Waymo's self-driving Jaguar I-Pace begins testing on public roads

[diplomat33]

"Now, it appears self-driving Jaguar I-Pace vehicles are finally being tested on public streets around Waymo’s headquarters in Mountain View, Calif., TechCrunch has learned. A self-driving Jaguar I-Pace, with a safety driver behind the wheel, was spotted Monday morning. Waymo has confirmed that testing has begun."
Waymo’s self-driving Jaguar I-Pace vehicles are now testing on public roads – TechCrunch

 

[Daniel in SD]

"Now, it appears self-driving Jaguar I-Pace vehicles are finally being tested on public streets around Waymo’s headquarters in Mountain View, Calif., TechCrunch has learned. A self-driving Jaguar I-Pace, with a safety driver behind the wheel, was spotted Monday morning. Waymo has confirmed that testing has begun."
Waymo’s self-driving Jaguar I-Pace vehicles are now testing on public roads – TechCrunch

I don't really see the point of this. Their Dodge Caravans are way more comfortable to ride in and look better IMHO. I guess they're desperate to have an EV. They should troll Tesla by using a Model X. Haha.

 

[diplomat33]

I don't really see the point of this. Their Dodge Caravans are way more comfortable to ride in and look better IMHO. I guess they're desperate to have an EV. They should troll Tesla by using a Model X. Haha.

What is kinda weird to me is that the Caravans have all the hardware integrated nicely whereas these I-Pace cars appear to have the hardware tacked on. So it would seem like they don't have I-Pace cars with actual FSD hardware integrated in them yet. Why not do that first and then start testing? Furthermore, does the I-Pace not have its own FSD hardware?

 

[Eno Deb]

I don't really see the point of this. Their Dodge Caravans are way more comfortable to ride in and look better IMHO. I guess they're desperate to have an EV. They should troll Tesla by using a Model X. Haha.

You have to look at the big picture. Waymo is looking for partners in the automotive sector that could potentially become customers and adopt it's self-driving technology at some point. JLR is interested in learning from Waymo's technology. That's probably what's behind this partnership, plus some money. Besides, their press release last year said that the Jaguars will be prepared for self-driving from the factory, although it didn't mention specifically what that includes (at the very least it will be an interface to the car's control system).

 

[Tam]

...does the I-Pace not have its own FSD hardware?

I don't think so because I don't its LIDAR anywhere.

 

[electronblue]

Given that Waymo is developing its own autonomy suite I’m sure anything Jaguar has would be disabled anyway (except any ”preparation” type of arrangement that was mutually agreed upon, like Uber with Volvo)... so I don’t really see why or how Jaguar’s possible other autonomy efforts would be included here.

 

[CarlK]

Even if this eventually will work in some capacity it will be just self driving taxi while TeslaNet is self driving Uber/Lyft. Waymo will still need to acquire the very large fleet and taking car of their maintenance, storage, other overhead and all logistics. We all know the outcome of taxi companies against Uber/Lyft and reasons why. It's a losing proposition against Tesla's business model before it even started.

 

[electronblue]

Even if this eventually will work in some capacity it will be just self driving taxi while TeslaNet is self driving Uber/Lyft. Waymo will still need to acquire the very large fleet and taking car of their maintenance, storage, other overhead and all logistics. We all know the outcome of taxi companies against Uber/Lyft and reasons why. It's a losing proposition against Tesla's business model before it even started.

There is absolutely nothing stopping Waymo from licensing its autonomous driving system to car manufacturers though, once completed. Who gets where first depends on how well and fast they can make their system work.

Tesla’s deployment advantage is real but also dependent on them making the current suite work as an autonomous car.

 

[CarlK]

I think this mickey mouse thing is just so Waymo executives can have something to show to the board. Just what impact "up to 20,000 modified I-Pace vehicles" could make in the robot taxi space? Uber/Lyft have a combined 5 million drivers at the moment already. You'll need millions of autonomous taxis to replace them. Only Tesla has a business model that will work.

 

[electronblue]

I think this mickey mouse thing is just so Waymo executives can have something to show to the board. Just what impact "up to 20,000 modified I-Pace vehicles" could make in the robot taxi space? Uber/Lyft have a combined 5 million drivers at the moment already. You'll need millions of autonomous taxis to replace them. Only Tesla has a business model that will work.

Tesla has a business model that can work, but do they have the technology to make it work? And where is the competition if/once they do? That is the interesting question indeed.

I don’t think we should be fooled by Waymo’s 20,000 number. These are prototype or pilot fleets still, as the technology continues to be developed and mature. In Waymo’s plan, there is no point in building a million cars yet.

If/when the technology is ready, there is nothing stopping Waymo building autonomous cars with partners at much higher rates.

 

[Daniel in SD]

Only Tesla has a business model that will work.

I think licensing your technology to companies that produce way more cars than Tesla can definitely work.
It does seem a little bit like putting the cart before the horse though. Maybe they’re actually close to removing the safety drivers from their Arizona service?

