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.
EM thinks Tesla will have “Level 4 or 5” this year 🤣


Elon has no credibility. He's been consistently wrong now for years about FSD. And he admits that he is speculating. Why should we believe speculation?

But the "level 4 or 5" really makes me laugh. A big difference between L4 and L5 is the ODD. L4 has a limited ODD, L5 has an "unlimited" ODD. So which is it, Elon? Is the ODD limited or not? Apparently, he does not even know if the ODD will be limited or not!!! LOL.
 
But the "level 4 or 5" really makes me laugh. A big difference between L4 and L5 is the ODD. L4 has a limited ODD, L5 has an "unlimited" ODD. So which is it, Elon? Is the ODD limited or not? Apparently, he does not even know if the ODD will be limited or not!!! LOL.
The only way for it to reach L4 by the end of the year is to limit the ODD to whatever commute to Tesla HQ Elon drives on a regular basis that gives him such confidence to speculate.
 
I don't understand why people here are so versed in the SAE levels of autonomy, but then pretend that Level 4 must be flawless.

If Tesla added one additional feature to FSD Beta, where it's able to recognize a situation it cannot safely drive and is able to pull over and put the hazard lights on by itself, that could be Level 4. It may pull over often, but autonomously reaching a minimal risk condition is all that's needed.
 
  • Funny
Reactions: kabin
The only way for it to reach L4 by the end of the year is to limit the ODD to whatever commute to Tesla HQ Elon drives on a regular basis that gives him such confidence to speculate.

Yeah. Maybe your system can't do unsupervised FSD everywhere but could do unsupervised FSD under specific conditions. That's why the ODD really matters. Defining the ODD makes a big difference. In fact, I think defining the ODD would go a long way to addressing Tesla/Elon's FSD credibility gap. For example, if Tesla/Elon announced a plan to offer "L4 - highway" and laid out a clear plan and timeline to get there, I know I would take Tesla and Elon more seriously on FSD. And I suspect the AV industry would also take Tesla's FSD more seriously. But Elon sounds like he is totally winging it and that hurts credibility, especially since his "winging it" has been wrong year after year. I have a big problem when the CEO sounding like he has no clue and is faking it until he makes it. And to be clear, I am not saying that the Tesla engineers are winging it or faking it. They are doing excellent work. They have internal project roadmaps and deadlines. They are not winging it. FSD Beta has made good progress. Elon should just tout the great progress and keep his mouth shut about when FSD will be solved or unsupervised.
 
I don't understand why people here are so versed in the SAE levels of autonomy, but then pretend that Level 4 must be flawless.

If Tesla added one additional feature to FSD Beta, where it's able to recognize a situation it cannot safely drive and is able to pull over and put the hazard lights on by itself, that could be Level 4. It may pull over often, but autonomously reaching a minimal risk condition is all that's needed.

I don't think anyone is claiming L4 must be flawless. If Tesla designed the vision to handle the entire OEDR in a limited ODD and added the MRC feature like you suggest, it would be L4 It might still have a bad intervention rate but it would be L4. If you look at the CA DMV report, you see plenty of companies testing L4 with safety drivers that have bad intervention rates. So L4 definitely does not need to be perfect. The problem is that Tesla is not doing that. They have not added MRC. FSD beta is L2 yet Elon just speculates that in 6 months, it will be L4 or L5 (he is not sure which LOL). That's a problem for me.
 
  • Like
Reactions: kabin and DWtsn
haha. still no statement on FSD liability. still no effort for regulator approval. does not reduce insurance premiums. keep pumping that stock. I purchased a product over four years ago and have giving up on Elon about 2 years ago. Still no product! Is he saying the beta will end by the end of this year and it will be a feature complete product as sold. No, he is just pumping that stock! Tesla has not even defined the FSD or advanced summon yet. What is it. what are its capabilities and when will it be a product? woke minds want to know
 
I don't understand why people here are so versed in the SAE levels of autonomy, but then pretend that Level 4 must be flawless.

If Tesla added one additional feature to FSD Beta, where it's able to recognize a situation it cannot safely drive and is able to pull over and put the hazard lights on by itself, that could be Level 4. It may pull over often, but autonomously reaching a minimal risk condition is all that's needed.
So why don't they do it? They would start making tens of billions.
Oh wait they cant. But keep preaching.
 
If Tesla added one additional feature to FSD Beta, where it's able to recognize a situation it cannot safely drive and is able to pull over and put the hazard lights on by itself, that could be Level 4. It may pull over often, but autonomously reaching a minimal risk condition is all that's needed.
Reductionism suggests that a system that only pulls over to the side of the road is a Level 4 system. Oh, and the manufacturer must accept liability for anything that happens while it pulls over.

In case it wasn't clear, I'm not a fan of the SAE autonomy definitions.
 
So why don't they do it? They would start making tens of billions.
Oh wait they cant. But keep preaching.

Removing drivers in a way that would cause the cars to repeatedly stop in the middle of the road wouldn't make tens of billions. Cruise has thoroughly demonstrated that.

No need to rush SAE levels just for bragging rights. As I said, you can trivialize the levels pretty easily. Tesla has no incentive to push a poorly performing Level 4.
 
I don't understand why people here are so versed in the SAE levels of autonomy, but then pretend that Level 4 must be flawless.

