Tesla autopilot fail
- Thread starter Great White Drake
- Start date
Agreed. Until the government (NHTSA, DOT, etc.) create standards designed with self-driven vehicles in mind, the automakers are "playing with fire". There are just too many unique road conditions to cope with.
This is an Autopilot 1 car, like mine. It's cruise control with lane keeping. People like to assign it more responsibility than it deserves so that when it fails in a circumstance it was never intended to be used in they can pull out their old man card and say how much technology can't be trusted.
That setup uses a front facing radar, 12 ultrasonic sensors and a single camera to do everything. As far as identifying lanes, all of that falls on the camera and it would have kept the car in what it perceived to be the lane. So it's actually working exactly as designed. The driver just didn't see the wall because he wasn't paying attention.
I feel like people should be forced to have hands on time with the system before they criticize it. I've counted a couple times that I would have at the very least had an emergency stop, where the car identified a car 2 cars in front of me slamming on it's brakes and in response slowed the car down before the car in front of me ever saw the car in front of them braking. It is so amazingly good at what it does that it's hard to put into words. Now if we're talking about an Autopilot 2 car after they've gone fully autonomous then I'll step right up and criticize.
The other thing people need to realize is Tesla stories = page views = revenue. It's the Apple mentality. They want so desperately to make a story out of anything. Antenna-gate wasn't really a thing. I mean sure you could grip your phone in such a way to make the signal disappear but in practice nobody actually held their phones that way. But that doesn't matter. As long as you can make the connection that something can happen even if it very likely won't then you can create enough of a commotion to make a story.
Antenna-gate didn't die down because Apple gave free rubber bumpers to everyone; it died down because using the rubber bumpers prevented the problem. Nobody actually used the rubber bumpers. Same with the purple flare in the photos of the iPhone 5 or 6. I don't remember which. Same with fires in some of the early Model S's which I believe happened once or twice but were huge stories. So huge in fact that Elon came out and compared the handful of Model S's that had experienced a fire under any circumstance with the hundreds of internal combustion engine cars that went up in flames every day for a litany of reasons and received no press time. At the end of the day, a fancy new electric car possibly being a ticking time bomb sells more ad space than another Grand Prix going up in flames because of leaking valve cover gaskets.
And the same will apply to the next generation of Autopilot as it's rolled out in the next couple of years. They will relentlessly look for ways to make something out of nothing. Personally, I take comfort in the fact that as hard as they've tried over the last year, they still have nothing. There hasn't been a single substantiated example (that I can recall) of Autopilot truly failing. Even the dude that got his head taken off. The NHTSA opened an investigation into that accident and found that Autopilot was not at fault. The guy had 7 full seconds to see the big rig turn in front of him and I don't believe he ever even attempted to stop. He just wasn't looking.
Other than that I have no opinion.
That setup uses a front facing radar, 12 ultrasonic sensors and a single camera to do everything. As far as identifying lanes, all of that falls on the camera and it would have kept the car in what it perceived to be the lane. So it's actually working exactly as designed. The driver just didn't see the wall because he wasn't paying attention.
I feel like people should be forced to have hands on time with the system before they criticize it. I've counted a couple times that I would have at the very least had an emergency stop, where the car identified a car 2 cars in front of me slamming on it's brakes and in response slowed the car down before the car in front of me ever saw the car in front of them braking. It is so amazingly good at what it does that it's hard to put into words. Now if we're talking about an Autopilot 2 car after they've gone fully autonomous then I'll step right up and criticize.
The other thing people need to realize is Tesla stories = page views = revenue. It's the Apple mentality. They want so desperately to make a story out of anything. Antenna-gate wasn't really a thing. I mean sure you could grip your phone in such a way to make the signal disappear but in practice nobody actually held their phones that way. But that doesn't matter. As long as you can make the connection that something can happen even if it very likely won't then you can create enough of a commotion to make a story.
Antenna-gate didn't die down because Apple gave free rubber bumpers to everyone; it died down because using the rubber bumpers prevented the problem. Nobody actually used the rubber bumpers. Same with the purple flare in the photos of the iPhone 5 or 6. I don't remember which. Same with fires in some of the early Model S's which I believe happened once or twice but were huge stories. So huge in fact that Elon came out and compared the handful of Model S's that had experienced a fire under any circumstance with the hundreds of internal combustion engine cars that went up in flames every day for a litany of reasons and received no press time. At the end of the day, a fancy new electric car possibly being a ticking time bomb sells more ad space than another Grand Prix going up in flames because of leaking valve cover gaskets.
