Daily insights for city builders, delivered every morning at 6 AM ET. I’m Brandon Donnelly — a Toronto-based real estate developer and founder of Globizen. I’ve been writing here since 2013.
This is an interesting video by Phil Andrews of Maxinomics talking about the economics of Uber and what it could stand to gain from autonomous vehicles.
As part of this, he touches on the exclusive partnership that was announced between Uber and Waymo back in September. That was a big deal.
I find this topic fascinating because it’s hard to imagine it not reshaping the landscape of our cities. And it continues to get more real by the day.
One of the reasons why I’m interested in the autonomous vehicle space is that I know our built environment is sticky. And because we’ve designed so much of it around the car, it’s hard to imagine this dependance going away anytime soon.
In fact, there’s a very real possibility that autonomy leads to further decentralization. That is typically what happens when we make it easier and cheaper for people to travel longer distances. They sprawl. So we ought to start preparing ourselves for the positive and negative externalities.
I’m not a deep expert on autonomous vehicles, but as an interested observer, Waymo appears ahead of Tesla in delivering this future. Waymo has been offering fully driverless rides since 2020 and Tesla is still at L2 autonomy, which means a driver needs to be present in the vehicle.
I hear from lots of people that their Full Self Driving (FSD) software is pretty good, but according to some studies, it can require human intervention as often as every 13 miles. This doesn’t necessarily mean that Tesla won’t be first to “solve autonomy,” but their Cybercab isn’t here yet.
If you’re interested in this topic, here is an article by Timothy Lee summarizing a discussion that he recently had with the co-CEO of Waymo, Dmitri Dolgov. Some of it is a little technical, but it does offer a comparison between Tesla and Waymo.
That is, how their approaches differ and what the future might look like.
I watched Tesla’s We, Robot event last night. As many of you know, Elon and his team showcased a Cybercab, Robovan, and a humanoid robot that dances funny, all of which will be available in the market for purchase at some unknowable date in the future. What was obvious is that Elon himself has no clear idea of when this will be.
What I will say, though, is that the designs look cool. The Cybercab looks like a Porsche and a Cybertruck had a love child, and the Robovan looks like an Art Deco rendition of what the future is supposed to be like. I first wondered why they’d create a robotaxi with only two seats. But thinking about it now, most Uber rides probably only have 1-2 passengers.
Despite these pretty designs, the overwhelming reaction to the event seems to be one of disappointment. We’ve heard what was said before. Public transportation is bad (I disagree). Autonomy will free up your time and remove unnecessary parking spaces from our cities (allowing for more public space). And soon you’ll be able to put your under-utilized car to work and earn extra cash.
Cool, but when?
Waymo and Uber are not, as far as I know, hosting similarly flashy events. But as far as I can tell, they’re making meaningful progress in advancing toward full autonomy. As of June of this year, Waymo had already logged over 22 million rider-only miles. And in September, they announced a partnership that would bring AVs to Austin and Atlanta by way of the Uber app.
At this point in the hype cycle, I don’t think anyone is interested in hearing promises about what the future of autonomy will be like, especially without any firm dates. They want to know: Are we there yet? So I think it’s no surprise that people, including investors, weren’t all that pumped up by the event.
On a more important note, Tesla had bicycles with brightly illuminated wheels circulating around their event set (at Warner Bros.) to presumably demonstrate that their Cybercabs can successfully navigate around moving objects (when brightly illuminated). If you missed them, look at the 29 second mark in the below video:
I can’t be the only one who thought: “What are those? Now, that’s what I want!” So I’ve asked Elon when they’ll be available and when I can buy one. I’ll keep you all posted on his response.
Waymo has started releasing statistics for its autonomous vehicles. Here’s the link.
There are a number of important considerations when comparing human-driven and autonomous vehicles. For instance, the two have different definitions of a crash. AV operators have to report any kind of physical contact (property damage, injury, or fatality). Human-driven cars, on the other hand, don’t typically report accidents unless it was bad enough to necessitate a police report. So there are nuances to keep in mind.
That said, there is an argument to be made that AVs are already safer than human-driven ones. Through to June 2024, Waymo had already logged over 22 million rider-only miles. And here is what it is now reporting in terms of airbag deployments, injury-causing crashes, and police-reported crashes:
All of them are lower than their respective benchmark crash rates.
A few days ago, Waymo announced (on X) that its robotaxis are now doing more than 50,000 paid trips every week across Phoenix, San Francisco, and Los Angeles.
This means that the company is getting an average of 300 bookings every hour or five bookings every minute. And if you add in Austin, where it’s currently offering a limited number of rides, the company has completed a total of over one million rider-only trips.
In the announcement, Waymo also went on to say that “fully autonomous ride-hailing is a reality and a preferred mobility option for people navigating their cities every day.” All of this is something.
But perhaps the most important takeaway, right now, is that the company continues to claim — by way of a study from Swiss Re — that its robotaxis are already significantly safer than human-driven vehicles.
I don’t personally know if this is true, but it’s not hard to believe. I mean, human drivers suck. And assuming it is true, we should all want more robotaxis on the road, because statistically, we would be significantly safer.
The problem, though, is that autonomous vehicles suffer from a perception bias. We’re all looking for them to fail. If a robotaxi gets into an accident, it’s news. But if a human driver gets into an accident, it’s standard operating procedure. It’ll be interesting to see how and when this flips.
Remember Wuhan? Well, it turns out that it is emerging as an important hub for driverless vehicles. Right now it is home to the largest fleet in the world:
In Wuhan, 500 robotaxis, mostly run by Baidu, China’s rival to Google, recorded more than 730,000 ride-hailing trips last year. That compares with combined orders of more than 700,000 last year in Phoenix, San Francisco and Los Angeles, according to Waymo, the self-driving car developer of Google’s parent company Alphabet. Waymo told the Financial Times that it had “a couple of hundred cars” in each of the three fully autonomous zones.
One of the things that is allegedly helping Chinese companies is that they have access to more data. The networks of cameras and other infrastructure that make Chinese cities the most surveilled in the world are, coincidentally, also good for training machine learning models.
This has some industry experts speculating that China could reach an autonomous vehicle “tipping point” sometime around 2027. Meaning, the technologies will be significantly safer than human drivers (at least 10x) and ready for mass adoption.
I don’t know if this is the right timeline. There have been many forecasts made over the years. But I do know that competition is good for progress and that having a rival can be an important motivator. And right now, this is yet another example of the US vs. China.
One of the realities that we will have to face in, oh I don’t know, 5 or so years, is that there will be a mixture of different cars on the road. Some will operate with drivers. And some will operate with no drivers. Assuming that the cars with no drivers do well at their job, I would imagine that this will become the default. But in the interim, it’ll probably be useful to know which is which. And that’s why Mercedes-Benz (and probably others) has been working to establish a new internationally-accepted signal for computer-driven cars.
The decision so far: turquoise lights.
The company has just received permits from the states of California and Nevada for its Drive Pilot system, and as part of this, turquoise lights were earmarked for this exact purpose. Supposedly turquoise was chosen because it’s distinct and because there’s nothing else on the road that uses it. But I think the real reason is that it looks cool and kind of cyberpunk. So I hope this does become the standard way that we all visualize our shift toward autonomy. I can already imagine the long-exposure photography that will follow of our roads.
Images: Mercedes Benz via The Drive
