Waymo has no viable business model(provscons.com)
provscons.com
Waymo has no viable business model
https://provscons.com/why-will-waymo-fail/
62 comments
Wasn't part of the reason for LIDAR getting cheaper because of tech that Waymo developed?
The article makes more sense if you 100% buy all of Tesla's marketing.
The patents too. An IP lawyers dream...
These are reasonable points, but there's no hard numbers showing why Waymo cannot scale and why it can never be profitable.
In my opinion, despite the cost of the self driving hardware, which is much more than just Lidars, it can still be profitable. Before uber, taxi medallions could fetch $1.5M in New York City. Certainly the costs for Waymo are far lower and scale better with volume.
3D mapping is another cost, but it's also potentially doable since it's a big upfront cost with small ongoing costs. Taking it one (or a few) city at a time could well work.
At least nothing I've read here convinced me otherwise.
In my opinion, despite the cost of the self driving hardware, which is much more than just Lidars, it can still be profitable. Before uber, taxi medallions could fetch $1.5M in New York City. Certainly the costs for Waymo are far lower and scale better with volume.
3D mapping is another cost, but it's also potentially doable since it's a big upfront cost with small ongoing costs. Taking it one (or a few) city at a time could well work.
At least nothing I've read here convinced me otherwise.
Not to mention that most hardware has this magical deflationary property - gets cheaper over time. Look at lithium batteries or infrared sensors, for example.
This is ultimately the bet. Even if Waymos taxi business does nothing more than start scaling up hardware manufacture and building trust.
This isn’t a cheap strategy, but it’s believable that waymo could own most IP for self driving 10 years from now if they just outlast everyone else. They’re already on year 14 what’s wrong with another 2 decades?
This isn’t a cheap strategy, but it’s believable that waymo could own most IP for self driving 10 years from now if they just outlast everyone else. They’re already on year 14 what’s wrong with another 2 decades?
I upvoted your comment because you make great points, but...
> Before uber, taxi medallions could fetch $1.5M in New York City
I'm not defending the medallion system and it was corrupt af, but there were ordinary decent people who were destroyed by the subversion of this (completely corrupt and anti-competitive) system.
As for the software, eh, the Google Streetview vs. Waymo requirements were one of the weaker points of this article. We don't actually know what data Google collects with their cars (we do know they collect WiFi data[1]), and it isn't inconceivable that they were collecting precise data. It is true that this data ages faster than Google can refresh it, though.
I really believe that the software can scale but it is the real world/physical issues that will prevent this from ever becoming a real business.
[1] https://www.wired.com/2012/05/google-wifi-fcc-investigation/
> Before uber, taxi medallions could fetch $1.5M in New York City
I'm not defending the medallion system and it was corrupt af, but there were ordinary decent people who were destroyed by the subversion of this (completely corrupt and anti-competitive) system.
As for the software, eh, the Google Streetview vs. Waymo requirements were one of the weaker points of this article. We don't actually know what data Google collects with their cars (we do know they collect WiFi data[1]), and it isn't inconceivable that they were collecting precise data. It is true that this data ages faster than Google can refresh it, though.
I really believe that the software can scale but it is the real world/physical issues that will prevent this from ever becoming a real business.
[1] https://www.wired.com/2012/05/google-wifi-fcc-investigation/
Lidar isn’t expensive.
You can easily get the cost savings elsewhere in the vehicle or just target the premium market.
You can easily get the cost savings elsewhere in the vehicle or just target the premium market.
The article leads with "Waymo already solved the self-driving problem."
I don't dispute all their points but I sure disagree with that one.
I don't dispute all their points but I sure disagree with that one.
You have a fleet of self-driving cars at level 4 that have multiple cameras, and you claim it's going to fail because the system can't keep a street map updated? Does the author even understand how software and automation works?
I’m not aware of any HD maps built entirely automatically. As soon as manual steps creep in, the cost skyrockets. The backhaul and processing power required for updates can also be staggering.
The wages paid to professional drivers is also staggering.
