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

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

This site may earn commission on affiliate links.
I guess I’m one of those on the other side. My S was undrivable on FSD. Tesla checked the software, cameras and whatever they could . They didn’t ship me another build until .4. That was not useable on I95 or around town. I got . 6 and now usable on I95. I’ll go out at 4:00 in the morning and see what happens around town. However I drive a lot of two lane roads and that’s a problem. It doesn’t know to stay way back from orange trucks, and it can’t pass safely since the cameras are centerline and would expose too much car to oncoming. A long range camera in the driver side mirror would be wonderful. My daughter has a coworker with FSD on his Y. It works almost perfectly. You don even know the car is driving itself. Same build. Go figure.
I think there may be something to which car. There are far more Ys and 3s and maybe they are better because of more data? If so, that does not speak well for being able to license to other manufacturers without first having to get reams of data for each model. OT and FWIW, in my MYLR I drove from Oakland to Santa Barbara - or rather the car did - with zero disengagements. I requested a couple lane changes because FSD just isn’t that strategic in looking down the road, but no brakes and once a tap on the accelerator to not annoy the person behind me. I also got kicked out once while trying to find a podcast, so had to stop and restart the drive. I honestly believe I could have napped the 6 hour drive and arrived just fine, except for the one charging stop.

Edit: FSD 12.3.6
 
Weekend side topic, but still Tesla related:

I belong to other forums on the net, non-Tesla forums, but several of them have threads discussing Tesla cars and Elon. I'm always amazed how many people absolutely hate Tesla cars and despise Elon, and yet reading what they post about them I realize they all bring up FUD and inaccuracies about both Elon and the cars, constantly. Many of them don't know what they are talking about:

- Elon lies all the time and is a psychopath.
- Elon is a nazi.
- Elon inherited a family fortune to build his empire with.
- Elon is greedy.
- Elon hates humanity.
- Elon is a racist.
- Tesla wouldn't exist if not for the government handouts keeping them going even today.
- Tesla is a ponzi scheme.
- Tesla cars have terrible quality, even today.
- Tesla Autopilot and FSD are dangerous and kill more people than cars without them.
- Tesla shareholders just voted to give Elon $56 billion in cash.
- Tesla shareholders are all idiots who belong to a cult.
- Tesla's can't drive very far and their batteries don't last.
- Tesla's are not made in America.

It's wild, but it seems to me like most of it is emotional rather than grounded in fact. Its like some people hate Elon and Tesla simply because of the things the media tells them, and they simply parrot that info and sentiment rather than investigate for themselves.

What I've also noticed is how quickly they disregard anything to the contrary, like they can't even consider any possibility their beliefs might be incorrect. Anytime anything goes against their collective group hate its immediately labeled as cult mentality or "stanning" for Musk and consequently tossed aside so they can resume riding the hate train.

I always question everything, constantly, I have my entire life, but it surprises me how many people stop questioning once they get into a comfortable and easy position. The media is not always correct, they twist and lie often, and there are a LOT of people out there with strong false perspectives on both Tesla cars and on Elon too.

Sometimes I just don't understand society or humanity. People's behaviors often confound me, and it often makes me fear for the future of our world. 😔
 
I think its comical how people are so determined to cheerlead waymo here. Feel free to sell Tesla stock and invest in GM and Google in the hope that their dumpster fires of cash suddenly miraculously turn around.
Because we all know that GM are masters of software, and that google, never, ever just suddenly cancel a project. Like, they are famous for having never done so...


This is pretty funny. These "cheerleading" comments as you interpret are far from that, as I've seen these posters elsewhere on the selfDrivingCar internet and they are far from Waymo supporters....They are just trying to be realistic, regardless of company. We aren't bullish on Waymo lol.

Waymo saw the object but misclassified it as a "low-risk" object (like a plastic bag). This forum takes that example to the extreme to try to show it's LiDAR and HD Map aren't scalable. Meanwhile, Tesla FSD would run into objects at a higher rate / mile if not for having a driver.

"Just need more training data and that will solve it".

TSLA investors should actually be concerned that FSD can't stop running into objects yet. This shouldn't be about data - Tesla has plenty of data about the physics of the world. If it's not learning a robust representation of physical shape / distance / lighting to apply to all objects based on the vast amounts of data it already has, why would you think another 2x-5x of data will solve it? Tesla can't rely on memorizing every occurence of every object it will possibly see, it has to generalize like humans.

Again, there isn't any reason Waymo cannot do with Tesla is doing for perception. Waymo has plenty of raw data to build a camera based perception model. And adding LiDAR properly should only help, never hurt.

Mocking Waymo over a perception mistake is silly imo.

If you want to argue Waymo doesn't have enough data for planning / prediction, that would make more sense.
 
