DoorDash 的 Tony Xu：熬过1,000天创业地狱

I want to start with the fact that you said that PaloAltoDelivery.com, which was DoorDash before DoorDash, was the most minimal version of a minimum viable product. Can you explain how you built it?

Whenever you can ship something in 43 minutes to test your idea, I think that's pretty good. This was 12 or 13 years before the rise of LLMs and AI tools made it so easy to do that. Basically, the 4 of us wanted to test the idea that if you wanted to offer delivery from places that had never offered delivery before, what was the fastest way to see whether consumers would care?

At the end of the day, delivery is not a new idea. We thought one of the reasons why delivery hadn't been around in 2013 was simply because nobody wanted it. So we shipped PaloAltoDelivery.com. That alias was available for $9, so that's why we got it. It wasn't a very scalable URL, but we were able to get it.

It was a static page where you saw 8 PDF menus from restaurants that we frequented in Palo Alto. The only way you could order was to read through the menus and call a Google Voice number that would ring the cell phones of the 4 founders. One of us would pick up, take your order, place the order on your behalf, go and get it, and deliver it to you.

I used to be an intern at Square, and I had these card readers, which were one of their earliest products—these wide dongles that you could stick into the audio jacks of iPhones. That's how we collected payment.

Something I didn't remember until now, because it feels like DoorDash, Uber Eats, and everything else have been around forever, was the state of delivery. There were other delivery companies, but you essentially created the market for this. Can you explain that?

When I was telling people, “I'm coming—I'm really excited. I'm going to go speak to Tony from DoorDash,” they were like, “I can't believe he survived in this competitive market.” They just assumed there were already apps out there delivering for restaurants that didn't have a delivery fleet. That didn't exist then.

No, actually, yeah. I think one of the biggest misconceptions when we were founded was just how wide open the space was. There were about 1 million restaurants in the States, and maybe 20,000 to 25,000 of them offered delivery. Most of them were pizza shops, places in New York City, some in Chicago, and some in big city centers.

Outside of pizza places and maybe a few Chinese restaurants, nobody offered delivery. The real grand question, or experiment, of DoorDash and PaloAltoDelivery.com was, “What about everyone else? What if you can enable everyone to offer delivery? What would that take?” First of all, would people care? That's really why we shipped something so quickly—to see if people would actually come and place orders.

What were the existing companies doing then?

They were mostly faxing orders, believe it or not. They would have a website that received orders, and then they would fax the orders into machines that sat near the kitchen or the point-of-sale systems inside these restaurants. The restaurants would then go out and do the deliveries themselves. So they were lead-generation companies at the time.

I've heard you talk about developing this last-mile logistics network. Did you think about that back then, or were you just thinking, “I'm going to try to expand the market for food delivery”?

We did. Before we shipped PaloAltoDelivery.com, or even before we figured out how to get there, my co-founders and I really got connected because of an interest in small businesses. I think my story I've told publicly is that I grew up coming to the States as an immigrant from China. My mom put food on the table by working 3 jobs a day for 12 years.

One of those jobs happened to be at a Chinese restaurant where she was a waitress. I got to hang out with her and wash a few dishes when she allowed me to. That's how I grew up while my dad was getting his PhD at the University of Illinois. That was the first 10 years or so of my childhood in the States.

That experience always gave me a deep appreciation for what small-business owners represent. To them, there's no such thing as work-life; it's all the same thing. There's no concept of a weekend or a Saturday. Saturdays and Tuesdays are exactly the same days. You get into this process where that becomes your identity.

It's one of the most fascinating things I find about the great experiment that is America. Because it becomes this all-consuming thing, one of the nice positive derivatives is that small-business owners don't just create great experiences, like a restaurant, bar, furniture store, or T-shirt shop. They actually create the GDP for all the cities that we live in.

That GDP is what allows us to have great neighborhoods, schools, and all the positive things that happen in a local community. That was always my fascination with it. We had no idea, though, when we were looking at starting DoorDash what the problems of these business owners were.

My co-founders and I spoke with maybe 300 businesses up and down the Bay Area, from San Jose to San Francisco—restaurants, retailers, and service businesses. It was actually a baker who showed us a 3-ring binder of delivery orders she had turned down. She was a one-person shop with no ability—or, frankly, no desire—to fulfill all those orders.

That was a very strange moment for us. I said, “Delivery is not a new idea. It's 2013. No one offers delivery. Why?” That's what prompted us to launch PaloAltoDelivery.com to see if people cared.

To your question about logistics networks, we said, “If the first place in which we can help local businesses is by building a logistics network, we have to pick a place to start.” This is where the math brain comes in for me. When we studied every category of local retail to decide where we would start—whether it was deliveries for restaurants, grocery stores, convenience stores, or retail shops—

Those are all options?

We looked at all of them. We had this hypothesis that if you wanted a chance of creating a logistics network that could actually be successful, very fast, and very flexible—meaning it could deliver in 30 minutes or longer than that—you needed network density. You needed the greatest number of connections between consumers and stores.

We targeted restaurants because there were 1 million of them. If you compared that with the number of grocery stores, there were maybe a couple hundred thousand grocery stores. When you looked at other categories of retail, restaurants had the highest count of stores.

Very quickly, we made the assumption that if there were any vertical to get started in doing deliveries, it would be restaurants and prepared meals. That would give us a chance to build the highest-density network so that one day we could deliver everything else.

I want to tell you about the presenting sponsor of this podcast, Ramp. I have been reading a lot about SpaceX lately. SpaceX is one of the most valuable private businesses in the world, and there's an idea from their history that more companies should use. From the very beginning, SpaceX was constantly attacking and questioning their costs. RAMP helps many of the most innovative businesses in the world do exactly that. The median company running on RAMP cuts their expenses by 5%. And the important idea that I found by reading about SpaceX is that a religious dedication to controlling costs helps increase revenue because you can pursue opportunities you couldn't otherwise. We see that in the RAMP data, too. The median company running on RAMP also grows their revenue by 16%. So, when you're running your business on RAMP and your competitors are not, you have a massive competitive advantage that compounds over time. Ramp is the only platform designed to make your finance team faster and happier. Many of the top founders and CEOs that I know run their business on ramp. I run my business on ramp and you should too. Go to ramp.com today to learn how they can help your business save time, save money, and grow revenue. That is ramp.com.

There were other people who had maybe a similar idea, but I heard you tell the story once where you said they actually went into city centers. One advantage—you started in Palo Alto instead of New York City. I don't even think that might have been accidental. Can you talk about why that was important?

Starting in Palo Alto was not a conscious choice. It was just where we were students at the time. One of the earliest experiments we ran at DoorDash was doing deliveries in Palo Alto versus doing deliveries in San Francisco, a city center, if you will, that was close to where we started the company.

One of the fascinating things we found, and didn't initially understand why, was that we were completing deliveries faster inside Palo Alto than we were inside San Francisco. Obviously, San Francisco is a denser place. But one of the things we learned early on was that, in Palo Alto, you had much easier parking.

You also had a lot fewer apartment complexes where you had to go up and down the stairs and figure out where the lobby was or find the right elevator entrance and things like that.

