Should-Read: Charles Stross (2010): Insufficient Data: “So. I ask: how many people does it take, as a minimum, to maintain our current level of technological civilization?…
…I’d put a lower bound of 100 million on the range…. The specialities required for a civil aviation sector alone may well run to half a million people; let’s not underestimate the needs of raw material extraction and processing (from crude oil to yttrium and lanthanum), of a higher education/research sector to keep training the people we need in order to replenish small pools of working expertise, and so on. Hypothetically, we may only need 500 people in one particular niche, but that means training 20 of them a year to keep the pool going, plus future trainers….
You can’t simplify a complex society that runs on just-in-time delivery and a host of specialities. You need a huge training back-end to provide for the thousands of skilled graduate-entry niche occupations. You need an efficient just-in-time delivery system to keep everyone supplied with food, water, power, shelter and whatever else they need — it’s that, or accept huge inefficiencies in your supply chain that wipe out the gains produced elsewhere…. Seemingly similar artefacts (cars, phones, airliners) have invisibly accreted complexity… [that] makes them better (safer, more economical, more luxurious)… but vastly more difficult to engineer; stuff that used to be fixable by shade-tree mechanics and jobbing electricians has receded over the horizon. Back in the early 19th century, the complement of a sailing ship could expect to maintain the ship in every significant way using tools and expertise that they could carry aboard the ship. Today in the early 21st century, that’s not an option with airliners or probably even automobiles….
Space colonization? Get back to me when you’ve tracked down how many people it takes to design and build a space suit. (The number is in the hundreds, if not the thousands.) More realistically, we won’t have autonomous off-world colonies unless and until they can cover all the numerous specialities of the complex civilization that spawned the non-autonomous, dependent-on-resupply space program. Or, to put it another way: colonizing Mars might well be practical, but only if we can start out by plonking a hundred million people down there.