This metric says nothing about self driving capabilities. In fact, I'd argue that FSD supervising the driver (and doing things like limiting speed before corners) would make their cars safer.
To me this metric shows that their cars are very high performing, and for most drivers they're probably the fastest accelerating cars they've ever driven. Tesla should probably default them to 'chill' mode and provide a warning about how fast the car is when you switch out of that mode.
It is fantastically optimistic to attribute Teslas horrific stats to the cars being speedy
For instance the model y had a fatality rate of 10.6 per billion vehicle miles 4x the average.
Its also seems unreasonable to suppose that they are poorly suited to survive a crash as this doesn't seem to be indicated.
A more logical conclusion is that a box with a giant flashing distracting tablet in the center which lies and says it can drive itself gets crashed more because people are functionally incapable of going from passenger to driver at random intervals with no notice.
Teslas also tend to attract people who hate driving and are bad at it. Yes, this is anecdotal, but - several friends and acquaintances said something along the lines of "my Tesla is the best, self-driving helps so much, I hate driving and I can't wait for them to fully automate it".
I wonder if segregating bad drivers into a separate population affects those fatality statistics.
To me this metric shows that their cars are very high performing, and for most drivers they're probably the fastest accelerating cars they've ever driven. Tesla should probably default them to 'chill' mode and provide a warning about how fast the car is when you switch out of that mode.