-
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
willow_hiller
Well-Known Member
7 down, only 27,474 left to go (19,495 incorporated cities, towns, and villages in the United States and over 7,986 cities and towns in Canada).
For Tesla 0 down, only 27,481 left to go (19,495 incorporated cities, towns, and villages in the United States and over 7,986 cities and towns in Canada).
willow_hiller
Well-Known Member
For driverless autonomy, sure. But for useful automation? Nope. I'll take something that improves my life now over a hypothetical robotaxi that may not include my area in their geofence until 2030.
diplomat33
Average guy who loves autonomous vehicles
To be fair, Cruise only has a meaningful presence in 2-3 cities (SF, Phoenix and maybe Austin), the other cities are new testing or newly announced and don't have a public ride-hailing service yet. Cruise's scaling strategy seems to be to deploy small in a bunch of cities first and then build up the ride-hailing service later. Cruise has some big issues like frequent unexpected stops and they need to expand their ODD to highways. We will see how well they are able to scale.
Cruise and Waymo decided to do driverless first and then scale while Tesla decided to scale first and then try to get to driverless. We will see who gets to driverless "everywhere" first. I like that Waymo and Cruise have actual driverless that is starting to scale. Frankly, "V12 end-to-end" feels like Elon's latest Hail Mary attempt, another moonshot that he hopes will "solve FSD". If it works and Tesla is able to go driverless everywhere, they will obviously win. But the chances of it working like that, especially in the short term, are vanishingly small IMO.
I think it is more likely that Tesla will not achieve driverless but will eventually achieve passable self-driving everywhere that still requires driver supervision. And that might be ok. I've said before that perhaps Tesla does not need to do driverless, that just doing self-driving everywhere with driver supervision would be good enough. If Tesla gets to say 1,000 miles per intervention, that might be good enough for most Tesla owners. Meanwhile, Cruise and Waymo will press on, scaling driverless to more and more places. In any case, I am very excited for the future of autonomous driving.
Last edited:
What does that even mean? All progress toward automation and autonomy is useful. You might not experience it in your lifetime (highly unlikely) but it is progress for future generations. Just because you have not personally experienced it doesn't mean it is not useful. It is not a hypothetical robotaxi, it is an actual service being used today in various locations.
willow_hiller
Well-Known Member
diplomat33
Average guy who loves autonomous vehicles
Robotaxis are not hypothetical. They exist now in multiple cities. I think it is mistake to treat a tech that is not personally useful to YOU yet as if it does not exist. That is a rather selfish view. A tech can still be useful to a lot of people, even if it is not useful to YOU yet. The fact is that there are plenty of people in Phoenix and SF who say that Waymo and Cruise robotaxis have changed their life. Robotaxis exist and are already very useful to plenty of people.
And by the way, usefulness is relative. For you, Tesla FSD is useful despite not being driverless. But I have a friend who has a physical disability and cannot drive. For them, Tesla FSD is not useful since they can't use it. The system has to be driverless to be useful to them. So only robotaxis are useful to them. So usefulness will be different for different people.
But this is why trying to debate who is better or who is winning is silly. And that is why the SAE says that the levels of autonomy are not a ranking from worst to best. Tesla's AP or FSD is still L2, but it can still be very useful and improve people's driving experience, like you. And that is great. L4 can also be useful as well as we see from testimonies of people who say Waymo and Cruise have changed their life. How the tech impacts a person may be different of course, but no less meaningful to that person.
Last edited:
What part of "All progress toward automation and autonomy is useful", makes you think I don't believe FSD is useful? I've always been a fan of all automation systems from regular old Cruise control to Lane centering control that prevent people from driving off the road, Traffic aware CC to ADAS systems like Supercruise, FSD, Bluecruise, Pro Pilot etc.
There is this rather inane argument, which I know where it comes from, that you either have autonomy everywhere at the same time or it is not useful or scalable to take a methodical one city at a time approach to validation and safety. I haven't personally experienced it, so it is hypothetical, and it doesn't exist.
Yeah, by same standard (just having testing) Waymo has over 10 (San Francisco, Daly City, Mountain View, Sunnyvale, Los Altos, Palo Alto, Chandler, Phoenix, LA, New York, Seattle, Novi; probably more, these are just the ones media reported from a quick search).
Tesla is aiming for scale door-to-door L2 first and then transitioning from that to higher levels (although it may be a dead end approach). To the lay person completely unaware of nor cares about SAE levels, it looks sufficiently close in functionality (car does all the same actions as a L4 car with safety driver; with lines further blurred if hands free attention detection is rolled out in larger scale). There's plenty of people in less popular cities where the current L4 efforts seem like pie in the sky, given due to the commercial approach, it may not reach them until 10-20 years later, if ever.
