Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

AutoPilot Almost Killed Me!

This site may earn commission on affiliate links.
Not me, guy who bought it.
He wants his $8k back .
Can you help him?

Can’t help him get that $8k back but I can help him (and so can you since he’s your friend) by telling him we all make mistakes and it’s important to learn from them. If he made a decision to spend $8k and now regrets it, he should evaluate what led him to do so (perhaps some irrational exuberance) in order to recognize similar situations in the future so he will make a better decision next time. Or put more succinctly, live and learn.
 
  • Like
Reactions: DanDi58
I’ve found auto pilot works good when you have plenty of distance between everyone. If your driving in a congested area with ton of traffic at a high rate of speed then auto pilot is not very good. I also watch for people who tailgate which is an issue anytime you drive.

A good example is I95 in downtown Miami so many cars going in out traffic there is no way a person can figure out what someone is doing never mind a computer. You never know what is going to happen.

If you have heavy traffic where everyone is stuck in a traffic jam then autopilot is god send and works well in that situation.

When heavy traffic you will want to use Autopilot. If you happen to be on it and able to travel at the speed limit you do not want to use autopilot since drivers will be crazy.
 
I can understand your concern, but if your ask is 100% reliability from any driver assist then you might as well sell your car .. it will never happen, not ever. The goal of any driver assist system (or, eventually, a self-driving car) is to be safer than a human driver. And let's face it, that's not a very high bar in some ways. Does that mean the car won't make mistakes? Of course it will, but if it makes fewer than a human (and, probably, FAR fewer), then that's a good thing.

In 2.5 years and 18,000 miles I never, ever experienced a phantom braking event on the freeway in my Outback with Subaru’s Eyesight system. That’s a camera-only system too; some how little ol’ Subaru has figured it out.

Tesla is supposed to be on the cusp of FSD but they can’t figure out how to solve freeway phantom braking after all these years? Stop it.
 
You have to set expectations correctly. I've had a couple of those, but I'm not scared of it, you gotta learn it. I've seen one where a yellow sign with a flashing yellow near an intersection, FSD wants to treat it like a yellow traffic sign and really hesitates to brake and then continues.
Sorry I think this is dangerous, actually quite dangerous. We drove interstate 640 miles yesterday, and enhanced auto pilot brake checked us 4 times, quite annoying. For comparison, we have a new Toyota RAV4 Prime and I have never experienced a brake check. So it can be done.
 
Last edited:
  • Disagree
  • Like
Reactions: evfeh and Brown1428
Subaru’s Eyesight functionality was limited compared to AP, without question, but the adaptive cruise feature was extremely predictable and reliable.

As others have said, once you learn the limitations of your system you can recognize situations where you know it’s going to struggle and intervene accordingly.

Unlike Eyesight, the trouble with AP is its difficult to predict what situations are going to lead to failure as the criteria seems to be ever changing. Phantom braking is the best example of this.
 
In 2.5 years and 18,000 miles I never, ever experienced a phantom braking event on the freeway in my Outback with Subaru’s Eyesight system. That’s a camera-only system too; some how little ol’ Subaru has figured it out.

Tesla is supposed to be on the cusp of FSD but they can’t figure out how to solve freeway phantom braking after all these years? Stop it.

I have a 2019 Forester with about 17000 miles on it and I’ve had several phantom braking events during adaptive cruise. They almost always involve an overpass or bridge and that’s why I suggested the shadow was being recognized as an obstacle.

I have learned where these trouble spots are (garden state parkway overpasses at exit 145/route 280) and I just simply disable it in that stretch.

It appears to only trigger when you’re on active adaptive cruise - it doesn’t happen when you’re not using cruise control.
 
I have a 2019 Forester with about 17000 miles on it and I’ve had several phantom braking events during adaptive cruise. They almost always involve an overpass or bridge and that’s why I suggested the shadow was being recognized as an obstacle.

I have learned where these trouble spots are (garden state parkway overpasses at exit 145/route 280) and I just simply disable it in that stretch.

It appears to only trigger when you’re on active adaptive cruise - it doesn’t happen when you’re not using cruise control.

OK, fair enough, but from the sound of it your Outback still phantom brakes a lot less than your Tesla, correct?

I find that I tend to leave my foot hovering over the accelerator when using TACC or AP, something I never felt the need to do in my Outback.
 
Don’t own a Tesla. Hoping to order a Y relatively soon (within the next 8 months)...

I don’t know if the master cruise button on the Subaru wheel will cancel it or not but usually if it says obstacle detected you need to override it with the accelerator.
 
Don’t own a Tesla. Hoping to order a Y relatively soon (within the next 8 months)...

I don’t know if the master cruise button on the Subaru wheel will cancel it or not but usually if it says obstacle detected you need to override it with the accelerator.

Yes, that’s true, and pressing the accelerator on the Tesla will have the same effect.

Don’t get me wrong, I like the Model Y a lot overall, but AP needs lots of work before I would feel comfortable using it as regularly as I used Eyesight. I don’t use it at all when my family is in the car unless we’re in really light traffic.
 
  • Like
Reactions: saltsman
I have a 2019 Forester with about 17000 miles on it and I’ve had several phantom braking events during adaptive cruise. They almost always involve an overpass or bridge and that’s why I suggested the shadow was being recognized as an obstacle.

I have learned where these trouble spots are (garden state parkway overpasses at exit 145/route 280) and I just simply disable it in that stretch.

It appears to only trigger when you’re on active adaptive cruise - it doesn’t happen when you’re not using cruise control.


This.
I recently returned from an 1100 km trip where I experienced phantom braking at least 30 times when on ACC; sometimes in heavy traffic and sometimes when I was basically alone on the road. Meeting large vehicles in left-hand curves seemed to be the main trigger, but I also experienced it when cruising on a two-lane road with no other vehicles in sight.
This was with my brand new M3.

I've been driving a Hyundai Ioniq EV for the past 3 years and have NEVER experienced phantom braking when using ACC.

The problem here is that Tesla is bundling ACC with Autopilot and not giving the buyers an option to use ACC without the Beta functions.
 
I've driven a few Teslas before, and never used AP. It's an unfinished product, and I am unwilling to beta test a product with that level of risk.

I'm willing to beta test for them, IF they pay me. Consumers shouldn't have to pay the company, and assume all the risks, for beta testing. If the car causes an accident, like getting rear ended by stopping too fast too suddenly, Tesla will have no accountability because of all the disclaimers. So the owner assumes all the risk, AND pay Tesla more for this "privilege". Lol. Something is not right with this scenario.
 
I'm willing to beta test for them, IF they pay me. Consumers shouldn't have to pay the company, and assume all the risks, for beta testing. If the car causes an accident, like getting rear ended by stopping too fast too suddenly, Tesla will have no accountability because of all the disclaimers. So the owner assumes all the risk, AND pay Tesla more for this "privilege". Lol. Something is not right with this scenario.

If you get rear ended, the fault lies with the person who hit you roughly 99.99999% of the time. It’s called following too closely. Or as they teach you in defensive driving, distance is your friend.
 
If you get rear ended, the fault lies with the person who hit you roughly 99.99999% of the time. It’s called following too closely. Or as they teach you in defensive driving, distance is your friend.
What will be really interesting if the car behind is also a Tesla on AP with following distance set to one car length, if the teslacam recording shows the car in front applied brakes going at 70mph, should be easy to prove in court case that it was car in front that was to blame for the collision