Lyft CEO David Risher on paying drivers more and the shift to robotaxis

Today, I’m talking with David Risher, who is the CEO of Lyft. I’ll just say from the jump: I think you’ll like this one, since David is refreshingly direct and doesn’t pull a lot of punches.

He has been on the board of Lyft for years, but he only stepped in as CEO just a couple of years ago to help turn it around. He’s done pretty well with that so far, but he’s pretty straightforward about how the company wasn’t doing well, and he had to make real changes to fix it. That also means he has a clear thesis about what kind of company Lyft really is — a service company that operates in the real, physical world, as opposed to a tech platform, which is very much how its big competitor Uber sees itself. 

Uber comes up a lot in this conversation, actually — the competition between the two is just as fierce as ever, and you’ll hear David make a lot of references to “the other guys” throughout this episode. But it’s not just competition for riders and drivers that Lyft has to deal with. It’s the future of transportation itself, and new AI tools that might take apps like Lyft out of the equation entirely.

David and I talked a lot about autonomous vehicles and how they’ll impact riders and especially drivers. I always ask my rideshare drivers what I should ask the CEOs of the platforms when I do these interviews, and the only thing they ever ask is simple: when are you going to pay us more? So I asked David straight up: can Lyft pay drivers more money, especially when the promise of autonomy is to replace the drivers entirely?

You’ll hear David point out that it’ll be a long time before Lyft or anyone else gets to the point where self-driving cars are the default. So for now, Lyft is still a service company with humans doing the work. But the transition to a world of robotaxis is going to upend that system over time, and David has a lot of ideas about how it might play out. 

Then there’s the other kind of problem with AI, what I’ve been calling the DoorDash problem. In a world of AI agents going out and booking cars and ordering sandwiches for you, apps like Lyft and DoorDash might just turn into commodities, not companies anyone interacts with directly.

Lyft, in particular, is the exact kind of service that seems really susceptible to this problem, given how many people will just flip between Lyft and Uber based on which one’s cheaper on any given day. So I really wanted to dig into that with David, to see what he thought a service platform like Lyft could do to retain loyal customers — customers they can sell subscriptions and other services to — when users might not be opening apps at all in the future.

There’s a lot going on in this one, but I also have to point out that David is one of the only Amazon or ex-Amazon people to give an original answer to the standard Decoder question about decision making. Like I said, he’s pretty direct. 

Okay: Lyft CEO David Risher. Here we go.

This interview has been lightly edited for length and clarity.

David Risher, you’re the CEO of Lyft. Welcome to Decoder.

It’s great to be here.

I am very excited to talk to you. It feels like I’m having a lot of conversations with various service providers, I would say, across the industry about how AI might be changing, how they get customers, how the platforms themselves are changing, the nature of the people who work on the platforms and provide the services, and Lyft has been there since the start.

It’s one of the very first app economy apps, right? It’s the progenitor of the gig economy; it started, I think, with Uber and Lyft. You have been turning the company around. You’ve got some new ideas. There’s quite a lot to discuss, so I want to start at the start. You’re a new-ish CEO, I would say, a couple of years into it. I think most people are familiar with Lyft in the popular conception of Uber and Lyft. I think I have a sense of what Lyft is. I use it quite a lot because I have a credit card that gives me rewards when I use Lyft. I price-match the two all the time. 

I’m just curious, what is your conception of Lyft today? There’s Lyft in the popular culture, there’s the Lyft many people have experienced, and there’s what you, the CEO, think it is and what you might want it to be. What do you think Lyft is today?

You know what, actually, I’m going to start with what I want it to be. What I want it to be is a way to serve and connect you better than you’ve ever been served before, and connect you to the real world. And so let me say a little bit about this. In a world where the virtual technical world is bigger and more powerful every single day and seductive, I want to be the one that gets you out and makes you part of the real world and maybe connects you to the best Lyft ride you’ve ever had because you have an incredible conversation with your driver or maybe you meet your future spouse in the bar you’re going to. 

That’s really what I want. I really want us to be the physical glue that holds our society together, and do it in a way that blows your mind from a service perspective.

When you say physical glue, do you mean transportation, or do you mean other services? There are a lot of ways to interpret that.

Yeah, for sure. Yeah, no transportation, that’s our bread and butter, that’s what we do. We do it mostly in cars 800 million times a year. If you live in New York City, we do it on Citi Bike; if you live in San Francisco, we do it on Bay Wheels; if you live in Chicago, we do it on Divvy. Increasingly, we’re doing it overseas as well through Freenow. So yeah, it’ll be through transportation, but transportation is a big deal. If you’re older, it’s how you stay connected to your grandkids. If you’re younger, as they say, it’s how you get to work every single day. It’s part of your daily life, and I don’t see that going away anytime soon.

When you say transportation, again, most people today think of…

Yep, rideshare.

Honestly, when you use these apps, a Toyota Camry shows up. What we have developed with all of this technology, billions of dollars in investment in fiber optics and wireless, and 5G, is that you can push a button on your phone and, to a high degree of certainty, a Toyota Camry will show up, which is pretty amazing. That was not true before all this investment. That’s one version of it. 

There’s another version where a robotaxi shows up, or you get a bike, or a shared service shows up of one kind or another. Are you thinking that broadly, and are you thinking about the transition from there’s a driver and a car to maybe it’s a robot, maybe we’re telling you to take a train? There are a lot of ways to think about that. How are you framing that in your mind?

So I would say… So great question, and maybe a nuanced answer. I actually think most people, when they pull an app like this, they kind of know how they’re going to get from A to B, they kind of know that already. So I’m not super focused on maybe it’s a train, maybe it’s a vertical takeoff aircraft; I’m pretty focused on a car, or maybe a bike, is going to be what you’re going to use.

Now, it’s going to change. So let’s use bikes, which is not the place that most people start, but e-bikes today are going bananas, absolutely bananas. And you can feel it, you can feel it in New York or San Francisco, where a couple of years ago, biking was a niche thing, and now it’s a huge, huge mode of transportation, and e-bikes are the reason for that. On the car side, AVs, autonomous vehicles, are going to be a game changer, a game changer. You can sit in the back seat, you can snooze, or you can go into party mode. Maybe there’s a car tender in front who’s making a drink while you’re driving, all kinds of crazy stuff.

But I’m pretty focused on people jumping in a car. Today it’s a Camry, maybe tomorrow it’s something else. We can talk about that. Today it’s driven by a driver, probably tomorrow it’ll be driven by a driver, but also driven by a robot. I’m a, let’s say, an advocate of focus in the technology, but where I’m expansive in my thinking is all the different cool things that you should be able to do by getting out of your house and not just sitting on the couch and watching Netflix and getting a food delivery.

A version of that that I’ve heard several times, most notably, I think, from Brian Chesky from Airbnb, who was on the show, was that we should also start selling the experiences. We want to get you out of your house; we want to get you doing things. Airbnb launched an entire platform that was bigger than just house rentals, all experiences, like where you can have a private chef. That’s a big expansion of the platform. Brian is very convincing when he talks about it. Is your head there? You should open Lyft, and we should send you to a concert?

Maybe. I mean, we’re earlier, I think, in that journey than he is, but I think that the destination is pretty similar. Literally, I actually went to a Dua Lipa concert a couple of nights ago, and it was so fun, and it was great. I can listen to Dua Lipa on my AirPods, and I can do it when I’m walking down the street. She can be with me all the time, but it’s 10000 percent better if we get you there, and we put you in the right seat, and we make sure we pick you up at the end, and we encourage you to do it. And maybe if you’re a Lyft member, you get some sort of special service: you go to a restaurant, you get a special dessert that’s not on the menu, or maybe you get special access to a lounge at the airport. Yeah, I think you’ll find us doing more and more of that.

I don’t want to over-rotate. Look, just getting you reliably hundreds of millions of times a year to where you want to go, picking you up instantaneously rather than having you wait five minutes, not having the driver cancel on you, all of these things. Making sure you get your points, making sure you can spend your points if you want to, all of those basics really, really matter. But I think over time, we need to be advocates for the physical world because the digital world is fighting pretty hard for your attention, and I don’t think that’s a great place if that’s where we end up.

