I would question the underlying assumption of your math. The number of fatal crashes should always compare the same environment and conditions. For example, if you compare with the number of accidents of a Waymo in SFO downtown, one would need the number of crashes in SFO downtown caused by human drivers as well. The scenario when crashes contribute to the statistics should be taken into account.
That said, I guess it would take Waymo even more than 20Years to proof safety.
Henri -- You are right that it will take even more miles to do an apples to apples comparison. But then you also go down the road of looking at the detailed contributions with dramatic variations based on zip code, road type, time of day, etc. It's so complicated I wrote a whole book on that topic (https://safeautonomy.blogspot.com/2022/09/book-how-safe-is-safe-enough-measuring.html)
The way to interpret the simplistic calculation here is that it will be a long LONG time before we know how safety at the fatality rate will turn out, and any statements by the AV industry about already saving lives are hype.
I would question the underlying assumption of your math. The number of fatal crashes should always compare the same environment and conditions. For example, if you compare with the number of accidents of a Waymo in SFO downtown, one would need the number of crashes in SFO downtown caused by human drivers as well. The scenario when crashes contribute to the statistics should be taken into account.
That said, I guess it would take Waymo even more than 20Years to proof safety.
Henri -- You are right that it will take even more miles to do an apples to apples comparison. But then you also go down the road of looking at the detailed contributions with dramatic variations based on zip code, road type, time of day, etc. It's so complicated I wrote a whole book on that topic (https://safeautonomy.blogspot.com/2022/09/book-how-safe-is-safe-enough-measuring.html)
The way to interpret the simplistic calculation here is that it will be a long LONG time before we know how safety at the fatality rate will turn out, and any statements by the AV industry about already saving lives are hype.