This really belongs in Strategy & Tips as it's not an article.
-- optimal number and density of cities?
Fewer numbers, as cities cost but improvements don't. I'm finding 8-10 to be about right fairly early in the game on a standar map. As opposed to 15-20 and always wanting more from C3 (any version). But as you improve your cities, that number grows. So, over time, adding cities is worthwhile -- almost any city will EVENTUALLY be valuable...but the payback window may be long.
-- how does shorter game effect play?
Shorter game? What do you mean? Lower max turns? I never reach 2050 in either game. Total non-issue. Nothing to unlearn/relearn here. Maybe if you're a score-monger, but I don't know that game at all.
-- best initial tech paths?
MUCH more variable. MANY good options. Researching worker techs based on what's around. Big question of going for a religion or not. You HAVE to research tech, as you can't trade techs nearly as easily and buying them is essentially a joke. The lack of need to spam cities means Pottery is much less important. This is still an issue of much debate, too. At this point, I don't see a clear winner.
-- difficulty levels similar to C3 or different scale?
Hard to say. Not as many people are playing the really high difficulty levels, but strategies are all in their infancy. Give it time. I think the scale is different, but the game is very different. There are more of them, that's for sure.
Unknown. And dangerous in what way? Heck, I'm not sure how I'd answer that for C3 even -- depends on the goal. That "depends on the goal" fact is probably even more true in Civ4, but it's still not clear what traits work best for what strategy.
-- other changes we need to consider?
Who's "we"? Combat is much slower. In general, it appears all wins come a fair bit later. The SOD of artillery is significantly curtailed. The days of 50:1 kill ratios appear gone for good (thank goodness). Combat, on the whole, is quite a bit different. And I think it's for the better, but it takes some getting used to.
Wonders build MUCH faster now. I built the GL from scratch in a SG, so in less than 10 turns. And, no, it wasn't obsolete. They are fairly different in scope and effect, too. I'm still getting a grasp on them. They're quite different.
The whole specialist/great person thing. A whole different type of (micro)management. Very different stuff. New dimension of gameplay. Not something to unlearn, per se, as much as something to learn. And I'm still learning.
Religion is like a mini-game (to steal from the wise Sullla). You can play or not. If you play, you can get significant benefits. If you don't play, you can do other things differently/better. Nothing to unlearn so much as learn.
Culture flips are quite different. They still happen, but you can see them coming. And they don't kill the army stationed in the city. Raze'n'replace still is viable, but it's more the exception than the rule, in my experience. Capturing an isolated (or soon to be isolated) city is probably better than razing.
And I'm sure I'm missing stuff.
(Hmm...maybe this is an article -- nah, dh_epic covers a lot of this more eloquently than I just did).
Arathorn