I don’t get this. In the small farm town where my in-laws live, there is a population of say 200. And probably 400 cars. But very little visible sign of traffic — those 400 cars spend the majority of their time parked. They could easily be replaced by a fleet of 40–50 and save millions of dollars in just that small town.
What’s even more of a killer app is because this small town has effectively no services, there is a bank, a post office, and a small bar. The nearest grocery store is 10 miles away — a 20 mile round trip. With a self driving fleet, instead of making that trip, a great deal of trips can simply be eliminated by placing an order to the grocery store which is then loaded into a vehicle which brings you the order. And these sorts of orders can be easily batched, so that one 10 mile trip from the story replaces 5–6 20 mile round trips.
And a lot of people in those towns own big inefficient Ford F-150 trucks, that they take everywhere because it’s that family member’s only car. In the fleet model, you take a dinky pod car for $1 to the coffee shop, and summon a big truck for the Home Depot trip (which as I note, you might not have to make).
This gets into more bigger problems of course, when the grocery store now becomes an Amazon distribution center and has a big impact on the number of employees of those stores. In the end the biggest rebuttal to your statement may simply be that those towns you discuss will cease to exist. I won’t say that this is a good thing, but that might be the future.