-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla haven't recognized a train
- Thread starter Nickage87
- Start date
Drive down a narrow street with cars parked and have one of them with the door opened. You will hit it. Piles of trash on the side of the road and you will hit them. Gigantic pot holes and it will hit them. Even cars that are parked a little far out and it will NOT go over the center line to go around them and looks like it will side swipe them.
Okay - zero context. Why system was engaged when these issues happened and how recently have you experienced them? How frequently do you encounter these issues?
Last edited:
Just to get back on topic. The FSD Beta (New code/4D) still doesn't identify trains yet. So no way the "old" code we are running will identify them. It sees them as trucks and the crossing as a red light. Also have watched most FSD Beta videos and it never slows or seems to see RR tracks or crossings. Of course this will be updated at anytime in a future FSD Beta. They are on #8 now.
EDIT: Just to add this thread has almost all the FSD Beta videos in it.
FSD Beta Videos (and questions for FSD Beta drivers)
EDIT: Just to add this thread has almost all the FSD Beta videos in it.
FSD Beta Videos (and questions for FSD Beta drivers)
Thank you for posting all of those instances where people were using technology incorrectly. AP and FSD and NOT yet capable of completely autonomous control of the vehicle: the DRIVER is still responsible for the vehicle. These were all instances of DRIVER failure. Had the driver's been paying attention like they were supposed to, none of these instances would have happened.
What else you ya got?
Except you keep blaming it as a Tesla error when it is not. All of these videos keep saying "Autopilot failure" when in fact they are NOT AP or FSD failure. You keep promoting that Tesla and its technology is failing, when in fact, it isn't.
Yes AP (and FSD) are capable of not recognizing objects and hitting them. And so are drivers. Personally, I feel that overall, AP and FSD do a FAR BETTER job than humans of keeping our highways safe. The people who focus on these types of situations (yourself included) are just continually engaging in logical fallacies. If you're looking for the perfect system, you will live a sadly disappoint life.
MarsOrBust
Member
Auto-Pilot both in planes and cars need to be monitored.
On a car there is much more traffic and obstacles to be aware of then on a plane in cruise flight.
Some aircrafts do have collision avoidance systems that communicate with other aircrafts that have the same system and this works rather well. If this system was added to cars then those cars would avoid each other however with the millions of other cars with out there without that system onboard it would render it useless. If all Teslas had it you would not hit another Tesla on autopilot but thats about it.
Currently FSD is a long long long way from RoboTaxi and I really don't see that for at least another 10 years. To many variables. It still is not happening in planes with passengers and I suspect it would happen there first before cars. Now a monitored auto pilot does not have to be perfect to be useful, it's an aid. Even a perfected autopilot system might crash but as long as it performs better than humans thats all that matters.
On a car there is much more traffic and obstacles to be aware of then on a plane in cruise flight.
Some aircrafts do have collision avoidance systems that communicate with other aircrafts that have the same system and this works rather well. If this system was added to cars then those cars would avoid each other however with the millions of other cars with out there without that system onboard it would render it useless. If all Teslas had it you would not hit another Tesla on autopilot but thats about it.
Currently FSD is a long long long way from RoboTaxi and I really don't see that for at least another 10 years. To many variables. It still is not happening in planes with passengers and I suspect it would happen there first before cars. Now a monitored auto pilot does not have to be perfect to be useful, it's an aid. Even a perfected autopilot system might crash but as long as it performs better than humans thats all that matters.
You also don’t need thousands of hours in the cockpit and years of formal training to get a driver’s license either.
In case anyone is wondering, that’s a Beechcraft Bonanza aircraft.
Old versions of this aircraft had a V tail, and was very popular with doctors, as it was considered the Cadillac of small airplanes. We used to call them “Fork tailed Doctor killers.”
There are scenarios where we are required to keep our hands on the yoke when the autopilot is flying. While it’s true that this isn’t 100% of the time, I think you’d be surprised at how often we have our hands on the yoke while the autopilot is flying.
I don't think I would be that surprised. Correct me if I am wrong, but the times you must have your hands on the control are the approach and landing if autolanding right ?
Now, how many times was it necessary to disconnect the automation cause it would have ruined your day ?
You need thousands of hours to get a pilot licence ? A commercial licence ? That's news for me.
Those are two instances, yes.
The second question I hesitate to answer, as I’d have to put each instance in full context; just giving a number isn’t a good indicator as to how good current automation is.
I think rather than an exact number, it’s better to say that automation technology is a great tool, when used correctly and with proper monitoring. It’s the proper monitoring part that I think is going to be a major cause for concern with autonomous vehicles. Believe it or not, but proper automation monitoring is a very large subject that is thoroughly taught in aviation ground schools.
The second question I hesitate to answer, as I’d have to put each instance in full context; just giving a number isn’t a good indicator as to how good current automation is.
I think rather than an exact number, it’s better to say that automation technology is a great tool, when used correctly and with proper monitoring. It’s the proper monitoring part that I think is going to be a major cause for concern with autonomous vehicles. Believe it or not, but proper automation monitoring is a very large subject that is thoroughly taught in aviation ground schools.
The numbers are actually published by the FAA, but they are absolute bare minimum values. Airlines typically hire pilots with far more experience than the FAA’s minimums.
At the time of the pandemic, even if you have a astronaut licence, they won't let you fly a 172
But in Canada, before the pandemic, they were hiring cadet with 250 hours for the 737....good old days !
Last edited:
I agree, proper monitoring is extremely important, cause no matter how good is the system...
But in a plane, I don't have the feeling that the plane is just waiting to do something stupid as I had wit h the car.
Of cours this is highly subjective.
(moderator note: I left this thread here originally, even though it is not a "model 3" topic at all, it is a "tesla autopilot" topic. Its now very much in the "general tesla / autopilot" discussion. I am moving this thread to the autopilot section, and out of the model 3 section. I will leave a pointer for 3 days so people interested can find it in its new location).
Similar threads