Palo Alto had the following: if you looked at places like Palo Alto, it represents, I think, most cities in the United States or a lot of the world. You have main streets, and then, in the spokes outside of this main-street hub of commerce, you have the places where people live. If you thought intelligently about what that really told you, you could build a very efficient logistics system if you understood how to manipulate some of these hubs and spokes.

This was one of the earliest hypotheses we had: that you could make a logistics business as efficient in a place like Palo Alto as in San Francisco, and that was guided by that experiment. But the second thing was just talking to customers. What customers told us was, “Look, in San Francisco, I can just take the elevator down and head out of the lobby, and we could probably find a few places to go and eat.” In Palo Alto, you’d be walking for miles before you could achieve something like that.

The closest set of restaurants near Stanford University, where we started, was 2 miles away on University Avenue, as an example. That’s true in a lot of places in America. If there was any place where we thought there would be the highest interest from consumers, and a possibility where you could actually make the math work, it was places like Palo Alto. The question to us was just: How many of them are there?

And the only people doing deliveries at this time are the 4 founders.

Yeah. In the very beginning, it was just the 4 founders.

Okay. So you had a line about this where you said, “It became obvious that the need was higher outside of the cities. We did not have the data to prove it at the time. We had the conviction that because we were doing the deliveries ourselves, this could be true.”

Yeah. One of the benefits when you do the deliveries is that you see how hard it is to actually bring someone a burrito on time, every time, and correctly. The second thing is that you get to see who the customer is. We saw that the customer was almost always a mom who had young children, didn’t have a lot of time, didn’t want to cook every single meal, and was looking for any solution to save time.

When we did those deliveries, we thought, “Wow, there are a lot of young families out there. Let’s go find out where they hang out. Let’s go find out where they live.” That’s why we had the sense that we could build a business with this audience to start.

Is that another unexpected benefit of starting in these suburbs, or basically outside the cities? Think about typical city populations: maybe more single people or just a couple, but not large families packed into these buildings.

Yeah. I think that was probably a derivative of the discovery. In the beginning, especially when you’re looking for product-market fit as an entrepreneur, you’re looking for someone who actually just wants your product organically.

We could tell very quickly that there were a lot of people with young children who maybe didn’t want to take a stroller, pack it up, pack all the things that come with the stroller, then put the stroller and the children into the vehicle, get them out, and somehow get inside a crowded parking lot or restaurant. If we could solve it for that group, then we believed we could build a business that could easily grow organically.

You’re right: there’s a second derivative, which is that there are more mouths to feed when you have a family than when you have 1 or 2 people living inside a city. But that wasn’t the first thought we had.

But even more than a second derivative, you were just explaining that if you’re delivering to somebody’s house, you know where to park. Whereas if you’re in a city, you have to navigate: Where’s the lobby? How do I get into this building? What floor do I go to? How do I access the elevator?

Yeah, totally. The presence of single-family homes made it a lot easier, for sure. That was one of the benefits of delivering to places like Palo Alto.

But again, I think it came from this very simple experiment, which had an anomalous finding: Why is it faster to deliver in Palo Alto than it is in San Francisco? Why is it faster to deliver in a less dense place, in other words?

Exactly. This is what’s interesting to me. It almost seems like your competitors did the most obvious, logical thing: “I need order density. Where are all the people? Let me just go to the cities.”

We chased where the demand was. I think when you’re starting out, the number one thing every entrepreneur is looking for is: Do you have something that someone else wants, and is it real? Meaning, it’s not artificially inflated with discounts, marketing dollars, and other ways to grow inorganically. Will people actually use it? Will they actually tell their friends about it if they like the service?

That’s what we found early on with places like Palo Alto.

Even when you were called Palo Alto Delivery. Money, right?

Yeah, we had no money. Exactly. We ran this out of my bank account, where I also had student debt at the time. Even though I was running it out of my bank account, it wasn’t going down every single week or every single month, so something was telling me that maybe this had a chance of working.

What were your costs at the time? You had the 4 founders as essentially labor. You probably weren’t paying yourself—exactly, you’re not paying yourself. Free labor, right? Just your time. You built a $9 website.

I heard something that was hilarious. We were like, “Well, we don’t have a sophisticated dispatch system, so we just use the Find My Friends app.” We used Find My Friends to track the drivers, which just happened to be all of us, our co-founders.

You have a Google Voice number. There’s no marketing or advertising, right?

No. We had no money to market or advertise.

So what other expenses did you have back then? Do you remember?

It was all kind of self-funded. This entire activity was self-funded until we had to start recruiting drivers and actually testing this out beyond just the 4 of us.

This is when you applied to Y Combinator?

Yeah, in that time period.

By the time you applied to Y Combinator, did you have more drivers than just the founders, or no?

We may have had 1 or 2.

Okay.

Very quickly, we realized that we were in class. We took turns doing deliveries while we were in class, but at some point it’s tough to be a student and do the deliveries.

How many years did you have left of business school? How many years were you in school while you were running this?

We had maybe 6 months left before graduation. We were effectively Stanford’s delivery service for the first half of 2013. Then we got DoorDash—the URL and the company name—and launched out of Y Combinator in June 2013.

Once somebody starts using DoorDash, or when I start using DoorDash, I’m like, “Oh, this is very convenient.” I just keep using it over and over again. Did you see that same behavior pattern back then?

Yeah, with a very small group of users, because in the beginning we actually did not have high volume. It was probably 10 orders a day, something like that. Our high day was maybe 21 orders a day.

Most of them, however, were done by a small group of users at Stanford. When you see that the same customers are ordering again and again, even though it wasn’t growing like wildfire, and our bank account also wasn’t getting depleted, it gave us enough conviction to keep going.

What were the conversations among the founders when you were saying, “Let’s keep going”?

I think we viewed it as a project more than we viewed it as a company. In fact, we were barely incorporated. We were not incorporated when we were running this at Stanford University, and then we got incorporated when we got into YC.

At the time, it was just, “Let’s see what the next phase should be.” I think sometimes when you start these projects, you absolutely should have a point of view on where this can go in terms of going the distance. But the most important thing is to just get started and then have a sense of what the next 2 or 3 steps are.

No one is able to know everything about the future. For us, the summer was really instructive. Doing the deliveries ourselves for the first 6 months gave us the clarity that the summer was really about answering 3 questions: Would consumers pay us $6, which is what we charged? Were there restaurants who would be willing to partner with us for 15%? And could we afford a wage that we could pay Dashers—the drivers—for the service?

That was it.

That was the entirety of the YC summer. It was not about Demo Day, raising the most amount of money, or becoming the most popular at some event. It was just about answering those 3 questions. If we had enough conviction in the answers to those questions, then we’d keep going.

Again, you told this hilarious story where, during the summer, some of your classmates were like, “Yeah, I’m going to go skiing or something like that. What are you doing, Tony?”

“I’m delivering hummus in my Honda.”

Yes. That was it. Look, I think we had a lot of classmates at Stanford who looked at us and just thought, “I thought they were smart, but I guess they want to spend their time doing this.”

