It sure feels like, barring some major regulatory setback/surprise, that within 5 years it will be funny to remember getting into an uber or cab driven by a real person when in any major city.
The biggest challenge overall is going to be overcoming the issues with autonomous driving in inclement weather. But I’m sure it’s doable eventually.
Last year, I went to a AI / labor roundtable discussion event at the university where I work. One of the folks attending was a state congressman or something like that, and the topic of autonomous vehicles came up. He said his goal was to ensure that there would NEVER be fully autonomous vehicles in Indiana. I got the impression that the rest of the room agreed.
I was pretty baffled by this, but figured that attitude should be expected at a "labor" event, so I brought the topic up with my IT colleagues, and many agreed that they would never trust a driverless car.
I agree that we will eventually get there, but I think major regulatory setbacks should be expected, and a 5-year timeline is pretty optimistic.
They'll do what we've been so-well-trained-for at self-checkouts: the passenger might be given control of the vehicle in such situations (e.g. loss of network connection).
>within 5 years
As for predicting the future: I think within five years ALL vehicles on public roadways will require a beacon for travel, simply to alert other self-driven vehicles & intersection controllers. This process will be gradual, akin to "insurance discounts" for allowing driving habit monitoring (e.g. without a beacon, traffic lights won't auto-cycle for you).
> As for predicting the future: I think within five years ALL vehicles on public roadways will require a beacon for travel, simply to alert other self-driven vehicles & intersection controllers. This process will be gradual, akin to "insurance discounts" for allowing driving habit monitoring (e.g. without a beacon, traffic lights won't auto-cycle for you).
I personally think within 5-10 years, forward facing cameras will be mandated on all vehicles, starting with commercial/heavy duty vehicles. Yes, there's some degree of a privacy argument with it all, but the safety and liability boosts dramatically outweigh them.
Assuming that happens, either your suggestion could also happen, or a visual solution to the same thing could also be in play.
> They'll do what we've been so-well-trained-for at self-checkouts: the passenger might be given control of the vehicle in such situations (e.g. loss of network connection).
Who handles liability if you get into an accident? When does it take control of the vehicle away? The most likely scenario is the car pulls over safely and asks you to get out or wait.
There are _many_ people who cannot or will not driver. They won't just take over control of a vehicle they're unfamiliar with. Teenagers, the elderly, people without a driver's license, people injured, ...
Not going to happen, too many risks. If anything, the future is autonomous vehicles without a steering wheel. I am expecting Waymo to invest into that next, after they finish expanding into several states and prove their growth and monetization strategy.
This five year prediction seems odd to me. Autonomous vehicles have to navigate in a world that was designed for human driven vehicles. If you were to design an autonomous transportation system in the middle of New Mexico from scratch it would almost certainly look nothing like what we have today which is comprised primarily of visual cues designed to be highly noticeable to human eyesight. That’s to say nothing of cars themselves being designed around the assumption of a human driver.
I guess what I’m trying to say is that it’s very difficult to imagine that if AVs take over let’s say over the next 30 years that the end result will still be running on the same infrastructure we rely on today and that if we will still be a society where most people drive most of the time that AVs will really be all that transformative in the first place since they would be filling a sizeable but still relatively small use case- taking a cab around a major city. I can’t tell whether it’s the amazement of seeing a driverless vehicle or the actual utility of it that generates the excitement
In the United States? That discounts how resistive to change and fragmented the various municipalities that own all the roadways are. The various states can't even mandate a shared speed limit, they are not going to all agree to (and pay for) retrofitting all of their traffic lights to a single standard.
Yeah, a beacon isn't that useful unless you want to require that all pedestrians wear them as well. The only self-driving-car-related serious injuries/deaths I can remember are Uber hitting a pedestrian and Cruise running over a pedestrian.
I think the biggest issue right now is economics and the logistics of expanding and getting a lot of vehicles, but this is easy compared to building a proof of concept of a robotaxi service. It's now just a matter of iteration and economies of scale.
They can already handle quite heavy rain and fog, the only thing missing is snow.
The economic side of Waymo still makes absolutely no sense to me. Uber, with no heavy capital investment, has been pretty much unprofitable for over a decade. The car companies are low margin. The taxi companies are low margin. Alphabet has like 30% profit margins. Waymo's business proposition is 1) to own a very capital-intensive fleet of bespoke cars with no resale value, 2) employ an extremely expensive team of engineers to develop autonomous driving for them, 3) use this incredible capital investment to try to undercut an industry that is already barely or not at all profitable and earn back their 10s of billions of dollars in investment (and 10s of billions more of costs in the future), one taxi ride at a time.
And that's completely ignoring their issue with surges, as their robot car supply is inherently fixed - they either don't intend to handle surges, or intend to have low utilization of their fleet.
