How about putting it this way, one person is killed in the US for every 100 million miles driven. Pretty amazing, now tell me how many miles were driven by autonomous cars on public streets before someone was killed? I'd bet it's a lot less than 100 million, so yeah it's pretty newsworthy.

I/A, I'm sure you think you've stumbled upon a profound idea. The humanist in you is like "what about everyone else dying, why is this news!" I assure you, you haven't stumbled upon a profound idea and your argument is flawed. First, it depends on the position that we don't make a big deal out of other accidents. Obviously not all accidents are covered. But lets experiment here. KATU: two top stories - Gresham teen injured on Division street and pedestrian killed on Sandy. KGW: one of the top stories - traffic cameras going into affect on Marine drive (to regulate speed and prevent accidents). KPTV: same two stories as KATU. KOIN: All the same stories. That's just this mornings sampling from our 4 outlets. And this is 7 days a week. Never mind the fact that the story you're talking about has already petered out from local news, and never mind that it's not even a yahoo news top story this morning, and never mind that out of the 20 slides of updated news stories on MSN it isn't represented at all.

Finally, you're missing the point on the importance. Autonomous cars are being posited as being safer because they remove the human element and (are supposed to) utilize advanced technology that anticipates and prevents accidents with a faster response time than people. This is a major benefit they are advertising in the autonomous market, and when the cars keep crashing, it brings doubt to the goal of this research.

Yes, Velma, but the difference is that sensors, algorithms, and everything else about self-driving cars will continue to get better and thus decrease the accident and fatality rate, while human driver performance has peaked and will likely stay the same or even get worse as distractions, phone addictions, etc., perpetuate.

That current human safety number is also only the current one - it is way better than it used to be before we mandated airbags, seatbelts, and before the development of braking assist, back-up cameras, and a whole host of other "electronic assist" safety features.

Eventually the safety numbers will cross over to where machine driving is demonstrably safer across the board statistically, and then only the rich will be able to buy the insurance necessary to drive their vehicles themselves.

So Flavio, how many people will die testing out autonomous vehicles? What is the acceptable number of deaths before they get it right? But it is still a newsworthy information.

42.