I was stuck in traffic the other day (a rarity in Alabama) and got to thinking about the prospect of automatic cars. It’s always struck me that if we could just do something in a car besides drive it wouldn’t be important to get to our destination fast, which would mean bad traffic wouldn’t be so bad, and cars could operate at slower speeds and greater safety and efficiency. Even if this weren’t a legislative nightmare (consider the fact that Segway failed in large part because of sidewalk right-of-way issues in its largest potential markets) it turns out that automatically controlling cars is a Difficult Problem.
In his novel Snow Crash, Neal Stephenson (rather insightfully, I think) suggests that synthetic speech will prove intractably difficult even when otherwise very capable VR technology is commonplace. (We still can’t do a great job of synthesizing monotone speech, forget singing and acting.) The solution in the book is a worldwide system of real-time arbitrage (very similar to the system by which ads are placed on web pages, for example) which selects voice actors for parts based on fee charged, past performance, customer satisfaction, and so on.
Imagine if cars were remote-piloted and drivers were allocated to vehicles in much the same manner. Now combine that idea with the zip car.