If amortized cost of hardware + backhaul & processing costs < professional driver wages, then the model works. This is especially true as tech improves and wages increase. There will be a crossover eventually.
If amortized cost of hardware + backhaul & processing costs < professional driver wages, then the model works. This is especially true as tech improves and wages increase. There will be a crossover eventually.
Back of the envelope calculation for HD mapping cost: $5k / km * 4,000,000 km of road in the US = $20bn / yr on HD mapping (assuming once a year updates, which is actually quite infrequent). And that’s not considering all other costs involved such as $500k in sensors per vehicle, safety techs (teleoperations are still super common), etc. It has to come down in cost a lot to compete with the price of unskilled labor.
$500k in sensors per vehicle is off by a substantial margin -- nearly an order of magnitude. I also seriously question your $5k / km "napkin math" without any elaboration.
I would also point out that there are 2M truck drivers earning $47k per year [1]. That's $94B annually in pure labor cost, and doesn't even account for rideshare & taxis.
[1] https://www.bls.gov/ooh/transportation-and-material-moving/h...
I would also point out that there are 2M truck drivers earning $47k per year [1]. That's $94B annually in pure labor cost, and doesn't even account for rideshare & taxis.
[1] https://www.bls.gov/ooh/transportation-and-material-moving/h...
Sure, but the point is that the car can flag something that doesn't look like the map so a human can review it. ie. closed road, new road, construction markers, etc.
That's way less intensive than having to build the whole thing from scratch or having it be fully automated.
Humans are good at exceptions. Let the humans do that while the machines do the rote task of "Yep, road still looks like the map. Nothing to see here."
That's way less intensive than having to build the whole thing from scratch or having it be fully automated.
Humans are good at exceptions. Let the humans do that while the machines do the rote task of "Yep, road still looks like the map. Nothing to see here."
I think there’s way more hand correction needed than people are estimating. Google Maps (the non hd version) employs a small army of 6k contractors [0] to correct way simpler data. Getting cm accurate lane markings and correctly classifying permanent and transient features of a scene is extremely error prone and I would guess at least an order of magnitude more manpower to make this happen at scale.
[0] https://www.businessinsider.com/apple-has-7000-fewer-people-...
[0] https://www.businessinsider.com/apple-has-7000-fewer-people-...
Why do you think the cost of building the HD maps is prohibitive? How much do you think it would cost to say, keep a square mile of a dense urban HD map updated for a year? It's probably way lower than you think.
I didn’t say it was prohibitive but it’s certainly not cheap. The cost is per km rather than per square km. DeepMap charges $5k / km for example (the big companies don’t publish their costs, but it’s widely known in the industry to be on par with this). I can’t seem to find the stats again, but map invalidation happens on the order of weeks to months.
I found myself agreeing with the cons mentioned in this article -- but, where is the insight? All of these challenges must have been known since before the company was formed. For example, people didn't suddenly learn yesterday that lidar has trouble in rain & snow. I suppose there may be value in stating all of the downsides in one place (as this article does).
Analysis that covers points that were almost certainly on the whiteboard of Waymo of risk factors before the first line of code was written is weak analysis.
I’m gonna go out on a limb and guess they considered this a decade ago and have a reason to think these aren’t a long term problem.
I’m gonna go out on a limb and guess they considered this a decade ago and have a reason to think these aren’t a long term problem.
Perhaps, but 10 years ago you didn't have viable competitors with pure or near-pure camera based self-driving systems.
Now we have little lightly funded companies like comma producing quite compelling experiences using cheap hardware.
Not to mention tesla with near-unlimited funding, leading the way in the space.
comma.ai is a joke. nobody in the industry thinks about them because they are so far from viability.
the idea of tesla having unlimited funding is a joke. They are barely spending across r&d for the whole company. what portion of their r&d is dedicated to self-driving do you thing? even if 100% is self-driving it's still less than the real players in the space.
the idea of tesla having unlimited funding is a joke. They are barely spending across r&d for the whole company. what portion of their r&d is dedicated to self-driving do you thing? even if 100% is self-driving it's still less than the real players in the space.
