Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register
This site may earn commission on affiliate links.
The next big milestone for FSD is 11. It is a significant upgrade and fundamental changes to several parts of the FSD stack including totally new way to train the perception NN.

From AI day and Lex Fridman interview we have a good sense of what might be included.

- Object permanence both temporal and spatial
- Moving from “bag of points” to objects in NN
- Creating a 3D vector representation of the environment all in NN
- Planner optimization using NN / Monte Carlo Tree Search (MCTS)
- Change from processed images to “photon count” / raw image
- Change from single image perception to surround video
- Merging of city, highway and parking lot stacks a.k.a. Single Stack

Lex Fridman Interview of Elon. Starting with FSD related topics.


Here is a detailed explanation of Beta 11 in "layman's language" by James Douma, interview done after Lex Podcast.


Here is the AI Day explanation by in 4 parts.


screenshot-teslamotorsclub.com-2022.01.26-21_30_17.png


Here is a useful blog post asking a few questions to Tesla about AI day. The useful part comes in comparison of Tesla's methods with Waymo and others (detailed papers linked).

 
Last edited:
yes I am unhappy
yes I was lied to
it was sold as a product not as a dream
Elon promised early access
The BMS manages the battery and many people on the forums blamed apparent range degradation with only the BMS being recalibrated and not real degradation
Charging would be nice to know to properly maintain the monist expensive component of the car
3rd party replacement - Historically owning a computer, cameras and batteries over five years is not a good investment. A Tesla is mainly a computer and batteries. It is software also but with over the air updates, it is still limited by the hardware. I believe there will be many Tesla is the junkyard premature do to computers and batteries.

bottom line - my happiness does not revolve around my car or this forum or Tesla shills. I will live my life and continue to point out fake engineers and stock pumping kickstarter cheerleaders that are fine being lied too and taken advantage of by the smartest and the richest guy in the world who knows more about manufacturing than anyone.

80% for daily driving is optimum for what? range, $/kw, battery health, battery degradation. When should daily driving be changed to a trip? only when additional range is needed to get to a supercharger or is the supercharging worse than charging to 95%. You might have all you answers with 80% but I still have questions and a forum would logically be the place to get answers, our at least others owners opinions, but I guess it is just an Elon circle flirt.

this is intended to be genuine advice and criticism of you and your opinions
Someone had to say it. There is a part if the Isaacson book where Elon admits needing to create some fantasies for fear of TSLA losing analyst/investor mojo. Apparently that's when Elon began the robotaxis distraction. From then it seems to be on steroids with every new TSLA venture (4680, CT, semi, optimus, dojo, AI, v12, ...).
 
Last edited:
I've had the .2 release for a full 48 hours now. So far, zero dry wipes (knock on wood).
Can I knock on carbon fiber? Replaced the wood dash. Also added a tilting/swiveling screen mechanism, $155 from RPM, and 20 minutes labor, very cool. As everyone else said, didn't expect to like it as much as I do. I outright dig it. I have had no recent wiper problems but still on 11.4.4. I have to say this is the best FSDb version I have had. Very few mistakes. Not looking forward to the 11.4.7.x updates.
 
Last edited:
I read this reaction a lot from people. Honestly, who cares that it drives center on an unmarked road if it detects an oncoming car and then moves to the right? Is there a reason you drive on the right side when no other cars are around?
If no other cars are around then I don’t care, but it isn’t confidence inspiring at all. My issue is that I live in a large residential development with no center lines anywhere. I need to test when it starts to move over for oncoming cars again. Previously it was reacting a bit late for my preference. Generally speaking though, the car should behave as a reasonable driver would behave and driving down the middle of a somewhat busy road isn’t that imho.
 
yes I am unhappy
yes I was lied to
it was sold as a product not as a dream
Elon promised early access
The BMS manages the battery and many people on the forums blamed apparent range degradation with only the BMS being recalibrated and not real degradation
Charging would be nice to know to properly maintain the monist expensive component of the car
3rd party replacement - Historically owning a computer, cameras and batteries over five years is not a good investment. A Tesla is mainly a computer and batteries. It is software also but with over the air updates, it is still limited by the hardware. I believe there will be many Tesla is the junkyard premature do to computers and batteries.

bottom line - my happiness does not revolve around my car or this forum or Tesla shills. I will live my life and continue to point out fake engineers and stock pumping kickstarter cheerleaders that are fine being lied too and taken advantage of by the smartest and the richest guy in the world who knows more about manufacturing than anyone.