In the beginning of a lot of these entrepreneurial ventures, nothing looks that amazing. We were working out of an apartment. We had Dashers in that apartment, and we had the cofounders living in that apartment. We worked from 10:00 a.m. to 2:00 a.m. every single day, but it wasn’t this glamorous exercise.

Nor did we seek that. We were just trying to answer those 3 questions that summer. We didn’t care that much about what our friends were doing. Clearly, we thought it was interesting enough to keep going, and that if we could answer these questions, we were onto something.

We just had Marc Andreessen on the show, and he’s got this great line where he says, “I firmly believe that people that do great things are doing them for the first time.”

Huh. Did anybody have any restaurant—or actually, not even restaurant experience, because you weren’t even in the restaurant—any delivery experience? Did any of the founders have anything to do with logistics or delivery?

No. It’s actually why we had to do the deliveries. The reason we were so hell-bent on doing the deliveries, besides the fact that we had no idea whether we had any business recruiting other drivers, was: How does this work? How should it work?

I think DoorDash, early on—even to this day, but especially early on—was so hard to explain. Even to build the MVP to test it, it was just this website, PaloAltoDelivery.com. But we had to build 4 things. We had to build a website for consumers, an app for the restaurants to receive the orders, an app for the drivers—the Dashers—and then a dispatch system that could oversee all of the operations.

Even in the very beginning, we realized, “Wow, this is actually pretty interesting.” It’s such a fun problem that, in order to bring you a burrito, you have to build these 4 things. To do it really, really well, that’s why we did all the deliveries: to figure out how you actually do that.

So, you were misunderstood back then. You just said something interesting. Do you think that’s still the case to this day?

Absolutely. Most people—and I totally get it—think of DoorDash as a consumer app. Most people think of us as lunch and dinner, and I think what they don’t see is everything behind the scenes.

A lot of times, you can look at products like ours, especially as a consumer, and say, “Wow, this looks like any other product. There are so many of them.” But then I would ask the question: How come one just gets used more often than the next or the others? It comes down to everything that you can’t see.

One of the things we say a lot internally at DoorDash is, “It’s always the data that you can’t see that kills you.” If you can see a truck coming at you, you’re just going to dodge and get out of the way. But if you can’t see it, you’re dead. It’s no different with our business. Our business is one where all of the magic, or the secret sauce, if you will, is in things that you cannot see.

No consumer is sitting there while they’re ordering DoorDash thinking about what the Dasher experience should look like, what the operations should be to get the best-quality experience at the most affordable price, or what the ways are in which you take out every single friction and cost with a restaurant or a retailer and make sure that all the items are actually there, even when they’re not there.

I think all of these things are what make DoorDash special and make DoorDash an end-to-end experience that’s very difficult to replicate. We knew that early on because we did all the deliveries.

You know who knows it? Your competitors. You’re not going to like this because, in my opinion, you’re really humble—probably too humble for my liking—but people in your industry are afraid of you.

I have to tell you a personal story that I don’t even think you know. I had heard about you before. I didn’t really use DoorDash, but I never thought about it. Exactly what you just described is exactly my experience. I was just like, “I have a magic button that brings me a burrito.”

Exactly.

Okay, I love that magic button. Don’t take that magic button away from me, whatever you do.

I was in Stockholm about a year and a half ago, and Daniel was very kind to host me and a handful of European founders. One of the European founders sitting next to me and Daniel at dinner was somebody I had never met before. It was Miki from Wolt.

Okay.

Right. Cool. But he told me something interesting. Basically, the story was, “Listen, I built the DoorDash of Europe,” I guess is how it was described.

He said, “I always thought of myself as an entrepreneur. I never thought I would work for anybody.” He was like, “We were in a head-to-head battle, and I had a term sheet in front of me. If I remember the number correctly, it was for about a billion dollars. I had the ability to raise another fresh billion dollars of capital.”

He was looking at the term sheet and thinking about signing it. Then, involuntarily, something came out of his mouth. He said, “I can’t beat him.” He was like, “I can’t beat him. I cannot believe that came out of my mouth.”

Then he looked down and said, “I could either light this money on fire, or I could sell my company for life-changing money, go work for Tony, and learn a lot.” I think to this day, he still directly reports to you. Correct?

Yeah. He runs all of our European business.

He was trying to explain to me and Daniel how hard it is to compete against you. It’s very similar to what we were talking about: all the magic is in the stuff that you don’t see.

You had another interesting quote I want to read to you. You said, “The way that DoorDash has achieved so much success is tens of thousands of experiments, 95% of which never even make it to the customer before they fail. The way to get more accurate on a delivery probably requires some level of detail that is lower and deeper than you realize.”

Can you explain what you meant by that statement?

This again starts from actually doing the work ourselves and realizing that if you actually want to get something on time, it’s very easy to think about it when you’re just intellectualizing it from the outside, especially when you’re getting started. “Maybe there’s a traffic issue, or maybe the food is taking longer than it should.” Whatever the reasons might be, you have no idea what all the sources of delay in an order are until you actually go and do the work.

Sure, there might be some of the issues that you can think about from the outside, but very quickly you realize that there are a lot of seconds of delay in every motion. In fact, there are about 20 steps you can decompose a delivery into, and there are delays at each one of those moments.

That’s even more complicated if the deliveries happen outside of restaurants, inside shopping contexts like groceries or retail items, or inside malls that are multistory—sometimes below ground, sometimes above ground. One of the things you start realizing is, “Wow, there are actually a lot of causes for delays.” There’s no way that you’re going to know about all of them until you literally encounter them for the first time.

A lot of what’s difficult about DoorDash is that we’re trying to build a structured dataset in a world that is chaos—the physical world. One of the reasons why there are all these sources of mistakes, delays, and costs that ultimately result in good or bad experiences for customers is because there is no data that exists. There is no nice dataset that a company like Google or somebody else has organized for you, because it’s all physical information and it’s changing all the time.

When you go into a grocery store and somebody moves an apple from aisle 6 to aisle 8, is that always going to get documented? Of course not. Those are the kinds of things we have to work on every single day.

You wouldn’t know that. What if I told you the cause for a delay was that somebody was homesick that day? How would you know that until that event actually transpired? And what would you do to respond to that event if it were to occur? That happens every single day when we’re doing millions of orders every single day.

The 1-in-1,000,000 event happens a lot, and the 1-in-1,000 event happens way more than that. Building a system that can ideally detect and prevent these issues requires a very fast-twitch muscle to build almost like an emergency-response system when something actually goes awry, so you can fix it.

That requires doing the work over and again and building a system that can learn over time to get better and better and better. Most of the time, we have no idea. We start with these experiments, and that's why most experiments fail. But when you get enough goodness out of it, if you can get the 5% out of tens of thousands of experiments to work in 1 year, that has a benefit for all of your audience for the next year. And then you just keep going, and that adds compounding surplus for all of the audiences.

