[peatmoss]

We still have geometric problems with everyone driving. Namely, cars take up space in two or three useful dimensions on roadways that are fixed in length, and have limited capacity to grow wider or higher. In short, self-driving cars won't handle peak demand and the accompanying congestion.

Where I am more excited about self driving cars is in their potential to change vehicle ownership economics in slightly denser areas. A car2go that delivers itself makes it pretty hard to justify buying an expensive capital asset that sits idle an average of 95% of the time. In other words, a rental service that is able to utilize its vehicle fleet more fully, should be able to make a handy profit while reducing travel costs for many people. The opportunity frontier is enormous here.

Once more people switch to user-fee based travel as opposed to all-you-can-eat-buffet travel, people all of a sudden start choosing how to travel rather than automatically hitting the default option. That opens the door to better public transit, which can handle peak demand, as well as more walking and bicycling.

And, we should have room for more and better bicycle/pedestrian facilities, since we won't need so much land to store all our mostly-unused boxes of steel and plastic.

I think self driving cars will be transformative for cities, but it won't be some Jetsons vision of the future where we simply upgrade suburbia with tech. That will probably happen too, but those changes will be rather less profound.