There is also the blessing/curse of automation. Any "menial" jobs would quickly be automated in an OP colony, resulting in the need for fewer workers.
I personally don't see any room to increase the efficiency of a building by adding more workers/scientists, because in an OP-style game, all structures would be optimized for efficiency beforehand.
For example: Agridomes and Residences both take only 1 person to operate. That's one person who can take care of 25-50 people or 50 units of food. I imagine that the jobs of those workers in fact consist mostly of watching displays and surveying the operation of the structure while robots do the actual work, maybe occasionally making a repair or stepping in when bots can't handle a situation. There would be no "floor mopping" or "grunt work" to speak of, merely supervision.
Because OP structures would be so optimized for efficiency, two supervisors could supervise twice as much, but could not really increase efficiency by supervising one thing twice as well. You would have to make a new building in order to overcome purely physical limits.
However:
I do see merit in the basic idea, not for getting more efficiency out of structures, but for emergencies when the population is low and there is not enough labor to go around.
What I am envisioning is a situation where in emergencies, one worker can serve multiple structures at reduced efficiency.
Lets say 1 worker can do 10 units of labor, you could divide that between 1 agridome and 1 residence, reducing the agridome to 50% capacity and the Residence to fewer people.
Lets also not forget that in OP3, structures may have other attributes that a decreased number of workers could affect, such as potential for accidents of breakdowns due to incomplete supervision. Maybe making your structure factory run at full capacity is worth the increased risk of a breakdown that could disable your smelter for a few marks, but might not if you are lucky.
Under normal circumstances this process would not be very useful, and values could be left at default. The usefulness of the process would instead come in those few desperate situations where you need every ounce of efficiency you can get, and losing 30% production in an agridome is worth it to get your vehicle factory running at 50% speed instead of 40% speed.
Now that I think about it, here is a related possibility off the top of my head:
Some people like to get as much efficiency out of their colonies as humanely possible, or play neck-and-neck games that are decided by mere seconds, while some people don't like that level of stress or micromanagement. If OP3 is as modular/adaptable/flexible as they say it will be, what if the code for this kind of worker-juggling is not hard-wired into the game, but rather an extension that the most obsessive of us can add to our OP3 copies while others ignore it? (It should go without saying that in multiplayer you could decide beforehand whether to let your opponent use it).
Come to think of it, the extensibility concept could apply to a lot of interface ideas around here...