Seems this thread's expanded a little to cover colonization rates in general. In a way, I feel a little guilty seeing the thread name still up here since the creator(s) worked hard and made a nice product. Unfortunately, major changes are still needed to make the game AI competitive along the lines initially outlined and expanded upon by others here. Played another game with fewer, larger AI empires. All of the original problems listed are still problems in 1.049. Targetted colony expansion into enemy held systems in order to sell colonizes nets ~170,000 - 250,000 per colony in trade value immediately following colonization, and turn arounds can be fast enough to avoid relationship penalties. As others have stated, eliminating tech trading by any of a number of methods is still a huge issue for difficulty (as I and I'd imagine others have no real desire to show restraint ingame). I prefer enemy AIs trading with each other only because I've seen it work profoundly well before (though for the life of me I can't recall the game it happened in - AIs traded tech with each other almost as fast as I could, and I had to jealously guard new tech / trade it to multiple AIs immediately to get a return, and couldn't trade exchanged techs as a result as they'd already been shared long since with harder AI settings).
On the issue of colonization, I'd say the best serviceable example would be Stars! (circa 1995). Homeworld populations were minimal (say 1/10th of maximum planetary population starting out, with variance based on growth rate), and while some races could expand like a plague immediately, others had to choose their colonies carefully and move limited populations from the homeworld to colonies, which each often required several years (turns - maybe 7-10) of expansion before they were viable enough to send even a single colony ship of their own off, which would in turn require spreading a limited population. Starting population for a new colony in Distant Worlds is what, 10 million? Make each colonizer require 10 million people from planetary population then (you don't need to bother with the story fluff justification until the rule set's in place). Expansion from Homeworld can be addressed seperately if the rate's still too great (lower homeworld population caps while preserving income?), but this will at least slow down the hub and spoke effect to something the current AI can compete against, and shut down at least part of the indy. race colonization problem noted by others in this thread and elsewhere.
For making the AI truly aggressive, I'd expound first by guessing at limitations. If ships are treated individually, or only reluctantly grouped into fleets, I can see where true AI aggressiveness would be a system hog. Have the AI form fleets at the design stage instead with homogenous speeds and fuel constraints, with specific instructions for Troop Transport behavior so that they'll always be present and go for the throat the moment the coast is clear. For competing against human ship designs, there are only two real rules. If the AI has the tech, make AIs design ships to fight players that have greater range (assuming any kind of damage potential at range) and greater speed. If the AI doesn't have the tech, load down AIs with a greater ratio of short-range firepower + shields and greater speed than human ships, however many engines it takes. If necessary, balance changes can be made to let laser ships go faster than torp ships (less mass carried or somesuch fluff). Whatever the case, AI ship design should be based off of human ship designs at least with respect to firepower, durability, and speed, perhaps with a cost function for efficiency. Generics simply do not cut the mustard against human opponents unless the generics are very, very mean and efficient. In the event that an AI is hopelessly outclassed by human, or possibly even other AI opponents, consider giving them an option to surrender, or ally themselves to greater AIs in a vassal manner. If England invades, ally yourself with France.