I think folks are missing the key here. If waymo offers its own app and there aren't any available vehicles when a user wants a ride, user trust is lost, and people don't bother trying it next time.
With the Uber app, Uber will show both options when they're both available. It allows waymo to offer a lower SLA/SLO while still gradually fielding a service. It also allows for easier fallback if a waymo vehicle is taking too long to arrive.
I see the benefits from both for now, but I’m not capturing how they could agree on the long-term negotiation.
There is so much upside in a few months for only one side of that deal… That was always the issue with Uber promising self-driving cars: human drivers have an unknown, uncontrollable deadline.
They could buy a lot of Teslas or partner with a Waymo competitor, but I doubt anyone with a working automated fleet will need Uber’s human drivers. Acquiring riders is cheap—good PR, a handful of promo codes: branded cars all over the city will do the rest.
That partnership makes sense in the transition period when Waymo can’t buy enough cars to scale for football matches, handle edge cases, and build a larger fleet operation team. I think the individual Uber talent (operations, software, etc.) are great. Their experimentation team, for instance, is 100% something I’d recommend acquiring wholesale. Even with a thousand better options at Google, I’m sure their back-end solutions are work acqui-hiring. But Uber itself? The confusing everything app? It’s a bad brand whose toe-stepping values are misaligned with what anyone is trying to do (except maybe Tesla) and with a ton of tech debt. I know there was a lot of effort to escape it, but I don’t see great synergies. With Booking, maybe?
Also Tesla has backed off on claims their cars will be self driving without a human safety driver. They’ll get there eventually but Waymo is far ahead of them on actual deployment of this tech.
With the boss mentioning multiple times that he wants end to end AI - ie. pixels in, steering angle out, I think this may change quickly.
Nobody else has managed to get end to end AI working, but this sort of stuff scales with training data, and tesla has the opportunity to collect far more than any other manufacturer, with possibly 1000x the number of cars collecting video data of the next closest competitor.
And the thing about end to end AI is I suspect that if anyone can get it working at all, it may not have the issues with the long tail of unusual situations that other approaches have. Things like "Someone has drawn a picture of a bear on the road in chalk, should I drive over it or around it?" might be answered better by the end to end system than a rules based system.
If all of those things work out, then Tesla will win this race. Otherwise, they'll lose.
As you say, no one has figured end to end out. The way AI works some tasks get figured out quickly and some take a very long time. Elon has been talking about an end to end system for a very long time, and I’ve watched every lecture from Andrej Karpathy on their architecture. I like their plan, but they have been working towards end to end for years and I see no reason to believe anything will “change quickly”. AI is very much a “don’t count your chickens before they’re hatched” kind of technology. People can promise anything about AI, because when it works it feels like magic. But it’s not magic, and we can’t expect every goal to be met just because someone is using AI. Nobody knows what challenges lie ahead for an end to end self driving system precisely because no one has done it. And it should be noted that with their “hydra net” architecture it is relatively end to end but it is not a single large network, but an engineered assortment of many small networks. Quite a lot of human choice goes in to decisions around how to design that, and this also leads to edge cases. But even more general systems like GPT-4 suffer from issues with edge cases.
We still don’t even know how plausible a camera based system is. Human eyes and the human brain are so much more advanced than a computer and cameras, it may not follow that “images are all you need”.
Very interesting question. Certainly I’ve never heard them discuss this in any of the AI system lectures. It’s always stated as “we know humans can drive using only our eyes so we have an existence proof”.
Plus, a human could drive in a variety of conditions but there’s just so many possible scenarios the car could encounter, and humans crash too, so the real question is what is the relative performance between a human and the machine across thousands of potential or actual crashes. Kinda hard to measure.
True, true. We may see a change of heart for them if solid state LiDAR ever actually hits the market, but last I checked Velodyne/Ouster keep announcing things and then letting them disappear. I did see them saying they have samples available for 2025 model years cars so perhaps Tesla could be testing with that now. I don’t see them deploying mechanical LiDAR but solid state is similar enough to a camera they may do it.
Edit: then again for such a critical component Tesla might want to make their own solid state LiDAR, or at least license someone’s technology so they can produce them on their own terms.
LiDAR also has FoV issues and is sensitive to lens debris among other things. Also, the addition of another layer of high resolution data may make it impractical for the compute power requirement to synthesize it all.
Well they currently devote a lot of compute power to building a 3D occupancy map of the world from cameras so it’s possible that fusion with LiDAR could be done with less compute. Also their computers keep getting more powerful. But generally you’re correct, we don’t know how it would go.
Lidar only works to a certain subset of distances, so you need cameras anyway to infer long distances beyond the scope. Also, since distance alone isn’t sufficient, you need cameras anyway for inference of other things.
Lidar works up to 300 meters. That’s plenty for highway driving. Everyone using lidar also use multiple cameras so they are getting best of both worlds (using sensor fusion).
Compute for full self driving hasn’t been solved, I hope it is, and I don’t know which strategy will work. It’s entirely possible LiDaR makes it work better but requires more GPUs on vehicle :)
This is fascinating, obviously it makes a tonne of sense for uber as a platform to offer waymo's autonomous vehicles. The first question I had though was "who extracts value here". Does Waymo get to demand human-driver level prices and therefore capture loads of value or does Uber actually capture the value because they're where the customers are. But then I realised... it doesn't matter. Waymo is part of Alphabet, and Uber is sitting at around a $70Bn market cap, surely if this is starting to show signs of success Alphabet just swoops in and buys up Uber, vertically integrates, gets a great platform, great customer base, has all the customer service infra that google sucks at and they just slowly phase out the human drivers by rolling out autonomous vehicles as they're legalized in each jurisdiction.
I am pretty confident google could spin up an org to replicate the uber app in a few months/maybe a year. However, Uber obviously cannot spin up an org to replicate self driving vehicles in a few months. By that logic I would say Waymo has all of the power since they could develop their own app and undercut uber on pricing. The customers would instantly leave as it is a race to the bottom with ride share.
Spinning up the Uber App is most likely trivial.
Spinning up the legal structure, figuring out all the local regulations and getting permits (it's not 2010 anymore! Many countries and cities regulate ride sharing pretty heavily) in a sane timeline is close to impossible.
And it might be not something that an Alphabet company wants to do risk/ reputation-wise. Better to offload bad press (both related to ride sharing in general, and autonomous vehicles) onto an entity which is already "hated".
It's not like Waymo/Uber can roll in with autonomous taxis in every city just because Uber figured out the local regulation - many cities would have explicit regulation for anything driverless. I think it's the later point - they want to stay in the shadows until they are super confident. I wonder what will happen to Uber eventually though.
It'll be interesting if Cruise ends up partnering with/acquiring Lyft. Although Cruise seem to be still betting on their Origin vehicle.
give me 3 good software developers and I can give you an Uber app maybe in 3-6 months? but what I can't give you is uber app being on every iPhone in the country.
> you can, but I am used to using uber app to call an uber.
yes, waze had 100M+ users and still failed to have a successful launch with waze carpool.
part of the problem is Google is too successful.
If a business makes USD 10M a year and takes up a month worth of attention every year from the Alphabet CEO, they will probably shut it down before we can say Google Product Graveyard.
When I worked at Google our team came up with a data product that would have been pure margin and $75m/yr with customers lined up -- and I couldn't find a VP interested in cashing the checks because it wasn't on their annual OKRs.
Google is a great company but it has never been truly hungry because it has never faced an existential threat.
It's not just OKRs. That $75m/year consumes attention all up and down and across Google. Lots of things that are slam dunks at a small or medium company are just more trouble than they're worth at most really large firms.
What's an example of something for consumers that Uber 2023 does a lot better than Uber 2010, if the drivers behave entirely cooperatively? (since we're talking autonomous vehicles here)
route planning / fleet management i think is a huge optimization problem. perhaps you can cut it for MVP but it would require having a big multiple of drivers to provide the same service
another big deal is gps accuracy in cities, which is a pretty non trivial problem but at this point a consumer expectation https://www.uber.com/blog/rethinking-gps/
Ahhh yep. I remember Uber 2010 unopinionatedly just setting the pin directly on whatever the reading was from GPS. Stale or inaccurate GPS readings led to so much confused communication between drivers & passengers.
That's a good one. I remember lots of shenanigans about putting pins next to airports & then contacting the driver to get picked up at the airport on early Uber.
Not sure this particular one would affect Waymo if they were going it alone, since their rollout to each service area would involve designation of pickup & drop off points at airports and collaboration with airports, which they'll have to do either way, but it's definitely a significant difference between Uber 2010 & 2023.
Uber is losing money and is frequently accused of undervaluing the deprecation and maintenance of the drivers' vehicles.
If Google was to enter a market at a loss in an effort to use its market dominance in Android to drive out a competitor that would potentially be something that would be considered anticompetitive and US regulators would likely be interested in it.
I don't believe that Google can compete with Uber losing money as it is nor that Google has an appetite for purchasing the necessary vehicles and maintaining them outside of a few test cities for a technology demo. Having driverless taxis also means that they would need to do a better job with end user support (compare the urgency of "help, I'm locked out of my email" and their current resolution time to "help, I'm locked in a car and can't get out!").