Deal will help your business hire, pay and manage any worker anywhere in the world. Deal is the best company in the world at building infrastructure for global hiring. Deal is one platform for payroll, HR, benefits, and device management across 150 countries. Deal gets you everything you need to run a high-performing global workforce on a single AI native platform. From first offer to final offboarding, Deal handles the complexity so you can stay focused on your business. The best founders and operators in the world have one thing in common. They control as much of their business as possible. And the founders of Deal do exactly this. When you use Deal, you aren't using a third-party payroll processor or a messy network of in-country providers. Deal built and owns the rails. That means faster speed, better service, and total accountability. The founder of 11 Labs, who I use to make transcripts for this podcast, has a great description of the value that Deal can give your company. He said, "We built 11 Labs to break down language and communication barriers. With Deal enabling us to hire and support exceptional talent anywhere, we can accelerate our innovation and bring more voices, stories, and ideas to every corner of the world." Deal is trusted by over 40,000 customers and growing fast. Learn how they can help your business by going to deal.com/enra. That is deal.com/enra.

How do you do that many experiments? Is this on a yearly basis? Is this over the history of DoorDash? Are you running thousands of experiments every year, ideally?

Yes. I think when we are at our best, that's what's happening. But it starts with actually building a system that wants to learn. If you think about why we have to learn, it's because the physical world is not structured. It's not documented anywhere. You can't scrape it. It's constantly changing.

There's a winter storm right now, for example, in the Northeast. These are all things that happen differently.

It's beautiful here in California.

Yeah, I know. We would have no idea here. We're spoiled here in the Bay Area. But in general, all these things are happening every hour of the day. There's going to be some missing item today. There's going to be some order that took a lot longer today. There's going to be some incorrect gate we enter at an apartment complex. There's going to be some Dasher who's going to get lost coming up the stairs of an office building. That will happen, guaranteed.

The question is, it would be impossible to try to figure out all of that if you can't build a system to learn how to do this. The most important thing is building systems. Building a system at DoorDash really starts with testing things in a very operational, hacky, do-things-that-don't-scale kind of way, and then taking the things that ultimately work—the ideas—and actually building products around them. Then you engineer the ones that actually work so that you're very efficient with this learning loop and can go from learning to shipping something that actually works.

It's a resource constraint with how many engineers we have and how many things we can actually ship, especially when the stakes are high. You want to make that loop as tight and as fast as possible. That's how you build a system in which you can learn thousands of things. You just have to keep doing it over and over.

Our business is one where we believe we have to earn the right to serve you the next day. Even though you ordered with us today, thank you very much for your business—we have to earn it again. The scoreboard goes back down to 0 tomorrow, and we have to do that all over again.

Where did you learn the importance of that?

Of what?

Of starting over again every day. I've heard you say that before, and I love that idea.

Very early at DoorDash, we learned how hard it is to keep someone's trust and how easy it is to lose it. I think I may have said this before, but there was a Stanford football game during which we lost a lot of trust. We were late on every single delivery because we didn't have enough drivers on the road. We had no ability to shut down the website.

Pause there. Where in DoorDash history was this game?

This was the third month of our operations, September 2013. It was a Saturday. We had no ability to fulfill the orders that came in, and we had no ability to even shut down the website, so we couldn't stop the floodgates.

Why were you having floodgates 3 months in?

We had floodgates on that day specifically because of when the game ended. People wanted to order DoorDash for dinner in Palo Alto. For whatever reason, that volume spiked pretty hard. We had no ability to turn it off and no ability to fulfill the orders, so we were late by at least 1 hour on every single delivery.

I think when you go through experiences like that—not just once, but we've had a lot of those kinds of experiences at DoorDash—you realize that you can lose someone's trust on 1 order, and you have to earn it again the next day. There is no such thing as a set-it-and-forget-it mentality. That came a lot from the early days, but I think this daily reminder when I do customer support is also another great reinforcing function.

So what happened that night of the game?

We were late on every single delivery. I think it was probably somewhere around 10:00 p.m. when we were tallying up all the refunds it would cost us if we wanted to make things right and give everybody their money back.

Were the customers asking for the refunds, or was no one asking?

No one was asking for anything. The night was over. We finished our last delivery and said, “Okay, that was a terrible night. What are we going to do about it?” We could complain about the orders or something, but at the end of the day, within a very short period of time—15 seconds—we decided, “Okay, we have to make things right by the customer. We have to refund everybody.”

The complication was that we had no money at the time. I was having a hard time raising capital—I mean, this is a pattern for me. I've had a hard time raising capital for the company in the earliest years, and that started right from the beginning. We had maybe 2 or 3 weeks of cash, and this refund would have cost us about 40% of the bank account. It would have taken those 2 or 3 weeks and shrunk them into even fewer days.

But you're right: nobody asked us for the refunds. I'm sure they were pissed, but nobody asked. We did the refunds right away, and then we stayed up that night baking cookies. We delivered those cookies at around 5:00 a.m., before we thought customers would wake up.

The idea was, we'd rather die trying to be excellent—or at least die trying to do the thing that we want to stand for—than live to be mediocre and not be something we'd be proud of. That's what we did.

That's excellent. Tell me more about building the system, this self-reinforcing learning system.

These things happen in steps. It started with the 4 of us doing the deliveries. We could keep doing the deliveries, but at some point we were going to start running into scale issues. 4 people can only do so many deliveries, so of course we were going to start recruiting Dashers, recruiting consumers, and selling to restaurants.

As you do the deliveries, you start noticing that you have to build products to scale yourself. Second, you start noticing all the problems. Whenever you see a problem recur more than once, you say to yourself, “Aha, maybe that's an example of a problem that we should build something for or actually run an experiment to see if we could solve it.”

Very early on, the bias for action turned into this experimentation mentality. We didn't have any organizations at the time or anything like that. It was just a few of us in my apartment. It wasn't as if there were a rigorous system that I'm talking about.

Those were probably the earliest inklings of how we thought about going from doing things that don't scale, to identifying hypotheses to test, to running experiments, and then to shipping products. That was probably the earliest time—the first year of the company.

Fast-forward maybe 1 year, as we started launching into multiple cities. All of the general managers of the different cities—you could be running Boston, someone else was running Dallas, and someone else was running a different city—would report to me. You start seeing that patterns actually emerge from city A to city B to city C, but they are still quite local.

For example, in Boston, there aren't a lot of cars. Car ownership is one of the lowest in Boston in the United States versus other places. There are some strange setups because of the historic nature of the city, in terms of that hub-and-spoke nature I was describing, that actually violate that setup.

There are local nuances, and you start realizing, “How do I actually teach this way of doing things that doesn’t scale all the way to shipping a feature that we know is going to work to each one of these people, so that we can run more experiments at the same time?” Then we would build more products that would go across all of these different patterns.

That’s how this thing has morphed over the years. You basically start with some basic scientific process, if you will. You reach a point where you have to figure out the next iteration in order to scale that process, and then you just keep that going. You’re always testing against whether or not you’re delivering better for customers. That’s always going to be the North Star metric of whether or not this process is actually making a difference. Is it better for customers if it’s faster, cheaper, or more efficient?

