You say this corrola is 5-10 years ahead of everyone else. Well simple question, when will this corrola do what every one is already doing.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
When Tesla will catch up with Waymo and Cruise at delivering L4?
- Thread starter t3sl4drvr
- Start date
TresLA
Member
See the multiple other threads complaining about these SAE levels and how they force a linear model that doesn’t fit all AV approaches.
Tesla is at L2’ish and seems like they’re going a different route that wouldn’t ever be L4. They’re just going for L5. L3 may even be “skipped”. Again, since these levels aren’t necessarily progressive/linear, it’s apples and oranges. Like previous commenter said, Waymo has never produced an L2 product but that’s because it doesn’t fit their AV approach, not because they haven’t “caught up” to Tesla. Likewise, Tesla probably won’t ever produce an L4 product, because it doesn’t make any sense for Tesla’s approach. Both companies may produce an L5 product (not addressing timeline here) because that does fit their business models.
Last edited:
TresLA
Member
You’re not listening and therefore not understanding. No contradictions here because calling people “Tesla fans” is just an irrational label to dismiss someone that doesn’t agree with you and/or to ignore logical counterpoints.
Cases in point: Tesla probably never attempted an L4 product, but is trying for L5. These SAE levels just don’t fit the linear progression of many AV companies. Waymo never produced and sold an L2 service/product, but that makes sense for their AV approach.
you are changing the topic. Conflating it into other things that it’s not. This has nothing to do with SAE levels.
It’s quite simple. When will Tesla have a car that can drive people around in city streets without a driver?
TresLA
Member
It’s not changing the subject. It’s pointing out the flawed logic/assumption of the question. I’m not sure how to make this more clear than others and I have repeated ad nauseam: Tesla seems to be going for all cities, not geo-bound/select cities.
Perhaps you’d understand if you think about the question: “It’s quite simple. When will Waymo or Cruise have a car that can drive people around in any city’s streets without a driver?”
I am not that lucky "one"!
And even if I fly to San Francisco tomorrow, I doubt I will be able to hail a Robotaxi upon arrival. As I understand (as stated in the article you linked to) only "small number" of riders have access to the service.Ship to who??? It appears to me they are still in testing phase, not ready to "ship" to any third-party operators and certainly not to individual owners.
I have purchased the FSD with full understanding that it is Level 2 and will remain at Level 2 for awhile. Level 4 was never explicitly promised to me. FSD is defined at Tesla's website as an explicit set of Autopilot features, most of which I already have, some are still in Beta and are improving. The SAE Autonomy Level classification is not so much about the capabilities of the car, but more about driver responsibility. Today my car can take me from home all the way to work with zero disengagements, but technically it is still Level 2, because I am still responsible for everything as a driver.
While Robotaxi is the eventual goal for Tesla, starting Robotaxi service Beta-testing in limited areas is not Tesla's initial priority. Tesla is working on a much bigger goal: mass-producing fully self-driving cars and developing universal (not geofenced) autonomy, essentially shooting for Level-5 autonomy in a mass-produced car. In that regard, I think they are way ahead of Cruise, Waymo or anyone else, even though technically they are still at SAE Level 2.
Now, while I disagree with some of your premises, I'll try to answer the question of your poll, the way I understand it. I think what you are really asking is this:
When will Tesla begin the first limited public testing of its Robotaxi service?
I think it will happen next year. In 2023. There is some slight chance it could happen before the end of this year, and there is also a possibility it could slip to 2024. But 2023 looks most likely to me, based on my personal experience with FSD Beta and its progress.
Last edited:
So which is it? You people want to have your cake and eat it too. Pick one!
With robotaxi businesss model (where one cannot just go and buy a car), only sure way to say if the car works in a particular location is to have the robotaxi company choose to operate there. To answer you question, no robotaxi operator will ever choose to operate in every location ("any city").
