Autonomous Car Progress

[Knightshade]

To avoid AV testing regulations.

Except, again, FSDBeta is missing numerous elements required of an L3 system- let alone L5.

And the OEDR is demonstrably missing vastly more than "ice cream trucks" to be complete for its ODD.


You keep making up an imaginary narrarive of Tesla secrectly testing an L5 system but calling it L2 when literally every piece of evidence that exists---Actual behavior and capabilities of the SW as demonstrated by testers, in Teslas legal filing to government agencies- in Teslas public statements- and in [B}the actual code itself[/B] says you are 1000% wrong.

It's a pretty weird hill to die on. And gets weirder the more you double down on it.

It seems like when Tesla decides to report that they are testing AVs has nothing to do with “how soon robotaxis”. Remember that they did actually report AV testing before they discovered this one weird trick to avoid it.

No, they didn't.

They've only reported it twice, ever.

Once when making the 2016 demo video, and once when making the autonomy demo video.

In neither case BTW were they running the code (or even the same general architecture) they released as FSDBeta.


Whenever Tesla begins to test (in CA) their actual new autonomous driving code-- which again FSDBeta is not-- they'll again need to report it.

I'd once again encourage you to read J3016, since you keep demonstrating a woeful ignorance of basically everything it says while continuing to try and discuss specific details of it..

 

[bradtem]

Whenever Tesla begins to test (in CA) their actual new autonomous driving code-- which again FSDBeta is not-- they'll again need to report it.

Many argue that Tesla is breaking the rules with this interpretation, and the DMV is considering making the same ruling on Tesla that it did on Uber, namely that "If you are trying to make a self-driving system, the fact that the prototype needs human supervision in no way excludes you from the regulations for testing AVs, you can't just call it ADAS."

They said that to Uber and told Uber to get in line or get the licence plates revoked on all their vehicles. Uber backed off. They have not had the guts to say that to Tesla but they might. "FSD beta" for all but a handful of registered cars would end.

 

[EVNow]

https://twitter.com/x/status/1508152648922402816

 

[Knightshade]

Many argue that Tesla is breaking the rules with this interpretation

Only those who don't understand the rules

, and the DMV is considering making the same ruling on Tesla that it did on Uber, namely that "If you are trying to make a self-driving system, the fact that the prototype needs human supervision in no way excludes you from the regulations for testing AVs, you can't just call it ADAS."

They said that to Uber and told Uber to get in line or get the licence plates revoked on all their vehicles. Uber backed off. They have not had the guts to say that to Tesla but they might. "FSD beta" for all but a handful of registered cars would end.

Except Ubers system was actually designed and intended to be >L2. FSDBeta is not.

There's actual, significant-to-the-rules, differences in functionality and capability.

Which is exactly why they did that to Uber, but not Tesla.

 

[Bladerskb]

Only those who don't understand the rules

Except Ubers system was actually designed and intended to be >L2. FSDBeta is not.

This is simply not true.

There's actual, significant-to-the-rules, differences in functionality and capability.

This is also simply not true.

Most of the time you are just making stuff up. You don't know the actual "design" of the Uber system. Nor its "functionality or capability".
You post as though you are god and get it wrong most of the time and keep double downing. its ridiculous.

 

[Knightshade]

This is simply not true.

This is also simply not true.

Most of the time you are just making stuff up. You don't know the actual "design" of the Uber system. Nor its "functionality or capability".
You post as though you are god and get it wrong most of the time and keep double downing. its ridiculous.

it's weird you insist I'm wrong, but then provide no evidence of it other than "nu uh"

Not surprising, just weird.

Meanwhile back in reality Uber themselves admitted their test cars were (at least) L3- with the car automatically doing driving, and the human there as a safety driver (ie performing the fallback safety task).

Which, again, is at least L3 (or L4 if it's the case Uber expected the car to be able to do fallback as well but wasn't' yet confident in it)

Tesla, on the other hand, only operates L2 systems where the human is always doing at least some part of the DDT, and thus needs no permit.

 

[diplomat33]

Interesting article on why it is taking so long to deploy self-driving cars:

It’s the last 20% that the AV industry is stuck on, especially the final 10%, which covers the devilish problem of “edge cases”. These are rare and unusual events that occur on the road such as a ball bouncing across the street followed by a running child; complicated roadworks that require the car to mount the curb to get past; a group of protesters wielding signs. Or that obstinate cow.