80% for daily driving is optimum for what? range, $/kw, battery health, battery degradation. When should daily driving be changed to a trip? only when additional range is needed to get to a supercharger or is the supercharging worse than charging to 95%. You might have all you answers with 80% but I still have questions and a forum would logically be the place to get answers, our at least others owners opinions, but I guess it is just an Elon circle flirt.

this is intended to be genuine advice and criticism of you and your opinions

Tesla owners must arbitrate false advertising claims over autopilot feature- ruling​

Daniel Wiessner
Mon, October 2, 2023 at 11:18 AM EDT·2 min read
0

7655ed12fb62711247bf34873aebc0bb

FILE PHOTO: Tesla Model 3 drives on autopilot along California freeway
By Daniel Wiessner
(Reuters) - A California federal judge has ruled that a group of people who own cars made by Tesla Inc must pursue claims that the company misled the public about its autopilot features in individual arbitration rather than court.
The ruling marks a significant victory for Tesla, as it means the company will not have to face class action claims on behalf of much larger groups of vehicle owners.
U.S. District Judge Haywood Gilliam in a decision issued on Saturday said four Tesla owners who filed a proposed class action last year had agreed to arbitrate any legal claims against the company when they accepted its terms and conditions while purchasing vehicles through a Tesla website.
A fifth plaintiff who did not sign an arbitration agreement waited too long to sue, Gilliam ruled in dismissing that plaintiff's claims.
Tesla and lawyers for the plaintiffs did not immediately respond to requests for comment on Monday.
The lawsuit accuses Tesla of repeatedly making false statements indicating that its advanced driver assistance systems (ADAS) technology was on the verge of delivering fully self-driving vehicles.
The plaintiffs all said that they paid thousands of dollars to purchase the optional ADAS technology when they bought Tesla cars between 2017 and 2022.
But instead of delivering on its promises, Tesla's technology has been unreliable and has led to accidents, injuries and deaths, the plaintiffs claimed.
Tesla has denied wrongdoing. The company moved to send the claims to arbitration, citing the plaintiffs' acceptance of the arbitration agreement.
Gilliam on Saturday rejected claims by the plaintiffs that the agreements signed by four of the plaintiffs were unenforceable.
The decision came in the midst of the first U.S. trial over allegations that Tesla's Autopilot feature led to a death because it was based on untested experimental technology that should not have been sold to the public.
The plaintiffs in that trial in California state court allege the Autopilot system caused a Model 3 to veer off a highway near Los Angeles at 65 miles per hour, strike a palm tree and burst into flames, killing the owner and injuring two passengers.
Tesla has said the accident was the result of driver error.
(Reporting by Daniel Wiessner in Albany, New York, Editing by Alexia Garamfalvi and Deepa Babington)
 
I read this reaction a lot from people. Honestly, who cares that it drives center on an unmarked road if it detects an oncoming car and then moves to the right? Is there a reason you drive on the right side when no other cars are around?

I don't necessarily mind that it drives in the center, I care about how it behaves when it does so:

1. It waits too late to move over for oncoming traffic. The goal is not "barely dodge the oncoming car", the goal should be "move over courteously way earlier, so the other driver doesn't freak out and think you're playing chicken with them".

2. It adjusts back into its own lane when it reaches a stop sign, again at the last minute. However, there are many scenarios where it can get stuck stopped in the middle, too. For example, there might be several cars in front of you, all heading towards the same stop-sign. When the lead car stops for the stop sign, and then everyone else behind comes to a chain-reaction stop, FSD often ends up stopped still in the middle. Now you're stopped in the middle, can't move forward (car in front of you), and an oncoming car that wasn't present before can now suddenly appear.

When you consider all the possibilities that happen in traffic, it would be better to default to hugging your own side of the road, and then only move into the center when it's clear that both (a) there's no oncoming traffic close enough + (b) there is definitely room, no matter what the other cars in both directions may do in the future, to get back over to the right before having to stop.
 
