diplomat33
Average guy who loves autonomous vehicles
Tesla is just missing one really important ingredient: they don't have any robotaxis. Tesla does not have reliable autonomous driving.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Waymo vs FSD 11.3.6
- Thread starter Porkalletta
- Start date
Tesla is just missing one really important ingredient: they don't have any robotaxis. Tesla does not have reliable autonomous driving.
Maybe you are not understanding me. IF the task is easy for humans and super difficult for computers, then you better figure out a way to overcome the difficulty cheaply.
There are many things a computer can do with easy like calculating complex math that humans cannot, so the cost per calculation is MUCH lower than the cost per calculation of a human. A human can take hours to do one calculation a computer can do within seconds.
However a robotaxi is flipped. It takes much more effort for a computer to drive than for a human, so you have to figure out a way to offset the cost of the human when you are ridiculously behind. Waymo/Cruise I don't see will get there, so that's why robotaxies are a money losing business because the competition(humans) is too cheap.
The point is Tesla is taking the long grindy way to get there because they have to figure out a way for a computer to offset the cost of a human in the task of self driving. SO yeah they don't have a robotaxi because they know forcing one using a bunch of complicated system with operating cost attached will just lose you money like Waymo.
diplomat33
Average guy who loves autonomous vehicles
No, I get where you are coming from. That is why you like the Tesla approach so much. You feel like the Tesla approach is lowering that cost to solving the difficult problem of autonomous driving. Certainly, if Tesla is able to solve autonomous driving with their low cost approach, it will be a huge win and a game changer. The big problem for Tesla is that the low cost approach makes solving autonomous driving even more difficult. It will likely take Tesla far longer to solve autonomous driving with this approach. It is also very possible that the current approach won't work and Tesla will be forced to add more sensors and raise the cost in order to solve autonomous driving. What we are seeing with Waymo is that the cost is higher, but it has enabled them to make faster progress in solving autonomous driving. Waymo's challenge will be to lower costs.
It is basically a race between starting cheap but taking longer to solve autonomous driving (Tesla) or solving autonomous driving faster but then having to lower costs later (Waymo).
Honestly if Tesla's approach doesn't work then robotaxies will need to be abandoned. This is a SUPER commoditized market. Not only can most people drive with ease for pennies on the dollar, the other competition consist of public transport and human transport. This is why I am in the camp of Tesla or bust. Bust is not a 0% probability. There are many innovations people created to replace humans which ends up falling on their faces. Musk unwilling to abandon their way of doing this for the same reason. It's not like they couldn't geo fence/HD map and start opening up a service. There's really nothing magical about it as many companies have done it. However making it profitable is pretty much impossible.
Not so sure about that. We know people pay for rides (e.g. taxi cabs / Uber / Lyft / etc). The drivers of those cars make a salary, and can only drive safely for a limited number of hours per day before needing rest.
Like all software, you can build it once and sell it infinitely. With autonomous driving, the first company to have fully working software (and hardware) can launch across the globe and replace all of those driver salaries.
They can also push the cost of driving services down to the point where less people would buy cars and instead woyld just take a cheaper robotaxis. They can compete with other public transportation.
The only reason these companies are willing to spend a BILLION dollars a year to achieve autonomy is because they expect big profits in the long term. Waymo says they're cost per mile is around 30 cents (and this will likely come down over time), while Uber and Lyft's costs are around $2-3 per mile (and this is likely to go UP).
Well if you buy a car for 40k and drive 150k miles, then you are at 3.75 miles per dollar+energy/gas. So roughly it cost a typical person with a car about 2 miles per dollar. So this is your biggest market and your biggest competitor. If your commute is 40 miles/day, you are looking at 20 dollars/day for your robotaxi to be equivalent in cost. Currently waymo is charging 43 dollars for 22 miles, or about 80 dolllars/day if you were to use the service.
Waymo is currently losing 2B a year charging 4x what it cost for a typical person to drive themselves. Yes you get to read a book, but is it worth first waiting 10-20 mins for your ride to come to your house and then pay 4x more?
This is why a robotaxi business is kind of DOA until
1. The scale for robotaxi is so big that it only takes 3-4 mins for one to arrive at your house, not 10-30.
2. It is so cheap that you are willing to abandon your car
3. It is so cheap that you are willing to abandon public transportation
Uber and Lyft are still losing money and THEY DON"T EVEN NEED TO PROVIDE THE CAR, do any geo fencing, have any engineers test driving cars and expanding areas of use. You see how far Waymo is from profitability? They are charging 4x what the majority of people are willing to pay to give up their car and they need to provide all the hardware, software, and operators. Whoever thought this can one day make money should buy a bridge that is for sell.
