Once built you can't move your capital buildings, so that is not a strategy option. The technology capital is the only one that gives a bonus that can't be recreated by another building, so don't be shy about building the other two capitals early.
Specialized worlds will rarley be operating at 100%, but neither will non-specialized worlds. Just for examples sake, say you set everything to 33%. If you mix labs and factories on a world, you still are only getting 33% of the value of each lab. If you make lab specialized worlds, it is the same as having many worlds with just a few labs. What sets specialized worlds apart are the bonuses. Only one world can get a +175% research bonus (tech capital, research coor, and omega research). While all of those labs are only operating at 33%, think about the number of tiles you are saving... or that those labs represent. Each lab built on a world with a huge research bonus is equal to about 2-3 labs built on other worlds.
The same with factories. True, if you put 100% into one of the sliders, then you are going to get worlds that are producing very little (never nothing, the basic colony gives some points of everything)... but that is a trade off you have made. Essentially, you are just taking average factories from several worlds and boosting them by concentrating them on a world with a large manufacturing bonus.
It is possible to play the game with 0% research and just get all of your tech through invasions. This is the only way I know of to have every building you build always operate at full capacity. This is a pretty restricting game style, however. As it stands, you have a basic choice to make about what percentages to put into each area.
In short, specialized worlds without large bonuses from somewhere are no better or worse than many worlds with no bonuses and an even mix of buildings. However, you will get more out of fewer tiles if you specialize the few worlds you find with large natural manufacturing or research bonuses.
Hope that helps.