head: take the AI out of server space into a separate process(e.g. a separate executable). this way each AI will(can) have a separate world state it can play with. now the AI can "think" on other players' turns
some division of decision logic is needed as well. i won't go into this because no one cares anyway, but i will provide an example:
AI goal? offensive war on a neighbor -> set all "high" production cities to produce offensive units -> set
Rome to produce warriors/swordsmen/whatever
the trick is that until the
goal changes,
Rome should produce warriors!
no need to check each city, no need to pointlessly move units around. all should be goal-centric. a SoD entered our territory! is it a threat? would i [AI] alter my goal(s) to deal with it(a new threat)?
The entire problem here is that it's subjective time. You're losing it, others are not. I'm not losing any time playing, I'm making good use of my time and am still struggling to understand why you're holding onto those 30 second periods as if you were going to invent free limitless energy with the time.
losing time waiting for your turn is objectively time lost. it's time spend not playing. you can choose to ignore the issue and/or spend the waiting time doing something else, but still... anyway a lame analogy is playing a shooter with 20 fps. is it playable? sure, but not so fun
That doesn't make sense. I wouldn't hire a programming team to develop an AI engine and then set them to basically setting up new civs. You can get extremely junior programmers to incorporate new civs, the fundamental AI engine underlying the game shouldn't care what Civ it's working with. So new additions are blinking coloured lights, they're not the tree.
providing a mediocre AI
and a crippled multiplayer is the decision that should not make sense.
next up is the boy genius saying somewhere a long time ago that firaxis values universal programmers. yep, this means that firaxis prefers to simply shuffle people around instead of specializing

last, but not least is say firaxis did have a decent AI engine. juniors would only cut it, if the newly added civs have generic traits (aka civ3/civ4 style). where juniors will horribly fail is adding civs with a unique terrain improvement(s), totally unique traits, traits that grant some units and/or unit types unique promotions, etc.