Yeah, it’s all of the above. Customers are always going to want the widest available selection. They want every item they can get delivered, the lowest possible price, the fastest possible delivery, and obviously no mistakes. They absolutely expect it to be on time, and if something were to go wrong, of course they deserve to be treated correctly. We get judged on all of those things on every single order.

So this is the idea that you can build a business around things that don’t change.

Yes.

What are the things that don’t change from the customer’s perspective for DoorDash?

Customers are always going to want more and more selection. They’re always going to want more and more affordability, and they’re going to want faster deliveries.

Is it like Amazon—almost the exact mirror of what Amazon is?

When you just think about what people want, I actually think it’s pretty easy because we can play that role ourselves. I think you just ask very basic questions about the direction of travel of certain things. For example, do you think people are going to expect more convenience or less convenience, especially in a world where you think people are earning more—today versus the past, tomorrow versus today?

What do you think they’re going to do with those dollars? Is it going to go more toward consumption? Are they going to expect or demand more convenience or less? When you start asking questions out loud, you get the common-sense answers that you can build a business around.

We were talking about this with the crew at breakfast. The cornerstone of their business is a trait in human nature that’s never going to change: We want more convenience.

Yeah. Always. It’s not rocket science. I think the rocket science is actually how you make it happen.

I love this idea of hiding the complexity. I spent several hours one-on-one with Jeff Bezos, and I’m obviously a massive fan of his. I’ve done about 15 episodes on him, and he listens to my other podcast.

I told him, “Dude, you know how crazy it is that they put a genie in front of your house in Washington? You made a magic button I can press, and anything I want in the world shows up at my house in 2 days. Now it’s a few hours, and all I do is press the button while you handle all the other complexity behind it.”

I was like, “You deserve all the money. I hope you have all the money.” He just laughed and laughed and laughed.

You said something: You’re doing customer support every day. Is this customer support through emails? What is this—emails or chats? Sometimes phone calls? Every day?

Yeah. Say more about this.

Well, why do I do this? I was saying earlier that, for a few reasons, so much of the magic or difficulty of building a company like DoorDash is in all the things you can’t see.

The first thing you have to do is build observability everywhere. Of course, there’s observability through dashboards and systems and, increasingly, AI tools. But I can also see the inbound messages from customers who write us, whether it’s a consumer, a merchant, a Dasher, or an advertiser.

I can choose to ignore them, but those are freebies. How lucky am I to actually have a product that people care enough about to let me know when something goes wrong? Usually, they’re not very positive emails, but they care enough to tell me something went wrong in their experience.

I think the greatest killer of a business is usually silence. Here, they actually care enough to let me know. I owe them certainly not just a response, but the responsibility of actually solving that problem, ultimately.

Second, it’s something I want the rest of the company to do. One of the easiest things that happens as companies get a little bit bigger and perhaps earn a little more success is that more obstacles come between them and the customers, or between them and the jobs to be done.

For example, when you become a company, all of a sudden the only things that get spotlighted are the financial metrics—your revenue and your profits—none of which are metrics that customers care about. There are no metrics in what we report as a public company that customers probably know about or frankly care about, and that’s always quite bothersome to me.

It’s because of our ability to serve customers that we can hopefully achieve the strong financial metrics that investors care about. A lot of what I’m trying to do is build as many reinforcing and repetitive mechanisms and motions, including things that I do individually, that will allow this company to always recognize that the number one job—and the only religion at this company—is to solve problems for customers.

What do you do when the data and the anecdotes conflict?

It’s a tough one. Usually, there’s always an element of truth in what customers are saying, and it usually becomes a trade-off discussion for different teams.

The reason why it’s a tough decision is that it’s so easy to veer toward the data. Almost always, when a customer notices something that’s wrong, or there’s an anecdote that may be a quote-unquote edge case, it’s usually at some tail of a distribution: a distribution of customer-support wait times, how friendly we were when we took the call, how on time or late we were, how accurate we were, or the number of items—or types of SKUs—you care about in a particular category of lettuce. Just lettuce, not vegetables, but just lettuce, right?

The data is probably always going to win when it comes to some sort of prioritization discussion. But when you think about how to make a product better, it’s almost always, by definition, about improving the edges.

That’s why I personally love spending time with a lot of our power users, whether it’s the top Dashers, the consumers who order most often, the merchants we’ve been doing business with for a very long period of time, or new users. They’re at the tails of the distribution of almost every outcome.

A new user who has never touched DoorDash before will tell us how easy or difficult it is to place their first order in a way that someone who’s been used to all the things we’ve been training customers on and figuring out may not. A power user also sees all the issues because they have the most shots on goal for some chaotic event to happen in the real world that we couldn’t capture.

Those edges of the distribution are almost always where the anecdotes are the most valuable. You have to pay the most attention to them because they almost always disagree with the data, and they are probably worth the most in terms of improving your product.

So let’s say you find one of these edge cases while you’re doing customer support every day. What’s your next step?

The ones I love the most are actually the really long ones—the ones where there’s a lot of gold. It’s probably like the research you do on founders: The longer it is, the better, because you get to study the distributions. When it’s a short email about something you already know about, there may not be as much interesting material in it.

I love the 2,000-word emails, especially from Dashers who will give many use cases of why the logistics algorithm broke for them. It becomes almost like a debugging exercise: physical-world things that have occurred, things about our systems that probably broke, and things in our products that couldn’t interface well enough between the physical world and our systems.

Then I go into our debugging tools and literally track it. I watch every single step.

Personally? You’re doing this personally?

Yeah. Once I start figuring out potentially where the sources of error are, I'll either generate a hypothesis and call or email the Dasher, depending on the best way to reach them, or contact the consumer. Then I'll find out whether there's a nugget of insight there—something we actually could improve. Put a different way, can we put a spotlight on an anecdote that improves the product? That's the opportunity I'm looking for.

I've heard you describe this as an eternal mission. How would you describe the eternal mission of DoorDash?

The eternal mission of DoorDash is to grow and empower local economies. We say this a lot, and the reason why it's eternal is because I think it's a fight worth fighting for, or a cause worth fighting for, forever. The best way to grow the GDP, happiness, or safety of a city is by making the small, medium, and large businesses in that city successful. They produce the vast majority of the jobs and consumption dollars for the economy, as well as the money for the police department, the fire department, the parks, the schools, the hospitals, and so on.

The question is, how do you actually make them successful? One of the most positive tailwinds of why this could be a very fruitful eternal mission is because the physical world is always changing, right? It's hard to just scrape it, and that's one of the things I love the most about it. It's hard to scrape all that information, say the job is finished, and then put it through some LLM or something.

The data is always changing. It's not organized at all, and it's not just a relationship between a text, an editor, and a person. I mean, there are a lot of people. There are 3 people involved in every single order at DoorDash: a consumer, a Dasher, and a merchant. Given that we do more complicated things, there are even more sometimes.

For those people, this could be their identity. Going back to what I was saying about small-business owners, what they do isn't an office job or something they just use to earn money so they can spend consumption dollars on something else. This is their livelihood. This is who they are. When I think about those kinds of people, I want those people to win.