> Spinning up the legal structure, figuring out all the local regulations and getting permits (it's not 2010 anymore! Many countries and cities regulate ride sharing pretty heavily) in a sane timeline is close to impossible.
This is nonsense. Google has more than enough legal resources to handle this. So much so that it would be nothing more than a regular 3-6 month project that would be assigned a small team of lawyers inside their org. It's no different than any other regionally-regulated product launch at large companies.
Not only that but Google has enough money that they could just hire away some folks that work at Uber/Lyft to get a huge jumpstart on any such project. Even if they didn't need such expertise they'd probably still recruit and hire those people just to be safe; to get the perspective of people with experience.
Big companies have big legal organizations and they don't actually operate that differently from the IT orgs that so many of us on HN are used to. For example, they would probably divide up the task on state-by-state basis with lawyers certified to practice law in each respective state assigned as SMEs or to-be-SMEs. They would do their research and for bigger states they might even have county-level SME lawyers or at the very least paralegals that focus on that particular region. Big cities would probably be assigned their own teams of SME lawyers as well.
It's like any given complex programming task: Divide and conquer. Rolling out a new highly-regulated product across a large region is a highly parallelizeable procedure.
Evidence suggests that none of this is true. Just look at how Google Fiber wasn't able to navigate the myriad regional regulatory roadblocks thrown up by local governments when lobbied by incumbent ISPs. It's reasonable to expect exactly the same scenario to play out if Google tries to make a competitive ride sharing service from scratch.
That's a pretty good example. But in the case of Fiber, part of the problem was that other ISPs own/control infrastructure (like poles) to which they're legally obligated to give competitors access...and they would just not comply, or drag their feet on it. Fiber physically couldn't execute because of this. Of course there was corruption involved, as you pointed out, but in the case of ride share apps I can't immediately think of anything analogous, where Uber could physically prevent Waymo from competing with them.
Google has tons of legal resources for their current business. But that doesn’t mean they can spin up the legal docs and relationships needed to run an Uber-like business at the drop of a hat. The type of risk that Google’s lawyers are used to mitigating are simply not the same, and their lawyers are for the most part not the type of lawyers that you’d want to set up a transportation company with international operations. There are some relevant lawyers there, but the vast majority are specialized in software/cloud, not wheels-on-the-ground.
Of course, it would be easier to put together the necessary legal structure now that Uber and Lyft have already both done so. It’s easier to be a follower than a leader. But it would still be much harder for Google than building a mobile app like Uber, which would be right up Google’s alley.
As others have pointed out, Google could scoop up tons of riders in a heartbeat, by announcing a ride share platform publicly (which would get tons of news coverage) and by advertising the feature within Google Maps, Waze, etc.
Google only needs to put together a one-sided marketplace, where Uber/Lyft had to build a two-sided marketplace. And Google already has a connection to billions (at least hundreds of millions) of potential riders.
It would also need to invest in the vehicles and the technology and the maintenance and the garage to store them in and pay the insurance on the vehicles.
All this to try to compete with Uber (which isn't profitable) in a few areas where the legal, political and climate ("we have cars in Chicago but they only run from March 1st to November 1st") options are favorable.
The marketplace is not the biggest issue that Google would need to overcome.
I don't know a lot of people who love Uber, but I know a lot of people who are only ever going to use one ride sharing app. If Waymo's app is limited to a 180 sq. mile territory and only half of their trips are in that territory, they won't bother. The first time the Waymo app tells them they're going outside their territory, it will be back to Uber forever.
> This is nonsense. Google has more than enough legal resources to handle this.
I don't think this is even close to correct. Lawyers were hit especially hard in the January layoffs, way more than 6%, and many teams lost their dedicated lawyers entirely (e.g. my team had one lawyer laid off and the other was reassigned to a basket of products instead of just our own).
Google has enough resources to in theory thrive in any business they attempt. Yet, they fail in so many businesses they attempt to get into. https://killedbygoogle.com/
> Big companies have big legal organizations and they don't actually operate that differently from the IT orgs that so many of us on HN are used to.
Well, if you are referring to IT organizations in firms in the tech industry where that’s the core business (i.e., a value center) you are wrong, they are run more like IT orgs in non-tech firms (cost centers), vigorously minimized to be a bit smaller than they should be to handle routine day-to-day needs effectively, with new, emergent, and non-routine work outsourced.
Sure, technically that's correct. But it's google we're talking about. The company with a tremendous list of failures that should have been easy slam dunks.
> Spinning up the legal structure, figuring out all the local regulations and getting permits
Is experience doing this with human drivers all that helpful for building a driverless service? I'd expect the legal and regulatory questions would be very different?
> figuring out all the local regulations and getting permits (it's not 2010 anymore!
Hm. This is like --
It's not 1066 any more, there are property rights now.
The original regulatory capture.
Step 1: Ignore the law; establish a fait accompli.
Step 2: Sanctify the new status quo.
Sad for a rule-follower to figure out that this is how things work.
On the other hand, I guess rule-following is low risk. For each William the Conqueror there were thousands of warriors who just got killed on the battlefield. And I'm sure the guy had to work really hard, you know, like nights and weekends.
> Spinning up the Uber App is most likely trivial.
as someone who has worked at Uber, this is a gross misstatement. It takes a lot to run Uber, far less even to run it at scale globally and with viable unit economics
Realistically though they don’t have to figure out the legal problems immediately, they could and likely would start in their own backyards and expand from there
I think you're underplaying the difficulty of what Uber has built. Uber is an operations-heavy business. Determining pricing strategies, handling incidents, and working with local governments is at the core of its service. Google has traditionally shied away from businesses that require big operational investments in real-world assets the way that, say, Amazon or Airbnb have. These take time to build properly, and is a big piece of the puzzle that I can't see Waymo wanting to take on. It makes sense for them to want to partner, especially now that Uber has spun out Aurora.
> I am pretty confident google could spin up an org to replicate the uber app in a few months/maybe a year.
Like how they can "just spin up" a messaging app?
Building companies and customer bases is really, really hard. Microsoft tried like three times to build a Github competitor. They still have at least one of those attempts (Azure DevOps). They still bought Github. An interesting product technically, sure, but: its where the people are at. It was worth it.
Spinning up a messaging app without a compelling differentiation is hard. It's much easier with a compelling differentiation. Waymo could just set prices at half of Uber and boom instant market share.
Using Google's dominance to expand into other sectors (at a loss) is a perfect example of anticompetitive practices that regulators would likely take a keen interest in investigating.
Using Android's market share to push or encourage the use of a Waymo robot taxi booking application that is in competition with Uber and Lyft might be seen as anticompetitive.
Other comments elsewhere are "No but if you're Google you can put it on literally every android phone quite quickly." and "All Google would need to do is push a notification to every Android phone that has Google Maps (so, all of them) that says "book your next trip through Google maps and the first ten rides are free!" and they're in business."
That sort of approach would very likely be examined by regulators.
Followed by a slap on the wrist of a fine? How many tens of millions of dollars do you think the TAM of NYC taxi service is, vs a single digit million dollar fine from regulators?
Most days, fines are just the cost of doing business. They cut into your profits, sure, but as long as profits >> fine, just pay the fine and keep doing it.
Instant market share if they're willing to lose 2x the money that Uber is losing. It's not like drivers are going to be falling over themselves to sign up if it's a 50% paycut.
The power of Uber doesn't lie in it's app,
Google famously has been forever trying to have a chat app and we all know how many times it's failed.
Then comes Google+, Google's attempt to get into social networking.
Unless Google came up with something novel or something different from incumbents it never ever succeeded.
Any SWE worth his salt can build a ride booking app in a month or two, but the traction and confidence in using that app both by riders and drivers is what drives its success.
I'm not sure what else Uber could possibly have. Name recognition? More people know Google than Uber. Install base? More people have Android phones than have the Uber app. Drivers? Uber is trying to eliminate its drivers, as this deal shows. Uber is defenseless against Google.
Name recognition and installed user base in a 2-sided marketplace. That's huuuge.
Google owns the action "to google something". Uber owns the action to "get an Uber".
Also, don't underestimate the value of a working business. Google has only ever been successful with a search engine and an ad platform. They've never been success in a consumer product/service. I don't think their corporate structure and business culture makes them likely to succeed in B2C.
Uber has it's flaws, but they've been very reliable as a consumer. And have responsive and helpful support. Imagine having to rely on Google to provide support to a stranded user or resolve a dispute.
The idea is to make it a 1-sided marketplace. Uber's entire business model is focused around squeezing vendors - now they think they can squeeze Google if Google becomes their vendor instead of a bunch of independent contractors who are legally inhibited from collective bargaining? Uber has no idea how to be successful when the seller side of their marketplace is organized and plays hardball at the negotiating table.
Tesla and Waymo are arguably pretty close in self-driving technology, so there's already another player. Uber could just buy a bunch of Tesla's once their tech is good enough for autonomous driving.
The car companies will do everything to prevent Google from having that monopoly. E.g. Google could be forced to license their software. Then, Uber gets their 2-sided marketplace back.
That's what they thought about Google Video, but then Google decided to buy Youtube. Technical advantage (which Google has in strides) does not always translate to business advantage (which Uber has done remarkably well for itself).
I think you’re right here. You can recreate the technical aspects of YouTube relatively easily. You can’t recreate the brand.