There’s an obvious comparison to Uber that’s going to come up over and over again. I’ll pre-apologize for the obvious comparisons to Uber that come up over and over again, but there’s one here. Dara was just on the show. I saw him again recently. They had a big announcement: they’re becoming like a work platform. What they, I think, abstractly see Uber as is that there’s supply, there’s demand. We are really good at matching supply and demand. We can do that for cars, we can do that for Waymo, we can do that for food. What if we just did it for everything?

And I think the first thing they’re going to do is AI training, which is a wild first thing to do, but they’re like, “Yeah, we got a bunch of drivers who are looking for work to do, opening an app every day, and maybe we can just put other kinds of work in front of them, not just driving cars.” That’s very different from “we’re going to send you to the Dua Lipa concert”; it’s a very different point of view. Did you evaluate a similar idea? Did you say, “We don’t want to just be a work platform; we want to be an experience platform”? Because that feels like a very big decision.

I agree. Look, I won’t comment on those guys, but what I will say is I really like… Okay, I’ve been in the job for two and a half years now, and one of the things that I said from day one is we’re going to be customer-obsessed. And I know you think a lot about Jeff Bezos. I worked for Jeff for a long time, so I don’t have to tell you all the reasons why I think that’s a good idea. There are two passengers, excuse me, two customers in every car: a rider and a driver. And I want to do everything we can to get to know our riders as well as possible and our drivers as well as possible and understand what it is they want.

And so I’m super excited about looking at the world through that eye and say, okay, from a driver’s perspective, I want to make more money, I want all sorts of different things, but I may be a little less enthusiastic about the concept of becoming “a technology platform for everything,” because I think what that tends to do, at least in my experience, is make you less focused on the customers and what it is they really care about and more about let’s build this cool tech to do a whole bunch of different things.

Let’s talk about that two-year journey. You joined the company in 2023, and you made a lot of changes at the beginning. I would say Lyft was not doing well. Immediately, you laid off more than half of the company, which you said, there are quotes, you said that it was very hard. Why’d you make that decision? Why’d you have to slim down right away?

To be able to pay drivers what we needed to pay them and to be able to charge riders what they wanted to, what they could afford. So again, if you start with the idea that customer obsession is what’s going to drive our profitable growth… And that was the thesis. The thesis… I can tell you the whole conversation about how I got selected for this job and how I said no to it, but eventually said yes. And the yes really came to, look, if the board of directors believes that customer obsession is what’s going to drive our profitable growth, then maybe I’m the guy. And if you believe it, there are a whole bunch of implications that come from that. And the first thing is our cost structure — it doesn’t allow us to do what we need to do, which is to pick people up highly reliably, yes, but also at a price that they can afford, and so on and so forth. So that was that, that was that, full stop.

Take me to that room. Vanishingly few people ever get to go interview with a board of directors to be the CEO of a big public company and say no and get called back in. Walk us through that. What was that actually like?

Sure. Here it is, and I’ll go step by step. So I had been on the Lyft board for a couple of years. John and Logan, the co-founders of the company, had invited me to be part of the board mostly because I think they had a very interesting observation, which is that boards don’t tend to think a lot about customers; they think a lot about strategy, they think a lot about finance, but they’re pretty far removed from customers. And I had come up, as you know, I worked at Microsoft in the early days and then for Jeff for a long time, even Worldreader, the nonprofit that I founded, all of these, customer obsession was right at the center. So they’re like, “Look, David, how about you join the board?” So this was in 2021.

In 2023, at the end of the year, John and Logan decided to step back. They had been doing this for a long time. They were about to turn 40 years old. The only thing they’d ever done. Time to turn it over to somebody else. And so the board did what they do, which is just formed a committee, looked at a bunch of candidates. I wasn’t part of it; I was just observing from afar.

And then one day, it was actually Valentine’s Day, I remember it very clearly, 2023, my phone rings, and the board chair, Sean Aggarwal, is on the line, and he says, “David, we’ve got an offer we think you can’t refuse.” And I’m preparing myself for, “We want you to be the chair of the audit committee.” Some terrible thing that he’s trying to butter me up for, whatever. And he’s like, “No, John. Logan and I have been thinking, and as we’ve been looking at all these external candidates, we’ve been evaluating in the back of our heads, maybe the right guy is sitting right here next to us. And David…”

And honest to God, and this is not… I said, “No, that’s ridiculous. I don’t even know what you’re suggesting, but I can tell you it’s not… I’m very focused on getting kids reading. I’ve been focused on that for many years with the Worldreader, which is a nonprofit I started. And you need to hang up the phone immediately, and you can get back to work, do something which has a higher likelihood of success.” But Sean said, “Why don’t you think about it?” And so I did. I literally took a walk around for about an hour, and I thought, and I kept hearing myself say, “Hmm, interesting.” As I mentioned, it was Valentine’s Day, so this became the topic of conversation between my wife and me that evening. And she said, “David, I think you should give it a try or go for it.”

So anyway, then John and Logan, a couple of days later, came over, and they sold me on the idea a little bit. And then they did something, which I don’t think they were being clever; I think they were just being honest. They said, “Just to be clear, we’re not offering you the job; we’re offering you the chance to apply for the job.” And I’m like, “Hold on. Now I’m getting competitive.”

So anyway, over the next… It was about a six-week process. I literally put together a 100-day plan. I talked to every individual board member. Some of them thought it was an interesting idea that I was applying. Some thought it was a crazy idea, like, “This guy? That doesn’t make any sense.”

But anyway, I put together a 100-day plan. I still have it. It’s actually interesting. I was looking at it recently, and the thesis of it was that I want Lyft to lead, and the kicker at the end was, and I want to lead Lyft, and everything in between the two were all the things we needed to do including lay off a big part of the company, changing the composition of the team, starting to innovate again around customers, and on and on and on. Anyway, that was it. One thing led to another, they offered me the job, and I started on April 17th, 2023, and I’m having the time of my life.

Again, vanishingly few people ever get to do this. So I have some very weedy questions.

Do it.

What software did you use to put your presentation together?

That is a weedy question. Two. Google Docs. So first, it was literally a written document. I wasn’t at Amazon at the time when Jeff did the whole “everything has to be a written document and don’t use presentations.” That was before all that happened. But I have always… I like to write, and I express myself through writing. So anyway, I wrote a document that literally was a page of text, and then maybe 2.5 pages of outline, bullet points type thing. That was phase one. And then that turned into a slide show, Google Slides, yeah.

Google Slides. The reason I ask this is I think it’s such an abstract thing, but you sat down and opened Google Docs like anybody else would open Google Docs and thought of a bunch of ideas to turn around Lyft, and then you presented them, and there was some conversation. The board said, “Yeah, that’s what we want to do.” 

Where in that process did you think… Because you were on the board. Where in that process did you think, “Boy, this company has gotten too big and too unfocused, and I need to make these two big changes. I need to cut a quarter of the company and turn over its leadership.” Because somewhere you open Google Docs… Was that the first thing you wrote down? That mechanical writing and thinking process is just so fascinating to me.

This is super interesting. I hadn’t thought about it at any level of depth for a while. So I guess here’s what I knew: There were two things that I absolutely knew: that we had to focus on customers. Again, it sounds cliché, but I can give you an example, okay. This was something that I detected while I was on the board, but didn’t really understand until I was inside the company. We would look at service metrics. An example of a service metric might be driver cancellations. And this wasn’t something that the board typically would look at, but I would have a particular interest in it. So I would say, “Let’s talk about driver cancellations because I have this frustration. I open up the app, and some percentage of the time, I get matched with a driver, and then three minutes later, it says, ‘You’re going to rematch with a new driver,’ which I find irritating, and it also lengthens the process. I don’t like it.”

And so I found out, this was actually after I joined the company, but it still tells the story. This is a company that said they were customer-obsessed. I’m like, “Okay, let’s talk about what that really looks like.” So I said, “Okay, let’s look at…” And they said, “Well, yeah, okay. So it’s about 15 percent of the time that this happens, 15 percent. But the good news is, 95 percent of the time, people end up rematching and taking the ride. So no big deal. Most people are still taking the ride.” 

