[agent008t]

Note that all of your examples had working prototypes; it just took 20 years to make them affordable/scale them (along with some incremental improvements).

With self-driving cars we have no such prototype and I would not expect one in the next 20 years. Unless we are talking about driving on specially equipped roads/lanes or in some special situations, where we might see them much sooner.

[lrem]

The DARPA challenge mentioned was for providing a fully working prototype, all these years ago. And not a highway, but through an unmapped obstacle course.

[agent008t]

It was a very different problem from driving on public roads.