Just because it’s big doesn’t mean Google will inevitably succeed. Look at Google+, Allo, Duo etc etc. to my mind there’s no guarantee a Google rideshare app would capture mindshare.
YT was relatively cheap and much stronger network effects. Uber does not.
Google is so much more integrated in to people's lives than 15+ years ago.
Google Maps + Waymo or "Google Taxi" would have millions of customers immediately.
What Google did to Yelp is what might happen to Uber. I think the main point thing stopping is what others have pointed out. Is it worth the reputation? Instead of running their own service just take a nice cut from Uber and let them deal with bad headlines. Especially when the headlines will read Uber "fires" poor single parent uber drivers in favor of robot cars.
Yes, Google can copy Uber, but can they make it successful?
Uber has a lot of users and drivers, and Google has to convince them to use their app instead of Uber or one of the other existing apps in this space.
Getting into a new market didn’t work for videos so they bought YouTube, it didn’t work for their social network, and it didn’t work for a ton of other cancelled projects.
Replicating Uber is hard because it's a 2 sided market - you need to convince both users and drivers to sign up and neither will if the other isn't there already. But that's not a problem for Waymo because they control the driving side. They simply need to get users to sign up.
When you have something unique and useful, people will sign up.
> I am pretty confident google could spin up an org to replicate the uber app
Google has already done this. Waymo One is the name of Google's app, and it's live and serving customers already in Phoenix, AZ. Completely driverless - nobody but the passenger in the car.
You're pretty wrong. No organization can just spin up large international companies. I don't know where you'd get such an idea from. Even the technical parts of Uber alone would take years to replicate. The app frontend? Sure, with manpower you can get that out in a few months. The backend, supporting infra, legal systems, customer support, relationships with users, restaurants, etc? Years to decades. Google could never do it, they are too big to move fast. Think, when is the last time Google successfully launched a project of even a fraction of the scale of Uber?
Uber had to fight tooth and nail to get permission in numerous countries - Alphabet will not want that battle. Add to that battle they suddenly want to be autonomous too and they will really struggle rolling out. Also, would you trust an Alphabet Uber? It'd be torn down within a year like usual, based on precedent - looks at Stadia controller annoyed.
... and the government would (again) look into Google using anticompetitive tools to expand its market reach and stifle other companies in the area.
... having a "use Google/Waymo ride hailing" only in a few cities would lose a significant part of the network effect that Uber has (you can likely use Uber in any city - not just a limited few).
> I am pretty confident google could spin up an org to replicate the uber app in a few months/maybe a year.
The app, sure. But Uber also has the aggression necessary to expand across the world and bulldoze/ignore/skirt/rewrite local taxi laws fairly successfully.
There is no way waymo/alphabet would manage that. They'd nicely ask permission, and in most cities permission would be denied by incumbent taxi operators and uber drivers who don't want their jobs taken.
There's no network effect from having the Uber app. All Google would need to do is push a notification to every Android phone that has Google Maps (so, all of them) that says "book your next trip through Google maps and the first ten rides are free!" and they're in business.
I mean, I use mostly Lyft in the US. And if there were another app or two that had a reputation for being cheaper or better service even if they didn't service everywhere, I'd be installing that in a minute. Brand has some value and I don't want to be installing one or more apps for every city I visit. But it mostly works when someone is as good or better for most purposes.
The backstory that seems to have been forgotten here is that self driving cars are still a money bleeding endeavor. It's why Uber sold its own self driving division and it's something Cruise has said in the past still needs to be figured out.
Uber has pivoted to doing partnerships with any self driving player that is looking for a customer base. Waymo doesn't have the tech nor the desire to figure out logistics to compete with Motional or Cartken in the delivery space, for example. None of these players can instantaneously ramp up to millions of vehicles on the road; they physically don't have enough hardware and the age of money burning for growth at all costs is behind us. I can't imagine Waymo has bigger utilization than Bolt, or even Alto.
The way I see it, Uber is more like McDonalds: "easy" to copy from conceptual perspective but also a globally recognizable brand with an undeniably strong customer acquisition arm.
Waymo still has a lot to prove to themselves in terms of ROI. Customer acquisition would be even more expenses on top of the already expensive tech, partnering with Uber to offload these costs to a proven customer acquisition player makes sense for them.
> The backstory that seems to have been forgotten here is that self driving cars are still a money bleeding endeavor. It's why Uber sold its own self driving division and it's something Cruise has said in the past still needs to be figured out.
Uber sold their self driving division because they killed someone in Arizona due to negligence. And then Anthony Lewandowski happened. Those two incidents complete deflated Uber's hopes to have autonomous vehicle technology. I'm pretty sure they would still be pursuing self driving if not for those incidents.
I think the more likely here is that Uber sold self-driving group (alongside other "big bets" like flying taxi research group) because of existential threat to the business during the start of COVID-19 which made them focusing on the main business (+eats) and sell parts that have no profitability in sight.
That's exactly right. Uber self driving was burning through the tune of billions of dollars a year when covid reduced revenue by the tune of 70%. Not exactly sustainable w/ only a few billion dollars of cash on hand.
As terrible as a pedestrian death is, the reality is there's news about Tesla accidents and near-accidents on the news all the time and people still love the brand. Sometimes it's easy to forget that Uber/Amazon/M$/etc hating is not as widespread in the general population as it is among tech bros.
As for Otto, my understanding is Uber pivoted similarly on the freight side, with a partnership w/ Volvo Autonomous Solutions.
The self-driving division was done to attract investment because investors thought it sounded cool. It would have been bad for Uber's business to have them - Uber relies on not owning cars, which moves all the maintenance to the driver-owners. If you own a self-driving car you still have to maintain it.
> The backstory that seems to have been forgotten here is that self driving cars are still a money bleeding endeavor.
They also don’t work. At least not by any sane definition. Sure, if you’ve mapped the area in advance, nothing changed,the weather is nice and nothing unusual happens all might be okay but otherwise it’s a shitshow.
Human drivers are cheaper and more versatile. There is no point in this other than trying to persuade the markets Google hasn’t spanked millions on a dead end.
I wouldn't even get in a self driving taxi if it was free because of how cheap human drivers are. For me, to use a self driving taxi it would have to be so safe that we would be moving towards banning human driving anyway.
Otherwise, it is about as bad of a risk to reward bet as I can think of compared to human drivers.
If no one steps up to challenge Waymo in this space in a reasonable timeframe, Waymo wins. If they do, Uber wins.
Uber's play is as a platform. They win if there are several SDC providers (with working autonomy). As an analogy, they're Android; they want to own the network and have multiple operators plug into the Uber app, while the hardware makers don't want to do the work to spin up an expensive 2-way marketplace. Waymo is playing at the iPhone game. They want to be the biggest market operator, and can get away with not owning the top of the funnel for the time being. If other operators fall far behind in progress, Waymo will spin up the investment to vertically integrate and try to siphon Uber customers off that way.
The game theory isn't really winner take all. In a two-sided market like Uber (drivers and passengers) it's nearly impossible to bootstrap competition, but with autonomous fleets it's just a matter of capital. You can dump 100 vehicles into a market and be a local player pretty quickly, and the between-city efficiencies aren't super important.
Developing the tech is getting cheaper too (sensing hardware improvements, ML improvements) so the first-mover advantage likely won't lead to stable moats. Competition will be more like Apple/Android where constant improvement (probably on cost) is necessary to maintain a lead.
If another company solves self driving after Waymo, they can still enter the market. I would usually give new, cheaper competitors a try. Who wouldn't?
Uber is kind of an exception, i don't use their services because i consider them scum.
> I would usually give new, cheaper competitors a try. Who wouldn't?
Just want to point out that the failure case of self driving (if this product fails I crash into another car while going 80 mph) makes the race to the bottom slightly less plausible.
I just want to chime in to say that it is ridiculous that such information is not so easy to come by. Billion-dollar corporations should not be allowed to keep their external investments so opaque! Especially businesses that are publicly-traded.
Also: I don't give a damn if the business isn't publicly traded either. If they're that big they have enough influence over any given market that the public should have the right to know what they're up to.
That is not surprising, or in question. The claim I am questioning is:
> Alphabet has a large stake in Uber.
Since Uber is a publicly listed company, its top owners are public information. Which show Alphabet to not have a significant portion of ownership in Uber.
Sure, but there's massive advantages to bringing it in house. The only reason I could see for them not doing it is that as autonomous vehicles are deployed widely there's going to be a massive backlash from current uber drivers and strategically it might be good for Alphabet to be able to use Uber as a heat shield.
Uber have always wanting to go autonomous cars. They poured a lot of money to try to develop them in-house and I think that they were too optimistic on how easy and quick that would be.
But if/when fully autonomous vehicles become a reality they will shed their drivers, same as everyone else.
Please no. We don’t need the ultimate “startup that was unprofitable and never could be profitable that was just trying to get bought by big tech” example to fuel more nonsense.
Does Waymo get to demand human-driver level prices and therefore capture loads of value
I'm pretty sure the Waymo Driver costs more than a human driver at this point. Obviously they want to make it cheaper eventually but who knows when that will happen.