Would you also agree on my thesis for Phoenix (or) Chandler as a prime candidate to start RT? (Edit, removed timeframe.)
Elon isn't a fan of geofencing as it causes an unnatural reprioritization which lead you down the path of local maxima and longer training times. Tesla's solution and approach is gathering corner cases in the billions of miles and training those out. Training a model for only small area, the miles gathered would be so small you wouldn't be able to keep up with the trained model everywhere else. And you wouldn't train for just the geofenced area as compute training time is the ultimate bottleneck.

Waymo did Chandler due to HD maps.
 
Weekend side topic, but still Tesla related:

...


Sometimes I just don't understand society or humanity. People's behaviors often confound me, and it often makes me fear for the future of our world. 😔
Just for context. here is the 1847 perspective, costs US$0.99:
Every time I have the sentiments you describe, an almost daily occurrence regarding something or other, I go back to McKay. It really does not help a lot, for humanity's home planet is paying the price but it does help understand why improved communication allows faster spreading of delusions...and a bit of factual information too.
 
I think its comical how people are so determined to cheerlead waymo here. Feel free to sell Tesla stock and invest in GM and Google in the hope that their dumpster fires of cash suddenly miraculously turn around.
Because we all know that GM are masters of software, and that google, never, ever just suddenly cancel a project. Like, they are famous for having never done so...

The chances of me here, in the UK, getting the option of cruise/waymo before Tesla robotaxi are about a trillion to one. And when robotaxi is US-wide, waymo will still be bragging about the 2 or 3 cities they are running loss making pilot programs in.
Waymo knows they are 'most likely' on life support as their current solution is 10 to 50x (capex and opex) more expensive per vehicle per mile and cannot scale fast enough due to high reliance on HD Maps and sensor fusion. Assume they are working on a solution where the outfitted vehicle cost is much lower and only relies on vision as that will be the only way to compete. And also assume they won't announce it until Tesla announces the start of Robotaxi beta (i.e. ChatGPT and Gemini; Google is a fast follower innovator and why I left Google Maps back in 2016 as they wanted nothing to do with AI at the time and still haven't innovated as they have all the market share even though Apple Maps is pretty good now). Look at the difference of Google Maps Mobile Navigation between 2016 and 2024. No noticeable difference.
 
Traffic and pedestrians can still be moving around the vehicle.
While true, this would be similar to when the light turns green and the car stops even if you press on the accelerator (because it sees an issue). Don't know if you've had this behaviour, but I have. The car should do the same thing because even if the sending is late, the realtime evaluation is not.
 
Elon isn't a fan of geofencing as it causes an unnatural reprioritization which lead you down the path of local maxima and longer training times. Tesla's solution and approach is gathering corner cases in the billions of miles and training those out. Training a model for only small area, the miles gathered would be so small you wouldn't be able to keep up with the trained model everywhere else. And you wouldn't train for just the geofenced area as compute training time is the ultimate bottleneck.

Waymo did Chandler due to HD maps.

Agree, but I didn't mean for Training, only inference.

Just geofence the pickup/drop-off scope to areas that don't have "one of the 1,000 nasty intersections" for example, or crazy regions prone to burning RT in the streets. But keep it simple and just say "Chandler Az" for example. Let selection be driven by disengagement data and Islands of Safety. I'm certain Tesla knows where they are just as readily as their 1,000 intersection data.

What could be a flaw in my case is that if they were to choose Chandler, then what it the next update caused a new problem in Chandler. So while these character flaws exists, stability would not be guaranteed but could be generally improved.
 
  • Like
Reactions: Usain
TSLA investors should actually be concerned that FSD can't stop running into objects yet.
If I remember correctly from the shareholder's meeting, Elon now claims that static object avoidance is now completely solved. I'm not sure which version he was talking about, but it was definitely 12.4, 12.5, or 12.6.

Maybe someone here has the quote handy?
 
Weekend side topic, but still Tesla related:

I belong to other forums on the net, non-Tesla forums, but several of them have threads discussing Tesla cars and Elon. I'm always amazed how many people absolutely hate Tesla cars and despise Elon, and yet reading what they post about them I realize they all bring up FUD and inaccuracies about both Elon and the cars, constantly. Many of them don't know what they are talking about:

- Elon lies all the time and is a psychopath.
- Elon is a nazi.
- Elon inherited a family fortune to build his empire with.
- Elon is greedy.
- Elon hates humanity.
- Elon is a racist.
- Tesla wouldn't exist if not for the government handouts keeping them going even today.
- Tesla is a ponzi scheme.
- Tesla cars have terrible quality, even today.
- Tesla Autopilot and FSD are dangerous and kill more people than cars without them.
- Tesla shareholders just voted to give Elon $56 billion in cash.
- Tesla shareholders are all idiots who belong to a cult.
- Tesla's can't drive very far and their batteries don't last.
- Tesla's are not made in America.

It's wild, but it seems to me like most of it is emotional rather than grounded in fact. Its like some people hate Elon and Tesla simply because of the things the media tells them, and they simply parrot that info and sentiment rather than investigate for themselves.

