At first, sure. Eventually, though, self-driving cars really blur the line between public and private transportation. You could have a large network of self-d

At first, sure. Eventually, though, self-driving cars really blur the line between public and private transportation. You could have a large network of self-driving cars that you simply hail on your phone and quickly pick you up. It won't look like a 4-door car, but more likely a small bus like you find in many parts of the world. People going in similar directions travel together, only picking up and dropping off with minimal interruption. In lower periods of demand, these are stored in lots with no wasted space as only a LIFO 'stack' is needed. Furthermore, self-driving cars could be programmed to move in caravans with minimal following distance and aerodynamic efficiency to boot. Not terribly helpful in an urban situation (probably some minor traffic density increase), but wonderful for commutes which could dramatically decrease density. So, you'd not only increase occupancy (easily 8 people in a car-sized footprint, presuming electric cars), you could also increase traffic density. All told, an order of magnitude capacity increase seems within reason, and in a reasonable amount of time. That said, I think subways are still a good investment for the time being.