Personally I play on Huge maps, although my computer gets so slow on them by the late game that it becomes a pain and I stop playing by about 1800. And that's on Vanilla. My only experience with the really late game is when I mistakenly make Modern the start era
. With regard to the GP's, there's no real reason to forbid super specialists unless they'd be extremely hard to code - I just settle my GPs when there is no wonder I need to pop.
Diplo should be changed so that influence is based on average influence throughout the game, times the percentage of turns in which the player and CS were allies or friends, in order to reward long-term relationships versus spamming the 100000-gold treasury one turn before the UN is complete. Liberation should remain the same, as it is a difficult task which involves the taking on of great empires militarily.
Space should be done in stages as mentioned earlier, and could perhaps be accelerated by extraterrestrial contact (although this would make the two victory types so intertwined that the strategy for seeking either would be the same, making Contact eventually worthless).
A possible workaround for this could be a chance that the spaceship fails to reach its destination, due to various factors such as electronic malfunction, unhealthy spaceship politics, impoundment of the craft by extraterrestrial beings, or plain ol' asteroid impacts. This chance of failure, though high at first, could be decreased through alien technologies (in the case of system malfunction) or social policies (in the case of spacecraft politics ruining the mission). The impoundment risk will remain until contact and good relations have been established.
The alien tech and social policy trees should start out each as one core but develop two separate branches, one focused on the Contact Victory and one focused on the Space Victory. It should be impossible for any player who desires to win to pursue both branches at once. The Spaceship branch of the tech and social policy trees should be cheaper than the Contact Victory, but carry a risk of failure, necessitating the need for spare Spaceship Parts which are only affordable to more powerful civilizations. The Contact Victory pathway, however, would cost slightly more but victory would be certain once the project was completed, making it viable for desperate civilizations who will spare no expense in earning victory. In the unlikely event that there is a hyperpower on the planet,
the contact project and spaceship parts should not be simultaneously buildable.
After an improvement to Victory Conditions is a need to reduce the attractiveness of the ICS strategy. Will follow up on that later.