Tesla have access to boat loads of capital in the public markets, they are spending heaps on AI, developing their own silicon, not only for the cars, but for their training computer as well.
Who are the real players in the space? There might be bigger players focused more on general AI, but Tesla seem like the biggest player by far, focused on automotive mainly.
Who are the real players in the space? There might be bigger players focused more on general AI, but Tesla seem like the biggest player by far, focused on automotive mainly.
The premise of the article is incorrect. Waymo has not “solved the self driving problem”.
“You don’t need lidar to drive a car, so why should the car need lidar to drive the car?” -Elon Musk, like six years ago
I’m sure George Hotz also has some pithy takes on this.
Anyway, this all completely misses the point. Waymo was born of a DARPA challenge and continued to exist to search for even more DoD money. But we don’t need self driving tanks when we can pilot them remotely.
I’m sure George Hotz also has some pithy takes on this.
Anyway, this all completely misses the point. Waymo was born of a DARPA challenge and continued to exist to search for even more DoD money. But we don’t need self driving tanks when we can pilot them remotely.
So why don't we have remotely driven cars?
Because they don’t solve any real problems.
If AI is suitable for 99% of driving, why not use remote drivers for the remaining 1%? (eg. accidents, construction, adverse weather)
You could have a central office in a cheap location of the USA for professional drivers to take control as required. We already do it military for drone strikes...
You could have a central office in a cheap location of the USA for professional drivers to take control as required. We already do it military for drone strikes...
The biggest issue for all of the self driving taxi service ideas is fleet maintenance.
First of all, you need to own and insure a fleet of rolling depreciating assets. Maybe not as much these days with the crazy used car prices, but still, you have put these cars on the books. You can't even really lease the vehicles because of all of the LIDAR and other self driving modifications that are required.
Next you have to keep these vehicles fueled - either gasoline or electricity, so you are going to need to pay people to handle that. And you have ongoing maintenance for brakes, fluids, etc. And space to park them overnight and in low demand times.
And you also have the interior maintenance. Someone has to clean up the puke and garbage and just general messiness of having humans in the backseat of this car. Throughout the day, every day.
And this has to be repeated in every single city while competing against Uber and Lyft and traditional taxi companies that already know that the driver is probably the least expensive part of this whole system.
First of all, you need to own and insure a fleet of rolling depreciating assets. Maybe not as much these days with the crazy used car prices, but still, you have put these cars on the books. You can't even really lease the vehicles because of all of the LIDAR and other self driving modifications that are required.
Next you have to keep these vehicles fueled - either gasoline or electricity, so you are going to need to pay people to handle that. And you have ongoing maintenance for brakes, fluids, etc. And space to park them overnight and in low demand times.
And you also have the interior maintenance. Someone has to clean up the puke and garbage and just general messiness of having humans in the backseat of this car. Throughout the day, every day.
And this has to be repeated in every single city while competing against Uber and Lyft and traditional taxi companies that already know that the driver is probably the least expensive part of this whole system.
I'm not sure why you think having a large fleet of vehicles is problem. Rental car companies already exist.
As far as fueling/charging- Even if you had to pay a few people in the short-term to do the work, it isn't a high-skill job. Being generous, I can't fathom a situation where it could cost more than $2 per car in physical labor. Assuming you could drive more than 200 miles on a charge, that is less than a penny per mile in cost. Also, there are already gas stations in some states with people that will fill your gas tank for you. I am also sure an autonomous solution would be devised if that became a limiting factor. You could pretty easily design a purpose-built system for it.
I agree that regular maintenance needs to be performed, but I don't agree that the companies themselves need to solve the problem on their own. This seems like it would be something to easily contract out to a local company to handle. For standard maintenance and cleaning, there are plenty of companies that clean cars and plenty of companies that perform maintenance on cars already. It should just be a matter of getting contracts in place and making sure that they have software to do their jobs.