While humans are able to generalize from one scenario to the next, if a self-driving system appears to “master” a certain situation, it doesn’t necessarily mean it will be able to replicate this under slightly different circumstances. It’s a problem that so far has no answer. “It’s a challenge to try to give AI systems common sense, because we don’t even know how it works in ourselves,” says Mitchell.

The edge-case problem is compounded by AV technology that acts “supremely confidently” when it’s wrong, says Philip Koopman, associate professor of electrical and computer engineering at Carnegie Mellon University. “It’s really bad at knowing when it doesn’t know.”

“Safety isn’t about working right most of the time. Safety is all about the rare case where it doesn’t work properly,” says Koopman. “It has to work 99.999999999% of the time. AV companies are still working on the first few nines, with a bunch more nines to go. For every nine, it’s 10 times harder to achieve.”

How self-driving cars got stuck in the slow lane | Self-driving cars - Verve times

“I would be shocked if we do not achieve full self-driving safer than a human this year,” said Tesla chief executive, Elon Musk, in January. For anyone who follows Musk’s commentary, this might sound familiar. In 2020, he promised autonomous cars the same year, saying: “There are no fundamental...

vervetimes.com

 

[EVNow]

https://twitter.com/x/status/1508503622488928261

 

[Bladerskb]

Lol you just proved my point to @diplomat33. People were spotting Waymo minivans and vans in SF in scale in 2020 (just like I did personally). My point is the ramp-up in SF did not start up with the I-Paces in 2021, so leaving out the Pacificas is misleading. And they were testing in SF already even before then with the Pacificas in smaller scale (which should not be left out of the timeline).

PS: Mountain View is still in the Bay Area, read what I wrote. There is no evidence that they have done much of their testing in CA out of the Bay Area, so a bulk of their tests if not almost all of it is likely inside the Bay Area.

Are you still denying the clear obvious scale that is going on?

Taking our next step in the City by the Bay

This morning in San Francisco, a fully autonomous all-electric Jaguar I-PACE, with no human driver behind the wheel, picked up a Waymo engineer to get their morning coffee and go to work. Since sharing that we were ready to take the next step and begin testing fully autonomous operations in the...

blog.waymo.com

 

[stopcrazypp]

Are you still denying the clear obvious scale that is going on?

Taking our next step in the City by the Bay

This morning in San Francisco, a fully autonomous all-electric Jaguar I-PACE, with no human driver behind the wheel, picked up a Waymo engineer to get their morning coffee and go to work. Since sharing that we were ready to take the next step and begin testing fully autonomous operations in the...

blog.waymo.com

I don't see anything there that contradicts my point that they started testing with the Pacifica in SF, with people seeing multiple Pacificas daily in the links you posted yourself that came in 2020.

 

[EVNow]

But can FSD Beta handle this ?

https://twitter.com/x/status/1509253049440739332

 

[Terminator857]

Article states Tesla reasons for Toyota using camera only approach: low cost.

1. Toyota adopts Tesla's camera-only approach to self-driving development
2. Toyota joins Tesla in developing self-driving tech with low-cost cameras

Quote:
Toyota unit Woven Planet has joined Tesla in trying to advance self-driving technology without expensive sensors such as lidars.

Woven Planet told Reuters it is able to use low-cost cameras to collect data and effectively train its self-driving system, a "breakthrough" that it hopes will help drive down costs and scale up the technology.

Gathering diverse driving data using a massive fleet of cars is critical to developing a robust self-driving car system, but it is costly and not scalable to test autonomous vehicles with expensive sensors, it said.

 

[diplomat33]

Article states Tesla reasons for Toyota using camera only approach: low cost.

1. Toyota adopts Tesla's camera-only approach to self-driving development
2. Toyota joins Tesla in developing self-driving tech with low-cost cameras

Quote:
Toyota unit Woven Planet has joined Tesla in trying to advance self-driving technology without expensive sensors such as lidars.

Woven Planet told Reuters it is able to use low-cost cameras to collect data and effectively train its self-driving system, a "breakthrough" that it hopes will help drive down costs and scale up the technology.

Gathering diverse driving data using a massive fleet of cars is critical to developing a robust self-driving car system, but it is costly and not scalable to test autonomous vehicles with expensive sensors, it said.

Toyota still plans to include radar and lidar for their robotaxis and autonomous vehicles projects:

Toyota still plans to use sensors like LiDAR or other radar systems for robotaxis and autonomous vehicle projects, Benisch clarified. It is the “best, safest approach to developing robotaxis,” the report said.