This is an R&D cost thing and the right analogy is to the telephone operator. It obviously cost a lot more to build even the first electromechanical last mile exchanges (with Strowager / SXS switches) than to hire some young women to work as operators, let alone to design something like System X (a computerised telephone exchange used in most of the UK from the 1980s) but you don't have to keep paying for the R&D on an ongoing basis. The girls married, or got less boring jobs, or just left work, each time you must recruit and train replacements - but you design System X once and use it for forty years no problem.
There are still System X exchanges in the UK to this day, it works fine, so why not. Eventually as copper last miles are discontinued these exchanges will be virtualised, and in some sense System X is then banished to museums like the Strowager electro-mechanical systems I saw as a child.
By public accounts, the total hardware costs for one SDC is around $300k. If you amortize that over a 5 year lifespan & 50 rides a day, you get $3.28 of variable costs per ride.
$3.28 variable cost per ride is less than what Uber is paying its drivers (median pay is ~$10.88 per ride). I imagine it nets out significantly in SDC's favor after adding back in cost of gas, insurance, and maintenance. And, the $300k should decrease over time.
An SDC has many of the same operational costs (and maybe more) than a human-driven car. (Also garaging, cleaning, etc. As well as the support staff if a car gets in trouble somehow.) You can admittedly assume electric and assume operational costs may be lower than say a Prius today. But I don't consider it a given that a company which purchases and operates an SDC fleet from Waymo can operate profitably while undercutting rideshare as it exists today.
That includes electrical and maintenance. That's still less than a regular car, but it's still a larger number than most people consider.
It would be possible to say its less, though I'd tend to suggest it would be more than the Prius since there's no human custodian of the car to handle the "pull over, I've gotta puke" situations which can add unexpected expenses.
>surely if this is starting to show signs of success Alphabet just swoops in and buys up Uber
It's too late for that. The regulatory pendulum for mergers like this is swinging the other way. The public was indifferent to big tech gobbling up the competition when times were good. Now that money costs money again, people with normal (i.e., not overpaid tech) salaries are feeling the squeeze, which means politicians need a cause that makes their voters feel like they're fighting for them. The giants must be punished.
Uber is a very different market player. Waymo only does autonomous car rides. People pick one over the other for a variety of reasons. It’s why local taxi dispatches are still in business. It would be ignorance to think they are the same.
Uber opened up people to the freedom of going anywhere and getting a taxi. This took market from a lot of players. But there was always a market it could not take.
Waymo will have to do the same. Carve out a specific market share for itself, from a lot of transportation options.
They couldn't bang out a social network over the last decade(s), I hardly believe they could bang out anything than keep extracting more money from the dying ad industry
Very different. Social networks value derives from network effect - every additional user adds additional value to the ecosystem. Google simply started too late.
Taxis compete on price, safety, quality of service. Virtually all of this value derives from the quality of the FSD implementation. Of course, this will extend to all forms of road transportation eventually. It is going to very much be a winner takes all situation for a number of years (perhaps a decade). The next phase will be a period of high fragmentation as the technology becomes commoditized and a race to the bottom will ensue. The final phase will be consolidation with a few players being best at execution, cost control and finance operations.
>Google could bang out a ride booking app in a weekend.
Google's culture isn't really built around providing human support to users. I believe there is also lots of onboarding involved for restaurants so Google may struggle there.
With that said, I am constantly hit up by outsourced Google reps for Adwords + Firebase but I think the average customer has a much higher transaction volume with much better margins than a restaurant would.
According to Linkedin 54% of Uber employees (48,000) are in operations roles.[0]
I suppose some of these could be drivers, but I believe Uber's biggest department is operations, which is one of the reasons it's so expensive to enter new markets. Can't scale operations in the same way you can scale development.
Google would have to do something similar to launch their own ride-share app powered by Waymo, so could make sense to just keep Uber in the middle.
Maybe not, but a human at Uber responded within 15 minutes when a ride I had booked drove away from me just before pickup. They canceled the original booking with no fee, then sent another car to pick me up and gave me some Uber credits.
That's 1000% better customer service than I'd ever expect from big G. Google would probably charge me a cancellation fee for not taking the trip, shut down my account for having the temerity to complain I'd been stranded, and then followed up with automated email confirming they made the right decision after review.
Agreed. I’m not sure an acquisition makes sense for Waymo/Alphabet. Operating a robotaxi is technologically quite different from an Uber vehicle — mapping is different, routing is different, customer service is different (a Waymo is stuck and needs remote assistance), maintenance is different (charging, cleaning in depots). Uber’s platform wouldn’t help with any of that.
This seems like a play to attract users from the most popular ridehail platform and eventually funnel them to the Waymo app. I don’t believe users are loyal to ridehail apps like they are to social networks (I don’t use Uber because my friends use it) and would jump to a service that is cost competitive and autonomous. Or, as I said another comment, maybe they are setting the stage to be just a technology provider to ridehail companies and let them deal with all the operational overhead.
Seems like but then why the interest from Waymo in an Uber partnership? They already have the Waymo One app. I think it’s because getting users to switch or use multiple apps proves difficult.
Or put it in Google Maps if they had any sense. They should know when to recommend ride hailing with high accuracy to those who enable location history, and they can skip the clunky app switching that other ride hailing services require.
You say the answer is simple, but I can't tell what your answer is.
The first half of your comment implies that Uber has a really bad negotiating position, so Waymo must be extracting most of the value.
The second half says that Alphabet (the majority owner of Waymo) will "partner with literally anyone" which would presumably mean Uber is extracting most of the value.
The second half said that Alphabet had choices on who to partner with, so they could play them against each other, so Waymo had the upper hand in negotiations.
Got it. Maybe I was reading too much into the tone. I read "they'll partner with literally anyone to get distribution" as them being desperate to get distribution and therefore willing to take a bad deal.
This is America. When a "self driving" car kills someone, we blame the someone for daring to not pay due respect to the magic machine. They shouldn't have crossed the road in the dark! They should have been more visible! They should have been in their own big self driving box!
Many comments saying Waymo could easily spin up an Uber-replacement app are missing one key point: integrating with Uber allows Waymo to have a slow rollout / soft-launch.
They can start adding support one city at a time. And for the Uber user, the Waymo option only pops up for you if the ride you requested is within the Waymo range and they have cars available.
This way they can also collect tons of data about how users respond to the offers, affinity to driverless cars per region, price elasticity, etc. And then dial the supply up or down as they wish. They can even start covering a city with just two cars if they wanted to, and then build popularity and word of mouth.
On the other hand if they started with their own app, the lack of car coverage in most areas (due to low car supply, pending regulations, etc) would quickly frustrate users who would then switch to another app, so user retention would be a nightmare.
Not to mention side-stepping all the customer-facing operations of running such a business, which Alphabet does not have an affinity for.
I think my confusion is more about Uber’s incentives here. They’re providing a ramp-up platform for Waymo as you described, but as soon as it hits any scale Waymo can easily part ways.
Uber gets to introduce driverless cars finally (an old promise of theirs) without the costs of owning or developing hardware, and use that as leverage over drivers.
I don't think Uber has anything to fear from Waymo:
(1) It will take years for Waymo to ramp up to "independence" scale. How many cars do they have now, and how many would they need? How long will it take them to negotiate new regulations with every city and state? 5 years?
(2) This deal is probably not exclusive. Uber can strike a similar deal with Cruise as well. Uber becomes the Amazon of driving services, a platform gating access, with all the data.
(3) Having a big money company behind them is good. And if Waymo acquires Uber in 3 years, it's not necessarily a bad thing for Uber.
Waymo maybe could have gone with Lyft (or built their own app) if the deal with Uber fell through, which is probably why Uber accepted it. It's not "help Waymo get to market or not," it's "help Waymo get to market or risk somebody else (or Waymo themselves) passing on Waymo's value to customers and taking the market." Especially with the chance that Uber might get to play gatekeeper in the future, it's a hard deal to turn down.
> The collaboration with Uber gives Waymo’s self-driving technology a second path to commercialization. As Katherine Barna, head of PR at Waymo, told TechCrunch, Waymo is “building a Driver, not a vehicle.” That “driver-as-a-service” model is similarly how Waymo intends to commercialize autonomous trucks, and it means that the company can lease out its AV technology, rather than being the owner-operator of that technology.
Waymo wants to be both owner-operator of their vehicles for their Waymo One taxi service and be a “technology provider” for other ridehail companies. I wonder if the latter is a more lucrative business model. Less operational costs due to someone else running the service, ability pass on liability to them and they can just develop/maintain the tech.
i think the main goal of licencing would be less capital costs. buying cars isn't cheap, especially not at scale - licencing their tech to somebody else means every car they put on the road is $150k waymo doesn't have to spend.
but if uber spend a couple billion on cars, and lyft spend a couple billion, and didi spends a couple billion, etc, that's a lot more spend than if only waymo spends it.
From Waymo's perspective it is certainly desirable to have someone else paying for the cars.
The problem is that when others spend on it, Waymo has more leverage over them if the cars are only going to work with the Waymo driver. Perfect vendor lock-in on assets worth $150k each.
What's in it for Uber? Perhaps just a more favorable & longer term deal, but if so, they're essentially handing over the future of their business to Waymo.
Good point. That would also allow them to bring true Level 4 autonomous driving to passenger cars with sufficient miniaturization of their sensor stack.
