The first real large scale (and possibly positive ROI, not too sure yet) use case for autonomy will be in freight, not passenger cars. And that still has so many challenges beyond the tech (laws, liability, network inefficiencies, single digit margin business already) that it’s, in the practical sense, still at least a decade away.
Now, autonomy to make trucks safer: sure, keep it coming. The challenge will be making it cost effective enough.<p>I am willing to argue with myself here though: I do think there are very practical use cases near term for closed circuit autonomy. Things like senior living communities, or the Las Vegas strip (I told my wife a few weeks ago when there that I bet in ten years humans won’t be allowed to drive on the strip) are perfect for just closing down all human traffic and effectively allowing a giant trolley/shuttle service to run safely in the street.<p>Could have interesting implications for city/community planning over time. You could potentially narrow all the streets and make everything one way since the cars can just flow together better.