I’m like, “Okay, hold up, you’ve just glossed over the most important thing, which is that 100 percent of the time it happens, it’s a pain in the ass, and the rider is frustrated by it. And so I will guarantee you, you don’t have to go and do a bunch of research on this, I already know, people who have that experience are less likely to take rides in the future. So you can go ahead and decide if you want to look at that or not, but I already know the answer to that.”

So I said, “Let’s focus on this,” as an example. Early on, this was not my biggest decision; it was one of the smallest. But again, maybe it tells the story. So it was 15 percent of the time that this would happen, and I said, “Let’s focus on it like a laser. Let’s talk about what information the driver gets when they’re making this decision. Let’s talk about how big the font is. Let’s talk about whether we’re talking about it in dollars. Remember, what’s happening in the background is that a driver is deciding whether or not to take the ride, and then, for some reason, a couple of minutes later, deciding differently. So maybe we’re telling them too early, maybe we’re telling them too late, maybe we’re not giving the right information, maybe it’s up on the screen for too little time, maybe the font is too small.”

We looked at every single one of those things. When I started, it was 15 percent; a year later, it was 10 percent; three weeks ago, it was 5 percent. As of this last Wednesday, it was 4.5 percent, so a massive change. But that’s the customer obsession side. So customer obsession, that was the first thing I knew, was that we have to get really customer-obsessed, not just blah, blah, blah. And then number two is we don’t have the right people on the senior management team. And this is old-school Jim Collins, Good to Great, if you’ve ever read the book. If you don’t have the right people on the bus, it just doesn’t matter.

And so I asked our CFO to leave very shortly after joining, and that was its own thing. And then there were cost structure things and innovation things and so forth. But those were the two basics that I started with when I opened up that Google Doc because I knew we had to make changes in personnel, and I knew we had to reorient the company around customers. And then yes, I knew that in order to pay for some of what we had to do to reorient ourselves, we were doing too many things, and we had to cut a lot of staff.

This leads right into the Decoder questions. How is Lyft organized now? How have you structured the company now that you’ve been on a job for 2.5 years?

Sure. So let’s see. Well, I answered the question, but I’ll give you a little bit of context. The question is that we’re organized by customer, or excuse me, the answer is that we’re organized by customer. So we have a rider group, we have a driver group, and we have what’s called a marketplace group. That group is in charge of matching those riders and drivers in real time, 24 hours a day, seven days a week. So it’s operationally, but also computationally, quite complex. We have a group that is focused on our ads business, which is a relatively newer business, a relatively small business, but at a $100 million run rate with high growth and high margins, it’s awesome. They also do some other things that are newer product types. Of course, then we have a bunch of central functions like marketing and legal, and so on and so forth. But that’s really the primary. Oh, and we have a backend group that does a lot of the infrastructure, but really, I’d say the primary organizing vector is by customer.

So when you organize that way, some of those central functions can get pushed in different directions. Engineering is a central function, but if you want to build the concert experience, you’ve got to devote some resources to that versus bringing down rider mismatches, right?

How do you make those decisions? How do you balance that tension out?

We put a lot of our engineering into those customer groups. So those customer groups are full-stack groups. They’ve got product management, they’ve got engineering, they’ve got tests, they’ve got design, all the rest. So we accept the fact that there will be some redundancy and some distribution of talent. Then you have to ask the question: How do you maintain, let’s say, excellence across… functional excellence? How do you make sure your engineering is operating at top talent? And we’ve identified people or teams or whatever to drive that kind of horizontal excellence across the company.

But the trade-off is that there’s some redundancy, we’ve got some stuff happening in… And we have two apps: a driver app and a rider app. You can imagine a world where there’s one group that develops both apps using the same frameworks and all. We don’t have that; we have two different groups. They develop it using two different frameworks. It’s a pain in the butt sometimes, but you make it work; it’s better that way because you’re close to your customers as opposed to close to your technology. That sounds great until you realize you have lost track of what your customers care about.

Amazon famously organized this way. I can just issue some criticisms of Amazon broadly. I know what those trade-offs are. We talk about the structures on the show all the time. If you look at how Amazon is, they have lots of single-threaded owners of two pizza teams that make their own products. Those products rarely talk to each other. 

As you were describing, you end up in a lot of different… Google, the same way. You end up in a lot of different directions, and suddenly you’re like, “We’ve got to roll out AI across the company,” and you don’t have a common shared framework to do such a thing.

Have you run into this at Lyft? Are you aware of this trade-off? How are you managing that?

Remember when you asked me about my vision for Lyft and I talked a lot about rideshare and a focus on getting people around and making sure that… So I would say a strength that we have is we’re really quite focused on our customers and our use cases.

And so while yes, occasionally those issues crop up, it’s a small thing for us. And also I would say, and this is… I don’t know if this is a good or bad thing, but it just is. I am very involved with product decisions. Very, very involved. Just thinking about the last 24 hours and how I spent my time and product is just… And it’s, again, there are parts of the team that like that, there are parts of the team that find it a little frustrating, but I have no problem saying, “We don’t need to do these three different things; we’re going to do this one thing here, we’re going to do it super well. And that means that these other two teams that thought they were going to get to work on those things, we’re just not going to have that happen. Instead, we’re going to have them focus on something else.”

So I guess a little bit of, to a certain extent, we solve the problem by focusing because we’re focused on one thing and not… Amazon’s focused on many things, but we’re focused on one thing, and then second, I play a pretty big role there in breaking ties and…

That does seem like the way to make this structure work; you need to have the leader who’s just going to show up and break ties all day long. It also seems like scale is the other… that the leader can’t scale. At the same time, if you want to attract great people, you have to give them some autonomy. How do you balance this? What’s the cadence of letting your folks do what they want to do and then showing up and telling them they have to do what you want to do?

I mean, it’s such a classic issue. And so I actually wrote about this last year in the shareholder letter. I wrote about two things: I wrote about stratification, why products tend to get worse, and how we’re pushing things the other direction. And then this topic, which I called Falcon Mode. So the sort of visual that I wanted people to think about is the falcon, which is flying at 2,000 feet and hangs out up in the sky often because they need to see everything. They’re looking for, Where’s my next meal? And they’re pretty good at, even at 2,000 feet, seeing where that next meal is. And then they dive in deep, and they have to get the meal; otherwise, they starve and fall out of the sky. So this coming down and going back up and coming down and going back up, that’s the world that the CEO lives in.

It’s art. And part of it is expectation setting. Part of it is telling my team, “I am going to do this. I am going to talk in excruciating detail about this loyalty program that we’re in the process of developing.” And anyway, that’s going to be where I go Falcon Mode on you, but then I’m going to go way, way up, and I’m going to say, “Now it’s yours,” and you’re going to tell me all the ways where I got it wrong or you’re going to make it better or you’re going to push back or whatever it is. I don’t know. There’s no easy answer here, but I think a lot of it is just maybe being judicious because if you do it…

I’ll say a little bit of an adjacent thing, sorry for going into such detail on this. This is actually a comment that I heard Gavin Newsom of all people say, which I thought was very interesting. He was saying you have two types of power as a leader: you have positional power, and you have moral authority. And the difference is with positional power, the more you use it, the less you have of it. So if you use too much of it, you squander it, right? Because people eventually get tired of being told what to do.

Moral authority is a little different. If you say, for example, in my case, “We’re going to be a customer-obsessed organization, and we’re going to look at everything that lands, and occasionally, I’m going to come in and remind you what that really looks like, but then I’m going to back way off.” The more you use it, the more it creates itself, it reinforces, and people go off and they have their own amazing ideas and stuff. So anyway, that’s the mode I try to get to. You can ask people on my team whether I’m successful or not, but I try to be very, let’s say, deliberate about the balance between the two.

I’m always curious when you end up in that divisional structure, there’s an amount of just re-coordinating that needs to occur, and most tech companies have chosen against it. So it’s fascinating that you’ve chosen this way, and you are very clear that you actually need to do that specific task, because I have so many conversations with… I mean, the number of CEOs who are like, “I don’t do anything,” which is very funny, is very high.