I expected this deal to happen right after Dara became CEO, and definitely after Uber sold off their self-driving division. This is why I've owned Uber stock since the IPO, despite the ongoing volatility.
Uber is a software-powered company, but their business model is squarely a marketplace. Running a marketplace is an economics problem, not a technical one. The actual dollars are made from the neverending work of balancing cost of acquisition from both sides of the market, while keeping all the logistics from overwhelming the operational costs. This is true whether it's finance (e.g, health insurance), goods (Ebay), or services (Uber).
There's a big huge gap between this and say, SaaS (trading money for bytes in a specific order and occasional human help) or a super hard hardware problem. When a company tries to have both, they often get stuck with two mediocre businesses rather than one amazing one.
The other key component is just Dara's style specifically. His past successes comes from in other marketplaces (Expidia), and he knows that running marketplaces is all about shaping deals to feed both markets. Despite the history between the two companies, if a deal is obvious, he'll find a way to make it happen. Google isn't prepared for a huge push into a consumer transport service, and Uber isn't willing to bet $20B on a tech hardware moonshot. The gravity for this deal makes it inevitable.
From personal experience: I can say that I never worked for Dara, but I did work for his cousin, and the whole family has the same style. They are very effective at execution, especially when it comes to shaping deals like this. This will work to the benefit of both Google and Uber.
Google owns a bit over 5% of Uber at 70B and a very high percentage of waymo at 30B evaluation. (last measured) so incentive is for Waymo to succeed / grow.
It seems counter intuitive to me that they do a strategic partnership with Uber given low switching cost of using another app if it had meaningful cost benefits to the consumer. Additional Google owns the Android platform so could leverage that to it's advantages as well.
But Uber is already in a world of low margins marketplace matching. The deal is likely highly favorable to Google in that it need only compete with the the humans so can start to recover some of it's investment without having to directly compete on a platform basis which would be expensive per city high density rollout.
Google won't have the same driver density as Uber for a long time and for an app to be your go to ride system it had to always be fast to get a ride.
For high utilization on Google side also benefits being part of Uber platform of on tap riders.
All said, the existence of Lyft and numerous other transportation apps globally impacts the value Uber can extract. More autonomous vehicles will quickly follow Waymo and an integrated network will dominate at some point. Just like Amazon using its own delivery infra once it reached a certain scale.
I don't see this being a long term relationship. Once a given density is reached Google can leverage Android / Google maps / cost competition to push people over and why would Google / Waymo continue to pay Uber market place margin on "it's phone platform". Google will be leveraging said platform to compete with integrated networks like Tesla, Amazon zoomx, cruise, wherever apple lands and others globally.
Interesting thoughts. I was actually going to say that _Uber_ had the advantage in this deal, since it owns the customer relationship and can shape demand toward or away from Waymo.
Google has a distribution channel via Android, but I don't think it's strong enough (otherwise Stadia would still be around). They'd still have to build a destination for consumers to remember, in addition to the interface, matchmaking, billing, etc.
I guess in any deal like this, there's the potential for advantage and disadvantage from both angles. As Clay said, a good compromise is one where both parties are dissatisfied.
Are you sure? Below lotsofpulp pointed out that they're not listed as a major holder [1] and there was news in 2019 about them considering selling their stake [2]
Yes, but I think amelius is suggesting that Google could put a button in Google Maps that called a Waymo (and just a Waymo). That's what could run afoul of antitrust laws.
Not the critical aspect of this announcement, but they are explicitly expecting autonomous food delivery. That makes sense in principle, but in practice… This will be all the wrong kinds of exciting for a while.
If you are a designer with a taste for spilling soup, cold pasta, and involuntarily calzone pizza, I recommend you start looking at options there, because:
- Restaurants operators do not want to leave their kitchen; if there’s a terrasse, waiters already go near the street, but it’s not true for everyone: drive through probably aren’t close enough without the car having an articulated arm.
- Food is messy at any conceivable temperature, and anything that looks even slightly gross is a huge No-No: human drivers have to deal with so many ways to preserve the food not to get yelled at, and all that knowledge is going out of the window without them.
- Customers do not want to leave their sofa. Period. Sometimes, they can’t. Always, they don’t want to — “playing video games” is the traditional excuse for “I can’t right now!” but bowel movements, or other reasons for priority hip movements, or just plain not paying attention. Your car will wait while the food gets cold. And losing money because hangry people inexplicably can’t pay attention to their notifications is not the worst part: their expectations will always be worse. The worst Karens are not the ones at Starbucks, upset that their name is spelled wrong: it’s those you haven’t seen yet because they can’t go to Starbucks because their lack of emotional balance keeps them at home most of the time.
Anyway: exciting concentration event (Alphabet is buying what’s left of Uber for pennies in a year), but there are still miracles to pull off. I, for one, will bow in silence when I finally get to see how one of the few geniuses we have left will fix this.
I still don’t understand the strategy for Waymo. After running that taxi service in phoenix for this long they resorted to partnering with Uber, it’s a huge failure. What does Uber bring to the table at all? An app that is easily replicable and with google’s reputation they can easily get customers off of Uber to their own platform. It’s a colossal failure. I was really hoping they would use Waze for this purpose. In fact it already has ride sharing and so thought that is the next natural direction. The ceo needs to go at this point. He is really taking the ship down.
It's a signal that ride share apps won't be 100% autonomous anytime soon.
If you're a consumer and you want to go from point A to point B you don't want to have to think about whether your route will be in the AV coverage area, or if AVs are able to operate in current weather conditions, or if there are enough AVs available.
You just want to press a button and have some vehicle take you from point A to point B, whether that vehicle has a human behind the wheel or not.
This makes so much sense - and why they'd team up with Uber. Google/Waymo don't care about non-coverage human-driver rideshare. However the users definitely do. So this partnership makes a ton of sense - if you're in a Waymo area, you could have a separate app, but having a single app for both types of rides is definitely a win for both.
Google had this with Google Maps showing the fare for your route on both Uber and Lyft, but I guess Google was getting some sort of kickback because Uber stopped [paying] to be on there perhaps over a year ago.
getting their technology working profitably at scale is obviously the first step and it is taking much longer than expected.
> app that is easily replicable
Waymo already has their own app and will continue to. the problem is that if you only have limited service in one small market then nobody has that app installed.
business travel is a large and highly profitable segment of ride sharing and it is dominated by Uber because it has the largest install base.
> What does Uber bring to the table at all?
demand. Waymo only has to accept a ride request if it meets their profitability metrics so this deal is all upside for them. if they have more profitable demand through their own app then they use their fleet for that and if not they fall back to Uber and get some free advertising to Uber customers in the process.
Waymo is a technology provider. Uber has lots of existing users and no ability to do autonomous driving. This way, Waymo gets to enter their cars in Uber's app and make a similar kind of revenue as a driver would. All Uber does is take their cut and supply passengers.
So Waymo doesn't have to duplicate all of what Uber already does, Uber users just order their ride as normally and maybe get a little cost benefit out of there not being a driver or a need to tip them, and Waymo gets to focus on the cars and the driving and rolling out in more cities.
I don't see how Waze or Google factor into this. Not really a competitor to Uber or taxis in general.
I think they’re accepting the fact that it’s crazy tough to gain market share and requires MILLIONS in ad spending. Sometimes it’s better to have a small piece of a large pie, especially when trying to bootstrap its brand, instead of taking a much bigger slice of a minuscule pie
Google doesn't even want to have human interaction, and Google's reputation for sure didn't help it in chat and social network spaces either.
And I'm sure Google doesn't even want to work it with restaurants and drivers, because you can't just put an AI algorithm to deal with their issues.
Small businesses go through a lot to just edit their own listings on local businesses, I wouldn't wish Google to be the middle man for my deliveries.
Just spitballing - I'm sure they can do the app part easily, but I think the harder part is legislation. And I don't know if Google is in much of a position to just start popping up in cities as transport competition like Uber did, whereas it might be an easier sell if it's just a new service in Uber which already exists that has the technology supplied by Google/Waymo.
>What does Uber bring to the table at all? An app that is easily replicable
What? Building an app is the easy part of building an Uber. It's the difference between building a twitter clone (weekend project) and having it beat twitter (hasn't been done yet)
Uber brings the most successful marketplace for riders/drivers and operating with local regulations globally to the table
With Google's reputation, I'd expect them to provide terrible customer support and then also just let the whole thing stagnate after the initial splashy rollout and it had to be maintained.
Not sure if that is the case. Imagine all ride hailing and cab services in a few years are based on Waymo tech including Waymo's own fleet. Waymo would be the clear winner.
I guess Uber gave up on their own autonomous vehicles. Interesting that it seems like Waymo who took the most cautious approach seems to be the last real contender standing.
People here arguing about whether "Google could / could not bang out an Uber app in a weekend" are apparently completely unaware that Google has already done this.
Waymo One is the name of Google's app, and it's live and serving customers already in Phoenix, AZ.
Imo, the fact that Uber knows this, and is still partnering with Waymo is a sign of weakness bc Waymo's intention is very clearly to own the customer eventually.
It's live in SF too, and the overall user experience is in a different league compared to UberX. The rides are much more comfortable, and quieter. I feel safer because it drives carefully and I'm not getting into a stranger's car. Pickup times and ETAs are actually accurate. It's like upgrading from cable to streaming, or from a Nokia 3310 to an iPhone.
