NaGaHappen

This is a frustrating article for a lot of reasons, but the three that bug me the most are:

1. After outlining some of the reasons that self-driving cars is a very difficult problem, the article makes it seem like they'll still be here any day now.

2. The idea that self driving cars are going to save lives is almost tautological and so really quite stupid.

3. The insistence that this will be a huge financial boon for [Uber...among others] comes from...?

As for the first point, just no.  People who drive think driving is "easy" and so we should be able to have a computer do it even better, but driving isn't like doing algebra.  When we drive we are taking advantage of two things that humans do much, much better than computers: visual processing and pattern recognition (and extrapolation).  Computers have gotten much better at both of those things, but we do them so easily that we don't even recognize that we are doing them, and computers are still nowhere close to us.

The second point is one of those annoying points that isn't really a point.  Yes, if self driving works it will be safer as the definition of it working will be that it is safe (as computers don't drink, fall asleep...if it is safe it will be safer than human drivers) but this comes back to point 1: we aren't actually close.

The 3rd point is the underwear gnomes "profit" point.  It glosses over the middle point of how, exactly this will be profitable.  I'm sure it will, for companies that license the tech and maybe for auto manufactures who sell it, but for Uber?  Their current capital expenses are minimum, they're a middleman who skims off the top, and that is (potentially, generally) a very profitable position to be in.  If they plan on owning a fleet of cars, that they will have to maintain and replace, that changes the equation.  That will also mean a lot more [actual] employees to deal with these vehicles, plus probably land to park them and charge them up when not in use.  Now they do get to keep 100% of the fare in this case, but I don't think that equation will balance out well for them without magical hand-waving.

Behind all of this, and not actually stated in the article (other than a couple benefits) is a notion that self driving cars are actually going to be good for society, and I don't know how that is true either.  Yes, there are discreet cases where individuals would benefit from a self-driving vehicle in some way that they wouldn't from improved mass transit or existing taxi services, but for most people they'll be just another neat gadget...and in the case of broad implementation, they are likely to make traffic worse, not better.