Ask HN: Is anybody tracking Teslas self-driving progress over time?
I have been googling a bit, but not found anything:
Is somebody out there driving the same route again and again , tracking the number of disengagements over time?
Is somebody out there driving the same route again and again , tracking the number of disengagements over time?
1) I don't know if it's the general social malaise and depression around the whole Covid stuff the last couple of years or the possible impact due to other social factors like illegal immigrants having far easier access to driver's licenses, but it feels like people have become significantly worse drivers on average in the US in the last couple of years. I can't quantify this objectively, but it feels like more drivers are impatient, taking foolish risks, caring even less about signaling turns, etc. Stats that compare the last couple of years with a few years ago might not be an exact apples-to-apples comparison of driving difficulty.
2) At one point, the self-driving feature might have been an extreme novelty that few people really trusted. They might have only used it very cautiously on safer straight stretches of road while monitoring the road themselves with their hands right over the wheel. As they grew more comfortable with it working well, they might have relaxed their attention, possibly to the point where they're not actively watching the road, or even nodding off or playing with their phone and not actively watching the road, and using it in far more dangerous roads. The system might be working better, but with much worse human supervision, it might paradoxically lead to more accidents.
On the surface this investigation appears to be focussed on whether the design of "Autopilot" encourages drivers to do stupid things but it does also touch on the fact that some crashes occur when the the car is not under the control of "Autopilot" but in some cases "Autopilot" is seen to be forcing control back to the driver a second before impact so more "Autopilot" then driver.