28 comments

[ 3.3 ms ] story [ 57.0 ms ] thread
Just as with reporting on the J&J vaccine possible complications, I feel like reporting on self-driving car accidents deceptively obscures the reality that self-driving cars are already much safer than human-driven ones, on average. ("Yes, some people may have had blood clotting, but many more people have had much worse blood clotting from Covid.") For those who are unfamiliar with the cold, hard data on both subjects, such reporting paints almost exactly the wrong picture and should generally be treated as anecdata
>the reality that self-driving cars are already much safer than human-driven ones

[citation needed]

(comment deleted)
The data there does not prove it. If you think it does, quote the exact proof.
>In the 1st quarter, we registered one accident for every 4.19 million miles driven in which drivers had Autopilot engaged. For those driving without Autopilot but with our active safety features, we registered one accident for every 2.05 million miles driven. For those driving without Autopilot and without our active safety features, we registered one accident for every 978 thousand miles driven. By comparison, NHTSA’s most recent data shows that in the United States there is an automobile crash every 484,000 miles.

Alternatively, you could have just clicked through and read this yourself

He likely read that and it is not a proof that Autopilot is safer.

It would be if miles driven were selected for comparison to be of same type of driving. Right now you got Autopilot engaged miles that might be highway (easiest) mostly vs other types. We do not know.

That's not a relevant citation.
Obviously a biased source, but Tesla's Q1 safety report[1] (released yesterday) reports this:

>In the 1st quarter, we registered one accident for every 4.19 million miles driven in which drivers had Autopilot engaged. For those driving without Autopilot but with our active safety features, we registered one accident for every 2.05 million miles driven. For those driving without Autopilot and without our active safety features, we registered one accident for every 978 thousand miles driven. By comparison, NHTSA’s most recent data shows that in the United States there is an automobile crash every 484,000 miles.

[1] https://www.tesla.com/en_CA/VehicleSafetyReport

Ars addresses Tesla's misleading safety reporting here[1].

[1] https://arstechnica.com/cars/2018/05/sorry-elon-musk-theres-...

This article doesn't say anything substantive, just points out that Autopilot is technically a group of different safety features and says that Ars couldn't find any "industry experts" to opine that Tesla's findings are true. Very poor non-technical purely qualitative writing here
Then again, there are vaccines which don’t have this side effect, and are considerably more effective.
Doesn't matter. J&J is much safer than Covid. If, on average, people are healthier by administering the J&J vaccine than not, it's profoundly irresponsible to take it out of circulation
We have enough of the others, for now.
Well, if "we" is the United States, then the answer is yes. If "we" is the world, then the answer is hell no
The Tesla isn't a "self-driving car", autopilot is lane-keeping cruise control.
Sure, feel free to substitute whatever terminology you like into the above comment
Cruise control is not safer than a human driver, it's a convenience for highway driving, not a replacement for humans.
Are airbags and seatbelts also conveniences that do not make the driver safer?
While not as marketable, "enhanced cruise control" might be a better name to avoid these situations because people associate cruise control with still needing to steer and remain focused.
It’s more than that. It can stop and start at traffic lights, change lanes and maintain distance between you and the car in front of you.
> I feel like reporting on self-driving car accidents deceptively obscures the reality that self-driving cars are already much safer than human-driven ones, on average

Ars has a good article that tackles this false belief titled "Sorry Elon Musk, there's no clear evidence that Autopilot saves lives"[1].

[1] https://arstechnica.com/cars/2018/05/sorry-elon-musk-theres-...

This article doesn't say anything substantive, just points out that Autopilot is technically a group of different safety features and says that Ars couldn't find any "industry experts" to opine that Tesla's findings are true. Very poor non-technical purely qualitative writing here