The primary problem with the AI is that it isn't finished. There are numerous "placeholder" functions that are waiting to be filled in with actual instructions, plus other functions that were hashed in quickly. The difference in AI performance between SE:IV and SE:V has little to do with the change to scripting from hard coding, but just the degree of completion. Aaron is returning to AI work, particularly with the last two beta versions, so improvements are pending.
The actual level of control is superior to SE:IV for modders, but the learning curve is a little steeper. Of the testers I know that would-be AI modders, the most common response has been the fun factor for them with the game - which has hurt their desire to work on AI. The current condition of the stock game is discouraging those that are modding AIs from doing it with stock.
You would have been hard pressed to put together a mod list greater than 10 for SE:IV at this same point in the game's development - so I wouldn't worry about that.
In terms of performance the items hurting it most is the drawing of 2D screens and the pathfinding routines. Yes, multiple chase combats can extend turns, but that is much easier to workaround.
And lastly as Tim as nicely put and I said earlier, it's really too much for one person to achieve in the timeframe we demand.
