But I believe you're being far too pessimistic as to organizing it. The "omnipotent creature" to organize it just... city government. Roads and utilities are already built/managed well enough. You're right it won't "self-organize" -- the city has to contract and build it. It's paid for, like other large projects are, out of bonds paid off over deacdes by taxes, and/or contracts to operate. Nobody has to "outlaw the old system".
You just build it gradually, and UPS/USPS/etc. will take advantage of it wherever it can (for all addresses supported) because it will be by definition cheaper.
Governments manage big infrastructure projects all the time. This is how major bridges and tunnels get built. All of your arguments apply to pretty much most big government projects.
> Nobody is signing up for those taxes so they can get their pizza or Amazon delivery via a tube instead.
My point is that yes they absolutely will if it's cheaper in the end. Making up numbers, if you spend $400/year on last-mile shipping and delivery, it's financed by bonds up-front, and the extra taxes are on average $200/year/person for the next 40 years (but with progressive taxation)... then of course people would be signing up for that. Vastly better delivery service for half the price? Why wouldn't they be?
Nobody's saying it's easy. But stuff gets built all the time. Saying it's too hard in principle and couldn't happen seems demonstrably... false.
If you're including everything into the factoring: never.
> For places like Manhattan, it's just hard for me to imagine that over a decade, it wouldn't be cheaper to build and use a system like that
Allocation calculations never actually function that way. Your premise can only exist in a vacuum where humans and governments and corporations don't exist, only robots. There's no omnipotent creature that can forcibly pull the unit capital away from the army of tens of thousands of deliveries and centralize it to the tube system instead. It won't self-organize because the cost calculations for traditional deliveries are individualized (I pay $5 for my pizza or sandwich delivery; not $50,000 in personal taxes for a hopeful tube system I'll never get my money back out of in my lifetime), whereas to pay for the super tube system you have to slam everyone (have to spread it out to everyone) with a massive on-going cost until it's finished (which none of them will want to pay), for services maybe they'll use in the future (it'll take multiple generations to justify the cost, which you won't be able to convince individual taxpayers to fund when there is an existing functional system that is 85% good enough).
The cost requires an extreme centralized outlay of capital, which nobody can raise given the other factors; nobody can get the building done, and nobody can get through the zoning and regulations, and nobody can spare the 10x time requirement (10x worse than it should be, as all construction is these days) to get through all of it.
It requires a nationalized step-in by some large government authority. They have to simultaneously outlaw the old system (or hyper tax it to death, same thing in effect), while drilling people on taxes to supposedly build the new system (which will take forever, go drastically over-budget, and underwhelm for the value vs cost in the end).
1.6 million people live in Manhattan. Do the math for just 500,000 * $50,000 = $25 billion. That $25b figure is likely on the low side by anywhere from 2x to 4x minimum, if you attempted to tear up all of Manhattan to do it. The cost would be staggering, and then it would take decades to reach completion.
Nobody is signing up for those taxes so they can get their pizza or Amazon delivery via a tube instead.