Although San Diego and other cities seem to ignore the idea, the self-driving car is likely to be here soon. Google introduced a fully functioning prototype of its self-driving car (meant to be built with no steering wheel or other human controls) and has suggested they will be for sale by 2020. Other carmakers are generally predicting the same timeline. There are a lot of legal hurdles, as well as the obvious technological hurdles. Maybe it will take a long time for these cars to become ubiquitous, but let's just imagine the impact of mass adoption of these cars. It seems very reasonable to conclude that this would entirely disrupt the design of cities as we know it and no one seems to be planning for this.
First, parking would become nonexistent. Today, on average, cars are parked 95% of the day. This is hugely inefficient. If cars were autonomous, you could easily share a fleet of cars: you call a car when you need it, pay for the time spent driving, and then you get out and the car is off to pick someone else up (you also wouldn't have to chauffeur your children or elderly parents around, they could "drive" themselves to soccer practice or bridge club). This would be far cheaper than actual car ownership. So, if we no longer own cars and these autonomous cars are basically always moving, there would be little to no parking. No parking lots, no street parking, no garages, no driveways. Rather than paying for expensive parking in the middle of the city, it seems logical that the car companies would park their cars during non-peak hours in vast lots in the middle of nowhere.
Second, roads could be redesigned. Generally, most people expect that self-driving cars will talk to each other and could drive much more efficiently, closer together, without worrying about accidents. Some even think we could do away with signs and traffic lights. Google's prototype is a tiny car, like a smart car, perhaps only 5 feet wide (sure, they are tiny cars, but if ownership becomes nearly nonexistent, these two-passenger cars would compromise the vast majority on the road. If a user occasionally needs a larger car, you just rent one for that special occasion). Once all cars are self-driving and networked, these cars would drive within inches of each other. The result: our current roads would have way too much capacity and be way too wide. Even allowing for right-lane bus lanes up to 10 feet, other lanes could max out at 6 feet rather than the current 10-13 feet common in many cities (although it's easy to imagine that buses would be wiped out by self-driving cars, on both the efficiency and cost fronts).
So, now imagine a city with no need for parking and lanes that could be shrunk in half. Keeping our current roads, we could narrow most lanes and eliminate on-street parking to add wider sidewalks and protected bike lanes on every street. Without parking lots, development could become much denser (and there would also likely be a drop in property values once all the parking lots come on the market for development). We could eliminate driveways and their curb cut outs. Garages would no longer house cars (ok, many don't already). In a residential area, streets could easily be 10 feet wide, total (imagine a current single lane), with some road set aside for bike lanes and the rest deeded to the adjacent homeowners, perhaps to build accessory units, further increasing density.
it's also easy to imagine that autonomous cars could result in the end of pedestrian and cyclist fatalities. For this reason alone, self-driving cars should be supported by anyone that rides a bike or walks in a city.
There are a lot of reasons to be skeptical: once traffic is gone and the cost of driving drops, car use might jump. But the alternative is just as likely, that if cities move to convert excess road capacity to more density and walking/biking, many areas will develop the level of density that will largely erase the need for cars. Once this capacity is gone in the cities, it will be difficult to take it back to build roads.
Obviously, cities like San Diego can't start basing all of their planning decisions on the assumption that self-driving cars will take over in the next 10-20 years, but shouldn't someone be thinking about this? If done right, a proper focus could result in a vastly improved city premised on self-driving cars. If these cars are incentivized (i.e., only allow self-driving cars in the urban core, or dedicated lanes on the freeway), a city could lead the way for this upcoming revolution.