 

[diplomat33]

If/when the technology is ready, there is nothing stopping Waymo building autonomous cars with partners at much higher rates.

So far, we've only seen Waymo test and deploy their autonomous vehicles in small locations like one part of Phoenix. Even if the vehicles work perfectly and become L4 in that area, that does not automatically mean that they will be L4 everywhere. After all, the vehicles could encounter a different road configuration or different driving scenario in another part of the US that it is not able to handle. I am not sure how Waymo makes that leap from having L4 autonomous cars that work in one specific area to deploying the tech in millions of cars that drive all across the entire US. In certainly would not be a quick process. Waymo would need to reach a high enough level of confidence that their software is generalized enough to work everywhere and would probably need to do more testing in different areas.

 

[CarlK]

I think licensing your technology to companies that produce way more cars than Tesla can definitely work.
It does seem a little bit like putting the cart before the horse though. Maybe they’re actually close to removing the safety drivers from their Arizona service?

It's not simple like, say, VW buys Waymo's Lidar based system, when and if it can work, to put in a million cars and sell them for $100,000 each to Uber or people who want to be independent contractors. You're talking about a tremendous amount of investment and risk no CEO will be willing to take. No body will be interested if you just do mickey mousing thing like to produce 20,000 of them. Pretty interesting 20,000 seems to be the universal number people do mickey mouse things, like number of EV's they will make, like to use.

Tesla's model is just to sell you good cars will good value that are all FSD ready. By the time the system is ready to turn on there will be a million robot taxi ready car instantly. It's like having everyone to finance it so it costs the company practically nothing (except extra hardware people may not pay for). Even for the machine learning endeavor we are all buying test cars for Tesla and help performing tests for free. I can't think of a smarter strategy than this. NO ONE will be able to duplicate it or do anything that is even close to this.

Most amazing is all these were planned many years ago, from OTA to putting hardware in every car, but we only start to realize it. And still only smarter ones could realize it even now. It's an understatement to say Tesla is the most misunderstood company. There simply are not that many people who are that smart.

So far, we've only seen Waymo test and deploy their autonomous vehicles in small locations like one part of Phoenix. Even if the vehicles work perfectly and become L4 in that area, that does not automatically mean that they will be L4 everywhere. After all, the vehicles could encounter a different road configuration or different driving scenario in another part of the US that it is not able to handle. I am not sure how Waymo makes that leap from having L4 autonomous cars that work in one specific area to deploying the tech in millions of cars that drive all across the entire US. In certainly would not be a quick process. Waymo would need to reach a high enough level of confidence that their software is generalized enough to work everywhere and would probably need to do more testing in different areas.

Yes I agree. The chance it will succeed without deep machine learning is close to none. My post to DiS was just to state the strategic advantage of Tesla regardless of whether Waymo or anyone else could develop a working system.

 

[Daniel in SD]

here simply are not that many people who are that smart.

I think everyone here understands what Tesla's plan is. You really don't have to be that smart, it's not like it's a secret!
Now I suppose you can say people are dumb for thinking it probably won't work. I've always said that if people are so confident that it will work that they are willing to call people who are skeptical dumb then they should be placing huge bets on the stock price. The stock is incredibly undervalued if you think there is a high likelihood of Tesla's plan working.

It's not simple like, say, VW buys Waymo's Lidar based system, when and if it can work, to put in a million cars and sell them for $100,000 each to Uber or people who want to be independent contractors. You're talking about a tremendous amount of investment and risk no CEO will be willing to take. No body will be interested if you just do mickey mousing thing like to produce 20,000 of them. Pretty interesting 20,000 seems to be the universal number people do mickey mouse things, like number of EV's they will make, like to use.

I have no idea what you're saying here. Any automotive CEO would love to sell a million car at $100,000 each. If you've already got factories that can produce a million cars a year wouldn't it be smart to build cars that can sell for $100k in them?

 

[electronblue]

So far, we've only seen Waymo test and deploy their autonomous vehicles in small locations like one part of Phoenix. Even if the vehicles work perfectly and become L4 in that area, that does not automatically mean that they will be L4 everywhere. After all, the vehicles could encounter a different road configuration or different driving scenario in another part of the US that it is not able to handle. I am not sure how Waymo makes that leap from having L4 autonomous cars that work in one specific area to deploying the tech in millions of cars that drive all across the entire US. In certainly would not be a quick process. Waymo would need to reach a high enough level of confidence that their software is generalized enough to work everywhere and would probably need to do more testing in different areas.

Yes of course, Waymo too has their work cut out for them in expanding their offering. They do test in a dozen or so locations already, but certainly going global for example would mean a wide range of new challenges to solve. Then again, Google does volume and the world, so it isn’t unfathomable Waymo might succeed in being quite rapid about it once they are mature enough.