Tesla or bust. This needs to be a generalized "one button" activation of millions of robotaxi that scales to infinitely or DOA.
I am not very bullish on the idea of robotaxi being this infinite money printer. We only see Taxis making money when they charge a stupid amount of money helping people in need for a pinch. When the price became commoditized by uber and Lyft, we see how little money these companies are actually making. Robotaxi is suppose to commoditize this cost even further! So imo Tesla is the only company with all the right ingredients necessary that can give this a chance at profitability while all the other companies has laughable business models.
Success and downfall of robotaxies does not depend on the technology but depends on the business model. It is suppose to replace an already cheap transportation system profitability. This is extremely hard to achieve.
Last edited:
EVNow
Well-Known Member
LOL. Nothing can be farther from truth.
Not only is the software maintenance free - something that works in US won't work in China or India or even Europe, for sure.
Not to mention - there is zero chance of any government allowing a monopoly to replace all taxis all over the world.
ps : I expect Chinese AV companies to do well in China and Waymo will never even enter China. There is zero chance of AVs being allowed (even if they work) in India.
pps : No wonder there is so much stupid money going into these AVs. The "libertarian" investors must all be thinking like you.
So true. As you can’t Buy a Waymo I see no reason to compare it to FSD. Add to that if any solution requires an R2D2 full gear mount like that it will Never be accepted in the market. If you want a fenced in slot car ride in a predefined micro area Waymo is your answer.
sleepydoc
Well-Known Member
No! We’re in
EVNow
Well-Known Member
It's not surprising to find team Tesla at TMC. But it is strange to find Team Waymo ... Esp. considering it doesn't even help anyone in anyway even if it ever gets to scale up.
jeewee3000
Active Member
It's true Waymo will not stay in its current form. It is a cool technology and it's impressive to see them grow.
However, in all your pro-Waymo (and anti-Tesla FSD) posts you refuse to give Tesla the same credit. Tesla FSD won't stay in its current form either. What we are seeing now is NOT the final state of Tesla's FSD.
I'm excited about both technologies, and I see problems in both.
Where most Tesla FSD critics agree upon is that Tesla has inferior sensory input than Waymo. This comes down to the discussion of which sensors are needed/optimal (i.e. cameras only versus LIDAR/radar/USS/cameras/... combined) and the location or placement of said sensors (many argue Tesla HW3 cameras are not optimally positioned, Waymo sensors are basically everywhere).
I'm not sure about Tesla camera placement, but I am a 'believer' in camera only being sufficient. The only question is how much compute power is needed to transform camera images into proper steering/controls output. We might be 80% of the way, or we might only be at 0,5%. (Since required compute can scale exponentially) It's anyones guess really at this point.
Waymo has better sensory input, but where I strongly disagree with you @diplomat33 is that you seem to see no issues in scaling the current Waymo approach of mapping all locations where Waymo is active.
The Waymo website states:
Mapping out every intersection, sign, and signal
Before our Waymo Driver begins operating in a new area, we first map the territory with incredible detail, from lane markers to stop signs to curbs and crosswalks. Then, instead of relying solely on external data such as GPS which can lose signal strength, the Waymo Driver uses these highly detailed custom maps, matched with real-time sensor data, to determine its exact road location at all times.Given the pace of mapping Waymo has done until now, I see no signs of accelerating growth.
Is it possible for Waymo to eventually be active in all major (dense) cities in the USA? Sure. That way they can capture a high % of the population/customer base without the need for mapping rural USA.
Will Waymo be able to map the entire USA, North America or the World? Extremely unlikely. More unlikely than Tesla taking liability for HW3 FSD with Tesla Insurance.
Tesla may need to upgrade to HW4 or 5 (with better cameras in different locations on the vehicle) to achieve L5, but when they do, Tesla's approach can scale worldwide very quickly. (Yes, per country it needs to add some code regarding local road regulations/traffic signs, but this is trivial compared to achieving L5 in the USA)
Waymo however will never be able to scale worldwide. So even though Waymo brings autonomous ride hailing sooner than Tesla (it does, no discussion), it is a different technology with different goals.
