I think autonomous vehicles will help here eventually too. A lot of road maintenance is due to large semi-trucks. If those vehicles are operated autonomously, they can be programmed to drive more efficiently in regards to wear on the roads. Slower braking, slower speeds (since you don’t need stops for drivers to sleep to reach your destination in the same amount of time).

We’d also see a closure of many roadside stops since there are no longer people around to use those facilities.

Road wear is caused by weight. Driving and braking slower won't make a significant difference.

Road damage is said to be proportional to the fourth power of axle weight [0]. The link to reference 14 on the article is dead but the internet archive has a copy of it at [1].

[0] https://en.wikipedia.org/wiki/Vehicle_weight#Importance

[1] https://web.archive.org/web/20150709151027/http://www.nvfnor...

An autonomous vehicle can keep the tires to the same track within 3cm. So there is potential to build pavement just where wheels go and corners.

I don't believe it, but it is possible.

Might also enable more frequent but smaller trucks, which also would enable more point to point movement of goods and less of a hub and spoke model.

The interstates tend to go to the airports though.

Interstates go a lots of places, airports is just one. they also cross Nebraska for example.

I would need more information to flesh out how you think my default assumptions (that indirect costs are approximately proportional to direct costs) should be revised.

Facts and logic are optional, but at least a quantitative assertion that's falsifiable would be a start.