I feel that Sulla's analysis of Civ 5 was spot on.
http://www.garath.net/Sullla/Civ5/whatwentwrong.html
In a nutshell:
1UPT Problem
1- 1UPT forced unit build times to be increased to avoid map gridlock
2- But you cannot make military units multiples of hammers more than buildings, so the solution was also to reduce production yields generally from tiles.
3- This had the effect of slowing down build times for buildings and units.
4- Plus, they increased science output relative to production (which has the effect of increasing cost of units relative to science)
All of that leads to a system where production is intentionally limited and science zips on through. The benefits of buildings large cities are marginal after cities reach a size of 6-10 or so. That, combined with the change to global happiness below, results marginal returns for building cities beyond a core size.
Global Happiness
1- They switched from controlling REX/ICS by having cities cost gold to having a global happiness.
2- In order to make this work, they made cities free but buildings cost gold. This means that building is now harmful to your city at times. Again, city building is demphasized.
3- But global happiness is controllable by policies, buildings and resources. Settle a city near a new resource and add a building and viola, no unhapiness from that city.
4- In addition, keeping cities small allows you to decrease the impact of global unhappiness. Again, demphasizing city building.
I also think that the switch from focusing on commerce to global happiness decreased the incentive to specialize or micromanage cities. No cultivating villages-towns, no building commerce or specialist cities. Production cities make some sense, but not really since you are really not churning out as many units in Civ 5. Every city formula is the same, regardless of your starting location. Sure, some variance around whether you build a lighthouse or a granary, or whether you have wine (and a monestary) or gold (and a mint). But in the end, cities are all cookie cutters. And the empire is a collection of all of the same cookies.
Just think about it this way. Say you build your production city on the periphery in Civ 4 and the AIs gang up and take that city. You now need to sue for peace to buy time to get a second production city running. In civ 5, all of your cities are the same so what does it matter if you lose one?
And that makes it easier to program an AI. The AI can be programmed to target and disrupt production and ruin your capacity to conduct a war. Not saying Civ 4 did that well, but they could have had that choice had the kept city specialization which would have decreased the AI problems with conducting tacital moves on 1UPT. Your empire in Civ 4 could be more vulnerable, whereas in Civ 5, its all scaleable. Destroy production, and you do not need a massive invasion. Or if you have one or two commerce cities paying for everything else, the AI could target and pillage those cities and all of a sudden, your empire crashes because it is out of cash. Again, make it possible for the AI to be successful.
The whole thing is just a series of choices the designers made that in the end were really bad choices in so far as they limited the dynamic content of the game and hamstrung the ability of the AI to be successful.