If we have to eternally look for the edges of the distribution to keep improving the product, of course we will. If we can do that and make them successful, then they're going to make many things about the cities and neighborhoods that we live in continue to be sustainable and very, very thriving.

The alternative is terrifying. You have 1 or 2 big players.

I don't even want to think about the alternative. You're totally right. The alternative is a very robotic world where maybe we buy things in 1 or 2 ways, or from 1 or 2 places. That's not a world in which you're going to grow the GDP of these cities. It's a world in which you may take away some of the identity of some of the neighborhoods.

I think one of the reasons why people love neighborhoods, or why there's a certain neighborhood that they prefer, is because it has a personality. So much of that personality is given by who the businesses are. You and your friends want to frequent those places and hang out there, in addition to your homes and things like that. That's what makes it tick. That's what makes a place feel awesome—a city feel awesome. I think that's an eternal mission worth fighting for.

This is not something that you can accomplish in 1 year, 5 years, or 10 years. It's constantly changing. What do you do with all this data that you're collecting?

The first thing is that we have to structure it. One of the things that I think Google so brilliantly did was organize a lot of the information on the internet and make it searchable to everybody. Right now, we're still collecting lots of information, and we're trying to do 2 things with it.

The first thing is that we're trying to grow a merchant's business by allowing people to search for their products through our app. We'll bring them incremental business that way. The other way is that we're trying to make the information useful for them by giving data back to them.

Data about their own business.

Yeah, such as when they're out of stock of certain items, whether they're underpriced on a particular menu item compared with what they could be charging, or whether there's an opportunity to bundle certain items or create certain SKUs or new items on their menu or in their catalog, if they're a retailer. We think that would grow their actual business.

This is like Bezos's line about Amazon Prime. He said, “We want to make it so valuable that it's irresponsible if you're not a member.” It's insane. If you can have data for small businesses, medium-sized businesses, and even large businesses that they didn't know—like that pricing thing is interesting to me—where you're charging $15 for a plate of chicken, and we see all these other merchants on your platform where people are willing to pay $25 for that thing, essentially, it would be irresponsible not to partner with you if you have all those insights.

We can also take the same approach that we've built for ourselves—the scientific process, from doing things that don't scale to shipping things at scale—on your behalf. As a merchant, you can be running experiments too. Maybe you can't because you're a single person. You're literally 1 person, like the baker I was telling you about who inspired a lot of our discovery of delivery, and you don't have all the capabilities to run all these experiments. But why can't we do those things for you? Why can't we, for instance—

What do you mean, do them for me?

We could talk about simple things, all the way to more difficult things. For the simple things, we could change menu prices on your behalf. We can buy different kinds of promotions for you based on what return thresholds you want to achieve.

We can talk about more complicated things. There are certain merchants who actually want to grow tremendously. Why not, right? They don't just want to sell something; they want their identity—their passion project—to be exposed to as many people as possible. Some of those businesses find it very hard to grow from 1 store to 2 stores and then somehow to 2,000 stores.

But imagine if you baked cookies, as an example, and wanted everyone to have your cookies. Why can't we match your products with businesses that don't sell your product and create a supply chain in which you can sell those products in more places? You can literally make everyone win. The new business that's selling your product now has a new menu item called a cookie. You get to maximally increase your exposure.

There's a range of things we can do with the information if we knew what your goals were. A lot of what we're doing with businesses is figuring out, at scale, how to maximally increase your exposure and identity and achieve whatever goal you may have.

So that's with restaurants. Tell me some of the—

Or retailers.

Yeah, this gets really interesting when you expand out to every physical business. When I think about restaurateurs and retailers, to me, they're no different from me in the sense that they're entrepreneurs. They want to create something. They have an idea or a passion, and they want it to be exposed to the world. That gives them fulfillment in a variety of ways.

Let's say you want to make T-shirts and sell T-shirts. That's a passion project of yours. There should be no reason why you can't do that today—from testing that idea with the audiences we have, the warehousing and logistics inventory we have, and the ability to very quickly test in any neighborhood or any city, across the tens of thousands of different neighborhoods and cities that we serve and operate in. You can see whether or not you may have something before you actually spend a lot of money trying to open a store. There's no reason why we can't be your business partner for any future creation.

Dude, this is blowing my mind because I just think about DoorDash as a way to get food.

Yeah, I love the idea behind this. It's all about where you start and how you keep going, right? A lot of these ideas came to us from our customers. Going back to your question about why I do customer support, I learn a ton too. Of course, I learn about all the edges of the distribution.

What are some examples of things that customers have—

Okay. One customer in 2014, I'll never forget, was a farmer who runs one of the largest farms in the state of California. They run hundreds of trucks every day up and down the state of California, distributing their produce, meats, and other products to a variety of grocers, restaurants, hotels, and so on.

They've been doing this for 3 generations as a family. They did not start their farm to drive a bunch of trucks. That is not the business they aspired to be in or were passionate about. Literally, in our 2nd year of operation, they called me—or they wrote in, actually—and then we had a conversation on the phone about what they were interested in. They were curious whether we could solve that problem for them.

That's wild. They even asked you that?

This was the 2nd year of the business. I said, “Not yet,” at the time. Perhaps I should—I almost feel like I owe him a call, so this conversation is a good reminder. But the thing I think we've learned is that our goal over time is to be the first phone call for any business, for any issue.

Totally get it. Totally understood.

The number-one calls we get today are about delivery. Increasingly, they've been about other things: Can you actually help us build our app? Can you help us acquire customers? Can you help us analyze customers, retain customers, provide customer support, and store inventory? Those questions are coming in more and more, and that's why we've shipped a lot of the products that we have at DoorDash.

But I think, if done right, DoorDash can be your first phone call to start any business. That's really what we wanted, and we can do it in a way that's very low-cost, that doesn't have to scale if you don't want it to. Some people are very happy with 1 or 2 locations. Or, if you want to become the next McDonald's or the next Walmart, you can do that too.

One of my all-time favorite quotes is from the book Zero to One. It says the single most powerful pattern I have noticed is that successful people find value in unexpected places and they do this by thinking about business from first principles instead of formulas. This is exactly what AppLovin has done with their new advertising platform, Axon. Axon is the most powerful advertising platform in a generation. Axon allows you to capture undivided attention. Axon ads are full screen videos that are watched for an average of 35 seconds. Retention that blows other ad platforms out of the water and you can launch in minutes. You set the goal and Axon achieves it. No complex setup, no expertise needed. And Axon scales quickly. They can put your ads in front of over a billion potential customers. Other businesses have seen immediate results, scale to hundreds of thousands of dollars of spend per day and increase their revenue by millions. And most advertisers aren't even thinking about this channel yet. Less than 1% of advertisers have access to Axon. So, you want to get started quickly and you can do that by going to axon.ai/enra. That is axon.ai/enra.

Something I want to compare your story to again is Bezos, because every time I hear you speak, I hear a lot of Bezos. He obviously did a ton of customer support at the very beginning of Amazon. He publicized his email and made it public: “Email me,” all the time. He tells this great story in one of the books: when he realized they were selling—I think at the time, just books, CDs, and maybe DVDs—and he would send an email to 1,000 customers a day or something like that and ask, “What else would you buy?”

