Waymo

[Doggydogworld]

Waymo opens to the public. So if you find yourself in Phoenix, sign up and let us know how the ride goes.

A few people report still being on waitlist, but at least one of those got accepted after updating the app. So I think it truly is open. Clearing their waitlist with so few vehicles supports my thesis that they have a business model problem. Expanding their service area will increase utility, of course, since a very high percentage of trips starting inside the current area end outside it and vice versa. But will increased utility grow ridership enough to make the numbers pencil out?

 

[diplomat33]

Waymo opens to the public. So if you find yourself in Phoenix, sign up and let us know how the ride goes.

FYI, if you are looking for videos of rides, the youtube channel JJRicks Studios has a bunch of great videos of a lot of rides:

https://www.youtube.com/playlist?list=PL-13jt3ZPb7X6qJTo_MEnREMe-86I4S_q

A few people report still being on waitlist, but at least one of those got accepted after updating the app. So I think it truly is open. Clearing their waitlist with so few vehicles supports my thesis that they have a business model problem. Expanding their service area will increase utility, of course, since a very high percentage of trips starting inside the current area end outside it and vice versa. But will increased utility grow ridership enough to make the numbers pencil out?

Waymo's ride-hailing service is still young. They just really opened it up to the entire public a couple days ago. it is not unusual to have a few supply/demand issues when you just start a new business. Waymo has more vehicles that they can deploy. I trust they will solve the waitlist problem soon. If anything, the waitlist, is a good sign. It means that demand is high.

 

[EVNow]

They just really opened it up to the entire public a couple days ago. it is not unusual to have a few supply/demand issues when you just start a new business.

I actually don't know what Waymo's business plan is. Tesla's robotaxi "plan" (as envisaged by analysts) is that they can offer the service for cheaper rates than Uber and still make money because of low EV operating cost. How exactly does Waymo plans to make money ?

 

[Daniel in SD]

I actually don't know what Waymo's business plan is. Tesla's robotaxi "plan" (as envisaged by analysts) is that they can offer the service for cheaper rates than Uber and still make money because of low EV operating cost. How exactly does Waymo plans to make money ?

Same. Plus they plan on doing long distance trucking too.
Labor costs always go up over time while the cost of machines always goes down so it should work eventually!
Obviously if Tesla FSD works Waymo is done. There is no way they can compete with the cost of Tesla's system.

 

[diplomat33]

I actually don't know what Waymo's business plan is. Tesla's robotaxi "plan" (as envisaged by analysts) is that they can offer the service for cheaper rates than Uber and still make money because of low EV operating cost. How exactly does Waymo plans to make money ?

Waymo has 2 parts to their business model: Waymo One which does the ride-hailing and Waymo Via which does trucking and deliveries. Waymo One is already cheaper than Uber. And Waymo plans to have an all EV fleet. So with the lower EV costs + no driver to pay, Waymo will offer much cheaper rates than Uber. Waymo One also plans to lease robotaxis to consumers. Waymo Via can make money making deals with companies to retrofit their trucks with the FSD hardware and software.

 

[EVNow]

Same.

Waymo has 2 parts to their business model: Waymo One which does the ride-hailing and Waymo Via which does trucking and deliveries. Waymo One is already cheaper than Uber. And Waymo plans to have an all EV fleet. So with the lower EV costs + no driver to pay, Waymo will offer much cheaper rates than Uber. Waymo One also plans to lease robotaxis to consumers. Waymo Via can make money making deals with companies to retrofit their trucks with the FSD hardware and software.

But who has the best EV costs ?

Essentially, the "depreciation" part will be great for Tesla compared to any other EV maker ...

BTW, "Waymo One is already cheaper than Uber." - I'm not talking about cost to consumer, but to the business as well. Currently Waymo is probably losing a ton of money for every trip.

 

[EVNow]

Essentially, the "depreciation" part will be great for Tesla compared to any other EV maker ...

