Here's an economics analysis: self-driving cars reduce the effective cost of a commute, by allowing the passenger to focus on other things than driving during the commute (reading, watching TV, playing games, etc). Since a significant limit to sprawl is the cost, broadly construed, of long commutes, self-driving cars will increase sprawl.
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.
Duration set to six days to encourage Monday as first day.