(Side note: Tesla FSD will never operate in every location either. While today's reasons include limited technology, surely legal reasons will remain for the foreseeable future that would block one from doing a around the globe self-driving trip. Interpreting SAE definitions strictly, this would mean that Tesla would "never" exceed L4)
What we do not know is, how well Waymo and Cruise today would operate in a particular city if they would choose to operate there. Assumption that their system could not drive at all in a location without pre-made maps is naive. To my understanding, their system takes in input from multiple sources - including cameras, maps, lidar, ... and based on available input make driving decisions. Lets say that two of the inputs - map and lidar - would be blocked in a particular situation: can the system still operate better than Tesla? Unfortunately we dot have answers to that question due to not being able to test it out.
Facts are that Tesla has chosen to have fewer inputs (no lidar, no HD maps), has developed the system for a shorter amount of time (vs Waymo who started in 2009), and has limited themselves to a fixed computing power (FSD/HW3 being in so many cars that it would be too expensive to replace).
With these constraints, they still might end up developing amazing L4 someday. But odds of beating the competition in making L4 widely available are not getting better due to competition lacking the same handicap.
With these constraints, they still might end up developing amazing L4 someday. But odds of beating the competition in making L4 widely available are not getting better due to competition lacking the same handicap.
Last edited:
TresLA
Member
Okay, this right here is exactly our point: it’s apples and oranges. My flipping the same question about Tesla around doesn’t make sense for Waymo/Cruise model to their AV approach. But it is Tesla’s (high bar, lofty goal, who knows when yadda yadda disclaimer). Tesla’s goal IS to operate robotaxi in every city (and suburb and rural and highways in between). Unlike Waymo/Cruise, Tesla’s approach to AV has never picked a city to operate in, because it doesn’t make sense to their goal. From what I understand, the messaging is in short, “if we can solve for one city, then we can solve for all. We will only be unnecessarily prolonging our goal if we spend time on a tangent like the local maxima of specific cities/towns.”
One disclaimer before someone responds with “but semantics/technicalities!”: Tesla’s goal seems to be all cities in the world, of course, but most technical hurdles will be sufficiently solved for if all cities in the USA are solved for. So an “intermediate” step would be L5 in the USA first (regulatory hurdles, varying edge cases covered, etc.) Most would probably consider an autonomous vehicle that’s only autonomous in the USA to be L5, and not L4 just because “that feature is only geo bound to the US!”
The biggest mistake Musk did was to drop Mobileye (out of Israel) and attempt to do it with his own engineers. My old 2014 Model S with Mobileye software still IMHO provides a better autopilot experience than my X, Y or 3 did. And it didn't cost an extra penny.
Cybermoose
Member
All I know is for things like robotaxi or similar concepts, other companies might be ahead and that's great but it's not what i'm looking for. When I think about how great FSD could be when it's done, I think about being able to go anywhere on public roads and switch it on. I want to be able to do it in the city, country side, highways, and I want it to be available everywhere.
The infrastructure that is required for these other companies is not realistic to see this everywhere for a very very long time. I haven't looked into it but I'd guess that they are a very long time away from ever profiting from this technology in it's current form. I'd be surprised if these self driving rides are actually making money with the support network they have in place.
Tesla went in the opposite direction and even took out radar. They want to improve their system to a point that if a person can do a drive with vision, so can a Tesla. I like this approach because it doesn't restrict where it can be used.
If Tesla had gone the route of other companies and built a network in a few cities that completely supported the automated driving, then you could make these comparisons, but then Tesla would be very far behind where they currently are with their goals.
The infrastructure that is required for these other companies is not realistic to see this everywhere for a very very long time. I haven't looked into it but I'd guess that they are a very long time away from ever profiting from this technology in it's current form. I'd be surprised if these self driving rides are actually making money with the support network they have in place.
Tesla went in the opposite direction and even took out radar. They want to improve their system to a point that if a person can do a drive with vision, so can a Tesla. I like this approach because it doesn't restrict where it can be used.
If Tesla had gone the route of other companies and built a network in a few cities that completely supported the automated driving, then you could make these comparisons, but then Tesla would be very far behind where they currently are with their goals.
Certainly possible and this statement doesn't say anything about robotaxi although many will jump to that conclusion. They are different. Since this includes highway driving that makes achieving "FSD safer than a human" somewhat reasonable even including city/residential streets in the statistics. Elon is being very conservative here. Afterall they are likely already tracking statistics against this conservative goal and already have a good idea where they stand.
Similar threads
- Article
- Article
