Here's a thought: combine driverless cars and driverless buses for a commute. One of the chief problems with buses is the sometimes long waits when one needs to change buses, but if a company could assure that one never needed to wait more than two minutes to change from car to bus or bus to car, the advantages in reducing traffic might be well worth it.
Consider, a small queue of buses waits at an interstate entrance ramp - only two or three buses, not enough to waste much time but enough to be sure no one will need to wait long for a delayed replacement vehicle - and most of the cars that would have gone onto the interstate stop and their passengers get onto the bus. At two minute intervals, the bus hits the road - taking probably 30 and perhaps (if a double-decker) over a hundred cars off the road. If this is happening at rush hour and at every major intersection on main routes going into a center city, ten thousand cars could be replaced on the roads by one or two hundred buses. Aside from reduced parking, consider the reduction in traffic in city centers. Add in traffic lights (or other controls) coordinated on the fly with buses, and riders could be assured of a smooth commute into town almost every time.
Of course, not everyone would be going to exactly the same place, but walking two or three blocks is healthy anyway, and not much further from a destination than most parking lots - or for more spread out city centers, more cars could be waiting at the exit ramp - with less than 30 seconds to transfer. With reduced traffic and higher safe speeds, commute times could actually be reduced, and of course commuters could spend their time more productively than driving. Driverless services could include options for breakfast or a snack on the buses, or even bunks to take a nap on long commutes - and of course wi-fi and the like.