Let me say one little tiny thing about that. So Scott Cook, who I’m sure you know of, the founder of Intuit and still very active on the board, he’s just in the process. I just saw him a couple of nights ago, and he’s someone I’ve known for many years. He was on Amazon’s board in the early days, and we’ve reconnected over the last bit. Anyway, he’s actually writing an article that I think comes out any day in the Harvard Business Review about a study of a couple of companies where he makes the case that the best companies are the ones where the CEO focuses not just on the what, but actually on the how, actually gets involved in the how.

This whole “I don’t do anything as a CEO,” he’s like, “That’s bullshit.” If you’re running a company, you’re doing a lot. And a lot of it is not just the big ideas; it’s how are we actually going to organize? How are we actually going to get this thing done? So I don’t know, I’d be a little skeptical. I don’t know. I think the CEOs are saying that either… I don’t know what that’s all about. But anyway, it’s not who I am. I’ll just say it that way.

I feel like our producers and I could do an entire episode of Decoder just on why we think some people say some of the things they say. Speaking of which, we have a little side bet going on on how you’re going to answer the other Decoder question: how do you make decisions? What’s your framework?

Okay. I mean, the obvious thing, and it really is true, is that I start from the customer and work backwards. I don’t know whether that’s what you were betting, I would say, but that is actually true.

I’ll say maybe a different thing, though, that I haven’t talked about publicly too much. So I guess, okay, I am blessed in the following way: I don’t find making decisions super hard. And what I mean by that is I think there’s a way of… In a sense, all you’re doing as a CEO or any leader is making decisions, in a sense. Like yes, no, hire the right people, fire the wrong people, say yes to the good ideas, say no to the bad ideas. That’s the job. And there are a lot of decisions there. Is it the right person or the wrong person? Do I fire them or keep them? I love them like a brother, but maybe they’re not the right person. All these things. And then is this a good idea that’s going to scale and customers are going to love it, or is it a bad idea that was just dumb in the first place?

Okay, so that’s kind of a framework, I guess, but not really; it’s just an observation about the job. And then I get down one level and say, “Well, I don’t personally mind making decisions a lot, I don’t, but I am aware that every decision takes a certain amount of effort. It does.” And so what I try desperately to do is make the biggest decision I possibly can so that everything else just becomes almost a checklist.

Let me give you a personal example. Years ago, my wife and I sat down, this was back in the early 2000s, and said… We met at Microsoft, and we said, “It’s interesting. At work, we have these multi-year plans, but here we have a family.” We have two daughters, as you and I were talking about before we started. At the time, they were very young, and we said, “What’s our multi-year plan for our family?” And we came basically to the conclusion that we want to live outside the United States at some point. We want to give ourselves and our kids that experience. That was the big decision.

Then there are a bunch of questions about where, when, what schools, and how to get insurance. All these things. But we already made the big decision. And so everything else was just a checklist. And we ended up doing it. A couple of years later, we moved outside the United States. It turned out to be a very, very long and very important thing for our family to do. But I told that very long story to say, I try to hold myself to what are the biggest decisions that can possibly be made, where once that decision is made, everything else becomes just a checklist. And then frankly, I don’t worry a lot about… If things are then on track, I don’t have to worry too much about it, and I can go on and make the next decision.

Okay, we all lost the bet. Congratulations. I think we should send you an award. You are the first ex-Amazon person to ever say something other than that there are type-one and type-two decisions.

Oh, God, yeah. Oh, shit. Oh, yeah, yeah, yeah. I mean, yeah, sure.

Literally in the pre-production, the note was, “He’s from Amazon, he’s going to say there are two one-way doors and two-way doors,” and we were all like, “All right, we’ll just get through it.” You’re the first one ever, the first person who’s ever come within a 100 miles of Amazon headquarters, who did not immediately say one-way doors and two-way doors.

I feel proud. I have my own ideas. Look at that.

Very good!

Let me ask you about some stuff that is changing, that I think you’re going to have to make some decisions about. And honestly, it will stress some of your structure. AI is here, and it’s happening in a lot of ways. Every CEO of a service company, whether that’s TaskRabbit or Uber or whoever has come on the show, I’ve asked this question. I’ve been calling it the DoorDash problem. I should probably get the people from DoorDash on the show to actually ask them directly about this thing that I’ve been calling the DoorDash problem for six months.

But just a couple of days ago, OpenAI had DevDay, and they showed a bunch of integrations where you could ask ChatGPT to go do stuff for you, including booking an Uber. We’ve seen other agentic products. Amazon announced Alexa+, which will be able to book a flight for you and will traverse websites; it’s built into Chrome now. We’re going to traverse websites on your behalf and do stuff for you.

The backend of that, whether it’s you or Zocdoc or whoever else is, well, we have a database of information, we know where all the drivers are. If you want to buy a sandwich, we know where all the sandwiches are. And so your agent’s going to come and order a sandwich on our website, and we won’t get the customer. 

We will just become a service provider to some chatbot interface, and we won’t be able to do upsells. We won’t say, “Hey, there are Dua Lipa tickets,” or whatever we’re going to say, and that’s going to shrink our margins, and we’ll just become commodity service providers. This feels like a very big problem. I’ve been asking everybody about it. Does that feel like a big problem to you?

I mean, maybe for the reasons you just said, but I wouldn’t say it’s one of the top five that I worry about. And a big part of it is, first of all, remember what you’re doing: you’re trusting something. You’re trusting that this thing, this person, is going to come and pick you up, and they’re going to just be on time, and it’s going to be safe. And if I leave my iPhone there, I’m not going to get the thing stolen. All these different things. And it’s physical, it’s safety, and it’s real-world stuff. And so the most extreme version of what you’re saying is I go to ChatGPT and I say, “Please come pick me up.” And some rando comes to pick me up, and there’s no guarantee, there’s no service, there’s no… That would be bad. I don’t think a lot of people would be super excited about just some rando coming, picking me up in an unbranded service, and whatever it is.

So if it’s not going to be an unbranded, just a rando picking me up, then it probably has to be one of the guys who are doing existing rideshare, and that’s us. And then we’ve got all sorts of ways where I think we can compete. So we want to compete on relationships, by the way, not just on transactions. And what does that look like? That [is something] you already mentioned: you choose us, among other reasons, because you get points on your credit card, an unnamed credit card, when you do that. Well, that’s still going to be the case in the future. And so you might have a preference for us that you push through ChatGPT. If they try to disintermediate, you say, “Well, no, I actually have a preference here.” And we’re going to do a whole bunch of different things to make sure that you have a very, very strong preference for asking for us by name, not just saying, “I want to get to a place.” And then second of all, remember that-

Wait, can you tell me what those things are? Because right now on my phone, the apps are side by side, and I open them both, and I will… if it’s within $5, I’ll pick the credit card points, but I will almost always pick the cheaper one. And I feel like an agent going off onto the web and finding the cheapest one is actually the most direct threat to your margins, to everyone’s margins.

So I don’t think it’s a big margin threat because we already price… Let’s talk about price specifically. So you are not alone; quite a few people price shop. Interestingly enough, from my perspective, I wish everybody did. And you’re saying, “Well, that’s weird. Why?” Because remember, I have a 30 percent share, and the other guys have a 70 percent share. And we price almost at parity. In fact, I mean, our strategy is actually to price a little less when we can, but it’s really hard because we have costs. And those are real costs: insurance, driver pay, and all these different things. By the way, the other guys have pretty damn similar costs, which is why our prices are so close like this.

Now, we might have a slightly different strategy; maybe we compete a little harder at airports, maybe they compete a little harder at something else, but it’s marginal. So my point is, why do I say I want everyone to check both? Because if everyone checked both, I’d win probably 55 percent of the time as opposed to 30 percent of the time. So that’s great for me. So let’s just, first of all, let’s just… Step one, I don’t mind that.

That is the margin pressure. If ChatGPT is saying, “Here are the two rates,” and the strategy to win is to always have the lower rate, you will quickly begin competing in a way that, right now, maybe you aren’t competing all the time.

No, no, we are. That’s the thing, we already are. That’s the basic… I think the premise of the question. This is true in some industries where price makes less of a difference, and therefore, if you’re… And nobody wants to be reduced to competing on price. But the truth is, that’s our life every single day. Every single day, we wake up and we look competitively market by market, where we are high, where we are low, how we can get lower, and so forth. So I don’t worry a lot about someone else. 

