Mr. Roberts is president of Connect Ithaca, a volunteer group that works on community issues like sprawl, and the heavy traffic is one reason he was hooked when he heard about the next generation of automotive technology — personal rapid transit, more commonly known as pod cars.

I'm not sure how I missed this article a few months ago. I was really enamored with the idea of PRT for awhile but I now think that as it's typically formulated, it's a non-starter:

"Stations" are an Achilles' heal. They add a lot of cost esp. when you want to make them ADA compliant, etc. Since they're on rails one way or another, the blind octogenarian with a heart condition who gets her umbrella stuck in the door bogs everything down. If you put all stations on a siding, they cost that much more and take up that much more space. If you don't, then the problem with one car stops the whole line.
You can't possibly build enough stations to avoid a last-mile problem.
The guideways are an eyesore and people don't want them around.

I now think robotic taxis are the answer for urban areas. If we can figure out a way to power them in-situ, from overhead lines, say, like the buses in SF, all the better. Maybe just inductive charging when you're stopped at an intersection. You pull out your smartphone and push a button and the system dispatches one to your location. I don't know the breakdown but I would have to guess the vast majority of the cost of a cab fare is the driver and the overhead of the cab company, not gas or the car. And maybe these don't even need to go more than 30mph, can be a "skateboard" car with in-hub motors, very few moving parts, low-maintainence, low-power, etc.

Unfortunately, noone seems to be taking a pragmatic approach to autonomous driving. The academics, as usual, are chasing the "unobtainium" solution. The Grand Challenge attracts a lot more attention than something as simple and inexpensive as burying magnetic reference markers in the street. There actually is some work going on at Berkeley (PATH) and UC Davis under the heading of "intelligent roadway" but it doesn't seem to be nearly as flashy as Grand Challenge. And it's a bad chicken+egg problem: automakers aren't interested unless the infrastructure exists and road agencies can't afford to blow money on pie-in-the-sky visions like this much less even afford to maintain the roads and bridges they have.