Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register
  • Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
  • Tesla's Supercharger Team was recently laid off. We discuss what this means for the company on today's TMC Podcast streaming live at 1PM PDT. You can watch on X or on YouTube where you can participate in the live chat.
This site may earn commission on affiliate links.
How much higher can TSLA go when, let's say, they start licensing out FSD to Cruise and Waymo? I'm not exactly sure how to value the licensing of FSD data and/or software.
While this sounds good, it's a long way off based on my FSBb experiences. I know this was a talking point since 2019 and its one that led to me buying FSD. It does not resonate with me anymore. I did buy the dream that it was "right around the corner" at first. It's still there just around the Dojo corner, maybe.

I had FSDb for $2000 and did not buy it on my recent purchase based on its performance over 140k miles and 3 years. I did buy enhanced Autopilot, as that works pretty well on divided highways. Only 2 times has the car headed right for a concrete divider while in FSDb on a divided highway with high g-force intervention required to prevent a collision.

I bought the subscription for FSDb on my new Y and let it lapse last month. It was still unable to drive without phantom braking on a 2 lane-divided highway which is perfectly straight without mistaking a mirage of hot pavement for an imminent collision. This was the same as 3 years ago when the previous vehicle was equipped with radar it still had the same phantom braking issues through the desert. Yesterday it attempted to pull right into the traffic stream on a right-on-red situation.

I don't post here as often as I used to but in my lurking, I just had to say something about this comment. Waymo and Cruise have accomplished driverless robotaxis, though the sensor and compute overhead is too high at the moment.

I had much of my investment in TSLA but am slowly moving it into real estate at this point.
 
While this sounds good, it's a long way off based on my FSBb experiences. I know this was a talking point since 2019 and its one that led to me buying FSD. It does not resonate with me anymore. I did buy the dream that it was "right around the corner" at first. It's still there just around the Dojo corner, maybe.

I had FSDb for $2000 and did not buy it on my recent purchase based on its performance over 140k miles and 3 years. I did buy enhanced Autopilot, as that works pretty well on divided highways. Only 2 times has the car headed right for a concrete divider while in FSDb on a divided highway with high g-force intervention required to prevent a collision.

I bought the subscription for FSDb on my new Y and let it lapse last month. It was still unable to drive without phantom braking on a 2 lane-divided highway which is perfectly straight without mistaking a mirage of hot pavement for an imminent collision. This was the same as 3 years ago when the previous vehicle was equipped with radar it still had the same phantom braking issues through the desert. Yesterday it attempted to pull right into the traffic stream on a right-on-red situation.

I don't post here as often as I used to but in my lurking, I just had to say something about this comment. Waymo and Cruise have accomplished driverless robotaxis, though the sensor and compute overhead is too high at the moment.

I had much of my investment in TSLA but am slowly moving it into real estate at this point.
Maybe your bar is set way too high? Cruise certainly doesn't care to master phantom braking prior to rolling out their robotaxies. Wait sorry, you just need to slap a warning sticker on the back glass and call it a day.


Nor do they care to solve cross traffic red light runners even though it has side cameras and lidar. I get it if a human didn't see the incoming car but certainly a Tesla would not move into this intersection even if the light turned green.



Example of FSD reacting to red light runners.

 
Last edited:
Maybe your bar is set way too high? Cruise certainly doesn't care to master phantom braking prior to rolling out their robotaxies. Wait sorry, you just need to slap a warning sticker on the back glass and call it a day.


Nor do they care to solve cross traffic red light runners even though it has side cameras and lidar. I get it if a human didn't see the incoming car but certainly a Tesla would not move into this intersection even if the light turned green.



Example of FSD reacting to red light runners.

Perhaps it is, but on a perfectly straight road with literally no cars for 0.5 miles I am phantom braking from 85 to 70 mph at least 1 time per 5 miles.

My model Y would also have crashed, when it changed lanes from the fast lane, into the shoulder across the left yellow line on Highway 85, and was driving in the shoulder headed for the concrete divider that narrows with the overhead signs. This was at 2 pm in relatively light traffic.

When I say Waymo and Cruise are successful, I mean they are actually doing it.
 
Perhaps it is, but on a perfectly straight road with literally no cars for 0.5 miles I am phantom braking from 85 to 70 mph at least 1 time per 5 miles.

My model Y would also have crashed, when it changed lanes from the fast lane, into the shoulder across the left yellow line on Highway 85, and was driving in the shoulder headed for the concrete divider that narrows with the overhead signs. This was at 2 pm in relatively light traffic.

When I say Waymo and Cruise are successful, I mean they are actually doing it.
Have you tried recalibrating? I can count on one hand the number of phantom brakes in 40+K miles and still have some fingers left (not counting hesitations, which aren't even all that frequent).
 
Have you tried recalibrating? I can count on one hand the number of phantom brakes in 40+K miles and still have some fingers left (not counting hesitations, which aren't even all that frequent).
Same experience here. I used to get phantom braking on older versions, but in V11 phantom braking seems to be a thing of the past.
 
Yah it’s pretty good on the highway. But in the city it hasn’t moved the needle much in the last two years as far as improvements go. Until it gets cities figured out it’s not really a big deal in my opinion.

Jmho.
Agreed on highway, but on surface streets in the city, at least my car has improved remarkably over the past two years. Two years ago, I never used FSD on surface streets, now I use it most of the time (albeit mostly without setting Nav).
 
Agreed on highway, but on surface streets in the city, at least my car has improved remarkably over the past two years. Two years ago, I never used FSD on surface streets, now I use it most of the time (albeit mostly without setting Nav).
Two years ago, FSD was simple. This is our street... improvements?

1692801570109.png


I just watched our own video recording of a major release about this same time. I clearly forgot how far it's come. I was excited that it could see the lines on the road on city streets... or even drive on city streets at all! I had to tell it to proceed at green lights as I recall. Earlier versions wouldn't even turn in the city and we argued here on TMC whether we should use it for the city yet since it was "only for freeways" (prior to this, but could do some basic driving on streets).

There have been so many upgrades, I'm guessing about 20 or more since then, and each is a small improvement - 3 steps forward and 1 back was typical. Jittery wheel was typically on the big changes as I compare it to a fawn learning to stand up first time. Phantom braking is/was the same as well. Very minimal now, if at all most drives. But watch, our next big release and it could be back... but then goes away in the next couple revs.

So we don't see the change that most people witness when they see it today for the first time (or past couple of upgrades). This might explain how some are disappointed in FSD progress. You really have to step back to see the big picture and the value proposition here. For us, the trend line feels constant with leaps and setbacks in technology along the way.
 

I think the main point here is that the neural network will automagically discover any future improvements they could have found. Thus it will be hard to beat. That's the nature of neural networks, they will discover what they need to discover, as long as the network is large enough, you have enough data and you train them long enough. And Tesla has plenty of all of these.
 
Last edited:
V12 in training
V11.4.7+ minor tweaks, bug fixes
Hope we all go to mostly trained v12 in December

Btw, remember the recent global hiring of testers, seems that was 4D chess move as they are deep in as v12 training enablers