I literally don’t know how someone, a third party, not us or the other big guys, could underprice us consistently. People try, and they go out of business. That’s the way that works because they realize that the cost that they have is no less than the cost that we have. They just try to subsidize it through some other magical thing for a while, and they run out of cash. And we have a big scale and all these reasons why it’s hard to underprice us.

And then between us and the other big guys, again, there’s just not that much left. You know what I mean? So it’s like I don’t know how either one of us can underprice the other in a sustainable way.

But then back to the fundamental thing, of course, which is if you then believe that the price is pretty much the same, and again, if most people believe that, then I probably have a 50 percent share, not a 30 percent share, but I digress. Then it becomes: Who can get you the points? Who can allow you to pay with the points? For example, who can pick you up faster? Today, we pick you up about a minute and a half faster than we did a year and a half ago. And oftentimes, not always, but oftentimes, it’s actually faster than the other guys because we’ve got good algorithms, I’ll just say that. And we have drivers who like us a lot. That’s also a real source of strength. So anyway. And then there’s a service you can get in the car. What happens when you get in the car? Today, you might say it’s a little generic Toyota Camry.

But I don’t know. Maybe there are things we can do with the drivers there to make you feel a little bit more special. That might make you say, “You know what? I actually do have a preference.” Even if the price is exactly the same, I’ll still always ask for a Lyft over the other guys because I’m going to get a better experience from the driver. So that was a little bit of a roundabout point. But the reason I don’t worry too much about it is that I think it’d be very difficult for anyone to “go direct” to the 1.5 million drivers that we have. I think that’s very, very hard to do.

And then on price, I’m not super worried about being competed out of the game because I think I’ve got a pretty good cost position. I think it’d be hard for someone to underprice, and we already compete pretty directly. And so then I think it comes back to who can offer the better service. I feel really good about our ability to offer a great, great service.

So that your customer relationship will traverse whatever interface people are using? People will say, “I still have a relationship with Lyft.”

And remember, most people don’t change what they’re doing unless something else is 10 times better. Checking prices on airline tickets is hard. It’s dynamic pricing, huge swings. You’re planning ahead, vacation of a lifetime, all the things. But our staff is like… I’d say very few people ever send me notes saying, “I had a hard time with your app or getting a car,” or anything like that. So, in other words, I actually think that we already do a very, very good job. So I’m not sure of the additional value added. Again, if there was someone else who could come in and give it a much lower price, I might be worried. But I…

The other challenge there is that you might have a brand relationship, but they’re not actually opening your app. The app is where you might show them upsells to other products, or straightforward advertising, or explain and reinforce why your brand is more valuable than the other one, because it’s all happening in someone else’s app.

And so that would be a drag. Yeah, yeah, that’d be a drag. But I would say, at least what I’ve seen so far, just using ChatGPT as the example, they seem to be, let’s say, more accommodating around that than you might expect. In other words, they don’t… And again, it’s not that they can’t change strategies or that there’s no reason for them to, but at least the kind of apps that I’ve seen tend to be a little bit more… give more control over to the app developer than just a generic thing, because I think they also recognize brands matter to a lot of people.

My theory on that is that the agentic products that they are promising don’t work as well as they should, so they are forced to put app views in the chatbots. But when you hear them all talk about… across the board, not just OpenAI, but across the board, you hear them talk about their agentic products, there’s not a view that you’re going to get dumped into an app; there’s a view that the agent will actually do it. And once the agent starts doing it, your customer relationship starts to diminish. And that’s the heart of what I have been calling the DoorDash problem. Now it doesn’t matter where the car comes from.

Of course, it’s something we think a lot about, so I don’t want to diminish it at all. But if you have… This problem becomes a much bigger problem if A, you’re not used to competing on price already, or B, you don’t have a big supply that would be hard to get to directly. And it’s really hard to get to 1.5 million drivers and know where they are and all the different things about them in any direct way, hard to go around. But who owns the customer and all that? That is going to be played out. And part of my job, of course, is to make the Lyft brand so interesting and compelling that regardless of how you get here, you still feel it.

Remember, here’s the last thing I’ll say, people’s interaction with the app — and this is different from us versus some primarily tech companies — is relatively brief compared to the time they’re spending in the car. And so I would say part of my job is to figure out how to make that in-car experience even more interesting and, frankly, have a bigger place in your brain than just how you happen to get there.

Let’s talk about that for a second. You do have the other customer, the driver. Every time I get into a rideshare car, I ask the driver, “What would you have me ask the CEOs of these companies?” It is always the same answer. I’m sure you can guess what it is. They all want the rates to go up. I don’t think I’ve ever even heard another answer. They all just want the rates to go up. 

The idea that maybe the rates you charge to customers might change, and the world of AI or the margins might change if you’re not doing as many upsells in the app because of AI interfaces, is in direct conflict with, boy, the drivers want their rates to go up. They would like your cost to be higher. 

One, can you pay drivers more? This is the number one question I get, so I’ll just ask you directly: Can you pay drivers more? Do you see a pathway to doing that?

I mean, the short answer is we pay drivers as much as we possibly can. Our interests are much more aligned with drivers than I think the popular imagination and drivers would think. And the reason I say that is because we both have the same goal, which is to get as many rides going through the platform as possible and to increase the total volume.

Now, of course, at the individual driver level, they might say, “Well, gosh, I wish there were fewer drivers on the platform so I don’t have to compete with as many other drivers.” There are things like that. Riders would say the opposite: “I wish there were more drivers so that someone could pick me up faster,” but broadly speaking…

So let me back up for just a second because it’s such a big deal. Okay, first of all, let’s start with the super basics. How much do drivers make? Let’s actually talk about this for a second. Okay, we’ve studied this a lot. There’s a great white paper on our website that actually is super, super well researched and database, not just opinions.

Okay, broadly speaking, when a driver is driving on our platform, which means they’re either coming to pick you up or they’ve got you in the car and they’re dropping you off. They’re not waiting. So we’re going to come back to that in a second. When they’re driving, they’re making about gross nationwide, $30 an hour, 30, three zero. Now they have costs: gas, maintenance, cleaning the car, these things. Not insurance, we pay that, but repairs, things like that. If you take all those costs into account, it’s about 20 bucks an hour, 20 bucks an hour, so 20 bucks an hour. So let’s just say that. 

Now, so then you’re thinking, “Well, gosh, that doesn’t sound so bad.” Well, okay. But here’s the part that you also have to know, which is they’re not always getting paid that 20 bucks an hour because sometimes they’re waiting, and when they’re waiting, they’re not making money. Now, when they’re waiting, often they’re on the other guy’s app. And so net-net, they still might be making $20 an hour if they’re really good at flipping back and forth. But you can… I can’t guarantee demand. I can’t guarantee on my platform, on anyone’s platform, there’s always going to be demand. 

What do they get in return for all that, for not getting a guaranteed $20 an hour? They get the fact that they can turn the app on anytime they want, off anytime they want, they can pick up their kids, they can go on vacation, they don’t have to call in if they don’t feel like coming to work — that’s what they get. That’s the big trade. The big trade is: here’s how much you’re going to make, and we’re going to try as hard as we can to have you make more, as hard as we can. We’re going to build AI. 

You mentioned AI a couple of different times. A lot of our most interesting applications of AI right now are actually for drivers. We have a whole driver earnings assistant that allows a driver to go in and say, “I want to drive Monday, Wednesday, Friday, not Tuesday, Thursday. I don’t want to drive over the bridge. I got to go home, be there at five o’clock. Give me the best possible plan,” as an example. Or a driver reference letter. You started to drive 2.5 years ago. You’re one of our top five percent drivers. You’re super reliable. Here’s a reference letter you can literally take to your next potential place.

There are all sorts of things that we can do to… We can give gas discounts, we can do all kinds of things to try to make your earnings both on the platform and even if you decide to go do something else as high as possible. But the reality is, A, we can’t guarantee demand, and B, it is absolutely true that there’s a cap to how much we can pay based on 800 million data points a year of roughly how much riders are willing to pay.

