Hit a squadron of school kids because you were reading a text? That’s on you. Drive under the influence of alcohol and plow into oncoming traffic? You’re going to jail (if you survive, of course).
Self Driving Cars Are Not “Five Years Away.”
John Battelle

This sets up a false equivalency, as those are exactly the types of “accidents” that an autonomous vehicle will always avoid. A self-driving vehicle will never be distracted by texts or intoxicated with alcohol. The issue is that those are the types of incidents that kill the vast majority of people. Estimates attribute human error to 90% of road accidents. If we have the technology to reduce traffic accidents by 90% or more, there’s going to be more than enough motivation to tackle the admittedly hard problems of figuring out what happens when the unavoidable accidents do occur, though they will become exceedingly rare as time goes on.

My current thoughts are that autonomy will be limited in speed in more densely populated areas, much like Google’s fleet that is currently operating in the city of Mountain View, CA. This limits the potential unavoidable accidents as a slow moving vehicle will have more than enough time to stop for your scenario of a group of kids playing Pokemon Go in the street. As the technology improves, I imagine there will be traffic corridors that will be closed down to human drivers as autonomous vehicles will be the only ones able to safely navigate them while maintaining the desired speeds. This would be especially true where there is a traffic crossing with no traffic light. The autonomous corridors would have the potential to move significantly faster than current speed limits, increasing traffic flow and generally decreasing congestion, even for the remaining human operated paths, much like carpool lanes.

While I agree with you that there are significant challenges (legal, ethical, insurance, etc), the ability to substantially reduce the 38,300 deaths and 4.4 million serious injuries that occur annually in the US (2015 numbers) is well worth the hard work of making it happen. I think we get to a tipping point of driverless cars taking over when:

  1. It’s safer than driving yourself by a significant factor (say 2x)
  2. It’s cheaper than driving yourself

While those 2 factors alone won’t cause everyone to abandon their existing vehicles, as they replace the vehicles they currently own (or lease), they likely won’t be replacing them with another vehicle that they have to operate manually.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.