I just wouldn’t get hung up on any number of cars they tout today. They are not ready yet so they are not deploying volume, they are deploying test and pilot fleets. Once their technology is ready we can assess how fast they can deploy it.

 

[electronblue]

It's not simple like, say, VW buys Waymo's Lidar based system, when and if it can work, to put in a million cars and sell them for $100,000 each to Uber or people who want to be independent contractors. You're talking about a tremendous amount of investment and risk no CEO will be willing to take. No body will be interested if you just do mickey mousing thing like to produce 20,000 of them. Pretty interesting 20,000 seems to be the universal number people do mickey mouse things, like number of EV's they will make, like to use.

When someone has a mature technology for self-driving, they will sell every car they can make and change the world. Licensing technology to a mature car-maker certainly seems like a potential formula for this to me.

Tesla's model is just to sell you good cars will good value that are all FSD ready. By the time the system is ready to turn on there will be a million robot taxi ready car instantly. It's like having everyone to finance it so it costs the company practically nothing (except extra hardware people may not pay for). Even for the machine learning endeavor we are all buying test cars for Tesla and help performing tests for free. I can't think of a smarter strategy than this. NO ONE will be able to duplicate it or do anything that is even close to this.

Assuming their hardware suite ends up being up to the task.

Most amazing is all these were planned many years ago, from OTA to putting hardware in every car, but we only start to realize it. And still only smarter ones could realize it even now. It's an understatement to say Tesla is the most misunderstood company. There simply are not that many people who are that smart.

The OTA capability has certainly enabled many wonderful things (and some disasters) but it has also made Tesla vulnerable to an overreach nobody else is vulnerable to: The presale and the oversell. Once deployed, changing the suite is hard for Tesla and they can spend an immense amount of time trying to shoehorn their system to whatever hardware they had the foresight to deploy. It may work out, or it may not, but it is a vulnerability others don’t have because others can keep changing the hardware in-house until the software catches up.

Tesla pretty much gambled they got it right with the hardware — and did so with the barest of minimums. Those of us with AP2 learned quickly what Tesla felt they missed the first time around when AP2.5 rolled in. It remains to be seen if HW2.5/HW3 is enough.

 

[Eno Deb]

So far, we've only seen Waymo test and deploy their autonomous vehicles in small locations like one part of Phoenix. Even if the vehicles work perfectly and become L4 in that area, that does not automatically mean that they will be L4 everywhere. After all, the vehicles could encounter a different road configuration or different driving scenario in another part of the US that it is not able to handle.

The same applies to Tesla. And before you say "but their massive fleet", consider that their fleet is actually heavily concentrated in a few geographic areas such as California. There are large parts of the world where their fleet is thin or non-existant.

I am not sure how Waymo makes that leap from having L4 autonomous cars that work in one specific area to deploying the tech in millions of cars that drive all across the entire US. In certainly would not be a quick process. Waymo would need to reach a high enough level of confidence that their software is generalized enough to work everywhere and would probably need to do more testing in different areas.

And Tesla's "boil the ocean" approach is easier?

 

[Daniel in SD]

Tesla only produced 2,450 Roadsters, mickey mouse numbers

 

[diplomat33]

The same applies to Tesla. And before you say "but their massive fleet", consider that their fleet is actually heavily concentrated in a few geographic areas such as California. There are large parts of the world where their fleet is thin or non-existant.

I am not talking about self-driving that works in the entire world, I am only looking at self-driving that works in the US. And in the US, Tesla does have a lot more cars on the road than Waymo. And Tesla has cars pretty much everywhere in the US and growing. It is certainly much easier for Tesla to test a wide range of cases than it is for Waymo.

And Tesla's "boil the ocean" approach is easier?

Not easier, just different.

 

[CarlK]

I think everyone here understands what Tesla's plan is. You really don't have to be that smart, it's not like it's a secret!
Now I suppose you can say people are dumb for thinking it probably won't work. I've always said that if people are so confident that it will work that they are willing to call people who are skeptical dumb then they should be placing huge bets on the stock price. The stock is incredibly undervalued if you think there is a high likelihood of Tesla's plan working.

I'm in a pretty good mood today so I'll refrain from commenting on smartness or dumbness of people. BTW I do have a significant investment, investment but not bet, on the stock because I'm so confident of Elon's ability and the company's future. I agree one should put money where his mouth is.

I have no idea what you're saying here. Any automotive CEO would love to sell a million car at $100,000 each. If you've already got factories that can produce a million cars a year wouldn't it be smart to build cars that can sell for $100k in them?

Yeah sure the same story as they could just produce a million competitive EV's if they want to. Where are even those? It's not that simple. Even not too smart people should be able to see that

Tesla only produced 2,450 Roadsters, mickey mouse numbers

Totally irrelevant. Roadster was a sports car or semi-super car with a small market. Most cars like that sells few thousands or even just a few hundred a year, if not entire production.