And the last thing I’ll say is, you can see that if you look across the country, for example, in Washington state, Seattle in particular, rates are quite high there, quite, quite high. And average, let’s say, bookings $30 an hour versus $20 an hour because of local legislation that went in. And guess what? The number one complaint I hear there is, “I don’t get enough rides. I don’t get enough rides. What’s happened to all the rides?” I’m like, “Well, guess what? That’s what happens here if you don’t let the market set the rate.”

The other side of changing rates, pushing them higher, is that how you get more drivers on your platform versus Uber? The driver availability battles of the early rideshare days are pretty legendary. You had venture capitalists basically subsidizing both sides of the market.

That’s right.

Right. So you paid drivers high rates.

That’s right.

And then I remember, I just got around New York City for free on the bank of SoftBank’s money.

Basically. That’s right.

I don’t know what else the Vision Fund accomplished, but I traveled in style.

But you traveled in style. Yep, absolutely.

That was fun.

It’s awesome.

Obviously, that has all come to an end. Now the market is actually connecting supply and demand more directly. But that would be how you could take share from Uber. You’ve talked about taking share from Uber several times now. One way you could do it is to just say: we have more drivers because we pay them more money. Does that come out? Have you modeled that out?

We actually try to pay drivers more. Again, just like we try to charge riders less, but it’s very hard. This is not a high-margin business, but we look at it literally every single week or every single day, actually. Here’s what I can tell you that might be a little surprising: We have a 29-point advantage over the other guys on a dimension that I really care about, which is, “Which rideshare platform would you prefer to drive for?” We have 29 points in preference.

There’s actually another question that’s adjacent to it that we ask every quarter, which is, does driving for Lyft or the other guys give you a sense of pride? And we also have, just coincidentally, a 29-point gap in both of those for all sorts of reasons. Primarily among them, and all the way back to your first point, we do something the other guys do not do, which is we guarantee, we guarantee you will never, never, never as a driver make less than 70 percent of what riders pay after insurance is taken out over the course of a week.

And literally every week, we send out millions of dollars, millions of dollars of direct deposits to drivers to top them up to at least 70 percent. Usually, the number is about 85 percent, but 70 is the absolute floor. That is a huge, huge driver preference for us. The drivers that are maybe a little bit more thoughtful or nuanced in their answer to you when you ask the question would add… But I know a lot of them, they know it, but they don’t really say it because it’s not in their best interest in any way, they’re still whatever. I know that Lyft has actually done some real work here to make sure that our pay is at least a… Because that we can guarantee… We can’t guarantee demand; we can guarantee we’re never going to pay you less than 70 percent.

So that’s a very long way of saying it’s hard for us to consistently pay more. We try it sometimes, and sometimes market-by-market we do, but broadly speaking, again, this is a very, very efficient marketplace, and so what we try to do instead is all sorts of other things for drivers to make them feel appreciated, seen, well-paid, not unfairly paid, and so on and so forth. That’s kind of what I’m getting at.

The other pressure on drivers and rates on this entire ecosystem is autonomous vehicles. A lot has been said about autonomous vehicles and how they might displace drivers, how they might change those rates. The cars don’t quite drive themselves yet, right? There’s Waymo in a handful of markets, there’s whatever Tesla is doing in a handful of markets, but it’s coming. We can see it’s coming.

That’s right.

That requires an enormous amount of investment. Lyft, I would say, mostly you guys are in a partnership game across the board. That’s how you’re operating. How are you thinking about that, and how that might affect the drivers in your platform today?

Let’s start with exactly that. And then zoom out. So we’re actually very focused on something called Lyft-ready. So Lyft-ready… Well, this is too detailed. Let’s back up for one second. Yeah, driverless cars are absolutely coming. No question. And we have got some great partnerships. Everything from May Mobility, which is in Atlanta, which is a relatively small company, to Waymo, which is a very big company. We’ve said that we’re going to be working with them in Nashville next year, to Baidu, which is the Alphabet of China, where we’re working together in Europe. So from the absolute biggest to the smallest, yeah, we’re partnering.

And that makes sense. We’re good at supplying demand and matching with supply and pricing and mapping and lost and found, customer service, even fleet management, we do that. The other guys don’t do that. We have a whole subsidiary that does that, which is keeping the car serviced and cleaned and ready, but we don’t do AV tech ourselves, and we’re not an OEM, we don’t make cars, okay, so we partner.

Now, to answer your question, well, broadly speaking, actually, there will be two big sources of AVs. Some will be from, let’s say, people who have fleets of AVs. Maybe they bought a whole bunch of them, and they want to monetize them like that, the big fleets, and then some from individual owners. And I really do believe this, and I think this is where the intersection with your question gets so interesting. Today, we have 1.5 million drivers on the platform. What are they doing? They’re trading two assets they have: their time and their car for money. That’s the trade. I put myself in the driver’s seat, I drive around, I use my car, and I get paid for that.

Okay. Tomorrow, you can imagine a world where they can do the exact same thing. They can buy a car, it’s a self-driving car, because I think over time almost every car will be a self-driving car, and they can then flip a switch and make it Lyft-ready if we’ve done our job right, and put it on the Lyft platform. What does that mean? That car drives around, it picks people up, it drops them off, it uses all of our mapping and all of our pricing and all sorts of things. It also uses our fleet management to make sure it’s always cleaned and always charged by the time it gets back to you, and then shows up again when you need it again.

I think one way to answer the question is we want to make sure that individuals can continue to participate in this gig economy. It’s just that now they can do it in a different way. They don’t have to use their time; they can use their physical assets. The second thing is we are… I mentioned-

Wait, can I ask you one question about Lyft-ready?

Yeah, of course.

You’ve just described the same vision for Lyft robotaxis as Elon Musk has described for Tesla robotaxis. He has been less, I would say, clear that the cars would be cleaned when they return to you. He doesn’t seem interested. Have you talked to Tesla about saying, “Okay, you’ve got this big robotaxi idea. Do you want to put a Lyft app on it and just make this go?”

I don’t want to describe exactly the conversations that happen or don’t. I would say, in general, the vibe that Tesla gives off, I think, is representative of how their company goes, which is, we like to do things ourselves. So I’ll just say that as a generic thing. Fleet management… You’ve touched on three things all at once. I think there’s a question about what Tesla is going to do. Interesting question.

Well, they’re the only car company that’s selling cars to consumers today that can do what you’re describing. Even if you believe that they should or should not be doing it, no one else is selling you a car that can do what FSD can do.

One hundred percent today, exactly right, one hundred percent. So, my view there is that they will need some kind of fleet management. Unless you want to push all of this back to… Well, first of all, they’re going to need a whole bunch of customer service and a whole bunch of other things that people don’t think too much about because the scale is so small. But as you start to grow and people start to leave their umbrellas and iPhones in the thing, and you have on board and off board, there’s a whole bunch of infrastructure. Let’s put it this way: creating a rideshare business is not for the faint of heart. So the first thing that you might ask yourself is, is all of that a good idea for them to spend a lot of energy on, or should they partner with other organizations that are already doing that? Which you’re asking, and I’m declining to answer the specifics, but I think in general, it’s an interesting conversation that the companies should be having.

And then there’s this more subtle part, I think, which is, again, to your point, it’s all well and good to think that these cars just magically charge themselves, clean themselves, maintain themselves, but that’s not the reality. We have a whole subsidiary called Flexdrive that oversees 15,000 cars. We do it today, mostly for drivers who don’t want to use their primary car. And so we rent them a car, we buy these cars, we tell the drivers when they’re going to need service. We make sure that we’ve got sensors on the cars to bring… We know how long it’s going to take to service the car, all these different things, that very unsexy fleet management stuff. Gosh, it’s the difference between profit and loss on an asset, and we happen to be very good at that because we’ve done it for many years.

And so back to your point, I think part of the reason that other companies don’t talk as much about this as we do is that they don’t do it, but we do it, and so we know. And it is part of the reason why our Waymo partnership in Nashville is, I think, quite interesting. Waymo and us said, “Gosh, this is really important, and you guys do this and do it well. So how about you do it here in Nashville for us?” So I know we’ve covered five different things all at the same time there, but I think that’s part of the reason people are not focused on it as much, because it’s not a capability that many companies bring to the table, but we do. Yeah.

