30,000+ deaths annually in the U.S. alone is an awfully high price. Even marginal safety improvements would save many, many people. And that's not even counting the many more people who are mamed and left debilitated.

Then, there's the economics of car ownership. Cars are the second most expensive capital asset most people have after their house. They sit idle, depreciating, taking up valuable land 95% of the time on average. A self-driving car is very complementary to services like Uber. If self-driving cars could be shared, increasing vehicle utilization past 5%, then we could repurpose vast swaths of unproductive urban land.

Once fewer people are owning cars, and instead paying a la carte, it also restructures the choice architecture of driving. Without a default option sitting in the driveway, I suspect that people may opt to walk, bike, or take transit with more regularity as well.

As a planner, I'm slightly more bullish on the potential of self-driving cars to remake transportation than my peers are. It was only a couple of years ago that I attended a demand management conference keynote in which the well-known (in planning circles) speaker was cracking jokes about how self-driving cars wouldn't change anything. I disagree--I just hope the future gets here soon.