Funny you should say that (btw: I totally agree) as Reuters just reported that Tesla is scrapping the Model 2 and going all in on robotaxis. This won't end well.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Autonomous Car Progress
- Thread starter stopcrazypp
- Start date
-
- Tags
- Autonomous Vehicles
diplomat33
Average guy who loves autonomous vehicles
Personally, I think the best thing Tesla could do is announce a new Tesla model that has cameras and will use V12 end to end but also has surround radar and front lidar, as well as HD mapping, for extra redundancy. And then show how this new Tesla model can reach eyes-off in specific ODD. Such a vehicle would sell great IMO.
Elon says a lot of things, and frankly I think Reuters are usually right.
It's because you want it to be true. Rueters is wrong often.
Not more often than Elon is lying btw. You do you.
There's 0 backing to that and Elon disputed it instantly, unlike the other rumors that turned out to be true.
There's no reason to give that story backing, especially when he basically agreed with Sawyer saying that both RT and the low cost car would be the same platform.
The Chinese cars aren't even in the largest car buying market. Again, there isn't any backing to that story...just anonymous sources.
Reality check for Americans: China is the largest car buying market. Especially so for EV:s.
Ok, the 2nd largest car buying market. The Chinese cars aren't and won't be in the US for the foreseeable future.
Again, there's no substance to that. It could happen, but that's as reliable as the stories that Teslas/EVs are catching on fire everywhere that circles major news sites.
Rueters, like most news organizations, report a lot of stories that end up being incorrect.
It's fine to not believe Elon, but there's no backing to this story and he publicly said they were lying.
Lying is also a poor term. They potentially received bad information.
Daniel in SD
(supervised)
Didn’t read the biography but isn’t this exactly what happened with the Model Y? Elon wanted the design to go full robotaxi (I.e. no controls). Employees were able to convince him to change his mind. Now he’s probably even more convinced that robotaxis are right around the corner. So now skeptical employees are leaking to the press to try and force him to reconsider.
In the Biography it was the M2 that Elon had to be convinced, but I believe the MY was also rumored to not have a wheel.
Hi, DDW --
> MHO Tesla will never do Robotaxis.
Let me do some goal-post shifting. Really my 2 main concerns are:
1: Are the tracker's numbers any good? Let me make a couple of observations:
A: At some level, the answer is "No". I've never done the stats work on this, but the enormous variance.makes it clear that the sample size is way too small. I've mostly been looking at them for trends,, and think they show zero improvement prior to V12. My belief is that the systemic sources of bias would tend to overstate FSD's reliability, but -- User reports also show enormous variance, ranging from, 'Works like a charm!" to "Tried to kill me instantly." so really who knows? If their tester pool contains a couple more "Tried to kill me immediately" people than expected, it might be understating FSD's reliability.
B: My purely subjective sense is, I find something 80-100 miles/critical disengagement for v10/v11 consonant with my social media monitoring. I wouldn't be able to come up with any kind of hard number without the tracker, I'd just say, "Pretty bad". I regard 80-100 miles as "pretty bad," and thus am inclined to just use the tracker's numbers without deeper inquiry.
C: If the tracker's numbers are roughly correct then Tesla is presumably looking at similar numbers. This raises a whole host of interesting questions, ranging from ethical to Tesla's communications about the program's progress.
2: Here's the goal-post shifting: Whether Tesla can ever ultimately get to robotaxis or not -- and maybe they can't, due to reasons ranging from lack of compute/sensor redundancy, b-pillar camera positioning, whatever -- could they show rapid progress toward that goal as measured by the tracker? Is that possible? Is it likely? How would that work?
Guesses about FSD's likely progress range from, "Level 4 in ;24!", to, "Never". I'm just trying put some rough numbers around these issues instead of shouting back and forth. "MOAR DATA!" doesn't seem a plausible mechanism for rapid improvement, because 12.x presumably already embodies a lot of data, and returns from more data are likely to be sub-linear.
I wandered across a related discussion on Reddit:
The OP's estimates of historical progress are, I think, based on cherry-picked (I'm not saying intentionally!) stats from particular releases; my read of progress is "zero, but 12.x is a real step up." But there are some other suggestions there about why rapid progress might be acheivable.
That's the kind of conversation I hope to have. (Why am I not having that conversation on Reddit? Because I don't want to join every social media platform in existence to discuss FSD, and TMC seems to have a base of knowledgeable users who are Tesla-sympathetic and thus inclined to explaining the bull case.)
Yours,
RP
Daniel in SD
(supervised)
diplomat33
Average guy who loves autonomous vehicles
Classic stock pump. TSLA certainly needs something to stay afloat.
Waymo is not dead yet until we see Tesla actually show it can do driverless with superhuman safety. Revealing a robotaxi prototype is one thing. Actually, launching a robotaxi service in a geofence with reliable safety, reliable ride-hailing app, etc... is quite another. And Tesla does not even have a permit to test driverless, let alone get the CPUC permit to do a commercial robotaxi service in CA. They are certainly not going to get one by August. So maybe Tesla launches a robotaxi service in a geofence outside of CA? Still a long shot IMO.
My guess is Tesla will do a big shiny reveal event that will get the Tesla bros excited for sure. That way Tesla can say that they have a robotaxi vehicle, like the Cruise Origin or the Waymo Zeekr, even if the Tesla robotaxi does not get deployed for years.
But my guess is that the robotaxi won't actually enter production because A) Tesla's FSD is not reliable yet for driverless B) Tesla does not have regulatory approval to do a commercial robotaxi service in CA. It will just be a big show to get people excited. But the TSLA bulls will claim that it is "game set and match" and Tesla has won autonomous driving because they will product 1M robotaxis by end of the year. LOL.
I do wonder if the robotaxi will be that two seater that we saw some early mock up shots of, kind of like the Waymo firefly? But IMO, if the Tesla robotaxi is vision-only, it is pretty much dead on arrival because it won't have enough redundancy. If Tesla does produce a robotaxi, it will likely have a safety driver so it won't be driverless. But I think Tesla could do a geofenced robotaxi service with a safety driver, just to claim that they have launched a robotaxi service.
AlanSubie4Life
Efficiency Obsessed Member
Similar threads
- Article
