Discuss: Teslas Using Autopilot Are 10x Safer According to Latest Vehicle Safety Report

[Not a Tesla App]

Tesla is the safest vehicle in the world and it's even safer when Autopilot has control.… Read More

Teslas Using Autopilot Are 10x Safer According to Latest Vehicle Safety Report

Tesla is the safest vehicle in the world and it's even safer when Autopilot has control.

www.notateslaapp.com

 

[PrescottAZRichard]

That's great news, and I hope my insurance premiums are adjusted appropriately based on this info.

Sad thing is - one incident of phantom braking 'causing' 8 cars to wreck rules the headlines instead. That's unfair for several reasons including leaving a safe distance to react to cars stopping in front of you. It's also annoying because it (first and foremost) absolutely shouldn't happen.

I'm not a fan of derailing threads immediately with negative posts but this discussion was bound to expand to include this other story.

 

[waaahsabi]

I've owned a Tesla for about 4 weeks now. Within these 4 weeks i had 2 phantom braking incidents. One happened with active autopilot, the other with active cruisecontrol only. In both cases it was an object to the front-left far ahead (incident 1 a left-turner that wasn't in my lane but in front of me, incident 2 was oncoming traffic in a right-bend that was situated to the front-left of the Tesla). The 8-vehicle tunnel crash looks quite like that too. So this is a real issue, i daresay reproducible, and not limited to FSD.

So yes, this bug should absolutely get fixed with highest priority, because nothing keeps people from using cruisecontrol / autopilot and running into the exact same situations frequently. I tell my wife (main driver of the car) to NOT use even cruise control outside of freeways, because of that.

 

[SteveWin1]

That's great news, and I hope my insurance premiums are adjusted appropriately based on this info.

Sad thing is - one incident of phantom braking 'causing' 8 cars to wreck rules the headlines instead. That's unfair for several reasons including leaving a safe distance to react to cars stopping in front of you. It's also annoying because it (first and foremost) absolutely shouldn't happen.

I'm not a fan of derailing threads immediately with negative posts but this discussion was bound to expand to include this other story.

I think, in the other story, its not even confirmed that FSD/AP was activated. The driver blamed it, of course. I've done this too when pulled over for speeding. "No officer, I wasn't driving, the car was and I guess it didn't see the change in speed limit." Got off without even an official warning -- cop just thought my car was cool. Regardless, all systems (including human brains) make mistakes. The fact that FSD/AP makes mistakes isn't important or interesting at all. All that matters is whether accidents and fatalities increase or decrease when these systems are used. Tesla's data specifically mentions autopilot, but not FSD. I wonder if this means that FSD doesn't have very impressive numbers?

 

[RNHurt]

This report is good and I think I can agree with the results. I think most people forget how terrible human drivers are, yes even you sometimes. We all get tired, frustrated, bored, and distracted sometimes and we end up making mistakes. No, FSD/AP isn't perfect and yes it makes mistakes, but (according to this data) those mistakes are fewer that when we are in control.

It seem like everyone in all of these "auto-drive" conversations expects computers to be 100% perfect all the time and never, ever make any mistakes at all. That's just not reasonable. It might, MIGHT, be possible if/when all the cars on the road are computer controlled. But put a human in there and all bets are off.

As for the recent "pileup", yes it's terrible that people got hurt and it should never happen. However, how many of these same type of accidents have been caused by humans over the years? Has a human driver never stopped in the middle of the road? Has a car never broken down and not been able to get clear? No, those things happen every day. And we don't see them on the news precisely because they happen all the time; only new and unusual events get news coverage.

BTW: I had a tractor trailer stop in front of me on the expressway once because they had missed their exit and was trying to get over. Being completely stopped while cars are approaching you and whizzing by at 75MPH is terrifying. Humans are bad drivers. Full stop.

 

[SteveWin1]

You know what they don't mention in that article... If all the cars behind that Tesla were driving on autopilot, they would have been following at an appropriate distance and they would have actually braked in time. There would have been no pileup at all. But, because some people are willing to risk their lives and the lives of those around them by NOT using autopilot, this pileup happened. Obviously we need immediate legislation banning driving without autopilot. But seriously this is a more valid argument than the inverse.

 

[juren`]

Won't make any difference to the current NHTSA manager who will NEVER recognize Tesla for being the safest and best-engineered vehicles on the road

 

[Geekraver]

Autopilot is mostly going to be engaged on highways. Most accidents occur in parking lots, rural intersections, etc, when you won’t be using autopilot. That’s a much more plausible explanation than claiming it’s 10x safer to be driving with autopilot. Correlation is not causation.

 

[SteveWin1]

Autopilot is mostly going to be engaged on highways. Most accidents occur in parking lots, rural intersections, etc, when you won’t be using autopilot. That’s a much more plausible explanation than claiming it’s 10x safer to be driving with autopilot. Correlation is not causation.

Yeah, but there are only about twice as many accidents on local roads per mile driven vs highways. So, even if ALL AP miles were highway and ALL non-AP miles were non-highway (an extremely unrealistic assumption that would help your theory for explaining the difference), this wouldn't come close to explaining a 10x difference, or even the 3.5x difference between Teslas using vs not using autopilot. The more AP miles you think are non-highway (my entire commute is local roads and I'm on AP for 90% of that, every day) and the more non-AP miles you think are highway miles (most cars on the highway are not on AP), your point gets even weaker. It also wouldn't explain why the gap keeps widening as Tesla keeps rolling out updates.

 

[Coulomb]

This is good news but I wish they'd do the careful research to make apples to apples comparisons.

Comparing Autopilot on vs. off: Do drivers go manual in trickier conditions? Does it still count as Autopilot miles if the driver starts steering to try to avoid a crash?

Comparing Tesla vs. other cars: Need to compare approximately similar drivers; cars of similar age, safety features, & condition; similar road types (highway vs. local) & road conditions; and similar areas. The average car is much older, lacking recent safety features, and in worse shape.

 

[tomw8073]

I've owned a Tesla for about 4 weeks now. Within these 4 weeks i had 2 phantom braking incidents. One happened with active autopilot, the other with active cruisecontrol only. In both cases it was an object to the front-left far ahead (incident 1 a left-turner that wasn't in my lane but in front of me, incident 2 was oncoming traffic in a right-bend that was situated to the front-left of the Tesla). The 8-vehicle tunnel crash looks quite like that too. So this is a real issue, i daresay reproducible, and not limited to FSD.

So yes, this bug should absolutely get fixed with highest priority, because nothing keeps people from using cruisecontrol / autopilot and running into the exact same situations frequently. I tell my wife (main driver of the car) to NOT use even cruise control outside of freeways, because of that.

I personally find it less stressful too just to drive myself in traffic than be on edge waiting for a phantom braking issue to appear.
It has braked 5 or 6 times in the last 3mo.

 

[tunkoj]

I have driven both model 3 and now model S in total over 50 000 miles, mostly in FSD highway and local streets and found it incredibly safe. Only problem is that it will not exit freeways. I have to take over. See prior post.

 

[tunkoj]

Thanks, I hope you all read his post. I said to my wife the same awhile ago. Idiot drivers are everywhere. Thanks to Tesla, great safety records is great!