finished my game as arabia last night, some thoughts (big and small):
1) i achieved one-city cultural victory at turn 286 (1868 give or take) in mecca, earning a whopping 85 culture points per turn at the end. considering the AI reached 10k in mecca in another game, i think raising the limit is necessary. i could have reached 18k without any further cultural developments (ie wonders), so even 20k is probably not out of reach. that said, it might be too high for the AI, and probably has a lot to do with religious wonders (see below). assuming we modify these, i'd say 15k one-city is a stretch, but do-able.
2) the 3 islam wonders give a real cultural advantage if you can get all 3 in one city. not only are they worth a bucketload (10 culture all up - 3, 3, 4 - if i remember correctly), but they will easily hit the thousand year mark and double later. perhaps we should go through all the religious wonders and total up culture to be the same for each branch. for example, if the pyramids (the only ancient egyptian religious wonder) are worth 4, then perhaps the 3 islam wonders should be worth 2, 2, and 3 respectively - slightly more than the single wonder to account for the greater build necessary, but not so much more that it's unbalanced.
3) i wasnt pushing total civ culture, but i was pop-rushing monuments in every city as i founded/took it. so although my total civ culture was only around 23k, it could have been a lot higher if i had been going for it. that said, i've never seen an AI come close. i'd say 35k for total civ culture would be good, but that's just a guess/feel.
4) i'd agree with the other comments about reducing the domination victory conditions. in my game with germany, i had everything from london to lisbon, trondheim to constantinople, madrid to moscow and still wasnt even close. the problem isnt population, it's area. i'd lower world area to 20%, leave population higher though (40%?), as the high settler costs mean the AI is always short population (for some reason, they seem to keep around tons of extra settlers "just in case!")
5) i disagree that the primary consideration for victory conditions should be adjusted to make the AI more competitive. the sad fact is that the AI is completely non-strategic - it absolutely does not move towards any victory condition with the single-mindedness of a human. as such, we'll never be able to set victory conditions in a way that help the AI remain competitive without making it too easy for the human.
on the other hand, i have a personal dislike for the way c3c difficulty levels increase. above regent level the AI simultaneously gets bonuses in both static areas (eg corruption, production) and dynamic start (eg extra units). i'd much prefer the start bonuses to be smaller in favour of better static benefits for the AI. that way, you're not so much chasing at the higher difficulties, but rather planning and strategising to remain at parity.
6) on that note, i feel that the palace guard gives plenty of advantage to the AI to prevent early warrior rushes, and that the added penalty of no worker for the human is overkill. perhaps this should be linked to difficulty level: at regent and below, the human gets the starting worker.
7) i should also note that i absolutely love the naval movement change. i wound up in a colonial war with japan, where my singapore and australian colonies were under fierce attack, and having galleasses (galleai?) with sufficient movement to not only get there, but to search and engage in a single turn was just brilliant. it also gave japan an edge in starting the war because out of nowhere 3 ships dumped troops into australia and declared war before i could intercept their ships!
8) i also love the various unit upgrade paths. this has led to some real strategic decision-making and greater realism.
9) i couldnt hear the sound on the cavalry attack animation. i'm not sure if it was very soft relative to the defenders' sounds, or just not present. is there anything we can do to adjust this?
sorry for the longish post, hope this sparks some ideas.
EW