What I've also noticed is how quickly they disregard anything to the contrary, like they can't even consider any possibility their beliefs might be incorrect. Anytime anything goes against their collective group hate its immediately labeled as cult mentality or "stanning" for Musk and consequently tossed aside so they can resume riding the hate train.

I always question everything, constantly, I have my entire life, but it surprises me how many people stop questioning once they get into a comfortable and easy position. The media is not always correct, they twist and lie often, and there are a LOT of people out there with strong false perspectives on both Tesla cars and on Elon too.

Sometimes I just don't understand society or humanity. People's behaviors often confound me, and it often makes me fear for the future of our world. 😔
It’s sad. It's because we're ALL manipulated by the media. Some more than others. Few realize it and many have NO clue. It's especially true with politics, but Tesla and Elon are a popular narrative. The best we can do is realize it, seek multiple and lower bias sources. With blind sheep, one approach is to respectfully share truth, facts and counter points to FUD *AND* suggest that they question the source that fed them FUD in the first place so that they might "consider the source" next time. It's definitely an uphill battle.
 
This might sound FUD-like but this is a honest question...Just curious about thoughts here for older, non-RT vehicles being added to the network....have we seen a reasonable "work around" for the blind spots that are very close to the current gen vehicles? The same blind spots that prevent Tesla from stitching together surrounding images as parking assistance. We can't have a loved one (animal or child), sneak into that blind spot while the vehicle is "asleep". Maybe the vehicle can't ever sleep and must watch the cameras 24/7 to ensure nothing enters the blind spot without leaving? If it didn't leave, ask human to get out and verify? Maybe use the horn? I can think of a few problematic edge cases, like 2 small "things" enter the frame looking like 1 and disappear into the blind spot and then leave the frame, but in reality 1 left and 1 stayed behind? I assume Tesla owns that liability while in the network and that type of accident could cause public outrage. While I understand this question is totally an edge case, it's in the realm of possibilities...
 
  • Funny
Reactions: ShareLofty
Its like some people hate Elon and Tesla simply because of the things the media tells them, and they simply parrot that info and sentiment rather than investigate for themselves.

It's true there are a lot of misconceptions, however some of the animosity towards Elon/Tesla is also because of things Elon has posted on Twitter/X.
 
If I remember correctly from the shareholder's meeting, Elon now claims that static object avoidance is now completely solved. I'm not sure which version he was talking about, but it was definitely 12.4, 12.5, or 12.6.

Maybe someone here has the quote handy?

Elon also described self driving entirely as "basically a solved problem" back in 2016--which it very clearly was not- so I think believing any specific aspect is solved, rather than Elon being his overly optimistic self, without a solution actually being deployed to customers, is perhaps an overly optimistic feeling.

That said I'd love for it to be true as that specific thing is, to my knowledge, the one and only technical capability/safety thing preventing Tesla from offering at least L3 (if not L4) at least on highways.
 
It's true there are a lot of misconceptions, however some of the animosity towards Elon/Tesla is also because of things Elon has posted on Twitter/X.
Has hating on China stopped people from buying Chinese goods? Heck, if you lost your job because your factory moved to China - you/family likely are still buying the cheap options from China.

Looking at the conversations leading up to the vote and the results of the vote ... looks like small % of people create more noise. So the same actors (what we call trolls here) might be on these other forums as well ...

many 99%ers think Teslas are for the 1%ers - that will change once Tesla M2 comes out.
Looking forward to 2026 and beyond .
cheers!!
 
Last edited:
Elon also described self driving entirely as "basically a solved problem" back in 2016--which it very clearly was not- so I think believing any specific aspect is solved, rather than Elon being his overly optimistic self, without a solution actually being deployed to customers, is perhaps an overly optimistic feeling.

That said I'd love for it to be true as that specific thing is, to my knowledge, the one and only technical capability/safety thing preventing Tesla from offering at least L3 (if not L4) at least on highways.
Maybe I'll take the time to dig up the quote if nobody else has it. Maybe there is a transcript of the shareholder meeting somewhere?

From what I remember, Elon was not speaking in the abstract like he did in 2016. It wasn't, "It's a solved problem". It was, "We have solved it".

But I'd need to go back and hear it again to be sure.
 
Maybe I'll take the time to dig up the quote if nobody else has it. Maybe there is a transcript of the shareholder meeting somewhere?

From what I remember, Elon was not speaking in the abstract like he did in 2016. It wasn't, "It's a solved problem". It was, "We have solved it".

But I'd need to go back and hear it again to be sure.

OK, I found it. He said, "We're pretty close to having almost zero static object collisions. Now you have a gigantic improvement in miles between intervention."

So it's a little less definite than I remembered. It's more committal than the abstract "It's a solved problem". But it's somewhat less committal than "We have solved it."