If Tesla added one additional feature to FSD Beta, where it's able to recognize a situation it cannot safely drive and is able to pull over and put the hazard lights on by itself, that could be Level 4. It may pull over often, but autonomously reaching a minimal risk condition is all that's needed.

And what if life was all about the hokey pokey? There's so much more to achieving level 4 and FSD is struggling at level 2.
 
  • Like
Reactions: Goose66
Removing drivers in a way that would cause the cars to repeatedly stop in the middle of the road wouldn't make tens of billions. Cruise has thoroughly demonstrated that.

No need to rush SAE levels just for bragging rights. As I said, you can trivialize the levels pretty easily. Tesla has no incentive to push a poorly performing Level 4.

Ah but it didn't stop them from pushing a poor performing level 2.
 
Ah but it didn't stop them from pushing a poor performing level 2.

Depends on your baseline. Compared to a geo-fenced robotaxi service with limited vehicles, chase cars, and remote operators? Sure. Compared to literally any other ADAS on the market? Miles ahead. I can't think of any readily available in the US that are even attempting stop signs, stop lights, or turns.

And even if you think FSD Beta is trash, that makes it even more likely that Tesla will launch a level 4 system this year. Tesla doesn't let the perfect be the enemy of the good.
 
  • Like
Reactions: sleepydoc
Depends on your baseline. Compared to a geo-fenced robotaxi service with limited vehicles, chase cars, and remote operators? Sure. Compared to literally any other ADAS on the market? Miles ahead. I can't think of any readily available in the US that are even attempting stop signs, stop lights, or turns.

And even if you think FSD Beta is trash, that makes it even more likely that Tesla will launch a level 4 system this year. Tesla doesn't let the perfect be the enemy of the good.
Lol. Luckily it’s not up to Tesla if they release a half baked level 4 system on the public. Regulatory authorities need to approve it first…

Currently Tesla circumvents that by beta and you are fully responsible and hands at wheel at all times… level 4 wouldn’t require that
 
  • Like
Reactions: Genie
Lol. Luckily it’s not up to Tesla if they release a half baked level 4 system on the public. Regulatory authorities need to approve it first…

Currently Tesla circumvents that by beta and you are fully responsible and hands at wheel at all times… level 4 wouldn’t require that

Yep, you're right. Absolutely nooo half-baked level 4 systems on public roads right now. A regulator would never approve a driverless car unless it was absolutely perfect and incapable of mistakes. /s
 
Yep, you're right. Absolutely nooo half-baked level 4 systems on public roads right now. A regulator would never approve a driverless car unless it was absolutely perfect and incapable of mistakes. /s
Mercedes got approved for L3 in just one state and only up to 40 mph…. But sure thing Tesla will get approved in the entire country at all speeds by end of year. Lol. At L4+ 😂😂😂
 
  • Funny
Reactions: t3sl4drvr
Mercedes got approved for L3 in just one state and only up to 40 mph…. But sure thing Tesla will get approved in the entire country at all speeds by end of year. Lol. At L4+ 😂😂😂

If that's where you're moving the goalposts, then there isn't a single autonomous vehicle that meets that criteria. Tesla is by far the closest, however. Having advanced L2 operational in all 50 states and Canada. What's the next closest? 3-4 cities in 2-3 states max?

And anyway, SAE levels are based on design-intent, not regulator approval.
 
  • Funny
Reactions: spacecoin
Tesla is by far the closest, however. Having advanced L2 operational in all 50 states and Canada.
I think you need to look up the terms "driver assistance" and "autonomy" in the dictionary at ponder on the differences. Shipping an eyes-off system with manufacturer liability even at 40 mph is an impressive engineering feat. If a current Tesla on HW3/4 does the full OEDR in a meaningful ODD on existing cars before 2030 or ever, I think that would have been made possible by some serious break-throughs in computer vision in the coming years.

Wake me up when we have unsupervised diagnostics of cancer in radiology, because that will most likely be happing before computer vision (alone) is doing the full OEDR at high speeds for driving.

It doesn't seem to be Tesla's style to do anything so that it works 100% of the time, so I doubt we'll see them take liability for anything. They don't even do it for Autopark, Summon or anything else.
 
Last edited:
I think you need to look up the terms "driver assistance" and "autonomy" in the dictionary at ponder on the differences. Shipping an eyes-off system with manufacturer liability even at 40 mph is an impressive engineering feat. If a current Tesla on HW3/4 does the full OEDR in a meaningful ODD on existing cars before 2030 or ever, I think that would have been made possible by some serious break-throughs in computer vision in the coming years.

Wake me up when we have unsupervised diagnostics of cancer in radiology, because that will most likely be happing before computer vision (alone) is doing the full OEDR at high speeds for driving.

It doesn't seem to be Tesla's style to do anything so that it works 100% of the time, so I doubt we'll see them take liability for anything. They don't even do it for Autopark, Summon or anything else.

I just don't believe the approaches most companies take to geo-fenced L4 autonomy are generalizable. For the most part, it looks like they're overfitting their vehicles to a single city, one city at a time, and compensating for system deficiencies by using HD maps.

That's great if you live in the heart of a city, and never plan on leaving it. But for the vast majority of people, that type of autonomy is not accessible.
 
  • Disagree
Reactions: diplomat33