The business case of their robotaxi service is simple, they aim for substantially lower costs compared to traditional taxi services and personal car ownership.
The cost per vehicle could get to something like $50k in the long-term, labor costs could get to maybe 2 people per 10 vehicles and utilization would be substantially higher compared to personally-owned cars.
The vehicles are / will be designed to last for a long time, Dolgov used 400k miles in his napkin calculation.
They can charge more because the user experience is better, so not only lower cost per mile but also higher price.
Cost of developing the technology will become negligible with scale, it's a fixed cost.
They're not just competing with other taxi services, but with all types of road transport. It's a technology that has the potential to be used by most people on the planet, daily.
Regarding surges, there are multiple ways to deal with it, for example surge pricing (which would motivate people to use the service when utilization is low) or using free capacity for delivery.
Their unit costs are at least break-even in San Francisco, and this is with their 5th gen vehicle, 6th gen is supposed to be much cheaper.
I am expecting Waymo to launch their own autonomous vehicle with a simplified body, no steering wheel, etc. They can bring down the cost as they scale. Resale value is not something they should be concerned with, it's a specialized transport that can be utilized for a fairly long time.
The most interesting bit is that Uber is responsible for operations.
> Through this expanded partnership, Uber will provide fleet management services including vehicle cleaning, repair, and other general depot operations. Waymo will continue to be responsible for the testing and operation of the Waymo Driver, including roadside assistance and certain rider support functions.
I thought Uber just let drivers ("independent contractors") deal with the vehicle maintenance and cleaning up when someone pukes in the back. Do they even have their own depots?
100% this. Dara came in and pulled all the P&L levers he could- basically closing down their own autonomous unit and buying a stake in aurora technology to allow another company to lose money and not have it show on their balance sheet. Then raising prices and lowering pay as much as possible.
They are now entirely reliant on quarter to quarter consumer spending and the performance of three stock holdings which are completely out of their hands.
Expect more and more manufactured headlines to keep the AV story in tact.
yeah this is interesting. I forgot where I read recently but Google needs to build out big depots where the cars can come back and charge etc. Its a huge cost of operations.
Offloading all that to Uber who can basically become a marketplace of robotaxis and service many companies makes a lot of sense for both companies.
The economics of this seem highly questionable. After all how many robotaxi companies can there possibly be given the costs associated with vehicle ownership and autonomous software development? In a world where it costs nothing to create content or launch a drop shipping business aggregators like meta and Amazon make sense. But if you end up with like three AV providers, Uber seems like an unnecessary middleman in a business where margins are already incredibly tight.
You have to always remember that the competition is a guy in a car charging a couple bucks a mile
Can anyone explain how this makes any business sense? Doesn't Uber have a self-driving business that competes directly with Waymo? Wasn't one of Uber's early engineering motivations for running on-prem infrastructure (instead of cloud) to avoid being cut off from GCP should Google decide to start competing? And now Uber is directly partnering with Google to offer a Google service through their platform?
Uber doesn’t have a self driving business anymore. That was shut down after they killed a person in Arizona during testing. They divested much of it into Aurora, which I believe Uber still has a stake in.
How this makes sense for Uber is that they don’t want to be left out when robotaxis replace human-driven ride hailing vehicles.
Nah, they shut down their self driving car division because they needed a boring conservative safe pair of hands for operations and the current CEO wasn't much of a visionary unlike the founder. Say what you must about Kalanick being a frat boy but he embodied "move fast and break things" like no other.
That should require a huge introspection and if needed, shut down unsafe lines of business.
Their self driving car division was, by the accounts I read, never going to achieve self-driving and had serious problems.
If being a visionary means you are okay with killing people, then fuck being a visionary. “Moving fast and breaking things” should NEVER include killing people.
> That should require a huge introspection and if needed, shut down unsafe lines of business.
Which line of business would that be? Maybe the one involving human drivers that leads to ~75 deaths per year [1]? Though that's still a lower-than-average death rate compared to the general population (so shutting it down without a replacement could cause even more deaths).
You could argue that Uber has a moral imperative to replace its human drivers with automation as soon as possible.
The real head scratcher is why Waymo would partner with Uber aside from operational incompetence. Maybe Waymo decided that depot services are too hard to spin up, but Uber doesn't do depot services either, so they're paying for Uber to spin that up and a little extra for Uber to make a profit while introducing an intermediary for ride hailing instead of using the ride hailing app they already built.
Probably just comes down to less friction in getting paying users in their cars since they already have the app installed and these are small roll outs. From the announcement it sounds like the streets of Austin will be covered in waymos but after having been downtown a few weeks ago I can say there are like 10 of them. Partnering with a ubiquitous app makes getting real users in these small launches feasible.
