Changed planetary projects to be % bonuses instead of flats. For example the project that produced credits used to give a credit per 4 production, now it gives a percentage boost to how many credits that planet produces (so you are better off producing research on research focused planets, etc)
I would like to know the formula for this. The current model may be flawed, but it at least gives me a reason to build relatively balanced worlds as under the current model there are conditions in which balanced builds maximize a planet's output, and the current model's optimal slider settings are easy for single-output type worlds. The proposed model seems likely to remove any reason to build anything other than a fully-specialized world, and it seems likely to make the slider setting optimization problem much more difficult. I view this as a negative; playing with a slider to optimize output is not particularly interesting, choosing a build for the planet which optimizes its output can be.
Currently, if B is the bonus to the desired output type from structures and b is the bonus to desired output type from other sources that you can achieve on a fully specialized world, and m*B and x*B are the bonuses to manufacturing and to the desired output type from structures while bm and bx are the bonuses you can achieve to manufacturing and the desired output type on a mixed world, then a fully-specialized world will offer greater output than a mixed world if
B <= (4 - x*(1 + bm) - m*(1 + br) + sqrt((x*(1 + bm) + m*(1 + bx) - 4)^2 - 4*m*x*((1 + bm)*(1 + bx) - 4 - 4*b ))/(2*m*x)
Under the assumption that m = x = 0.5 (i.e. the total bonus you can get to all output types is constant) and bm = br = b = 0, B ~= 13 (i.e. a 1300% bonus). B = 0 if bm >= 3 (well, technically B is imaginary if bm > 3), but under the assumption that m = x = 0.5, even with bm = 2.9 and b = bx = 1, you still need to be able to achieve B = 438% or better for the generalized world to be superior to the specialized world.
If the problem which is supposed to be solved by this change is that generalized worlds are too often superior to fully-specialized worlds, it would appear to me that the appropriate fix is to adjust just how much of a bonus you can get from non-structure sources, not to prevent generalized worlds from ever being optimal outside of the "I want to have a bit of a manufacturing bonus on my research/purse worlds just in case I need it" situation.
Moreover, the proposed change shifts the problem from "what do I build to optimize output" to "where do I set the sliders to optimize output." Maximizing research output by building a pure-lab planet and then fiddling with the slider until whatever manufacturing bonus I have on the planet adds to the research multiplier in such a way that the research multiplier is maximized isn't an interesting problem.