The only problem is the fleet size; there's still a waitlist and it can take 20 minutes to get a ride. But that's just because they're waiting on permits to scale up and start charging for autonomous rides in CA.
It's not clear why Uber would choose to direct their customers to a superior product that they can't hope to beat on either price or UX -- except perhaps as a desperate move to squeeze the last dregs of revenue from the human-operated rideshare market before AV operators move in and clean house?
Except this is Google we're talking about. Expect the Waymo One app to be discontinued in favor of the Uber app, so Uber can "own" the customer service costs.
At the moment - self driving technology is as good as LLMs are. When they work, they work wonderfully and when they don't, they don't but the problem is that there's no sure shot way of knowing when they'll work and when they won't.
The problem is that, unlike with LLMs where a human always can and should fact check/reword before using, there is no such option with self driving. People's lives are literally directly at stake if things go wrong.
For limited scenarios such as geofenced to a city with good weather, it seems that self driving cars (at least Waymo's tech) is good enough, baring some small lapses which are more funny than dangerous. As a general thing everywhere? To me it doesn't look like we're even close (like 10 years) to being there. Also, the problems actually instead of supposedly being solved by self-driving are... very minimal. There will be no noticeable improvement in things that matter the most on the global scale (pollution, congestion, time spent), only marginal practical and maybe financial improvements (there might be more and cheaper taxis).
Self-driving electric trains are already a solved problem that could solve probably the majority of transportation needs, congestion and pollution all in one go.
Cheap taxis could lead to people abandoning car ownership and this a huge reconfiguration of American cities. (if they end up cheaper than self owned cars)
People in America mostly view cars as a status symbol and a part of their identity, which is why everyone buys giant trucks and SUVs. Why would a better taxi change that? Only "the poors" use taxis
If everyone still takes the taxi everywhere, what would be the reconfiguration about and where would it come from? Everyone will be still stuck in traffic due to the physical impossibility of fitting everyone in an individual vehicle on the roads at the same time.
It's been shown through computer modeling that road congestion emerges as a function of both vehicles/road segment/time unit and human driving behavior.
Remove the humans from the driver's seat, and your effective road capacity goes up significantly without doing anything else.
Do you have links to those computer models? Do they take into account induced demand?
And still, even if autonomous cars are drastically more efficient, with a passenger per car the space use is still drastically less efficient (even the biggest highways will have a fraction of the capacity of a proper metro).
You're more than capable of doing this research yourself.
> with a passenger per car the space use is still drastically less efficient
Sorry, this simply isn't the metric for success. It might be your metric, but it certainly isn't the population's metric. A lot of people want to travel comfortably in cars, "mindless pursuit of space efficiency" be damned.
Ubiquitous self-driving vehicles will allow us to eventually eliminate most congestion, and you're just going to have to accept that cars aren't going anywhere.
How would they eliminate congestion when they're so space inefficient? Even if they're more efficient at traffic flow, there's only so much physical capacity on the roads that people want to travel on.
> you're just going to have to accept that cars aren't going anywhere.
Cars aren't going anywhere, but they're the wrong solution for a variety of reasons (congestion leading to time loss, pollution (including noise pollution), land use efficiency or lack thereof, etc.) for a very big part of transportation needs.
> A lot of people want to travel comfortably in car
And a lot don't and aren't given that option.
> "mindless pursuit of space efficiency" be damned.
Mindless? There's only so much physical space available in the places where people live and work. American cities are often comprised of massive parking lots taking up valuable space that can be put to much better use, simply because most people are forced to commute by car. Give them another option and many (of course not all) would prefer it, which would also make driving much more pleasant for the remaining drivers.
> How would they eliminate congestion when they're so space inefficient? Even if they're more efficient at traffic flow, there's only so much physical capacity on the roads that people want to travel on.
Sure, but that's not the limiting factor at all, which you'd understand if you bothered to do the slightest bit of research outside of your comfort zone.
> And a lot don't and aren't given that option.
That's very true, but it's not one-or-the-other. I am a strong supporter for intensive funding of mass transit systems....for those who want to use them. But that doesn't preclude continued usage of cars.
> massive parking lots taking up valuable space that can be put to much better use
"better" is in the eye of beholder. Car users are generally okay with how space is currently allocated.
If you want change towards increased mass transit, I highly recommend not being hostile about it. There is no need for mass transit to be positioned as a replacement for car use. When you threaten peoples' way of life, that's how you get people actively working against you.
It should come as no surprise that I'm a huge fan of getting around by car, but I would still happily have my tax money go towards bolstering mass transit systems and walkable/cyclable infrastructure for the portion of the population that wants to commute that way.
Would this not lead to induced demand where more people decide to start traveling more frequently once they perceive the roads as having cleared up? An additional consideration would be people incapable of driving would now have greater access to freely take rides.
The notion of shutting down all improvements due to fear of induced demand is ridiculous. Let's also stop building sewage systems because of induced demand and all defecate on the ground. It will reduce water use and be good for the environment by enriching the soil.
That’s not what I meant to imply. I was simply asking if removing humans from the driver’s seat and replacing them with autonomous vehicles that are hypothetically more convenient to take would actually increase the effective road capacity.
I’m not so sure it would actually solve the problem. I think a more effective and easier solution would be to design cities to be more accessible without cars. That way, traffic would likely be reduced due to fewer humans in the driver’s seats even if they continued commuting. To summarize, in order to achieve a greater road capacity, I think we should focus on replacing the method of commuting that requires taking the road altogether. Subbing it for another method that still uses the road will invoke induced demand.
That's not actually a blocking problem, as you have the same issues with human drivers today. At the end of the day, just like literally everything else in life, self-driving tech is a statistical process with corresponding costs.
The bar for success here isn't perfection. Rather, it's whether or not self-driving vehicles are cheaper than human-driven vehicles. Part of the cost of a self-driving vehicle hinges on how often the vehicles end up in failure states necessitating remote human intervention and/or insurance payouts. It's fine if they sometimes fail, they just have to do so infrequently enough to be cheaper. And if Waymo is to be trusted, they already appear to be safer.
One thing that hasn't been addressed, given the limited test markets, is whether autonomous taxis actually will actually be popular and preferable among consumers. Theoretically having a car that drives itself seems like a great product/service, in practice it's more clunky and awkward than you would hope. From the experiences I've had, a human driver is undoubtedly the better option. Many things happen on the road were a driver may be forced to speed above the speed limit, circle around the double yellow line, reverse on a one way street. Be a passenger in an autonomous car within an urban setting and you'll know immediately why a human driver is the superior service (and likely to remain so). The "potential" future of every car becoming autonomous driving on highly controlled roads being able to drive in the highest level of efficiency isn't a likely scenario.
This is very true, but it sounds to me like a classic case of design iterations: find patterns, develop them among users, learn ways to introduce new users… A search engine, an email, chat… non of that was natural, and the first iterations where clunky, but you know can’t imagine they’d be any simpler or that anyone would want anything else.
So if we look exclusively at only successful software products for our comparison, we'll see success for a completely different product category with high capital costs--got it.
I imagine autonomous taxis will be significantly cheaper than driver taxis. From a quick search it looks like in general Uber drivers earn somewhere around $9 for a 30 minute ride [1].
And out of that $9, the driver has to pay for their gas (or electricity) and upkeep of the car. Uber apparently earns about $3-4 for their service. If Uber is now covering the costs of an expensive for now autonomous vehicle, it's not clear to me the fare drops at all.
So they're going to make the cars they're already running in Phoenix available through Uber. Is this because they don't have enough users already? Are they doing it so that they can then expand into other areas without having to spend money on advertising/operations? What's the game plan?
What does Waymo get out of this? It seems to me like if the value proposition in self driving is there (privacy, consistency, etc), people would be willing to download another app? Especially once word of mouth goes through.
Plus currently the Waymo app has neat features (when the car is on its way to you, it shows you stuff like when it's stopped at a red light or whatever) that Uber probably can't replicate due to API access issues.
To clarify: they have remote assistance teams, but they cannot “take over” and control the car at all. All they can do is things like plotting a path to go around if the vehicle gets stuck. They have no ability to do safety critical interventions, the vehicle is capable of reaching a ‘minimum risk condition’ all on its own.
Uber drivers would also need to be trained to do this and Uber needs to setup facilities for this. I don’t believe this is likely.
Waymo sends physical human drivers to a car if the Waymo driver (software) is stuck. Remote humans can reach into the model to fix e.g. label problems (e.g the driver thinks this is a pedestrian but it's actually a life size cardboard Captain Picard, thus it's not a problem to drive very close to it) but they cannot drive the car.
Big companies are very good at letting bygones be bygones when behaviors and circumstances change. Look at the changed relationships between any number of companies and Microsoft.
This may just mean that Waymo is outsourcing its Chandler AZ self-driving operations to Uber. That makes sense for Waymo. Have details come out about which way the money is flowing?
For Uber, it's all about maintaining the stock price.
I am happy to see movement from any big player on autonomous driving. I still believe Tesla will win this war but I am happy to see healthy competition in this space, happy that partnerships forming to make it happen.
