Can you elaborate? I don't think I got your point.

I'm intrigued by the difference between "understanding" a population, in aggregate, and having a fine-grained view of each individual while also knowing how each individual relates to the whole.

So, what might we learn about the galaxy if we knew what / where / when every single bright spot of star-ish size is, that we would not learn from an aggregated simulation? Rhetorical question, obviously.

I don't even know how many people live within 5km of me. How many are human-trafficked, held against their will? etc. We can make approximations, but there's a lot of interesting information at the margins.

Well, most of what you get is verification. When the problem is entirely local, you can verify separately calculated measurements with other measurements. This is why I think we'll be able to move much faster to work on problems of similar magnitudes locally than on relativistic scales.

While you don't know how many humans are within 5km of you, that's certainly easier to verify with confidence N than it is to verify how many humans within 5 light years.