I didn't do the math to figure out the cost of a human driver vs the total taxi cost, but even if true the cost isn't trivial. Truckers, for example, can make upwards of $80k per year and that's not even at 50% utilization for the vehicle. If you assume a taxi driver would make $40k/year and they have 50% utilization of the car (I'm making up numbers for the sake of discussion), that is still a significant amount of cost. I would be surprised if maintenance and fuel would be $40k per year regardless of vehicle.
As far as fueling/charging- Even if you had to pay a few people in the short-term to do the work, it isn't a high-skill job. Being generous, I can't fathom a situation where it could cost more than $2 per car in physical labor. Assuming you could drive more than 200 miles on a charge, that is less than a penny per mile in cost. Also, there are already gas stations in some states with people that will fill your gas tank for you. I am also sure an autonomous solution would be devised if that became a limiting factor. You could pretty easily design a purpose-built system for it.
I agree that regular maintenance needs to be performed, but I don't agree that the companies themselves need to solve the problem on their own. This seems like it would be something to easily contract out to a local company to handle. For standard maintenance and cleaning, there are plenty of companies that clean cars and plenty of companies that perform maintenance on cars already. It should just be a matter of getting contracts in place and making sure that they have software to do their jobs.
I didn't do the math to figure out the cost of a human driver vs the total taxi cost, but even if true the cost isn't trivial. Truckers, for example, can make upwards of $80k per year and that's not even at 50% utilization for the vehicle. If you assume a taxi driver would make $40k/year and they have 50% utilization of the car (I'm making up numbers for the sake of discussion), that is still a significant amount of cost. I would be surprised if maintenance and fuel would be $40k per year regardless of vehicle.
Rental car companies do not need to either pay for the labor or the fuel to power their vehicles, and actually make a lot of extra money with "prepaid" or "return it empty" schemes. By churning their vehicles on a regular basis, they avoid many of the ongoing maintenance issues that would impact Waymo with their customized vehicles.
You're right that a Waymo would probably outsource the interior and (probably) mechanical maintenance to local providers, but they would need to deal with that in every city and with the additional technology involved in self-driving, you're probably not going to trust that the local shade tree mechanic will be able to handle that. So you have at least two problems now.
Over The Road drivers do make surprisingly good money, but that is a skilled job with a significant amount of training and regulation. In-town cabbies are often people with few other opportunities and are easily exploitable. The taxi cab business is notorious corrupt and is a big part of how Uber/Lyft could easily replace these services.
You're right that a Waymo would probably outsource the interior and (probably) mechanical maintenance to local providers, but they would need to deal with that in every city and with the additional technology involved in self-driving, you're probably not going to trust that the local shade tree mechanic will be able to handle that. So you have at least two problems now.
Over The Road drivers do make surprisingly good money, but that is a skilled job with a significant amount of training and regulation. In-town cabbies are often people with few other opportunities and are easily exploitable. The taxi cab business is notorious corrupt and is a big part of how Uber/Lyft could easily replace these services.
Fuel alone is easily $10,000 a year for a taxi.
The biggest expense people are not seemingly considering is the depreciation of the vehicle. taxis don't last long because the drive so much and so much stop and go. There could easily be $10,000 in depreciation a year.
The diver does things a robo car will not be able to do like help people with their bags, clean the car, responds flexibly to situations as they arise.
The biggest expense people are not seemingly considering is the depreciation of the vehicle. taxis don't last long because the drive so much and so much stop and go. There could easily be $10,000 in depreciation a year.
The diver does things a robo car will not be able to do like help people with their bags, clean the car, responds flexibly to situations as they arise.
Using $10,000 as an example.
Assuming super expensive fuel and terrible efficiency...
Gas car - 20 miles/gallon, $4.00/gallon ($4.00/20 = $0.20/mile. $10,000/$0.20 = 50,000 miles)
Electric car - 200 miles/74kWh [Tesla M3], $0.25/kWh (74 * $0.25 = $18.50. $18.50/200 miles = $0.0925/mile. $10,000/$0.0925 = 108,108 miles)
So assuming a gas car since the miles are the most expensive there, you would be looking at 50,000 miles for $10,000 in fuel. That checks out for me.