Question here from someone new to FSDb: I thought that the problem of the lane change because of an interruption of a left-side barrier was solved.

Yesterday I was driving along the CA-37 - a long two-lane road, one in each direction mostly separated by a cement barrier - and at one of the barriers' interruption the FSDb thought that was a good idea to move to the imaginary lane on the left. Luckily the turn signal caught my attention because FSD was already steering in fractions of a second...

Note: the picture of the exact point attached is taken from Google Maps, I'm 90% positive that now there is a yellow solid line on the right over that new patch of tarmac.
 

Attachments

  • 2023-10-02_10-10-28.png
    2023-10-02_10-10-28.png
    2.7 MB · Views: 81
I don't necessarily mind that it drives in the center, I care about how it behaves when it does so:

1. It waits too late to move over for oncoming traffic. The goal is not "barely dodge the oncoming car", the goal should be "move over courteously way earlier, so the other driver doesn't freak out and think you're playing chicken with them".

2. It adjusts back into its own lane when it reaches a stop sign, again at the last minute. However, there are many scenarios where it can get stuck stopped in the middle, too. For example, there might be several cars in front of you, all heading towards the same stop-sign. When the lead car stops for the stop sign, and then everyone else behind comes to a chain-reaction stop, FSD often ends up stopped still in the middle. Now you're stopped in the middle, can't move forward (car in front of you), and an oncoming car that wasn't present before can now suddenly appear.

When you consider all the possibilities that happen in traffic, it would be better to default to hugging your own side of the road, and then only move into the center when it's clear that both (a) there's no oncoming traffic close enough + (b) there is definitely room, no matter what the other cars in both directions may do in the future, to get back over to the right before having to stop.
Centering makes sense on a straight, flat, and open roads, but I was about to reply with the exact same points BBTX makes, plus I'd add that it centers too much when approaching blind hills. I disengage at the same spot on every trip home because of this.
 
I don't necessarily mind that it drives in the center, I care about how it behaves when it does so:

1. It waits too late to move over for oncoming traffic. The goal is not "barely dodge the oncoming car", the goal should be "move over courteously way earlier, so the other driver doesn't freak out and think you're playing chicken with them".

2. It adjusts back into its own lane when it reaches a stop sign, again at the last minute. However, there are many scenarios where it can get stuck stopped in the middle, too. For example, there might be several cars in front of you, all heading towards the same stop-sign. When the lead car stops for the stop sign, and then everyone else behind comes to a chain-reaction stop, FSD often ends up stopped still in the middle. Now you're stopped in the middle, can't move forward (car in front of you), and an oncoming car that wasn't present before can now suddenly appear.

When you consider all the possibilities that happen in traffic, it would be better to default to hugging your own side of the road, and then only move into the center when it's clear that both (a) there's no oncoming traffic close enough + (b) there is definitely room, no matter what the other cars in both directions may do in the future, to get back over to the right before having to stop.
I was driving on a narrow parkway last week with no center line and FSD actually behaved very naturally. It would drive towards the center then move over in good time when it spotted an oncoming car, almost exactly how I would have done it. Past versions seemed to revel in playing chicken with oncoming cars but 11.4.7 seems to get this right in my experience.
 
Yeah but as a sound effect, proper is no fun.

Even Elon prefers gastric distress noises over the simple smooth hum of the world's most advanced drive motors...
Which leads me to a whole other point of disappointment with Tesla. When they were made to get rid of the clapping coconut sounds when you were in drive.

I swear to all that’s holy, driving around like I was on a quest for the holy grail was the only reason I got my S.
 
And the lord sayeth, that whichith be solved, becomith unsolved by the righteous hand of thine lord’s unhumble garden gnome, eshorticus minimus muskis.

FSDbj 3:16

Joe


Question here from someone new to FSDb: I thought that the problem of the lane change because of an interruption of a left-side barrier was solved.

Yesterday I was driving along the CA-37 - a long two-lane road, one in each direction mostly separated by a cement barrier - and at one of the barriers' interruption the FSDb thought that was a good idea to move to the imaginary lane on the left. Luckily the turn signal caught my attention because FSD was already steering in fractions of a second...

Note: the picture of the exact point attached is taken from Google Maps, I'm 90% positive that now there is a yellow solid line on the right over that new patch of tarmac.