I have contemplated having varying tiers of specialist, but I don't think I could get the AI to handle it in the current system. I will probably make specialists more rare anyway, saving their powers for industrial and beyond.
This sounds really good. It will definitely keep ai competitive for a longer period. I would be careful about how many slots are introduced at once, or else the effect can backfire. I think I a slow progression of 1 slot in industrial & modern and two slots for post-modern eras will keep the ai growth rate fairly stable.
I am considering removing the % yields from governments, increasing their improvement bonuses, and then adding specialist yield bonuses on top of the improvement bonuses.
But then won't that clash with the limiting of the specialist. Unless you mean for governments for the Industrial Era and later.
I am going to be working on changing some of the building/government yields around over the next few days, so this will likely change. I wanted to put more of the % changes on the government, but I don't feel like that has worked well enough. And per pop yields on buildings generally doesn't feel like an improvement, even when it is. That is my feel on it, not sure how you guys like it.
For buildings that modify production, I would definitely agree with going with percent improvements. But some things, like science buildings actually do work better on a per pop system. If you really want to get technical, per pop is the
same as percentage based; after all, what is a library except a 50% increase in science? It's just written as x yield per pop for later buildings like universities that modify
base science, not the overall total.
With that in mind, I think governments work best as percentage modifiers, but they should focus on the
total yields, and then modify those. For example, consider a size 24 city with library, abbey, and university. The science output is 168 (24+12+12+120). If I were in a government that increases science by 30%, it would (should) go to 218.4. Something like this is easy to see in a percentage type explaination. If you make the bonuses dependent on improvements, who knows how that will affect overall productivity?
BTW, I just PM'd you, so please check it when you can

.