*autonomous cars are just the test cases for the real goal here – autonomous trucks. The thinking is that if they can perfect the technology on more heavily congested and challenging roads and streets in urban areas, then adapting that tech to over-the-road trucks for long-haul will be a snap.
IMHO, I doubt autonomous cars will ever be widely used, if at all. But trucks on interstates, yeah, I can totally see that, and probably sooner than later.
Autonomous trucks are a simpler problem since they can be limited to highways outside of urban areas, where the problems get really hard. And there are simpler use cases like letting the driver sleep for a while on the long drives between dropoffs and pickups.
Autonomous cars are much harder, even beyond all the complicated cases that the article hints at. There’s a social element, and the problem that autonomous cars simply can’t move as quickly through an urban environment as a human driver who pushes the boundaries. As I understand it, car companies (other than Tesla) have given up on the idea of fully autonomous driving for now, and are looking at augmenting humans — alerting when the driver seems to be doing something dangerous, like drifting out of lane etc.