I personally much prefer the previous incarnation of the planetary projects. In that form, projects were generally useful and gave me a reason to build mixed planets. Under the current system, there's more or less no incentive to build a mixed planet; the maximum output mulitplier a fully-improved world can have is roughly equal to the sum of the output multipliers it does have, whereas its actual overall output multiplier is less than or equal to the maximum of the output multipliers it currently has. A fully specialized world producing only the type of output it was built for is therefore going to produce roughly three times the total output of a world with even output multipliers. Three specialized worlds can therefore give you about as much output as nine generalized worlds, and the only failing specialized worlds have relative to their generalized counterparts is in build up time.

The old system had its issues, but at least it allowed generalized worlds to be competitive with and, under the right conditions, superior to highly specialized worlds. It also kept its optimization problem in what you build and had trivially easy optimal slider settings (100% to desired output type if manufacturing bonus was less than some value, 100% to manufacturing otherwise), whereas the optimization problem under the new system is entirely in the slider setting and has a race condition - the more production you sacrifice to manufacturing to increase the project bonus, the more you penalize the output multiplier you're trying to improve with the project.

For the most part, I agree. However, the way the projects currently work, the project will grant at least 10% so long as your manufacturing output is positive (in the strict sense, not the non-negative sense). This means that for any world with a bonus to the desired output type of less than +890%, you will get marginally more output by setting 1% manufacturing, 99% desired output type and using a project than you will by setting 100% desired output type.