Using that assumption.
Uber charges $0.78/mile and $0.27/minute.
Assume only 75% of driving is charged to the customer (going to a fare, going to a charger, etc), that gives you 37,500 miles.
Assuming 45 miles per hour (probably too high) at $0.27/minute = (37,500/45)60 = 50,000 minutes of driving $0.27 = $13,500
37,500 miles * $0.78 = $29,250
So $42,750 in rev. - $30,000 in fuel, insurance, maintenance, and depreciation, you're still $12,750 in profit. Again, is pretty much giving every disadvantage possible. In an electric car scenario you could easily drive 200,000 miles on $10,000 in electricity. Which would multiply all numbers by 4.
I'm not trying to take your numbers too specifically, just illustrating for the sake of discussion.
Assuming super expensive fuel and terrible efficiency...
Gas car - 20 miles/gallon, $4.00/gallon ($4.00/20 = $0.20/mile. $10,000/$0.20 = 50,000 miles)
Electric car - 200 miles/74kWh [Tesla M3], $0.25/kWh (74 * $0.25 = $18.50. $18.50/200 miles = $0.0925/mile. $10,000/$0.0925 = 108,108 miles)
So assuming a gas car since the miles are the most expensive there, you would be looking at 50,000 miles for $10,000 in fuel. That checks out for me.
Using that assumption.
Uber charges $0.78/mile and $0.27/minute.
Assume only 75% of driving is charged to the customer (going to a fare, going to a charger, etc), that gives you 37,500 miles.
Assuming 45 miles per hour (probably too high) at $0.27/minute = (37,500/45)60 = 50,000 minutes of driving $0.27 = $13,500
37,500 miles * $0.78 = $29,250
So $42,750 in rev. - $30,000 in fuel, insurance, maintenance, and depreciation, you're still $12,750 in profit. Again, is pretty much giving every disadvantage possible. In an electric car scenario you could easily drive 200,000 miles on $10,000 in electricity. Which would multiply all numbers by 4.
I'm not trying to take your numbers too specifically, just illustrating for the sake of discussion.
> Next you have to keep these vehicles fueled - either gasoline or electricity, so you are going to need to pay people to handle that.
I don’t understand how someone who could solve the self driving problem could solve the “plug the car in via simple robotics after it parks itself” problem.
They are also able to park near the city where parking is cheaper, though I guess it could also turn out like it does in China (if anyone is out at 2AM and see all the taxis pushed over to the side of the road with sleeping drivers).
Anyone who lives in a city with a heavy taxi presence already isn’t going to be deterred by these challenges, just the Americans and some other drivin heavy countries will need to adapt.
I don’t understand how someone who could solve the self driving problem could solve the “plug the car in via simple robotics after it parks itself” problem.
They are also able to park near the city where parking is cheaper, though I guess it could also turn out like it does in China (if anyone is out at 2AM and see all the taxis pushed over to the side of the road with sleeping drivers).
Anyone who lives in a city with a heavy taxi presence already isn’t going to be deterred by these challenges, just the Americans and some other drivin heavy countries will need to adapt.
There have been modern electric cars available for, let's say, 15 years now and I haven't seen a single practical automated charging system.
There have been plenty of proposals, like Tesla's snake robot or inductive undercar charging plans, but no actual production-ready solutions. But even if you had the magical autocharging robot that will not help with gasoline powered vehicles (like all of Waymo's) and you still would need dedicated real estate and charging stations to recharge these vehicles.
There have been plenty of proposals, like Tesla's snake robot or inductive undercar charging plans, but no actual production-ready solutions. But even if you had the magical autocharging robot that will not help with gasoline powered vehicles (like all of Waymo's) and you still would need dedicated real estate and charging stations to recharge these vehicles.
I’m pretty sure once self driving cars are really a thing for all consumers, it will coincide with the move over to EVs en masse (2030?).
The only reason no one finds automated plug-in charging necessary right now is simply because those cars don’t really need it yet.
The only reason no one finds automated plug-in charging necessary right now is simply because those cars don’t really need it yet.