One guy said, “Will you sell me windshield wipers?” And Bezos was like, “Oh my God, we're going to be able to sell anything.”

Or everything in that.

So, yeah, I love that idea. What are these other products that you're building? You have to educate me now, because you need to do more podcasts, first of all. I've listened to all of them, and I didn't know some of the stuff you're telling me right now. I know you've launched a bunch of different products during the last 12 months. Tell me about 1 that you're really excited about.

Well, one of the things that we're trying to do is obviously deliver everything inside the city. To put some context behind it, there are tens of millions of items inside a city that you could deliver. DoorDash delivers a fraction of those items today.

What fraction do you deliver today?

A very small fraction.

Very, very small fraction.

There are many times more things to deliver than what we currently offer. But there are challenges in making these deliveries. For instance, how do you actually know what the catalog looks like for each city? How do you know if the catalog is actually accurate? What if the items are not available in a store but are available in a warehouse somewhere far, far away? There are a lot of these challenges you need to address before you can do something like deliver everything inside a city.

Do you talk about that internally—“We're going to deliver everything in a city”?

Yeah. One of the things that we launched last fall—it was actually in September—we announced DoorDash Fulfillment Solutions. For companies like Kroger or CVS, we'll carry their items, and you can order their items directly from our site. Sometimes they'll actually come from a warehouse that we're operating on their behalf.

That is an example of a warehousing, inventory-management, and logistics solution in which we're offering perfect accuracy and fast delivery in a way that retailers don't have access to or the capabilities to do today. That's part of how you can deliver throughout a city by bringing, aggregating, and making some of the inventory closer to where someone lives.

We're building autonomous vehicles as well. That's something else that we announced last year. It was a fascinating journey—again, candidly, mostly pain and suffering—but most of the journey was recognizing that you actually have to build a purpose-built or intentional product to do last-mile delivery in a way that's very different from, say, robotaxis or transporting humans.

You have to solve problems of getting products, for example, inside and out of the vehicle in a way that passengers can naturally do in a robotaxi, but items cannot do on their own. You have to think about what types of vehicles you may need for shorter-distance deliveries versus longer-distance deliveries, and heavier deliveries versus lighter packages. When I think about some of those products, that's all part of this mission of trying to bring you everything inside the city and giving every business a chance to win.

And are you making the hardware yourself?

Yeah, in some cases we are. We don't have—again, the only religion we really subscribe to is making customers win. We don't have a religion about whether or not we have to build the product or someone else has to build the product.

Actually, when we started the autonomous project—the autonomous vehicle project—we started with the belief that we did not have to build the vehicles. In partnering with a lot of different companies, we ultimately realized that nobody actually wanted to build what we wanted to build. That's ultimately why we decided to start our own project in 2019 and shipped it last year. That's 6 or 7 years—almost 7 years—of development.

Why did they not want to build it? They just didn't want to build what you wanted?

Yeah. Well, if you think about it, in the world of autonomous vehicles, a lot of the projects, capital, and attention are going toward robotaxis. That's just a very different solution and form factor, in our opinion, from what you need for last-mile delivery.

It's very hard, for example, to drive a robotaxi into a crowded hub of merchants—whether it's a mall or a main street—and somehow find parking and get access to the products by itself. I think that's a difficult endeavor to accomplish.

We built DoorDash Dot, which will travel on the road, but it can also travel on the sidewalk and in the bike lanes. It's a much smaller form factor. It doesn't go as fast, but it has the ability to get to the last 10 feet of solving the problem of last-mile delivery, which really is a last-10-foot problem.

Is this live right now?

It's in Arizona. It's in the Phoenix-Scottsdale area.

Is it true? I heard that you guys have partnered with Waymo. Is that true?

We do. We do partner with Waymo, and we do lots of things together.

Is it people just not shutting the door when they get out of a Waymo? Is that true?

One of the things that I think is fascinating about the problems that a company like Waymo—or a company like DoorDash—has to solve is that there are always these funny edge cases in the real world that are very hard to predict. Shutting doors may be one of those examples, right? But you actually wouldn't know about that until you literally read the logs of these customer transcripts.

Look, I think there's going to be lots of things that we could do together over time. But I think it starts with just building the foundation. The foundations you need to build one of these companies for the physical world are very, very different from the digital world.

And you know, that’s kind of the fun part of the exercise at DoorDash.

Okay. Let’s talk about the talent needed to do all the things that you’re describing. I heard you say that when you were recruiting, you looked for Rhodes Scholars meets Navy SEALs. What does that mean?

Yeah, this was a shorthand, I suppose, early on, when we were looking for the types of people that we thought would do well at DoorDash. I think it started because we did every job ourselves, whether it was deliveries, customer support, making menus, or selling to restaurants. We recognized the personality type, if you will. Yes, you needed to be smart, and you needed to have high processing power in terms of analyzing all the information, especially in a world that’s very unstructured.

But one of the things you really needed was the ability to just do things. I think what’s challenging and very different about the physical world, versus building software, is that you have no control in the physical world. We don’t get to control when you hit that order button. We don’t get to control whether or not a Dasher accepts or rejects an order. We don’t get to control how slow or how fast somebody makes an item, or whether an item is in stock or out of stock.

You have to be able to do things to go figure those things out. One of the earliest things I did, I remember, was an interview question. If you made it to the interview with me, your final-round interview was most likely a surprise, because our teams would ask you to answer some prompt about fixing a problem in a city or something like that. You would go out and do your analysis and come ready with a one-page set of notes or something, and then you would come to me thinking that the interview was to present that to me.

I would literally ask you, “Well, this could be a really long or really short interview. I’m going to give you 20 minutes, and you can ask me any question that you want. But after the 20 minutes expires, I’m going to give you $20 that you can use to go and acquire 100 customers for us, and you have 8 hours to do so. But here’s also a plane ticket. I know you traveled far to come to this interview, in case you want to quit the interview now and just move on and find somewhere else to work.”

That was the interview, because that’s the action part, right? So much of what we were trying to test for early on was someone who was going to do something to go and collect information, as opposed to someone who was going to collect data, scrape information from some internet protocol, do some magical analysis on it, and then ship code. What if none of that information existed? You have to go and do things in order to actually collect information.

That was one big behavioral bias for action that we were testing for. That’s the Navy SEAL part, where you have to be willing to do things and be accountable for things. A lot of that was on the non-engineering front.

On the engineering front, we looked for engineers who certainly were great at coding, but we also looked for engineers who would be willing to do deliveries with us. In fact, the interview with me, if you’re an engineer, was the final-round interview: We would go and do deliveries together. The interview would literally take place in my Honda, and we would be doing deliveries for maybe an hour or two. I’d walk you through the flow of the order and ask your opinion on how we could productize it.

In Silicon Valley, I think sometimes there’s a mythical obsession with the 10x engineer, right? I totally get it, and they absolutely do exist, but a lot of times that is about coding prowess. That’s great. We have a lot of respect for that. At DoorDash, we also need you to have problem-solving prowess. The coding prowess, in quotes, at DoorDash is about how you solve this end-to-end problem.