Just a guess but if that’s accurate and then waymo expands to where it becomes the de facto option and familiar for a large number of ride hailers you can rest assured the next step will be to take back operations and cut Uber out
other folks have responded about the state of Uber’s Autonomous driving program, so I just wanted to weigh in about the other question here.
GCP was not a strong competitor (in our eyes) when the on-prem decision was being made in the 2011-2013 timeframe, which was also before there was even an Autonomous program. The reason early infra was on-prem largely boiled down to a general distrust for the cloud from the Boss, who in my recollection had had some bad experiences with other portfolio companies suffering major downtime due to AWS outages.
Uber’s relationship with Google was great until that little legal matter. They were huge and very helpful investors in Uber.
So bynext year Waymo’s rollout will look like this:
* Los Angeles - Waymo One
* San Francisco- Waymo One
* Phoenix - Waymo One, Uber
* Austin - Uber
* Atlanta - Uber
It’s still early days for AVs, so I’m guessing Waymo is still experimenting with which model they want to go with. I wonder if we will see Waymo or Cruise launch on Lyft as well.
Wow, how’d I miss that? I can’t find any articles about the partnership ending. Did they ever reach the stage where the cars were actually available for riders? Or did it end before that?
It was a quiet ending; not surprised there aren’t any articles about it. Cars were definitely available, many rides happened. I think they had safety drivers sitting in the driver’s seat but wouldn’t swear to it. You can find Youtube videos of the rides around Chandler, AZ.
This is going to be the real test for Waymo. From anecdotal experience, Austin has more inclement weather, and its road infrastructure is lacking in some parts of town.
Having lived in both areas, overall I find it easier to drive in Austin (even the trickiest parts) than in SF.
However, in my experience, Austin roads are uniquely bad in one respect: surprisingly often, lane markings are so faded you cannot tell where your lane is. It's even worse when it rains.
Sometimes it is so bad that I know neither I nor anyone else can tell where their lane is, and I make a conscious decision to drive differently to compensate for that. I don't know if autonomous vehicles are going to deal with that very well.
Austin doesn't get hurricanes either. The tropical storms it gets have precipitation roughly equivalent to an atmospheric river. Austin gets snow once every two years, and it exceeded one inch only once since 2010. That amount of snow is hardly noticeable.
Another danger is cars getting swept away at low water crossings. You hear "turn around, don't drown" repeated a lot in Austin.
It is not at all unusual for people to get killed this way. They think their car can make it across, and they're wrong.
It's a problem in many parts of Texas because severe thunderstorms happen frequently and some of them bring heavy rain. But in Austin, it's especially bad because Austin is hilly, which makes flash floods worse.
Flooding on roads is not super common in Austin. Some areas have flooding, sure, but it’s extremely localized. I think you hear it a lot as outreach related to other Texas cities.
Houston, by comparison, might as well be a river for how often it floods and how widespread flooding tends to be there.
Except most pollution from modern standard vehicle is tire particulate dust which is worse than ICE vehicles due to increased weight and road friction. Long story short we need more walking, cycling and trains, cars aren't the answer. https://www.theguardian.com/environment/2022/jun/03/car-tyre...
> The wear rate of different tyre brands varied substantially and the toxic chemical content varied even more, he said, showing low-cost changes were feasible to cut their environmental impact.
And also
> Dr James Tate, at the University of Leeds’ Institute for Transport Studies in the UK, said the tyre test results were credible. “But it is very important to note that BEVs are becoming lighter very fast,” he said. “By 2024-25 we expect BEVs and [fossil-fuelled] city cars will have comparable weights. Only high-end, large BEVs with high capacity batteries will weigh more.”
And
> The average weight of all cars has been increasing. But there has been particular debate over whether battery electric vehicles (BEVs), which are heavier than conventional cars and can have greater wheel torque, may lead to more tyre particles being produced. Molden said it would depend on driving style, with gentle EV drivers producing fewer particles than fossil-fuelled cars driven badly, though on average he expected slightly higher tyre particles from BEVs.
Once we've stamped out emissions from ICE vehicles, we have a new goal: stricter regulations for tires.
Developing countries often have less strict tailpipe emissions standards than developed ones. They'd also benefit from large-scale EV adoption.
I look forward to you convincing Americans to give up their cars. It’ll take decades to get the infrastructure once you do. I’m personally waiting for low-speed maglevs to catch on.
In the meantime, there are 16 million Ford F150s on the road. They’re heavy too, comparable to many EVs.
EVs turned political and people threw out the extra weight of batteries without comparing that to the weight of the current SUVs, etc that Americans drive.
You could look at Norway for an idea, (non-autonomous) electric cars have captured the market for new cars and for long enough that the older fleet is ageing out. They are at 26% of registered cars: https://elbil.no/elbilstatistikk/
The biggest challenge overall is going to be overcoming the issues with autonomous driving in inclement weather. But I’m sure it’s doable eventually.