Duh ... the other way round.

 

[ThomasD]

The average cost for an Uber is 2 dollars per mile Uber drivers clean their own cars and pay for their own repairs. and pay their own insurance. These vehicles will be dirty and get damaged so now they will have to pay someone to clean and repair them. I think they will be equal to or be a little higher than what it costs to ride in an Uber. Businesses are not going to subsidize rides for much longer

 

[Daniel in SD]

The average cost for an Uber is 2 dollars per mile Uber drivers clean their own cars and pay for their own repairs. and pay their own insurance. These vehicles will be dirty and get damaged so now they will have to pay someone to clean and repair them. I think they will be equal to or be a little higher than what it costs to ride in an Uber. Businesses are not going to subsidize rides for much longer

Uber also loses money every year.
A robotaxi operating 24/7 could probably do 50,000 miles a year so that's $100k of revenue per year. Long term it seems plausible that robotaxi operating costs could be low enough to make a profit.

But who has the best EV costs ?

Essentially, the "depreciation" part will be great for Tesla compared to any other EV maker ...

I don't think depreciation matters at all for robotaxis. These vehicles will be driven into the ground.
As I said this debate is silly. If Tesla's software works there will be no competition. Waymo and everyone else have way higher operating costs. They have a much more expensive sensor suite, have to do extensive verification for each market, require a crew of remote assistants, etc.
The GM Cruise Origin robotaxi can probably compete with Tesla on unit cost but will still have much higher support costs. The only question is whether Tesla can deliver robotaxi capability with their current approach.

 

[EVNow]

I don't think depreciation matters at all for robotaxis. These vehicles will be driven into the ground.

It always matters. You can use st line depreciation = cost/number of years of service as the depreciation to account for each year.

There is a reason Tesla wants to aim for a million mile battery.

 

[Daniel in SD]

It always matters. You can use st line depreciation = cost/number of years of service as the depreciation to account for each year.

There is a reason Tesla wants to aim for a million mile battery.

Ah. I was thinking you mean percent per year or something.
The Cruise Origin is designed to last a million miles.

“We built this car around the idea of not having a driver and specifically being used in a ride-share fleet,” he says. “This vehicle is engineered to last a million miles and all the interior components are replaceable. The compute is replaceable, the sensors are replaceable. And what that does is it drives the cost per mile down way lower than you could ever reach if you took a regular car and tried to retrofit it. The replacement cost and the upkeep of that would just kill you from a business standpoint.”

Exclusive look at Cruise’s first driverless car without a steering wheel or pedals

 

[DanCar]

Waymo has reverted back to safety drivers?

 

[diplomat33]

Waymo has reverted back to safety drivers?

In the comments section, JJRicks, the youtuber in the car in the video, thinks the reason for the safety driver is that it had just rained a lot. So I guess, out of an abundance of caution, since slippery roads from rain might make driving less safe, Waymo temporarily put the safety driver back in. You know how Waymo is about safety.

In the comments section, JJRicks says he has a video coming out soon that shows the car self-driving when it starts to rain. He says it is an interesting video. So we might get to see how Waymo's FSD performs in rain.

 

[Mardak]

Waymo has reverted back to safety drivers?

Looks like this case needed safety drivers both local and remote. I'm not sure if the "RAD MONITORING" message would show up if it weren't for the local safety driver, but it does show an interesting situation where the system realizes its desired path is blocked but is unwilling to choose a different path:

The line in front of the vehicle in the visualization seems to be a barrier placed by a remote safety driver which then forces it to find an alternate route.

I suppose down the line, Waymo can say the remote safety driver don't even need to be a licensed human driver as the problems for assistance are probably more of human reasoning, e.g., this parking aisle is closed for at least the next few minutes so treat it as a permanent wall. That does have some interesting implications as even a child could be making these decisions and would allow Waymo to maintain "If a licensed human driver is required, it's not full autonomy."