I think this makes sense for the both. Waymo's technological advantage won't last forever. Unless they can overtake the entire automaker industry in a foreseeable future (which is extremely challenging goal even with Google's money), it's probably better to put path dependencies on your potential future customers and make it much harder to change. Uber desperately want to reduce its spend on drivers, so this also aligns well with their goal.
Having a sole source partner with self-driving cars doesn't help Uber reduce spending much; Waymo isn't charging Uber cost-plus pricing, they're charging as much as they can get away with. And Waymo has more negotiating power for its car fleet than individual Uber drivers have for their one car.
I'm delighted to read these news. Uber finally found someone to partner with/get bought by. They offered their complete self-driving department for sale more than two years ago (I was part of a group to assess the deal with a possible buyer).
I think Waymo also realized that Uber's self-driving devision is no competition to the big players. So I guess Waymo is mainly interested in their user base.
I was a founding engineer on the first uber platform for third party autonomous vehicles. It was just ~4 people and we got a ton done, growing to ~60 people and onboarding a lot of partners before I left. It was a really interesting time, and I wonder how much of that original code is still running here, or if the marketplace team (where the project eventually transferred) ended up replacing it all.
Uber moving strategically into the beginning of their late game. Interesting times.
I suspect that Uber's main added value is a "safer" way for Waymo to increase the reach of their technology with lesser blowback against their own reputation (Uber did this, Uber did that ...), and to make it more mature, even more than the money potentially being exchanged here.
> a "safer" way for Waymo to increase the reach of their technology with lesser blowback against their own reputation (Uber did this, Uber did that ...)
They are still Waymo-branded vehicles. Attempting deflection wouldn’t go well. This is a risky integration pursued for scale.
I think it also helps Waymo fill in the gaps between markets that they support. Customers get an app that works pretty much anywhere, and Waymo can drop into any given market with 10k self driving cars pretty much overnight.
This. Spinning up an Uber app clone is easy, actually running the service and having it available everywhere is hard and not something Google is likely to be good at.
> proof that this means that they're ready for the late game, as opposed to just investors being desperate
Whose investors? Uber isn’t taking technology risk. And Waymo has worked in Phoenix for years. Partnering with Uber is a show of confidence in that it suggests their demand limited.
I had interviewed with Uber's ATG years ago and their pitch, even then, was that they were building a platform for whoever won the autonomous game to be available on uber, not just their own cars.
Interesting they kept that strategy even after spinning out that group. Curious if they managed to keep anyone from that team to help with this product.
After Uber hired and stole technology from Waymo, which they got caught and the engineer prison time I believe (?), looks like now they are going to go the route they should have always gone since they didn't have the technical capability to pull it off.. partnership.
>In 2019, Levandowski was indicted on 33 federal charges of alleged theft of self-driving car trade secrets. In August 2020, Levandowski pled guilty to one of the 33 charges, and was sentenced to 18 months in prison.
>He was pardoned less than six months later on January 20, 2021, the last day of Donald Trump's presidency.
I know - its the brashness, level of ease and amount of pardons by the Trump presidency that makes it different. Not that it isn't done by other parties.
I bet Alphabet has calculated that of all of the rides a typical ridesharing service offers only a subset is currently serviceable by autonomous vehicles. A deal (or buying) Uber allows them to offer the service on only the rides they are confident they can fulfill.
I've only seen 2 comments about Tesla on this thread. But, it seems people underestimate Tesla on here (I'm not the biggest fan). Though they are having great progress in their FSD beta- one way to mark this is by the number of users with access to it, which has increased significantly. This allows them to collect a lot more real world data then their competitors. Once this gets better and is able to prove better driving than a human, which seems possible in certain metros at least as Waymo is doing, then they can do an autonomous Uber network of their Teslas.
But I think this is a smart move to join forces, and will give the consumer healthy competition in this space.
1. Lack of LiDAR sensors. Vision is still not (and IMHO will never be) an adequate replacement.
2. Waymo/Cruise etc. can upgrade their taxis easily with capabilities over time. Tesla is restricted by their promise that FSD will work on previously shipped cars.
if we're talking about overall effectiveness, then I think it is. Maybe passing a minimum threshold and having a safety effective of say 95% vs 99% or something. It's obviously going to be a very high percentage threshold that needs to be crossed to allow the public to use it.
So, right now it's like a trial experiment with FSD, but I definitely feel like a data advantage would play out into a higher safety rating before it's in GA. Maybe eventually others will catch up, that could be quicker than longer, but it's a valid potential advantage for Tesla.
just to nitpick,
but with airline travel about five or six nines (99.9999%) safe, that is, one fatality every million miles of travel or so, 95% vs 99% is nowhere near good enough.
Yes it will be likely needing a 99.99%+ accuracy. I just made a vague suggestion, as I didn't have time to look it up. Elon mentioned it too, it doesn't just need to better than a human, but massively better. I don't know the exact accident stats for human drivers. But self-driving cars will need to be much higher.
Somehow this seems scarier to me than self driving cars.
1. Driving without the physical feedback that comes from acceleration, braking, taking a turn too fast, etc.. is really hard. Maybe they could overcome this with simulators of some sort, but I expect that would be too cost prohibitive.
2. I want the driver that can think to have skin in the game. I don’t want them to be able to drive off a cliff, or in to a crowd, from the comfort of their couch.
You could probably overcome this with some sort of hybrid model that simply doesn’t allow stuff like this. But at that point you’re still having to develop some pretty sophisticated self driving technology, and you still have to pay a driver. Not sure who is motivated to do this.
3. What’s the point? You still have to pay a driver either way. Sure you get some cost savings by being able to switch out tired drivers instantly. But in the end, the most expensive cost is paying someone to sit there and drive.
Yeah ultimately it could be a sort of hybrid where the AI drives until it encounters a situation it is high risk to handle, like crowded street, then a human is alerted to take over. A pool of humans could then manage exponentially far more than one to one vehicles depending on how good the AI is.
That's a huge money saver.
Additionally recruiting truck drivers is difficult to do so maybe having a work from home job would make it easier.
It's not that it's point less it's that some people just don't see the point.
I don't know just seems like an idea worth pursuing for a million reasons, including being able to outsource driving or find the best drivers in the world. Also.mikitary reasons. Save lives with a remote controlled tank.
Remote driving would likely result in worse safety with very little gained by a commercial transportation company. They still have to pay for a driver but now that driver is staring at a screen 12 hours a day. Humans are already mediocre drivers when physically present and that would get so much worse if they were doing everything remotely instead
What's the gain here? Can they pay you less than another driver who will sit in the car?
Because dealing with signal being lost sounds like a huge problem to deal with for very little payback.
in the typical scenario taxi/ride-share drivers are only driving paid miles for a small fraction of the time they are on the clock.
with autonomous/remote vehicles you have a fleet of driver-less vehicles that are capable of fully-autonomous operation in some circumstances but are remotely monitored/controlled for situations they can't handle.
this means that you can have a remote driver who is continually switching between driving different vehicles and only driving in the situations where actually needed and profitable.
with driver based transportation when you get into low utilization hours then drivers and cars have to be pulled of the street, the drivers who are working spend more time sitting unprofitably, and they have to driver further to pick up rides.
in a remote scenario you can keep all of your driver-less cars on the street (parked) and your remote driving staff takes control of the car closest to each ride request to pick up and deliver that ride so they spend much more time driving with much less mileage waste.
Drone planes are good for the military because they move all the risk from the pilot to the vehicle. Drone cars are bad for civilians because they move all the risk from the driver to the vehicle. What you’re suggesting would quickly replace America’s mass shooting problem with a VBIED problem. And it wouldn’t even save any time or effort! You’d still need to concentrate on driving the car.
At this point even multiplayer driving games haven't figured out how to not kill half the field every time there is a single late packet. You cannot drive in a concrete jungle and get stable, reliable, verifiably usable internet.
There isn't that much of a difference between the taxi driver sitting in the physical car or at home.
Only when you Location A is in low demand, you could "hop" into a car at Location B.
But the goal is to completely eliminate the driver, which increases efficiency a lot.
But that doesn't create any profit, it's only a convenience for some random number on a sheet.
Also managers probably don't like remote work in that industry :p
"autonomous" vehicles are remotely monitored and controlled. that is just not part of the marketing.
the big blocker, and why autonomous/remote vehicles are not being rolled out more widely, is that remote requires lots of very reliable low latency bandwidth.
5G (mmWave 5G in particular) is a key enabling technology. until you get mmWave 5G or some comparable technology the bandwidth for remote at scale doesn't exist.
The Waymo vehicles being discussed in the article are not remote controlled like the GP is talking about. Remote assistance isn't controlling the vehicle, the autonomy stack remains in control throughout. They're just giving it additional information that it can't determine for itself. They're also not a part of the typical operation, but rather for infrequent, exceptional circumstances.
Anyone else find this entire thing odd from a legal perspective?
* Google Ventures owns a non-trivial stake in Uber.
* Google infamously filed a lawsuit alleging that Uber stole trade secrets from Waymo.
* Uber made an early decision to run on-prem instead of on the cloud (GCP, AWS, etc.) because they feared, rightly so, that Google might launch a competitor to Uber.
* And now they are partnering and merging their core businesses.