I’m very curious about the “My car’s going to make me money as I sleep.” There’s a lot there, but one company that can actually sell that product today. But the rest of it, I always think about it as like, man, this really implies there’s a lot of demand in my sleepy town, while I’m sleeping, for my car. Is that going to work? Because I actually need my car quite often during the day. Let me ask you about-

I think that’s also fair, I think that’s also fair. I think that for many people, that won’t necessarily be their primary car. I think there will be entrepreneurs who go out and buy five cars just like people buy small fleets today for black car service, and that’s probably more likely in the near term.

Let me ask you about Waymo real quick.

Sure.

And I want to zoom out again. You’re Waymo’s partner in Nashville. Waymo is partnered with Uber in other markets. Waymo is running its own service in some markets. It does feel… And they’re obviously backed by Alphabet; they’ve been backed by Alphabet for a long time, and they’re going to spend a lot of money to win. It just seems very obvious. If you ask Sundar about it, he’s like, “We’re just going to keep spending money because now it’s very clear that we’re close to winning.” 

They’re going to spend a lot of money until they win. They’re looking for partners, they’re looking to see what winning looks like. You’re in a weird competition. You’re in a weird bake-off. You’re the partner in one market, they’re the partner in another market.

Totally.

They’ve got their own service in yet another. How do you perceive that competition? Have they told you what winning looks like?

I think they’re figuring it out. I really do. And I take them at their word that they… Look, I think a word that they use pretty often is optionality. They have something pretty cool. They’ve got technology that works about as well as anyone’s in the world. Again, I think really only Baidu would be the real competition there. I think everybody else is to some degree off where they are. And it works and it works well, and people like it. So that’s a good place to be.

Thus, if you’re in that position and you think, “Well, gosh, I’m not really sure how this is all going to play out.” A totally reasonable strategy is, “Well, let’s try a little bit of everything. Let’s try doing it ourselves end-to-end. Let’s try partnering in a certain way with one company. Let’s try partnering in a different way with another company, and we’ll see.”

When I fast-forward, I think a very likely outcome is they will realize, gosh, as I just said, running a rideshare business is quite expensive, and it’s very physical. It involves… Again, think of the scale: 800 million rides we do every single year, 50 million riders, 1.5 million drivers. I know those aren’t part of the picture, but someone’s got to own those cars. There’s going to be someone else in this picture who owns these things and wants these things utilized.

If you’re Google, do you want customer service? Do you want all these different things that you have to do to operate that service? Maybe you don’t mind it in a couple of markets. Maybe it’s cool. You can have direct access to your customers, and you can do certain brand-building and maybe get some data from it, whatever, whatever. All good. But do you really want to do it in 280 cities around the United States or all around the world? I don’t know. 

I think a very likely outcome is that they will be friendlies. Or what’s that? Co-opetition or whatever. They’ll compete in some markets, and in others, they’ll partner, and our job is to be the best partner they can possibly have so that they, over time, give us more of their business and stay focused in their own way in some small number of markets. And that’s great.

Lyft used to have its own autonomous car division that was sold before you became the CEO, but you have other partnerships that bring you closer to actually making the hardware. You just announced one, a company called Tensor. There’s some weirdness with Tensor. It used to be a Chinese company called AutoX, and we were told that it all got wound down. Have you ridden in a Tensor car? Do they exist?

They do exist. I have not ridden in a Tensor car myself, but colleagues and people on my team have.

I’ll tell you about Tensor just for 30 seconds. It’s a very interesting company. So they are also trying to create, somewhat uniquely in the market, a car that is a self-driving car and a robotaxi. So, in other words, drive me or drive somebody else from the start. They partnered with VinFast, which is a Vietnamese company. They have a small number of cars here in the United States. But I think what’s most interesting to them about VinFast is that it’s a very new company, and therefore, their assembly lines are quite new. They have very modern technology, and so on and so forth. They have outfitted their cars with a crazy number of sensors. Everything is redundant.

I mean, you talk to… The guy’s name is Professor X, who runs the company. You talk to him, and he’ll tell you how important it is to have redundancy around braking, redundancy around steering. He wants this thing to be absolutely bulletproof, and the steering wheel literally moves out of place if you don’t want it there. I mean, it’s a very, very interesting product that they’re designing. It’s very expensive. It’s $300,000, something like this. But, of course, it’ll come down. Still, it’s really meant to be a very bespoke thing. And as I said, self-driving from the start. He’ll tell you about lidar and how there are different versions of lidar; they’ve got the best, all this stuff.

Okay, why did I go into detail about this? I think over time — and you can see this through history — people start by retrofitting something existing as a way to get started, and then eventually they realize, “Gosh, this is a new thing, and so therefore, let’s create our own purpose-built thing.”And I think they just decided, in particular, “We’re going to jump over the bespoke or the take somebody else’s and try to retrofit, and we’re just going to go. It’s small-scale, it’s super expensive. We’re going to ride that cost curve down.”

Other people, obviously, Waymo’s done exactly the opposite. They use the Jaguar I-Pace, and they’re going to move to the Zeekr. Zoox is more… So anyway, long answer to a particular question, but I think Tensor is very interesting because it’s our first proof point for the Lyft-ready concept, even though it’s super small, super experimental, but really interesting. And I’d love the fact that they’re trying to innovate.

When you look at that bet, you’re describing the big disruptive bet: We’re going to take the new technology, and we’re going to start with that as the foundation of the product, and we’re not going to worry about all the stuff that happened before. Sure, that pattern repeats.

Right in the middle of that, just to bring it back around, is the driver. The idea that Lyft-ready will put a bunch of rideshare cars in people’s houses, or that it’s worthwhile to make a technology bet on self-driving cars in that specific way that Tensor might be making. All of that puts pressure on the driver. The logical end state of that is that one day, there will not be drivers on this platform at all. How long do you think that will be?

Many, many, many years, many, many years. Beyond the work span of 99 percent of drivers on our platform. Part of it is just basic laws of physics. Like at the Taylor Swift concert or at the end of the NFL game, or even at five o’clock every afternoon or nine o’clock in the morning every day, there just aren’t enough self-driving cars. And there’s certainly not enough self-driving cars being driven… Excuse me, there are certainly not enough riders who only want self-driving cars.

 A lot of them just want to get where they’re going fast and cheap, and therefore, they don’t really care whether it’s a robot or not, but they really won’t wait for 15 minutes when they could wait for two minutes for a human-driven car. And then there are people who want help with their luggage, and then there are people who want to have the conversation, and then there are people who just don’t like technology.

So there are all sorts of reasons. And then there are regulatory things, and then there are snowstorms, and then there are ice pellets, and all kinds of things. There are a billion reasons why, in the near term, the hybrid network is the better approach, all the things.

In the medium and long term, there will be fewer drivers as a percent, but remember 160… Let’s just remind ourselves of this crazy fact. Today, as I said, we do 800 million rides, maybe the other guy does 1.5 billion rides. So maybe 2.5 billion rides between the two of us every year in rideshare in the United States, I’m talking about. Okay, what’s the total number of rides that people get in the car and drive themselves? 160 billion per year. So 2.5 billion, 160 billion. So there is a lot of room between 2.5 billion and 160 billion for us to continue to grow and expand and have more drivers on that platform.

And by the way, all the estimates around the number of self-driving cars by, say, 2030, are 30,000. That 30,000 is a tiny thing. We have 1.5 million drivers on 30,000. Now they’re working 24/7, high efficiency, high utilization. So it’s not apples to apples, but still. That’s, again, sorry for all the verbiage there, but it’s a way of saying I think the hybrid network dominates for a long… Hybrid, meaning some driven by humans, some driven by autos, actually by robots, will dominate for a long, long, long time. And I think by the time we get to a point where there are relatively smaller numbers of drivers than there are today, gosh, I think we’re talking about… You might as well think of it as a generation. I don’t mean 25 years, but I mean most people who drive on the platform don’t do it for more than three, four, five years type of thing. At that point, it’ll be a whole different world.