Aren't these all issues that car rental agencies have?
Somehow they turn a profit on renting you a car for $80/day which I imagine is less than a tax takes in for fares per day.
Somehow they turn a profit on renting you a car for $80/day which I imagine is less than a tax takes in for fares per day.
They do, but they buy or lease ordinary unmodified vehicles for a year or two then flip them at auction. Zero overhead here except for the buying/disposing of vehicles.
$80/day is probably a decent estimate, but once you include the optional insurance upsells and other add-ons (full tank of gas?) I bet they make way more than $80/day per car.
They also have very limited liability for what the driver does in their car while Waymo basically has unlimited liability.
$80/day is probably a decent estimate, but once you include the optional insurance upsells and other add-ons (full tank of gas?) I bet they make way more than $80/day per car.
They also have very limited liability for what the driver does in their car while Waymo basically has unlimited liability.
I mean, some of them do. Hertz's recent bankruptcy[1] and outlandishly...audacious[2] stock offering is one counterpoint. But it was a crazy year+
[1] https://www.marketwatch.com/story/bankrupt-hertz-shares-soar... [2] https://www.bloomberg.com/news/articles/2020-06-18/what-it-m...
[1] https://www.marketwatch.com/story/bankrupt-hertz-shares-soar... [2] https://www.bloomberg.com/news/articles/2020-06-18/what-it-m...
I’m sure the costs Uber offer are accurate and not incredibly subsidized: https://www.uber.com/us/en/drive/vehicle-solutions/
So $260-270/week? Not sure if they have any extra fees applied, however.
So $260-270/week? Not sure if they have any extra fees applied, however.
Did you miss the /s tag?
You are actually trusting Uber's calculation, especially when they recently had to spend $250 million[1] to try to recruit new drivers?
[1] https://www.marketwatch.com/story/uber-spending-250-million-...
You are actually trusting Uber's calculation, especially when they recently had to spend $250 million[1] to try to recruit new drivers?
[1] https://www.marketwatch.com/story/uber-spending-250-million-...
I’m not sure what your piece of data has to do with how much Uber says you can rent a car for to drive for them. It seems like you want to really really believe that rental car prices advertised anywhere are complete fiction.
I clicked on the link and got "Sorry, there are no offers available in your area. Please try again later".
So their rental business is limited, at best.
So their rental business is limited, at best.
Now is a horrible time to rent a car since there is an industry wide shortage.
> Did you miss the /s tag?
My browser’s search function spotted only one instance of that string, and it was this question.
My browser’s search function spotted only one instance of that string, and it was this question.
All of these maintenance issues exist with regular taxi fleets driving equivalent miles, though. The main difference is the taxi driver salary vs the cost of the self-driving tech plus some support and extra interior cameras and this should be in favor of self-driving cars. From the customer's point of view, self-driving cars might be slower but safer.
It depends on who owns, maintains, and is paying for the vehicle and license/medallion to operate.
The whole system - whether Uber/Lyft drivers who "relied on their unfamiliarity with capital depreciation calculations"[1] or traditional cabbies who, let's be honest is not a glamorous job, is tenuous at best. Cutting out the lowest cost factors and replacing them with technology and also incurring the fleet costs doesn't really seem like a great business.
[1] https://pluralistic.net/2021/08/10/unter/#bezzle-no-more
The whole system - whether Uber/Lyft drivers who "relied on their unfamiliarity with capital depreciation calculations"[1] or traditional cabbies who, let's be honest is not a glamorous job, is tenuous at best. Cutting out the lowest cost factors and replacing them with technology and also incurring the fleet costs doesn't really seem like a great business.
[1] https://pluralistic.net/2021/08/10/unter/#bezzle-no-more
Yeah, can't agree more. The main thing stopping self driving taxi services is maintenance and cleaning.
I work in the space - whoever wrote this clearly isn’t familiar with how Waymo, Cruise, Zoox, etc. actually function. Just going to address a few particularly egregious takes:
>Waymo relies on 3D digital maps.