 

[diplomat33]

Looks like this case needed safety drivers both local and remote. I'm not sure if the "RAD MONITORING" message would show up if it weren't for the local safety driver, but it does show an interesting situation where the system realizes its desired path is blocked but is unwilling to choose a different path:

View attachment 617531
The line in front of the vehicle in the visualization seems to be a barrier placed by a remote safety driver which then forces it to find an alternate route.

I suppose down the line, Waymo can say the remote safety driver don't even need to be a licensed human driver as the problems for assistance are probably more of human reasoning, e.g., this parking aisle is closed for at least the next few minutes so treat it as a permanent wall. That does have some interesting implications as even a child could be making these decisions and would allow Waymo to maintain "If a licensed human driver is required, it's not full autonomy."

When the car first approaches the blocked path, we can see on the screen that it had the correct path to go around:

If the car had followed this first path it wanted to take, it would have been fine. It seems to have "second guessed" itself as it then switches to a different path which was blocked and then hesitates. We then see the path on the screen bounce around a bit which would seem to indicate that once the car had moved in far enough, then it got confused.

So this is what I think happened in the car's computer. It saw it's intended path was blocked and figured out the correct detour but then part of the algorithm saw that there was a more direct path so it tried the more direct path but then detected that it's path was now completely blocked without backing up and got stuck.

I am curious if the car would have been able to figure it out on its own eventually or not. Did the remote operator intervene to save time since there is a passenger in the car that paid to get somewhere or because the car actually could not figure it out at all? I think Waymo can fix this by tweaking their path finding algorithm. The car did have the right idea at first.

Also, I find it interesting that the local safety driver did not just disengage to get the car unstuck. Why use the remote operator at all when you have a local safety driver? Perhaps, the local safety drivers are trained to only disengage for safety reasons but to leave the FSD on as much as possible. And this was clearly a case where they did not need to disengage since the FSD could handle it as soon as the remote operator told it what to do.

 

[mark95476]

I don't think Uber and other rider sharing drivers were rolling in money either. They probably depend on tips like restaurant service staff.

Uber also loses money every year.
A robotaxi operating 24/7 could probably do 50,000 miles a year so that's $100k of revenue per year. Long term it seems plausible that robotaxi operating costs could be low enough to make a profit.

Who knows how much Tesla is going to charge to add your car to their robotaxi service and what you as a provider can charge. They're going to make the owner eat all of the depreciate, wear and tear, cleaning,...etc... My price will be $100,000 per ride--essentially you need to pay for a completely new car plus 50% extra for the convenience and you get nothing after the ride.

Personally, I like the idea of having my own private valet service where my car will drop me off and pick me up for work, shopping, airport, somewhere in the city, restaurants, ...etc... so I don't have to deal with parking, theft, and vandalism. Right now, autopilot/EAP/FSD requires you to have a person in the drivers seat.

I don't think depreciation matters at all for robotaxis. These vehicles will be driven into the ground.
As I said this debate is silly. If Tesla's software works there will be no competition. Waymo and everyone else have way higher operating costs. They have a much more expensive sensor suite, have to do extensive verification for each market, require a crew of remote assistants, etc.
The GM Cruise Origin robotaxi can probably compete with Tesla on unit cost but will still have much higher support costs. The only question is whether Tesla can deliver robotaxi capability with their current approach.

 

[diplomat33]

Waymo autonomous cars spied at the Las Vegas speedway parking lot:

It is hard to tell but it looks like they might be driverless.

 

[diplomat33]

Waymo has reverted back to safety drivers?

No. Here is ride going on now live that has no safety driver.

JJ Ricks is streaming live and taking questions.

 

[diplomat33]

Found another driverless ride from a couple days ago. You can see several other Waymo cars driving around too.

 

[diplomat33]

16 minute informative presentation on how Waymo measures safety for their autonomous driving: