The argument is that self-driving cars would reuse existing roads, fuel, service, and manufacturing infrastructures that have been developed for existing cars. And, adoption can be incremental. Start with cars which drive themselves highway only, and require a capable driver to be at the wheel at all times. That may be soon. Gradually add more use cases at the speed of engineering, regulation, and social acceptance.

Some of the hard things to change in the world are hard because they require many people to change their behavior in coordination. Cars to self-driving cars doesn't. Email to Email with innovative clients doesn't. Introducing a new communication protocol does (though of course it's not impossible). Of course, this is only one aspect of a complex world.

Let's consider just one aspect of this that I see as a show-stopper out of the gate: networks.

Driverless cars, even even in a narrowly adopted context, like a single highway, will need to be networked. Not only will they need to be networked during the time that they are on that highway, for cars in close proximity, it needs to be a network with 99.999999999% uptime. At a minimum, the cars will need to have a transponder with nearly zero bandwidth (like planes); but more likely, this connection will need to send more data than just position. GPS isn't nearly accurate enough unless cars can be spaced 100s of feet apart.

Do you know of any networking technologies that you would trust with your life?

Compare driving on a highway to being in a plane. In a plane, there are usually minutes following a worse-case equipment failure or total loss of communication. In a car, there are seconds or less.

Nobody suggests driverless cars are working open-loop from GPS coordinates; no current driverless cars work that way. Its generally a combination of gps, local sensors including sonar, radar and cameras, and car-to-car networks. The UofIowa Driving Simulator has done studies of trains of cars linked by short-range networks, where they coordinate braking and acceleration to achieve inter-car distances of a few feet.

Sure, in principle; my point is that even the best networks fail way too much for this to be practical. In the context of this particular thread, I'm saying that implementing driverless cars even in a narrow context is a non-trivial infrastructure upgrade, whatever form it takes. These networks will need to have reliabilities and uptimes on par with medical or nuclear safety systems.

We don't need any networks. Self driving cars will always need to be able to share the road with normal cars. That mechanism will support the case of sharing the road with other self driving cars too. Networks can add fancy features, but they can happen on their own time, incrementally.