And then the last thing I’ll say is that I mentioned it as sort of a joke at the beginning, but not really, the car tender idea. I think there will be fun things that people are going to be doing in the cars that are not just driving; it’s making drinks, it’s telling stories, it’s being the local guy, it’s, again, helping you with the luggage, it’s doing all kinds of other stuff that drivers do today on the side. And now, I don’t know, just being a… I don’t know. Who knows, who knows?

Are you going to expose that sort of driver individuality?

I hope so.

I mean, you could do it from the top down. You could say, “You’re all bartenders now. We’re doing Old Fashioneds in the car.” And maybe the platform will support that, and maybe the ecosystem will support that. But the other thing I’ve heard from drivers… Sometimes you get into a Lyft and the driver has set up 15 charging cords and will let you play with the music. And sometimes it’s just a guy in a Camry, and there’s no rate differential there, and the extra effort is not rewarded. 

Essentially, it’s a commodity. The point of the platform is to commodify the service. Do you think you’re going to let the drivers decommodify in that way?

I do, I really do. And I think, look… We just announced two days ago, I guess, the acquisition of a company called TBR, which is a very, very high-end, ultra-luxury, chauffeur-driven service. It’s for non-deal road shows and Super Bowl events and stuff like that. It’s a very, very bespoke thing, but one of the reasons we did it is because the level of service that they provide is unbelievable. And gosh, can we learn a lot about it? 

There, the drivers do all kinds of interesting things. I mean, they’ll get you your coffee before you get in the car because they know that you like a latte and not a flat white. And then when you spill the coffee on your shirt, they’ll pick the new shirt up while you’re in the meeting so that it’s ready for you when you come out. And they’ve already called the next person, saying, “We’re going to be two minutes late because the guy’s got to do something before he gets there.”

So I’m not saying you can provide that level of service to every single person 24/7, but I think the idea that… I mentioned briefly before this notion of enshittification, a service that starts out amazing, and then gets a little bit less so over time. I think rideshare for so long has been caught in this kind of binary of us versus the other guy, and very competitive-focused, and everything’s a commodity, and everything’s about cost, and so forth. And I think over time, you get to the point where it’s like, that’s no longer interesting. 

What’s much more interesting is the next 160 million rides that people are taking in rideshare that they’re not taking today, and how do you let that driver show up in a way where they become different from a robot, not just a robot with… a human robot, a flesh robot. So yeah, I think there’s a lot. And then the question is how do you do it at scale, and how do you do it in a high-quality way, and how do you do it in an economic way? But those are what… Aside from self-driving and all the rest, the human side is what makes this industry so interesting.

I do have to offer a plug here: Cory Doctorow, who wrote a book called Enshittification and coined the term, will be on Decoder very soon, and you can hear him talk about it.

Oh, fantastic, I’m a fan. I sent him a little note after I wrote this. So I wrote about enshittification last year, and I got introduced to Cory. He actually came to the office. He, at the time, said all kinds of nice things. So anyway, he’s a good guy. I hope you have a great conversation with him.

It’s Sarah Jeong on my team who did it, but she’s smarter than me, so they had a great conversation. Both of them are, so that’s just the way it goes.

Let me ask you just one question about the curve here, and then we can wrap it up. You’re saying in the aggregate that attrition will solve this problem. Drivers will graduate off the platform, they’ll find other work, or they’ll retire, or do whatever they do, and they’ll go. And as that happens, robots will come online, and you’ll find some sort of happy medium. That’s broadly the plan. We’re not going to replace a lot of drivers. Drivers are going to graduate. Robots might replace them. Some other drivers might replace those. Who knows? That’s the long horizon. And I can see how that might work out.

In the short term, what I hear very directly is, “Oh, the robot cars have come to my town, now the demand is moving away from this platform. I’m waiting around for even more rides than I was before.” Or the dynamics have changed, or the rates have changed, and it’s very individualized, and it’s hard to make an argument that says, “Well, look, one day you’re going to retire.” But you don’t get to make the long-term argument to the driver themselves when they’re faced with the threat of autonomy. How do you make the argument to the individual driver that this will be good for them?

So two things. I mean, I think first — and we look at this data a lot — I would say that the stories you just told are not really supported by the data. In other words, for example, we’re actually growing. Lyft is growing faster in markets where there are AVs, even when we’re not participating in the AV, faster than average. So our average growth rate is 15 percent. We’re growing faster than that in places like San Francisco and Phoenix and places like that.

I think there’s a false causality where people see a self-driving car and they’re like, “Oh, I didn’t get a ride today, and that must be that,” but that’s actually not really the case. In general, self-driving cars actually expand the market. They actually oxygenate the market in new ways, bring tourism, and all kinds of things. So point two, however, but that’s just data, which isn’t necessarily going to convince anyone who’s feeling this.

Okay, so let’s talk about the feeling side. We are very active in this. We have a whole driver roundtable thing, and the whole topic of that… Where we bring drivers together — we just did it again last month, we’re doing it in different parts of the country. The topic is let’s talk about what happens when AVs come to town and how you’re going to be okay. How are we going to collectively make something that is okay for you?

I mentioned this part very briefly before, but I’ll come back to it. One of the things I am super excited about is this driver accomplishment letter that we now let drivers create for themselves. So once you’ve driven a certain number of rides on the platform, you can go in, and thanks to AI, push a button that says David started driving in… This is actually true. I drive for Lyft, and I think my first drive was on April 13th or 14th of 2023. So anyway, he’s been driving since April 2023. He’s one of our top — now this part, I’m making up — five percent drivers. He’s super reliable. Here are a couple of representative comments. I would recommend David for any service-oriented business he might have.

And that credentialing… Because our drivers are in the service business. So I think one answer to your question is not just let’s wait for attrition, but let’s help drivers who want to use this as a sort of mobility thing.

By the way, remember that thing I just talked about with TBR, the company we just acquired? Those drivers make quite a lot of money, quite a lot of money. And so to the extent that you want to drive up in the ecosystem, black and higher level, that’s another possibility too. And that ain’t going away anytime ever. No robot takes the place of the chauffeur who gets you coffee.

So I think it might sound a little vague, but all I want to say is that this is not something we’re just saying, ah, too bad for them. We’re actually saying the opposite, which is, “How can we help this transition happen in a way that feels orderly?” Yes, how can we share the data of what’s happening, of course, but also how can we help drivers to transition?

And by the way, maybe we can hire them in customer service. This is something we talk about all the time, all of our driver customer service folks, how cool would it be if every one of them were ex-drivers, as an example? I think there are some answers to this question.

Well, David, this has been a great conversation. I’m excited to have you back as more and more of these things develop. What’s next for Lyft? What should people be looking out for?

So if you look to, maybe, six months ago and then six months from now… Six months ago, we were only a domestic company; now we’re an international company. We acquired a company called Freenow in Europe, which is awesome. Now, when you go to Europe and you open your Lyft app, you’ll get a note saying, ”Please download the Freenow app. 

Those are our partners. Over time, that’ll just be the Lyft app. That’ll be wonderful. We’ll be able to operate in Europe natively. If you’re a United customer, you’re soon, very soon going to be able to open up the Lyft app and get points from United Airlines. So put that in your cap.

If you’re a CEO, soon you’ll be able to use TBR as part of the overall Lyft ecosystem. So a couple of very obvious ways where we’re, you might say, expanding out and expanding up. We’re going up in terms of certain demographics and service level, and out in terms of geography. In more and more markets, you’re going to be able to take a self-driving car on the Lyft platform. Today, you often do an off-platform or on somebody else’s platform. So that’s super exciting. All those things are happening in the near term.

I think if you look at the medium and long term, I’ll come back to the beginning, our purpose is to serve and connect. And this is almost more of a philosophical thing than a company thing. Every minute you’re spending watching Netflix at home, every minute you’re spending getting food delivery, it’s awesome, those are great experiences, but gosh, I really do want to be a helpful force in your life to also make sure you’re connecting with people and getting out and about and experiencing the real world. And so as you look over two, three, five years, you should expect Lyft to become, I hope, more and more helpful in that part of your life as well.

Terrific. Well, I’m excited to get a ride in a Tensor car very soon. It looks crazy. I just want to see one in person.

It’s insane.

We’ll have to have you back. Thank you so much, David.

Yeah, it’s been a huge, huge pleasure. Thanks for the time and the questions.

_{Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!}