tech in the mid-to-late game

[Yosho]

If I can, I usually conquer and expand very quickly in the early game, grab a lot of land (and crash my economy in the process), spend the middle of the game developing my economy and all those cities in order to catch up and pass the AI's (quite possible with all that land), and then I'm in a good shape to either conquer (usually starting with the age of rifles/cavelry/grenaders), or go right for the space race or culture victory in the modern times.

That's not always what I do, and it's not always possible (on Pangea sometimes you seem to always be in at least one war or another the entire game, but can never focus on one enemy long enough to destroy them, so you never have a chance to take an economy-building "breather"), but when you can do it is seems to work well.

[uberfish]

8) Different AIs take different paths through the tech tree. Generally democracy will only be researched in a timely fashion by the AI if their favourite civic is Universal Suffrage.

[Thrak]

1) Acquiring some solution for happiness so your cities can grow bigger and more productive is a priority in the classical age. You will want at least one, maybe all of drama, monarchy or calendar.

See I nearly always grab Pyramids, so this doesn't apply (as much) and I flounder a little mid game.

Having said that I am polishing my early game again post chop-nerf.

[Son of David]

Don't let the AI build observatories before you do.

I used to make that mistake and wonder why AIs were steaming on through the industrial age while I was trying desperately to get out of Renaissance ....