Comment by rich_sasha
9 days ago
As much as I'd love to pile in on Tesla, it's unclear to me the severity of the incidents (I know they are listed) and if human drivers would report such things.
"Rear collision while backing" could mean they tapped a bollard. Doesn't sound like a crash. A human driver might never even report this. What does "Incident at 18 mph" even mean?
By my own subjective count, only three descriptions sound unambiguously bad, and only one mentions a "minor injury".
I'm not saying it's great, and I can imagine Tesla being selective in publishing, but based on this I wouldn't say it seems dire.
For example, roundabouts in cities (in Europe anyway) tend to increase the number of crashes, but they are overall of lower severity, leading to an overall improvement of safety. Judging by TFA alone I can't tell this isn't the case here. I can imagine a robotaxi having a different distribution of frequency and severity of accidents than a human driver.
He compared to the estimated statistics for non-reported accident (typically your example, that involve only one vehicle and only result in scratched paint) to estimate the 3x. Else the title would have been 9x (which is in line with 10x a data analyst blogger wrote ~ 3month ago).
> roundabouts in cities (in Europe anyway) tend to increase the number of crashes
Not in France, according to data. It depends on the speed limit, but they decrease accident by 34% overall, and almost 20% when the speed limit is 30 or 50 km/h.
They reduce accidents in general, but bring us some “entertaining” new ones where a (usually) drunk driver crashes into the statue/fountain/whatever in the middle or uses the little “hill” in the middle as a jump ramp…
As opposed to drunk drivers who treat a non-roundabout intersection as a game of Russian roulette?
>they tapped a bollard
If a human had eyes on every angle of their car and they still did that it would represent a lapse in focus or control -- humans don't have the same advantages here.
With that said : i would be more concerned about what it represents when my sensor covered auto-car makes an error like that, it would make me presume there was an error in detection -- a big problem.
Lapse in focus is such a great point. If we looked at the number of accidents caused by very human errors such as "lapse in focus" and "sudden medical events" etc. which we would 100% expect to go away when offloading tasks to any computer, the statistics of accidents remaining becomes the bare minimum for what computer-based automated driving must achieve.
This is compounded by the system mistakes likely being hard-errors. A computer hard-error vs. human lapse of judgement is potentially the difference between the vehicle slowly crushing a small child as they scream and beg for help vs. a human stopping as soon as they felt/heard something. Context matters.
The compared error-rates must consider if it could have been avoided or mitigated, the near misses, the human vs. computer type of error, and how hard-errors may lead to horrifying scenarios.
I wonder if slow speeds affect the detection?
A bollard at three feet might look like a grain silo at 400 yards. I could see angles getting to where the camera sees "beige rectangle (wall), red cylinder (bollard)" and it's basically an abstract modern art piece.
I see things on security cameras a lot that in low resolution are nearly impossible for me to decipher.