But when in +100 years all transport is autonomous and kids growing up don't know what companies "raced" to L5 autonomy in the early 2000's, the tech in those autonomous vehicles will IMO be a lot more similar to Tesla's general approach than Waymo's mapping approach.
I know you don't agree with this, but you have to admit Waymo cannot scale worldwide. Unless a breakthrough in mapping tech happens (for example the Waymo cars can map on the fly in a location they have never been in, such as Tokyo or Moscow) but this is not expected anytime soon.
diplomat33
Average guy who loves autonomous vehicles
I do give Tesla credit for their progress. FSD Beta has made huge progress since the first version was released. FSD beta 11 is radically superior to the first FSD Beta 1. And of course, we are not seeing the final state of Tesla's FSD. It will continue to get better over time. Will the final state of FSD be L4 or L5 and if so when is the question?
Sufficient for what? Vision-only is certainly sufficient for doing autonomous driving. Nobody argues with that. I certainly don't. The question is whether vision-only is sufficient for removing driver supervision all the time. According to Mobileye, carmakers won't remove driver supervision unless it can be proven to be 10x safer than humans. So far, nobody has proven that with vision-only. That is why Mobileye's approach is to do "eyes on" FSD with vision-only, like Tesla, but to add radar-lidar to make the same FSD "eyes off" (ie no driver supervision). The advantage of radar and lidar is that it adds redundancy in cases where vision might fail.
So my position is very simple: Vision-only is perfectly sufficient for "eyes on" FSD but it is not sufficient for "eyes off" FSD.
And autonomous driving is a software problem, not a hardware problem. We have all the necessary hardware now to do autonomous driving. Computing power is not the issue. Software is the limiting factor. We need autonomous driving to be more intelligent and be able to handle more edge cases. So we need to solve the software.
Thats's a strawman argument. No, Waymo will not scale worldwide. Of course, I admit that. But nobody is going to scale L5 worldwide. That is completely unrealistic. That is not going to happen. Just think of all the differences in road infrastructure, traffic laws in different countries. You are not going to get FSD that works with no driver supervision, in all conditions, in the entire world.
And I think you are overestimating Tesla's ability to scale L5 worldwide. Tesla can certainly scale FSD with driver supervision worldwide, sure. But removing driver supervision worldwide is not going to happen. Even scaling FSD with no driver supervision in the US is unrealistic. There are 4M roads in the US and 15.8M intersections. And then factor in all the different permutations like rain, snow, bad lane markings, construction zones, truck blocking the view, etc... Scaling FSD that can handle all 15.8M intersections and 4M roads, in all conditions, with no driver supervision ever, is insanely hard. If you think Tesla is close to removing driver supervision in the entire US, you are deceiving yourself.
And Waymo does not need to map the entire world to do useful autonomous driving. Waymo just needs to map the cities they want to deploy in. And Waymo is focused on a robotaxi service. It does not need to scale everywhere to be useful. So when I say mapping is not the issue, I am talking about mapping the relevant cities Waymo wants to scale to. And Mobileye has already mapped every road in the US. So no, I don't see mapping as a major issue. Waymo could certainly map the entire US if they wanted to but they don't need to. Certainly, mapping is a lesser issue than solving behavior prediction or planning.
Last edited:
jeewee3000
Active Member
@diplomat33 , thanks for the thorough reply. Now I can contextualize your opinions better.
We're on the same page on many things. Waymo is indeed trying to be a robotaxi service in the most dense of places, where the total adressable market is greatest compared to the area they have to map. And they are making excellent strides in doing so.
In my book that however is not "solving L5", which is Tesla's goal. That goal is further away, and a lot more 'sci-fi'/utopian.
Where you state : "If you think Tesla is close to removing driver supervision in the entire US, you are deceiving yourself." I'd like to point out I in fact do not think Tesla is close to removing driver supervision, not in the US, not anywhere.
Where we still disagree is that L5 worldwide is possible or not. You say nobody is able/will be able to scale worldwide. I think that is a gross underestimation of the advances in AI. We are only at the very beginning of the AI revolution.
Will it take longer than most expect? Of course. But not insanely long.
Personal feeling: maybe within the next 5 years (but low likelihood is my hunch), very likely within 10 years, and I'd be shocked if L5 isn't achieved in 20 years. (Not necessarily by Tesla).