The cost of HD mapping is vastly overstated in popular press and blogs. It’s not at all a limiting factor. You can update it using the same cars that are part of your service. I've seen hard numbers that back this up.
Google’s unable to keep street view up to date because they have a fleet of hundreds of mapping vehicles spread throughout the entire world, vs hundreds of cars in a small service area that Waymo has.
>Waymo is LIDAR depdendent.
This has been discussed to death - the weather point applies to both cameras and LIDAR, and you can train ML models to deal with noise from weather the same way you can with models running on cameras. Cost has also been addressed a million times before- it’s coming down drastically.
>Waymo is GPS dependent
This one is particularly inane. HD map based AVs localize themselves to the HD maps with their sensors - they’re not actually GPS dependent. The 2-4m of accuracy GPS has is not enough for AVs.
>Internet connectivity issues >Waymo depends on 3D HD maps. Therefore, the vehicle needs to have onboard downloaded maps at all times. Waymo must download these maps in realtime, or you have to download these 3D maps on your Waymo vehicle before you visit a new Area.
I don't know why the author thinks you'll be able to drive robotaxis from city to city. Regardles, AV companies design their cars so that they don't need a continuous internet connection to function. They know cell connections can be spotty or blocked while going through a tunnel. See https://readwrite.com/2017/01/11/google-waymo-security-tl4/
>No Viable Business Strategy >Waymo’s primary business strategy is to sell its technology to the vehicle manufacturer.
This is just wrong and I have no clue where the author got that from. The model is robotaxis, and maybe partnering with trucking fleet operators to outfit their cars.
>Simulated driving is not real world driving.
The data stuff has been discussed before in detail but suffice it to say AVs generally aren’t a data problem - they’re a robotics problem.
>To be successful Waymo has to work very well not only in the USA but also in Canada, Mexico, etc
The revenue from ridesharing services is hyperconcentrated in a few dense urban areas in the US. If you got a robotaxi service working well in just a few of those tiny areas, you could capture a huge amount of revenue to fund tackling other areas.
>Waymo relies on 3D digital maps.
The cost of HD mapping is vastly overstated in popular press and blogs. It’s not at all a limiting factor. You can update it using the same cars that are part of your service. I've seen hard numbers that back this up.
Google’s unable to keep street view up to date because they have a fleet of hundreds of mapping vehicles spread throughout the entire world, vs hundreds of cars in a small service area that Waymo has.
>Waymo is LIDAR depdendent.
This has been discussed to death - the weather point applies to both cameras and LIDAR, and you can train ML models to deal with noise from weather the same way you can with models running on cameras. Cost has also been addressed a million times before- it’s coming down drastically.
>Waymo is GPS dependent
This one is particularly inane. HD map based AVs localize themselves to the HD maps with their sensors - they’re not actually GPS dependent. The 2-4m of accuracy GPS has is not enough for AVs.
>Internet connectivity issues >Waymo depends on 3D HD maps. Therefore, the vehicle needs to have onboard downloaded maps at all times. Waymo must download these maps in realtime, or you have to download these 3D maps on your Waymo vehicle before you visit a new Area.
I don't know why the author thinks you'll be able to drive robotaxis from city to city. Regardles, AV companies design their cars so that they don't need a continuous internet connection to function. They know cell connections can be spotty or blocked while going through a tunnel. See https://readwrite.com/2017/01/11/google-waymo-security-tl4/
>No Viable Business Strategy >Waymo’s primary business strategy is to sell its technology to the vehicle manufacturer.
This is just wrong and I have no clue where the author got that from. The model is robotaxis, and maybe partnering with trucking fleet operators to outfit their cars.
>Simulated driving is not real world driving.
The data stuff has been discussed before in detail but suffice it to say AVs generally aren’t a data problem - they’re a robotics problem.
>To be successful Waymo has to work very well not only in the USA but also in Canada, Mexico, etc
The revenue from ridesharing services is hyperconcentrated in a few dense urban areas in the US. If you got a robotaxi service working well in just a few of those tiny areas, you could capture a huge amount of revenue to fund tackling other areas.
