EVNow
Well-Known Member
Waymo Driver is not for consumer cars and it will not work outside small geofenced areas.
For consumer cars you need the sensors to be not eyesores, cheap and to work everywhere. Waymo fails on all accounts.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Waymo
- Thread starter Daniel in SD
- Start date
For consumer cars you need the sensors to be not eyesores, cheap and to work everywhere. Waymo fails on all accounts.
willow_hiller
Well-Known Member
Plus, if these autonomous technologies are really as valuable as we think they will be, what financial incentive does Waymo have to license it instead of keeping the value for themselves? They could be one of a tiny number of operators of robotaxis in a few major cities, making money with every fare 24/7. Or they could license it and all the operators will have a race to the bottom on fares until it's just barely profitable.
You don't see Google licensing out their search algorithm.
You don't see Google licensing out their search algorithm.
The answer to this is Microsoft vs IBM in the 1990's; it's generally better to be a high-demand software vendor selling to a dozen PC Compatible hardware vendors competing to drive down manufacturing costs, than it is to try setting up your own protected vertical (see Apple during that same time period, before non-PC hardware became popular).
If L4 automation became an essential feature which only Waymo can provide then Waymo can name their price and every vehicle manufacturer will be required to pay it or close up shop.
EVNow
Well-Known Member
I know you are very new to this - look at my post. You will understand Waymo driver is a mainframe. Not PC.
It is not YET for consumer cars. If you have evidence for them not ever aiming for consumer cars, please share.
It's neither a mainframe nor a PC; Waymo driver has no physical hardware or manufacturing infrastructure connected to it. It's very much a Microsoft Windows (or Android).
They will include a published a minimum physical specification, a demonstration device, a list of vendors who can provide components (lidar and Tensor chips for example) at reasonable prices, and Waymo will provide software/insurance/legal/patent use/support for a monthly licensing fee. Yes, the spec will require lidar and wide-spectrum cameras but it won't require Waymo lidar or Waymo wide-spectrum camera.
Style differentiation will be left entirely to the vehicle manufacturer which definitely will not be Waymo unless something goes very wrong with the business plan. They'll need to meet spec though, which will be minimum coverage at a minimal sampling density.
Last edited:
Doggydogworld
Active Member
Waymo has said they will license to automakers. But they also think the Seba/Musk personal car replacement vision might still win out. It's too early to tell. I maintain they pursued car replacement in Phoenix. The 82k cars they ordered are enough to replace 2nd/3rd cars in the entire metro area. They never got far enough to really test the vision. Now they're just trying to find any business model that works in any market. Then they can go back to grand visions.
Never in history have electronics become smaller and cheaper! /s
Mercedes, Lexus, Nio and others ship cars with lidar today. Meanwhile Musk is in year 8 of fake it 'til you make it.
Not only that, but a hybrid sensor stack like Waymo's surely is able to operate with a subset of sensors: less sensors leading to a higher collision risk.
When NNs pre-processing vision in Waymo's stack becomes robust enough, they might just drop the requirement of needing lidar. IMO there is no reason to believe that Tesla would be ahead of game in getting to vision only: they just deprive themselves of the chance of getting to L4 first by not using lidar in addition to vision.
EVNow
Well-Known Member
I don’t see Waymo in that list.
BTW, it’s not just me saying it. MobilEye also thinks you can’t go from commercial taxis to consumer cars.
Moreover, Waymo isn’t creating HD maps of all the roads in the world.
EVNow
Well-Known Member
Surely you are wrong about Waymo. No HDMaps, no complicated hardware, no expensive compute = no autonomy.
You should listen to MobilEye presentation too …
"All roads": true. Nobody will ever map all roads until the roadway building equipment does the mapping.
Google street-view vehicles started installing Lidar (2 Velodyne Pucks, albeit mounted at an odd angle) in 2017 and will have hit a non-trivial %age of the developed world urban roads since that point. Where that data goes, we don't know, but I doubt it was to improve the accuracy of 3D rendered maps in the Google Maps app.
Waymo would have access to that data if they found it beneficial.
It's possible that's enough on it's own; or enough to allow a fully-automated Waymo Driver to collect a fully detailed survey driving cautiously which is effectively the same (bring 100 vehicles to city; automated mapping & testing for local issues over a month; start production service).
Last edited:
EVNow
Well-Known Member
They can barely drive in San Francisco and yet you are bending over backward to speculate on how Waymo tech can be used in consumer cars. Clearly the design intent is different. Do you have google stock ?
Its like the Tesla robotaxi network speculation.
Speculation.
Not going to waste anymore of my time responding to you.
diplomat33
Average guy who loves autonomous vehicles
Informative Motor Trend review of Waymo's driverless rides in Chandler. I found the info on the different driving behavior between the 4th Gen and 5th Gen interesting:
Last edited:
diplomat33
Average guy who loves autonomous vehicles
Not sure if Chandler is "upscale" but the current service area is common knowledge. Here is the map of the current Waymo One area in Phoenix:
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/23355027/waymo_one_Metro_PHX_map.png)
Waymo did recently start testing with safety drivers in downtown Phoenix and will be launching a ride-hailing area in downtown Phoenix soon.
diplomat33
Average guy who loves autonomous vehicles
I don't think we know yet. Waymo has not released a map of the downtown service area AFAIK.
But if I had to guess, I think it would make sense for Waymo to connect the two areas since it would allow customers in downtown to get a Waymo ride to a destination in Chandler and vice versa. So IMO it would be more useful to customers.
Doggydogworld
Active Member
Almost certainly not. It would violate their First Principle: Thou shalt do no thing which makes commercial sense
Practically speaking, it would mean adding highway service. I don't see that for years.
Similar threads
- Article