I am in the camp of "L5 will only be unique to one company for a very limited amount of time". Once one company shows they've done it, the door is open for others to follow suit. The second-movers will have a lot less legal hurdles than the first mover, and can start developing based on what is known about the tech of the first-mover (with the state-of-the-art hardware and software that the first-mover didn't have at the time of its development process).
That's why I think ARK's insane valuation of Tesla due to FSD ride-hailing is kind of unrealistic. Their argument however is that, even though L5 is achieved by multiple companies, the first 5-10 years there won't be enough robotaxis to fulfill demand for cheap transport, thus any player with L5 can profit.
Either way, we are living in exciting times.
We're on the same page on many things. Waymo is indeed trying to be a robotaxi service in the most dense of places, where the total adressable market is greatest compared to the area they have to map. And they are making excellent strides in doing so.
In my book that however is not "solving L5", which is Tesla's goal. That goal is further away, and a lot more 'sci-fi'/utopian.
Where you state : "If you think Tesla is close to removing driver supervision in the entire US, you are deceiving yourself." I'd like to point out I in fact do not think Tesla is close to removing driver supervision, not in the US, not anywhere.
Where we still disagree is that L5 worldwide is possible or not. You say nobody is able/will be able to scale worldwide. I think that is a gross underestimation of the advances in AI. We are only at the very beginning of the AI revolution.
Will it take longer than most expect? Of course. But not insanely long.
Personal feeling: maybe within the next 5 years (but low likelihood is my hunch), very likely within 10 years, and I'd be shocked if L5 isn't achieved in 20 years. (Not necessarily by Tesla).
I am in the camp of "L5 will only be unique to one company for a very limited amount of time". Once one company shows they've done it, the door is open for others to follow suit. The second-movers will have a lot less legal hurdles than the first mover, and can start developing based on what is known about the tech of the first-mover (with the state-of-the-art hardware and software that the first-mover didn't have at the time of its development process).
That's why I think ARK's insane valuation of Tesla due to FSD ride-hailing is kind of unrealistic. Their argument however is that, even though L5 is achieved by multiple companies, the first 5-10 years there won't be enough robotaxis to fulfill demand for cheap transport, thus any player with L5 can profit.
Either way, we are living in exciting times.
diplomat33
Average guy who loves autonomous vehicles
Thanks for the equally respectful and thoughtful reply.
The SAE essentially defines L5 as equivalent to a typical human driver. They may not always be good drivers but they can drive anywhere, anytime, as long as the roads are drivable of course. That is why the SAE specifically says that L5 only needs to drive when a typical human driver could drive. So for example, L5 is not expected to drive in a white out blizzard. L5 is certainly a laudable goal. I definitely see the appeal of L5, of autonomous driving that is like a human driver in that it could drive anywhere, anytime as long as the roads are drivable. And I agree that eventually AI will be good enough for L5. But here is why I don't think L5 is really needed. I see L5 as more of a long term end goal.
But note that the SAE does not say anything about how good the L5 has to be. The big challenge is not just developing autonomous driving that is capable of L5 but making it safe and reliable. And of course, to be useful, it needs to be safer than human drivers. Most carmakers want their autonomous driving to be at least 10x safer than humans. So there is a lot of testing and validation before you can deploy autonomous driving. Testing and validating L5 is exponentially harder than L4 because you have to test and validate everywhere and anytime. It is much easier to test and validate L4 since you only need to make sure it is safer than humans in a specific ODD. That is another reason why I see L4 as a better path to L5. You can start with L4 and expand the ODD over time until it is essentially equivalent to L5. I think that will be easier than trying to test and validate L5 directly.
This is also why I say "L5 worldwide" is not going to happen. Even if we get generalized AI that is capable of driving anywhere in the world, there needs to be a lot of testing and validation before companies would deploy it unsupervised. And testing and validating that autonomous driving is safer than humans on literally every road, every conditions, in the entire world would be insane. And carmakers won't deploy their cars in literally every single country. Furthermore, the SAE defines as L5 as per country or per region. So I think it makes sense to talk about L5 for a country. Carmakers will eventually deploy L5 in specific countries. Over time, more and more countries will get L5. But I don't know if literally every single country will get L5 to be truly "L5 worldwide". But perhaps you would consider it close enough to "L5 worldwide" once enough countries have it, even if it not every single country does?
Another important point:
L4 and L5 are actually the same autonomous driving. They are both full autonomy, meaning no human driver needed. The only difference is the ODD (when and where the system can be used). L4 is just L5 with a smaller ODD. But L4 does not have to be a tiny geofence. Technically, the ODD of L4 can be anything from a single street to 99.999% of L5. Also, the ODD limit does not have to be geographical. So for example, you could have L4 that works in the entire US, except for one city block, and it would technically be L4. or you could have L4 that works everywhere but cannot work when it rains or cannot work at night and it would be L4 even though it works in the same geolocations as L5. The only real advantage of L5 is that you don't have to specify the ODD since it works "everywhere" and "anytime". With L4, you do need to specify the ODD so you need to know what limits it has. But I hope you would agree that if L4 had a ODD that is say 99% of L5, for all practical purposes, that would be close enough to L5 for most people. And if the L4 can meet all the personal driving needs of an individual, it is essentially "L5" to that person since they never need to take over. So this is why I don't think we really need L5. we only really need L4 in a big enough ODD to be useful and practical and work towards L5 ODD.
With this in mind, I don't think it is quite accurate to say that Tesla is solving L5 while Waymo is solving L4. IMO, it would be more accurate to say that Tesla is trying to deploy L5 while Waymo is focused on deploying L4. But both Tesla and Waymo are trying to solve full autonomy. The deployment ODD is just different. I say Waymo is solving full autonomy because the Waymo Driver has complete perception, prediction and planning designed to handle all driving tasks. In fact, Waymo has demonstrated that their autonomous driving can handle a pretty wide ODD, from rural, suburbs to dense urban and highways and also inclement weather. Waymo is just limiting the full autonomy to a geofenced ODD, so we call it L4. But it is only L4 because the ODD is restricted, not because it is not full autonomy. Waymo is not interested in deploying L5 but they are interested in full autonomy and expanding the ODD over time.
Tesla and Waymo just have different approaches to solving full autonomy. Tesla wants to deploy L5 and their approach is to start with L2 but in a "L5-like" ODD, ie everywhere/anytime, and then work towards full autonomy. This makes perfect sense for consumers who want to go anywhere, anytime and will be in the driver seat anyway to supervise. The Waymo approach is to start with full autonomy but restrict the ODD and expand the ODD over time. This makes sense for a robotaxi model where you need to do driverless and only need to offer ride-hailing in a metro area where your customers are.
There has been a lot of debate about the SAE levels. Some want to get rid of L5 entirely since they see it as a uneccesary. Once "L4" has a ODD that is say 99% of L5, then doesn't it become indistinguishable from L5 for all practical purposes? In fact, wouldn't we probably just call it L5 at that point anyway since it is so close to L5?
So personally, I think it might be better just to think in terms of full autonomy, ie no human driver, and then define the ODD rather than talk about L4 or L5. The goal should be full autonomy with a big enough ODD to be practical and useful. That is one reason I really like Mobileye's taxonomy of "eyes on" or "eyes off" with well defined ODD. I think that is clearer to the consumer. Consumers basically want to know if they need to supervise the system or not and where and when can they use the system. And if the ODD is big enough to meet all their personal driving needs, they probably won't care if the system is technically only L4.
Sorry for the long post but this topic is a passion of mine.
Last edited:
jasonh4451
Member
Took my first Waymo rides this weekend in San Francisco’s Mission district and was super impressed. Tight urban area with lots of pedestrians, bikes, double parked cars. A normal busy bustling Saturday afternoon. Waymo was confident and smooth, very human like in tricky situations. I got home and took out my model 3 running 11.4.2 and did the same route that the Waymo had done. It was no comparison. FSD was jerky, nervous and I had to intervene several times. I believe the tesla camera suite and processor cannot compete with Waymo’s suite. It makes sense since the Waymo has $100k of hardware. That’s my very unscientific experience.
EVNow
Well-Known Member
Let someone drive and you sit in the backseat.
When you are the driver every little jerk feels exaggerated.
Like CYBRLFT taxi riders, you will hardly notice.
sleepydoc
Well-Known Member
As others have pointed out there are more difference between the two than just the camera suite.
EVNow
Well-Known Member
Its not the $100k hardware - its the $10B they have spent in the last 13 years concentrating on a small geographic area. If Tesla had done that they could have been very good in SF & Chandler (and probably a few more places since they aren't as conservative).
EVNow
Well-Known Member
If Google didn't have infinite monopoly money (or we had decent regulators in US), Waymo would have been shut down 5 years back. Its a money pit with no end in sight.
Ironically ChatGPT might force the closure of Waymo.
Similar threads
