Helmuth
Member
I agree!
It's scary to see where NOT to use the autopilot!
It's like a don't-do-this video...
I never had a problem using AP2.0 while using on streets where it should be used though!
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
FIRMWARE UPDATE! AP2 Local road driving...and holy crap
- Thread starter Skalinator
- Start date
It's scary to see where NOT to use the autopilot!
It's like a don't-do-this video...
I never had a problem using AP2.0 while using on streets where it should be used though!
You are, as a driver, still required to think and make the right decisions such us when to use AP2.
We have seen worse AP1 videos and I am certain people can make scary videos while using AP1 still.
Unfortunately I no longer have my AP 1.0 car. I sold it because my 100D with AP 2.0 should be here soon. However when David received his AP 2.0 car I did do a side-by-side on the same road (I was following him in my AP 1.0 car). However that was before anything was released for AP.
I do know a lot of other Tesla owners in Colorado with AP 1.0. I will see what I can do getting a current side-by-side video done.
Here is a link to that video:
Thanks Erik
I'm baffled Tesla chose to enable local road usage. 1) everyone is waiting on highway usage at adult speeds. And 2) it doesn't f'n work.
I could maybe understand if they had it working well and this could placate the owners waiting, but this seems like it's just going to further frustrate owners while showing that Tesla ISN'T actually being ultra cautious to release only safe functionality. (Which was previously their best excuse)
As usual, total head scratcher...
I could maybe understand if they had it working well and this could placate the owners waiting, but this seems like it's just going to further frustrate owners while showing that Tesla ISN'T actually being ultra cautious to release only safe functionality. (Which was previously their best excuse)
As usual, total head scratcher...
After AP1 came out and continuing to a lesser extent through today, there have been numerous TMC posts with videos or reports describing what sounded like alarming and "unsafe" behavior. This includes AP1 allegedly driving cars into ditches, crossing double yellow lines, crossing lane lines, abruptly stopping next to or not recognizing parked cars, attempting to drive into trucks, not recognizing cars stopped in the road in front of them, causing drivers to become distracted and not pay attention to the road, and on and on and on. Given the constant clamor about how dangerous AP1 supposedly was, I think many people were very surprised to learn that when the data was in, the NHTSA found that activating AP1 actually reduced serious accidents by 40%.
The NHTSA report suggests that posts on TMC and other social media (including video in many cases) showing or describing apparently "unsafe" behavior were not only wrong, but got it completely backward. Cars with AP1 -- warts and all -- appear to be far safer than those without. And if the data holds up, AP1-enabled cars are not only safer, but with a massive improvement that is potentially unmatched in the history of automotive safety, except possibly by the seat belt. And AP1 continues to learn and improve and will undoubtedly only get safer.
So before jumping to the conclusion that AP2 is unsafe based on similar anecdotal reports/videos, I think it is worth taking a step back to consider why the anecdotal reports on how dangerous AP1 supposedly was were so misleading. The answer is that anecdotal reports are not data, and they are subject to all sorts of biases that skew the data, especially toward dramatic events that make a compelling story. There are numerous other kinds of biases at play as well.
The video you posted is a good example. Correct me if I am wrong, but this is not your video -- it is something you found on the internet. I also saw you post it on reddit, and perhaps you also posted it on other social media. Those posts will tend to heighten awareness of this example where AP2 did not perform well, and will tend to skew perceptions about AP2.
You chose not to post video or reports of AP2 doing well, of which there are many on TMC and elsewhere. I am not criticizing you for posting it, but just pointing out that this is an example of what happened with AP1, and explains why so many people who relied on anecdotal evidence incorrectly assessed AP1 safety. From this video and others I have seen, AP2 (like AP1) appears to do better on well marked roads that do not have sharp curves. Like AP1, it has learning to do with intersections because there are no lane lines. But that does not mean it is unsafe -- just as AP1's limitations do not make it unsafe.
Tesla has just recently released AP2 autosteer for use on surface roads limited to a relatively low speed (35 mph) with the driver instructed to keep hands on wheel and maintain responsibility for safe operation of the vehicle. There have been many reports of AP2 working well on well-marked and on some poorly marked surface streets. There have been some examples (like the one you posted), where this is clearly room for improvement, but the driver remains in control of the vehicle and can take over when necessary.
Bottom line: I don't think you have any basis for claiming AP2 is "unsafe" based on the video and anecdotal data. If history is any guide, Tesla's own assessment of the data is a more reliable guide than anecdotal reports/videos like the one you posted.
Edit: what @Vitold said.
Last edited:
I only watched the first 5 minutes and I can say with certainty that AP1 will not work on that road with those center islands. Not the use case you want to test out now.
If AP2 does well on the highways, that is a big plus.
The point is that AP1 works flawlessly in those exact road types and roads.
and another...
Here we go again
@EinSV - I don't really care about numbers(I already proved that the results in the report are expected no matter what the car is if they have a decent collision avoidance system), I care about my personal life and not getting into an accident.
I don't consider a system safe when at some overpasses or some other metal structures that "bend" over the highway, I have to pay attention if someone is behind me.
Just the other day I enabled TACC (just for fun, and because the highway was almost empty) and just before a traffic camera/radar or something, the car decelerated from 75mph to 45mph and back to 75mph. I was kind of expecting it but it happened so fast that I did not have time to react, by the time I wanted to do something the car was accelerating again.
I'm writing this because I want new owners to be aware and consider the system unsafe, for now. You can always control where the car goes but you cannot always control when the car stops.
@EinSV - I don't really care about numbers(I already proved that the results in the report are expected no matter what the car is if they have a decent collision avoidance system), I care about my personal life and not getting into an accident.
I don't consider a system safe when at some overpasses or some other metal structures that "bend" over the highway, I have to pay attention if someone is behind me.
Just the other day I enabled TACC (just for fun, and because the highway was almost empty) and just before a traffic camera/radar or something, the car decelerated from 75mph to 45mph and back to 75mph. I was kind of expecting it but it happened so fast that I did not have time to react, by the time I wanted to do something the car was accelerating again.
I'm writing this because I want new owners to be aware and consider the system unsafe, for now. You can always control where the car goes but you cannot always control when the car stops.
Last edited:
Precisely.
The data that you posted simply tells us about the effectiveness of forward collision avoidance which every manufacturer has and is required by law.
Doesn't say anything about AP1 driving behavior or lack of thereof. The things people pointed out are valid. They are even listed in tesla warnings, things like the car may not stop for stopped cars.
There have been dozens of accident reported where the car doesn't brake for the car in front of it.
These things aren't anecdotal. They are facts. The stats you posted have nothing to do with it.
It has nothing to do with the fact that autopilot 1 can drive you flawlessly on the freeway for 100 miles straight and afterwards has a hiccup and drive off and hit a guard rail.
These are facts and these are facts that tesla acknowledge and these are facts that would have prevented that autopilot tragedy.
The bike lane had nothing to do with it.
How is it that you do not understand? When AP1 first arrived (7.0) these situations were a piece of cake. Surface road was easy pickings for AP1. Now you need to take over every 10 seconds or you're dead. The point isn't that ap2 hasn't reached parity with ap1. Its that 17.7.2 for surface roads is dangerous. that's literally from every 17.7.2 video I have seen.
Again Ap2 is good on highway/freeway. No one is denying that. but its beyond dangerous on most surface roads.
Similar threads
