> Now they use the money to get Starship flying, and in 1-3 years they'll likely get it to fly well enough. And they'll have the question what to do with that huge capacity, so they'll have to invent something which needs a lot of super-heavy launcher flights. Could it be a significant Moon base, for tourism, science, manufacturing, resource utilization?
While I agree that there will be payloads built specifically for Starship's massive capacity, most aren't going to need that much space.
Historically, an entity (company, academia, government) that wants to launch a satellite has had to not only buy/build it, but also buy the rocket, then wait months/years for launch. Some small satellites have been able to hitch a ride on launches of larger payloads, but even there the timing is out of the smallsat's owner's control.
With Starship, an entity only has to build a satellite. As it nears completion, the entity will go to spacex.com, find a day with a Starship launch with sufficient open capacity, and reserve a slot. If the satellite is delayed, the reservation can be moved to another day (perhaps with some percentage of the deposit withheld if too close to launch time). In other words, Expedia for space.
There will be as many Starship launches as needed. Maybe the combined Starlink+outside customer demand will at first result in one launch a week (less frequent than Falcon 9's launch cadence, because of Starship's greater capacity). Maybe over time it will rise to two, three, five, seven, ten, 20 weekly. The more frequent and more consistent the launch cadence, the more confident customers will be that a launch slot will be available when they want one.