While there is some lip service paid by those companies to releasing a general consumer version in the future, it's pretty clear they have their hands pretty full just with the robotaxi operation and the companies don't want to devote resources that might distract from that (as evidence by the ending of the trucking division at Waymo for example, even though it would seem to have a lot of commercial potential).
diplomat33
Average guy who loves autonomous vehicles
Which is why I see value in both approaches. Robotaxis can offer a safe chauffeur experience to people who live in cities. This can add great convenience since they don't need to worry about driving at all and can ride in the back seat and relax. Robotaxis can be especially valuable to people who cannot drive due to age, disability etc... I want to see robotaxis scale so that more people have access to that convenience and safety in cities.
But I also appreciate that for people who live in areas that may never get robotaxis, having L2 door to door, can also offer great value. Sure, they can't ride in the back seat but L2 door to door can still offer a similar experience if it is able to drive to their destination with no intervention. A good L2 door to door can make driving more relaxing if the car is able to safely do the driving tasks while the person just pays attention to the road. And it will work in areas that they need to go, where perhaps the robotaxis are not available yet. As someone who lives in Indiana and could be one of those people who never see a robotaxi in my area for 10-20 years, I definitely appreciate the value of a good L2 door to door. In fact, it's the whole reason why I purchased a Tesla with FSD. I am actually not a FSD beta hater. I just don't believe it will be driverless. But I really want FSD beta to be a good hands-free L2 door to door someday.
And eventually, in the future, I hope we also get "eyes off" systems on consumer cars too so that we can use our cars to go anywhere and have the added convenience of not having to pay attention (dependent on the "eyes off" ODD of course).
It might be wishful thinking on my part but I am still hoping that robotaxis are really just an intermediary stage on the path to consumer cars. Robotaxis do allow companies like Waymo to deploy their autonomous driving and improve it. And as robotaxis scale, the cost will need to come down which will also help make the hardware more affordable for personal cars. So I am hoping that once the robotaxis are safe and cheap enough in a wide enough ODD that the tech can be adapted relatively easily to consumer cars. So it won't be a matter of dividing resources, more like using robotaxis to develop the tech for consumer cars. And, I bet if carmakers said they were interested in licensing the Waymo Driver, and billions of dollars of revenue were on the table, Alphabet would tell Waymo to devote the resources to making it happen. My point being that Waymo will devote the resources if they see sufficient commercial potential.
Two lines in the recent Waymo announcement give me hope that I might be right and Waymo will eventually apply the tech they developed in robotaxis to consumer cars. Obviously, it is years away but it could happen.
1) Waymo says that they are building a generalizable driver and one of the applications will be personal car ownership. The fact that they mention personal car ownership means it must at least be a possibility on their minds.
2) They mention testing use cases that could be helpful to "partners" in the long run, making me think that they are using robotaxis to test for the use case of personal car ownership. Who are these partners in the long run they mention? Could it be carmakers?
Last edited:
willow_hiller
Well-Known Member
Take a quick step back from the polarized "us versus them" mentality:
On a scale from 0 to 100, where 0 represents a car driven entirely manually by a human, and 100 represents a car that can drive without a human in any place in any condition, do you honestly think FSD Beta represents a 0?
Because if so, then I don't think we have a shared basis of reality, which makes having a conversation almost impossible.
diplomat33
Average guy who loves autonomous vehicles
Brad Templeton and Prof. Raj Rajkumar debate the question: will self-driving cars go mainstream within 10 years?
0:00 Introductions
05:00 We start with opening statements and rebuttal on the core question, then do 3 quicker rounds on:
30:47 How Viable Tesla's approach is
40:33 Making Consumer Cars
50:45 Commercial Applications
0:00 Introductions
05:00 We start with opening statements and rebuttal on the core question, then do 3 quicker rounds on:
30:47 How Viable Tesla's approach is
40:33 Making Consumer Cars
50:45 Commercial Applications
New Model 3 owner here and this is where I'm stuck. Elon continues to say $15k FSD is a bargain in context to the robotaxi business case, but I never see myself ever doing that. Especially considering my car has HW3 and probably will never get there.
For me to get L2 door to door means what, that I pony up the $6k for EAP, at least for highway driving? And that assumes the FSD code base makes its way back into that product. I guess I'll be relegated to subscription-land for the life of this vehicle.
Knightshade
Well-Known Member
No, you need to pony up 15k for FSD.
EAP only does useful stuff on divided highways and parking lots.
diplomat33
Average guy who loves autonomous vehicles
The stuff about the cars increasing in value when they become robotaxis is BS from Elon to justify the $15k price tag. HW3 will not do robotaxis. "FSD" will only be L2 door to door. So the $15k only pays for L2 door to door.
KArnold
Active Member
Although, even if version X can do hands off for freeways with HW3 - I'd think my legacy MS without cam will appreciate in value. But that's a very big IF.
That's what I was worried about. I bought this car because I actually like driving...sometimes fast. But a L2 autopilot capability that can switch lanes on long road trips is definitely something I'd consider assuming it doesn't break the bank. It sounds like EAP is that product. Maybe I'll wait for the "mind blowing" FSD v12 announcement to see if that code base will somehow makes its way back to EAP since I don't want to be stuck on a legacy code base for the life of the car.
Last edited:
voyager
Active Member
Similar threads
- Article