Thank you for an informed take. I agree completely. There is no fundamental reason why Waymo can't work. Sure, there are challenges, but is servicing cars REALLY something Google can't solve?
How about servicing phones or google accounts?
LIDARs are not inherently expensive. The volumes are tiny. Continental was prepared to put a good, low-cost automotive LIDAR into production. They're an auto parts company. Someone has to order a few hundred thousand units.
Good color TV cameras cost thousands of dollars 25 years ago. The one in your cell phone costs about $10 and is much better.
The business model problem is real. A big disappointment has been the lack of self-driving shuttle buses. The cost problem isn't so bad, and at moderate speeds the sensing problem isn't that bad. Shuttle buses are expensive from a labor perspective - they don't carry that many people, and you want a lot of little ones running frequently all day and night. servicing airport parking lots and such. Not a big bus every half hour. There are a few companies selling them, and they work OK on fixed routes at slow speeds. But they're not selling well. There are a few in various places around the world, but not many.
A big name not well known seems to be set to change this. New Flyer is the largest big-bus builder in North America, and they have come out with an automated model.[1] This should be interesting.
[1] https://www.newflyer.com/bus/xcelsior-av/
Good color TV cameras cost thousands of dollars 25 years ago. The one in your cell phone costs about $10 and is much better.
The business model problem is real. A big disappointment has been the lack of self-driving shuttle buses. The cost problem isn't so bad, and at moderate speeds the sensing problem isn't that bad. Shuttle buses are expensive from a labor perspective - they don't carry that many people, and you want a lot of little ones running frequently all day and night. servicing airport parking lots and such. Not a big bus every half hour. There are a few companies selling them, and they work OK on fixed routes at slow speeds. But they're not selling well. There are a few in various places around the world, but not many.
A big name not well known seems to be set to change this. New Flyer is the largest big-bus builder in North America, and they have come out with an automated model.[1] This should be interesting.
[1] https://www.newflyer.com/bus/xcelsior-av/
Lately, I have been seeing lot of youtube videos of unruly passenger boarding uber. I wonder what happens in that scenario. It seems like it is common thing during weekends in big cities. Is car aware of the fact that some body puked or took a dump inside it?
Unruly passengers without supervision who destroy properties might just be the reason it won't be a viable business.
Unruly passengers without supervision who destroy properties might just be the reason it won't be a viable business.
There are cameras inside of the cars in AVs that face towards the passenger area - it's trivial to have someone check for crud in the back of the car before rides and send it back to the depot for cleaning. And automating that via ML would be easy considering the simple environment.
Doing that on interior camera that the cars are sure to have … that is some evidence that goes way beyond “he said she said.”
I hate that HN's voting system allows low-effort, factually incorrect, borderline spam articles like this one to get boosted right to the top because the title is catchy and supports people's biases.
I've always hated any karma-based forum. I just don't participate in it. I've gone as far as to block karma displays on Reddit and HN through element hiding and running custom userscripts. It allows me to type without trying to please random nobodies.
I never trust articles without a byline.
Who wrote this? Who funds them? How can I know that they aren’t introducing bias to their assessment in their own self interests?
Who wrote this? Who funds them? How can I know that they aren’t introducing bias to their assessment in their own self interests?
I am curious why this was flagged?
This wasn't convincing. Sweeping statements about viability and scaling do not constitute an argument, I think.
If I were a city, I'd want people spending all of their money on food, drinks, music and other attractions, right? Not necessarily parking and taxis. This might be off base, but I feel like some cities would happily help pay for the cost of having the mapping done and kept up to date.
The other points that the author raises are obvious things known for a decade, and Waymo doesn't even have to solve them, unless it really wants to.
- The cost of Lidar has fallen a lot already, and is likely to continue to do so.
- Waymo doesn't have to manufacture cars at scale. Carmakers would love to license FSD tech from Waymo and sell premium cars. I would happily buy such a car for $100k.
- Uber/Lyft will fight over themselves to license the FSD tech, Waymo has no need to build its own taxi services.