And the same will apply to the next generation of Autopilot as it's rolled out in the next couple of years. They will relentlessly look for ways to make something out of nothing. Personally, I take comfort in the fact that as hard as they've tried over the last year, they still have nothing. There hasn't been a single substantiated example (that I can recall) of Autopilot truly failing. Even the dude that got his head taken off. The NHTSA opened an investigation into that accident and found that Autopilot was not at fault. The guy had 7 full seconds to see the big rig turn in front of him and I don't believe he ever even attempted to stop. He just wasn't looking.
Other than that I have no opinion.
Cannot like this enough.
Another example would be be this uber Volvo crash. They make it seem like the company is doomed because they had one of their autonomous vehicles involved in a crash. It's always the little blurb at the end that says "oh by the way the crash was caused by another car running into the autonomous car so it wasn't its fault"
Another example would be be this uber Volvo crash. They make it seem like the company is doomed because they had one of their autonomous vehicles involved in a crash. It's always the little blurb at the end that says "oh by the way the crash was caused by another car running into the autonomous car so it wasn't its fault"
That Volvo accident happened where I go to school (ASU). They have those things driving every where out here. And the drivers on the road are complete fucktards. The worst I've ever dealt with.
This is an Autopilot 1 car, like mine. It's cruise control with lane keeping. People like to assign it more responsibility than it deserves so that when it fails in a circumstance it was never intended to be used in they can pull out their old man card and say how much technology can't be trusted.
That setup uses a front facing radar, 12 ultrasonic sensors and a single camera to do everything. As far as identifying lanes, all of that falls on the camera and it would have kept the car in what it perceived to be the lane. So it's actually working exactly as designed. The driver just didn't see the wall because he wasn't paying attention.
I feel like people should be forced to have hands on time with the system before they criticize it. I've counted a couple times that I would have at the very least had an emergency stop, where the car identified a car 2 cars in front of me slamming on it's brakes and in response slowed the car down before the car in front of me ever saw the car in front of them braking. It is so amazingly good at what it does that it's hard to put into words. Now if we're talking about an Autopilot 2 car after they've gone fully autonomous then I'll step right up and criticize.
The other thing people need to realize is Tesla stories = page views = revenue. It's the Apple mentality. They want so desperately to make a story out of anything. Antenna-gate wasn't really a thing. I mean sure you could grip your phone in such a way to make the signal disappear but in practice nobody actually held their phones that way. But that doesn't matter. As long as you can make the connection that something can happen even if it very likely won't then you can create enough of a commotion to make a story.
Antenna-gate didn't die down because Apple gave free rubber bumpers to everyone; it died down because using the rubber bumpers prevented the problem. Nobody actually used the rubber bumpers. Same with the purple flare in the photos of the iPhone 5 or 6. I don't remember which. Same with fires in some of the early Model S's which I believe happened once or twice but were huge stories. So huge in fact that Elon came out and compared the handful of Model S's that had experienced a fire under any circumstance with the hundreds of internal combustion engine cars that went up in flames every day for a litany of reasons and received no press time. At the end of the day, a fancy new electric car possibly being a ticking time bomb sells more ad space than another Grand Prix going up in flames because of leaking valve cover gaskets.
And the same will apply to the next generation of Autopilot as it's rolled out in the next couple of years. They will relentlessly look for ways to make something out of nothing. Personally, I take comfort in the fact that as hard as they've tried over the last year, they still have nothing. There hasn't been a single substantiated example (that I can recall) of Autopilot truly failing. Even the dude that got his head taken off. The NHTSA opened an investigation into that accident and found that Autopilot was not at fault. The guy had 7 full seconds to see the big rig turn in front of him and I don't believe he ever even attempted to stop. He just wasn't looking.
Other than that I have no opinion.
Where was this made a big deal? I didn't see it on the news or fb, or anything honestly. It was in a long chain of getting lost in YT vids. I was genuinely curious why this happened, if it's truly an autopilot 1 then it makes sense.
It was getting passed around everywhere when it first happened. It's the nature of the beast with new tech. I hate Uber as a company but they got the same unfair treatment. Huge stories this week about how one of their autonomous Volvos got into an accident and rolled over, most of which failed to mention that it was the car that got hit, not the car that did the hitting, and almost all of which brought into question the safety of autonomous cars.
Here's a post I made a couple months ago here:
That's huge and that's a fleet of cars of which nearly a quarter have no Autopilot hardware installed. Who covered it? Basically nobody. And they didn't cover it because it's not salacious.
I completely agree. The autopilot accidents get blown way out of proportion. They fail to mention how many go Daily accident free. Uber got the huge shaft with that news story.
Thread Info
-
Created -
16Replies
-
478Views
-
Participants list
