This is so strange. I have a '21 Model 3 and 12.5.4.1 is like yours. It is so so so good. A large majority of my drives are like the one you just posted. I might intervene to be polite or to reroute but otherwise, its all hands off.
You're not alone in your experience. My wife's Model Y has been disappointing with Full Self-Driving (FSD). I've heard from others who own Model 3 and Model Y vehicles and share similar feelings about FSD based on their experiences. It seems to vary from car to car at this point. My Model 3 performs remarkably well with FSD, except for the usual edge cases, but the Model Y is terrible, and unfortunately, there's nothing we can do to improve it. I would love to know how many others are affected by their cars not performing well with FSD.
Good video, I have a 2021 M3 with 12.5.4.1 and just did a perfect FSD drive with it yesterday but the previous drive did had trouble with mapping/routing (something I think TESLA is addressing) and confusion with poorly designed lane markings on the road. My BIG-GRIPE with FSD is how TESLA is handling the roll-out. I wish TESLS were more transparent regarding IMPROVEMENTS and KNOWN ISSUES with each release. They are asking us to be testers but are blindfolding us regarding the state of a new release and asking us to blindly test it. I wish they would disclose known-issues they encountered in their testing of the release -- that way we could be pre-aware of known issues with a particular release.
@@FSDdriver I did a reset/reboot last night and it took a few minutes for the screen to return to normal. We had a lunch meeting with friends today so used FSD on the way there and back. Much better at obeying speed limits (when it did read them) and not so aggressive but I did disengage when it decided (for no apparent reason) to move to the inside lane when we were within a mile of a right turn and there was traffic in the right lane behind us. I wonder if hardware 3 is ever going to be able to handle FSD properly.
I have a possible explanation. AI works basically on a statistical model that causes virtual neurons to fire. But remember this branch tree prediction is many if not hundreds or thousands of branches deep. Differences in lighting and traffic can cause slight variations, even on the same vehicle on the same route when performed at different times.
They can remove a lot of the stuff you would normally think of as "random" however there may still be other biases like model y vs 3 that exist. I think changes across model years and car model like 3/y/s/x etc may have a larger effect.
No, I can explain why its happening. Thats because AI is not intelligent, its stupid. Its just guessing. We need a huge leap in AI tech for it to work, it may never happen.
I believe this is a balanced perspective. Currently, AI acts more like an advanced search tool than true intelligence, gathering data from various sources to present to users. It can only work with information that already exists and does not generate new ideas independently. In terms of Full Self-Driving (FSD), we have well-established road rules and decades of traffic data to guide us. Progress should be steady, not one step forward and two steps back with every update. It's concerning that the system can overlook basic knowledge, such as recognizing that a green light means "go" and not confusing it with "stop." This suggests it may be receiving flawed data models.
![I.N "HALLUCINATION" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/n5B5q1Hwt_U/mqdefault.jpg)
![A320 BIA 🇫🇷 Bastia | LANDING 34 Visual Approach | 4K Cockpit View | ATC & Crew Comm [RE UPLOAD]](/img/1.gif)
This is so strange. I have a '21 Model 3 and 12.5.4.1 is like yours. It is so so so good. A large majority of my drives are like the one you just posted. I might intervene to be polite or to reroute but otherwise, its all hands off.
You're not alone in your experience. My wife's Model Y has been disappointing with Full Self-Driving (FSD). I've heard from others who own Model 3 and Model Y vehicles and share similar feelings about FSD based on their experiences. It seems to vary from car to car at this point. My Model 3 performs remarkably well with FSD, except for the usual edge cases, but the Model Y is terrible, and unfortunately, there's nothing we can do to improve it. I would love to know how many others are affected by their cars not performing well with FSD.
Good video, I have a 2021 M3 with 12.5.4.1 and just did a perfect FSD drive with it yesterday but the previous drive did had trouble with mapping/routing (something I think TESLA is addressing) and confusion with poorly designed lane markings on the road. My BIG-GRIPE with FSD is how TESLA is handling the roll-out. I wish TESLS were more transparent regarding IMPROVEMENTS and KNOWN ISSUES with each release. They are asking us to be testers but are blindfolding us regarding the state of a new release and asking us to blindly test it. I wish they would disclose known-issues they encountered in their testing of the release -- that way we could be pre-aware of known issues with a particular release.
Have the same experience with our 2020 Model Y. Just not able to use it. Will try the reset and see if FSD works better.
Let me know how it goes.
@@FSDdriver I did a reset/reboot last night and it took a few minutes for the screen to return to normal. We had a lunch meeting with friends today so used FSD on the way there and back. Much better at obeying speed limits (when it did read them) and not so aggressive but I did disengage when it decided (for no apparent reason) to move to the inside lane when we were within a mile of a right turn and there was traffic in the right lane behind us. I wonder if hardware 3 is ever going to be able to handle FSD properly.
I don’t think it will. Even HW4 won’t do it
I have a possible explanation. AI works basically on a statistical model that causes virtual neurons to fire. But remember this branch tree prediction is many if not hundreds or thousands of branches deep. Differences in lighting and traffic can cause slight variations, even on the same vehicle on the same route when performed at different times.
So completely random and humans don’t understand it and each out come could be different?
@@FSDdriver it's not entirely random, however, yes each out come could be different.
They can remove a lot of the stuff you would normally think of as "random" however there may still be other biases like model y vs 3 that exist. I think changes across model years and car model like 3/y/s/x etc may have a larger effect.
Things like not vision only system, not having a front bumper camera.
No, I can explain why its happening. Thats because AI is not intelligent, its stupid. Its just guessing. We need a huge leap in AI tech for it to work, it may never happen.
I believe this is a balanced perspective. Currently, AI acts more like an advanced search tool than true intelligence, gathering data from various sources to present to users. It can only work with information that already exists and does not generate new ideas independently.
In terms of Full Self-Driving (FSD), we have well-established road rules and decades of traffic data to guide us. Progress should be steady, not one step forward and two steps back with every update. It's concerning that the system can overlook basic knowledge, such as recognizing that a green light means "go" and not confusing it with "stop." This suggests it may be receiving flawed data models.