 

[Bitdepth]

Article states Tesla reasons for Toyota using camera only approach: low cost.

1. Toyota adopts Tesla's camera-only approach to self-driving development
2. Toyota joins Tesla in developing self-driving tech with low-cost cameras

Quote:
Toyota unit Woven Planet has joined Tesla in trying to advance self-driving technology without expensive sensors such as lidars.

Woven Planet told Reuters it is able to use low-cost cameras to collect data and effectively train its self-driving system, a "breakthrough" that it hopes will help drive down costs and scale up the technology.

Gathering diverse driving data using a massive fleet of cars is critical to developing a robust self-driving car system, but it is costly and not scalable to test autonomous vehicles with expensive sensors, it said.

The title does not make sense. It is a similar strategy Mobileye uses. They use cameras on consumer cars to harvest data they use to train their NN and build their maps but they also maintain that to deploy a safe ADS they will use multiple sensor modalities (LiDar, Radar, Ultra Sonic, Camera) to ensure a reliable safe and redundant system until a time when using just camera is safe enough. That is very unlike Tesla.

 

[EVNow]

The title does not make sense. It is a similar strategy Mobileye uses.

GM too …

Ofcourse who will click on articles that mention MobilEye / GM instead of Tesla ?

 

[Jeff N]

At the Austin Gigafactory Cyber Rodeo grand opening party tonight, Elon confirmed rumors that Tesla will build its own dedicated robotaxi vehicle after starting production on the Cybertruck, Semi, and Roadster 2.0 in 2023.

 

[diplomat33]

At the Austin Gigafactory Cyber Rodeo grand opening party tonight, Elon confirmed rumors that Tesla will build its own dedicated robotaxi vehicle after starting production on the Cybertruck, Semi, and Roadster 2.0 in 2023.

Not surprising. Building a dedicated robotaxi vehicle is a logical next step. But it shows that Tesla is behind the curve. Cruise and Zoox already have a dedicated robotaxi vehicle in pre-production. By the time Tesla starts production on their robotaxi, Cruise and Zoox will likely already have their dedicated robotaxi deployed to the public.

 

[Knightshade]

Not surprising. Building a dedicated robotaxi vehicle is a logical next step. But it shows that Tesla is behind the curve. Cruise and Zoox already have a dedicated robotaxi vehicle in pre-production. By the time Tesla starts production on their robotaxi, Cruise and Zoox will likely already have their dedicated robotaxi deployed to the public.

... they're behind on announcing it.

Everything else is weird assumptions. Even weirder if you actually watched how advanced a factory GT is compared to whomever is gonna make cars for the other companies you mention that don't own car factories.

 

[diplomat33]

Everything else is weird assumptions. Even weirder if you actually watched how advanced a factory GT is compared to whomever is gonna make cars for the other companies you mention that don't own car factories.

Who cares how advanced the manufacturing plant is? What matters is how capable the robotaxi is and how quickly you can deploy it safely to the public in a ride-hailing network. Sure, GT can probably mass produce 10M of these new Tesla robotaxi vehicles but if the Tesla robotaxi is not safe enough or if there is no ride-hailing network, what does it matter?

And the other companies have deals with the companies that do own car factories, who produce the robotaxis for them. And they are quite good at manufacturing vehicles. Robotaxi companies don't need to manufacture their own robotaxis, they can just have someone else manufacture the robotaxis for them.

 

[Knightshade]

Who cares how advanced the manufacturing plant is?

Anybody wanting to build vehicles in volume, at a profit.

I know Waymo is REALLY GOOD at losing money, but I didn't think they wanted to KEEP doing that?

What matters is how capable the robotaxi is and how quickly you can deploy it safely to the public in a ride-hailing network.

If you lose money on every one, and can't even make them very quickly, you just keep burning cash until you run out of it.

And the other companies have deals with the companies that do own car factories, who produce the robotaxis for them. And they are quite good at manufacturing vehicles. Robotaxi companies don't need to manufacture their own robotaxis, they can just have someone else manufacture the robotaxis for them.

Sure. Like GM that made 26 total EVs last quarter, and only a few hundred in Q1?

Using a third party to make your car means you lose $ to the middle-man of course, making a profitable business harder. Likewise if they can't actually scale production you have problems as well.


Not sure why this would be confusing to anyone. MULTIPLE elements are needed for a successful business here- not just "our tech works"