Outsider speculation follows: Uber under Kalanis brazenly stole Google tech, and then settled out of court and promised to be BFFs - with money generally flowing towards Google. Uber in return got a very damaging lawsuit dropped. Uber's self-driving project subsequently got shut down after it ran over and killed a pedestrian without braking. This is what's going on
Uber doubled down on the thing it was good at, the logistics of ride scheduling. That's a non trivial problem that benefits from a network effect and data. They're moving away from things they don't need to be doing, like developing self driving cars or running their own data centers.
its probably the best way to approach this. Waymo doesn't need to waste its time on all the complicated regulations and technology around just ridesharing + Uber gets a foot back in the door to AVs.
The whole arguing with your Uber Eats driver was bad
"please come to the door", "no, please come outside", "i can't leave my baby, please come to the door", "i can't leave my car next to a hydrant, please come outside"
There's VERY few population centers in the US where it's raining heavily >10% of the time.
The vast majority of rides in the US - and even in a place like Florida that's a swamp where it rains literally every day - happen under just fine conditions for Waymo.
This is fine for a taxi service.
If you make 10x as much profit in the vast majority of cases - you'll let someone else pick up the scraps in the edge cases.
It's almost as if Waymo & Uber considered weather...
> There's VERY few population centers in the US where it's raining heavily >10% of the time.
First, you're utterly wrong about how rainy cities are[1]. Most of the population of the US lives in a place that has more than 100 rainy days per year. The Northeast Corridor by itself has 50 million people and 120+ annual rainy days.
It also doesn't need to be "raining heavily". It can also be raining lightly, snowing, foggy, dusty, or smoky.
But even if you're right and it's 10%, the predictability becomes a huge issue. I've spent my whole life on the East Coast, and you never know if it's really going to rain until it does or doesn't. The forecast changes literally minute to minute.
> First, you're utterly wrong about how rainy cities are[1]. Most of the population of the US lives in a place that has more than 100 rainy days per year. The Northeast Corridor by itself has 50 million people and 120+ annual rainy days.
First, you are utterly wrong with your facts.
100 rainy days per year does not equal 2400 hours of rain per year.
It's closer to 500 hours of rain per year = 5.7% of the time.
Realistically, you're looking at the VAST majority of cities Uber being able to operate in 99% of the time.
Additionally - "rain" does not equal downpour. Waymo can drive in light rain.
And guess what - when it's downpouring, people crash MUCH MORE (fun fact - my father almost killed my hyrdoplaning in a downpour to get to my brother's basketball game).
Maybe it's not the worst thing in the world if people aren't on the roads nearly dying in the middle of The Perfect Storm (<1% of the time) to get places they probably don't really need to get to in the first place...
Automotive EE here, I could not disagree with you more on the state and expectations of self-driving vehicles.
I can tell you with confidence they have not really put effort into weather and it’s a problem “to be solved later”.
It’s all a joke. If we actually wanted this, it would be long haul A to B routes. Turning semi trucks into rail.
Self-driving was already slowing down before the era of free money. So I expect this is just PR to set expectations for Uber and prepare for a scenario where Uber is eventually swallowed by someone else.
The article to me (an Automotive EE/CE of 20 years) that “Fairy dust makers combine efforts to bring genie wish program forward”.
Even more so, if they "have not really put effort into weather" then apparently they didn't need to, there are videos of Waymo drives through SF at night in rain for example, it seems to work fine whether it took "real" effort or not
Ah, the internet where people who know absolutely nothing cite a PR blog directly from a company providing bold assertions that then give them 100% confidence.
I’ve been in at least four fully self-driving dev / S0 / drivable buck vehicles and they are absolute garbage imo. We had one planned demo that was delayed because of fog in 2020, not road fog, but crap on the LiDAR glass. How many have you directly worked with?
You’ll be happy to know it isn’t just weather that makes this all fake though.
Road crown, even minor, but especially major in the northwest. Reflections, refractions. Lights and electronic billboards. Potholes. Missing road paint. ANY sort of road work at all. “Unhomed” people washing windshields. Squirrels (day) and mice/rats (night). Other bad drivers. Just plain dirty vehicles, self and others. Dust.
It’s all garbage. If we were serious it would be long haul, V2V V2I (vehicle to infra/vehicle). But yea yea, Waygo says they solved weather so just post their claim and you win! Reality can’t compete with marketing, except it always does.
I'm not the person you're responding to, but I've ridden inside probably dozens of AVs at this point (and my work product is in hundreds). I can't relate to your experiences at all. That's weird, because if you're an EE at one of the major AV companies, we're likely within 1 degree of separation.
Every serious company has put significant effort into testing with weather (note: not all weather conditions and not the same testing as traditional automotive). They've also put effort into handling all the other scenarios you've mentioned. Waymo does pose estimation to better handle construction workers and traffic cops, for example.
Also the TAM numbers should make it clear why long distance freight isn't getting as much attention. Surely your employer has shared their outlook internally?
There are people taking rides in rain and fog, albeit only in SF and Phoenix. You can claim they haven’t put any effort into weather driving all you want, but the evidence suggests otherwise.
That blog is a great example about how they have no solution to the weather problem. Their heading image is a light drizzle, and they talk about "predicting" the weather, as if their cars being unable to manage an unpredictable situation is an okay product.
Mercedes got flack for it, but ultimately they had the right answer.
When a sort of pitch was made to an exec where the car has to crash into a tree or a cyclist, he said the bike. Because you aren’t going to pay $80,000 for your car to choose your death.
If you look at car ownership as owning a weapon, it does only make sense that you’ll have to prioritize for the owner.
But really, we aren’t even remotely to trolley problem. The systems BARELY detect and react properly to single issues, let alone detecting multiple and selecting between them.
I honestly think it’s a scam that was popular in the era of Free Money, and I’m so glad that is temporarily on pause.
The trolley problem is functionally irrelevant to AVs, but even if it wasn't, it's still prohibited if you ever plan to pursue type certification in Europe per 2022/1426, section 2.1.1.
There's no need because 1. cars have brakes 2. programming this into a car would make it more dangerous, because it might decide to trolley-program-sacrifice a grandmother when nobody was actually in danger.
Why not? It seems to work fine in rain - as evidenced by all the "bomb cyclones" in San Francisco this past rainy season. There are a lot of places they can go that don't have snow.
Can you please make your substantive points without fulmination or flamewar? That's always possible, and we're trying for something else on this site.
I know these topics are important and the feelings are understandable but the problem is that on an online forum, this kind of rhetoric becomes just-more-internet-drama-noise, which doesn't do anybody any good and drowns out the quieter, more curious conversation we're hoping for here.
Yea let's just keep workers in such a precarious state that they not only must sell their labor for pennies, but incredibly they must now supply the capital to the corporation, all to do a job that isn't even necessary for a human to do, is hazardous for the worker and the public, and generates tremendous amounts of pollution. Why the "optics" aren't already absolute dog shit is left as an exercise for the reader.
This is a strawman. He's advocating that things shouldn't become worse than they are, you are ignoring that point and instead just talking about how bad things are now (without any regard to how worse this would make it).
No. I'm arguing that a clean, electric fleet of autonomous cars, that would maim and kill far fewer people, where the capital is provided by the corporation instead of being extracted from workers through deception, would be a vast improvement over the status quo. We don't need to perpetuate a system of pollution, exploitation, injury, death just so workers can busy themselves performing unnecessary tasks for starvation wages.
Shunting away low-paying (usually with low education requirement) jobs doesn’t magically move people up the job totem pole. Those people need to be retrained, and during that training they still need to make a living, preferably above subsistence level.
On the other hand poor people who can't afford car ownership would get much more mobility and access to jobs and services if really cheap autonomous car sharing takes off.
Can you get a taxi to show up? The last two times I tried to rebel against the rideshare duopoly was the morning before international flights. Trips to the airport are the only use case I have for rideshare. Both times I scheduled the pick up by calling the phone number the day before, both times the driver just… didn’t show up. I panicked and downloaded Lyft.
I know some taxi companies have apps now, but if I’m going that route, I don’t see the point of not using Lyft.
Same. Every time we tried to call taxis pre-rideshare apps, we always got a "between 30 minutes and 2 hours" estimate if they ever showed up at all. Their CC machine was always conveniently broken unless you didn't have cash and the ATM was real far away. Good riddance.
The “sorry the credit card machine is broken” scam is about as ludicrous as the homeless people asking me for $1. Buddy, I haven’t carried cash since the Obama administration. Half the time I don’t even carry a credit card.
Sadly they have been squeezed out of many locations, and I can not always find one. I know where I can usually get them and try it as my first option. It depends on where you are.
I’m glad that works for you. But I’ve learned over years of discussing this stuff that there are so many different transportation situations experienced by different people that any universal rule is destined to become wrong.
This doesn't work in my city, taxis will literally just not show up. I know that sounds weird but you'll call, request a time and place and then be at the time and place and... nothing
To be fair my only experience of Uber was a friend calls an Uber for us, the Uber eventually cancels, she requested another one, they cancelled too, and then magically there are no Uber drivers in my city.
Conventional "radio controlled" taxi was the eventual solution I used to get us out of there.
With the Uber app, Uber will show both options when they're both available. It allows waymo to offer a lower SLA/SLO while still gradually fielding a service. It also allows for easier fallback if a waymo vehicle is taking too long to arrive.