It takes a certain kind of engineer who’s willing to do deliveries, and not just think about code all day or what the latest, greatest AI tools are, but ask, “Is what I’m going to ship actually going to solve a real-world problem? Is there going to be a real customer benefit, yes or no?” That was the type of profile, personality, aptitude, and attitude that we were looking for in engineers.

Was there a specific source where you were finding people like this?

Not really. In fact, to this day, I don’t really look at people’s backgrounds that much. I think one of the things I discovered along the way, probably in the 2015 to 2020 era, when especially DoorDash was building out its team, was that there were more attributes I was listening for than there were things on a resume that I was seeking. I was looking for a bias for action, and a lot of the ways I can tell in an interview is just what people naturally talk to me about.

For example, Christopher Payne, our first chief operating officer, I didn’t ask him a single interview question. After a 2-hour discussion about our logistics algorithm, he went home that night—it was Friday—and drove with his son for 4 hours doing deliveries. I didn’t ask him to do that. I also didn’t ask him the next morning to write me a 30,000-word email about why our logistics algorithm sucked, but he did it. That told me more than any set of interview questions.

This is certainly beyond the resume. We look for the ability to operate at the lowest level of detail.

You hadn’t even hired him yet.

No, I hadn’t hired him yet.

I remember my first one. It was actually supposed to be a coffee chat, not a so-called interview. It was scheduled for 45 minutes with our now president, then CFO candidate. He came to the coffee with his computer and this multimegabyte file, which was some projection of our financials somehow. I said, “What? This was supposed to be us getting to know each other.” But this is how he thinks, right? I didn’t need to look at or read the resume to decipher how this person worked. He showed it to me.

So he built a model and then walked you through it.

Yeah.

How long did that take?

We debated it for over 4 hours. It was literally going line by line to think through these kinds of examples. Ultimately, there were 3 or 4 other attributes that we looked for, and those told me more about how you operate, what makes you tick, what environment you’d be most successful in, and whether or not I thought it matched what was required. So there wasn’t a source. It wasn’t like, “Oh, yeah, we discovered the secret that it’s this company or this school or this background.”

But don’t they have to be very different people from those who would be satisfied working in a completely digital software company? You told this story one time where you took a small sample of 20 Dasher drivers and 20 UberX drivers.

I think the control of this experiment was that you were all getting a guaranteed $20 an hour. If you offered them more money, how many of these groups would switch?

Yes. They were making about $20 an hour at the time. I made a guaranteed offer of $25 an hour if they switched jobs. If the UberX drivers went to DoorDash, the DoorDash drivers would go to Uber. One out of the 2 groups of 20—so, 1 out of 40—made the move.

What conclusion did you draw from that?

This was very early. This was within weeks of the company getting started, because at the time, one of the questions we were trying to answer was whether or not we could acquire enough Dashers, enough drivers. If drivers only cared about money, we were ultimately going to lose, because obviously it’s more valuable to transport David than a burrito or a coffee. We were testing this—we were almost trying to confirm or deny this hypothesis.

What I learned was that they were actually 2 groups of completely different people. The DoorDash drivers were younger. About half of them were women, and they had all sorts of vehicles. Some of them drove motorcycles, scooters, bikes, and, yes, cars, but not exclusively cars.

The UberX drivers at the time were usually men in their 40s—almost exclusively men. Maybe there were a few women in the group, but almost exclusively men. All of them drove four-wheelers, and they viewed that job almost like a full-time job. In some ways, they were moving from taxi 1.0 to taxi 2.0, because some of them had formerly driven taxis.

The Dashers, on the other hand, came from a variety of places: schools, hospitals, restaurants, retailers, service businesses, and moms. If you look at it today, there’s almost no overlap, or very little overlap, between ridesharing drivers and delivery drivers. More than half of the Dashers are women today. They come from dozens of industries, literally every place. The average Dasher only does 3 to 4 hours a week, and 90% drive fewer than 10 hours a week. It just became a very different setup.

They kind of self-selected into what—

They self-selected.

I wonder if the insight you derived there is what I’m getting to: How do you find the engineer who is willing to go get in your shitty Honda—no offense—and do deliveries? I feel like that is such a different person from a software engineer at Google.

It’s a super pleasant person or something, for lunch.

Yeah, it’s super different. Exactly. No, I think you said it yourself. If you think about the early days, I remember we would take a coding break at 10 p.m. to take out the trash because it was an apartment. It wasn’t like an office building where there were janitorial services. We were janitorial, and so we would take out the trash.

It takes a certain kind of engineer and a certain kind of person to actually want to work in that environment. But I don’t think there was a specific background we were looking for. If anything, it was probably a personal background as opposed to a professional one. I think all of these people had a bias for action. All of them cared about the details. All of them had the ability to hold opposing ideas in their brains.

All of them had strong followership. They tended to move with others. Once they joined a company, a bunch of others followed them.

Oh, that’s an interesting trait to hire for. Say that about that again.

Yeah, it’s around followership. They had this ability where—and I didn’t even know why many times—but when you look at Company A or Organization A, wherever they started, they tended to have these groups that were attracted to them. They tended to be quite like-minded, and they always wanted to get better.

That was one trait that we discovered. You see this, and it’s not always professional, like trying to get better at some skill all the time. Sometimes they wanted to be the best burger maker, or they wanted to be the best karaoke singer. They would literally tell you about the process by which, on the weekends, they improved every single week.

That’s not that different from the scientific process that we’re recruiting for or trying to institute in our systems here at DoorDash. It’s actually very, very similar. There was almost an obsession with some activity, and there was a system they devised for themselves to get better. Those were the traits we looked for, as opposed to what company you worked for.

You have a great quote where you say, “DoorDash has always been a company where bias for action is the way we solve and settle debates. We don’t debate a lot. We tend to ship hundreds of thousands of experiments a week.”

You’re not sitting in an office or that conference room behind us, mapping things out and saying, “I have a hypothesis.” You’re just like, “No, we’re going to experiment. We’re going to get to the truth as fast as possible.”

Yeah. I think a big part of this is that, in the physical world, there’s so much that you can’t analyze. Sometimes there’s no analysis you can run on it, and it can be very counterintuitive.

For example, had we not done deliveries in both Palo Alto and San Francisco, maybe we never would have landed on the idea that you could deliver faster or more economically inside of a so-called suburb than in a city. It’s almost like an earned secret.

You read Sam Walton’s autobiography a long time ago.

Okay.

He has that line in there because his main competitors, Sears and Kmart, had people who came up with the idea before he did. They were in the city center. He was in Bentonville.

He basically said, “I was resource-constrained. I didn’t have a lot of money, so I had to start out in these small towns.” If he had never done that because he was forced to—because he didn’t have the money that Kmart or the other competitors had—he wouldn’t have realized how much business there was in these little towns.

The earned secret he had was that if he could compete on his organizing mantra, everyday low prices, people would drive vast distances to save money if he could sell them the same item more cheaply.

Yes.

Bernie Marcus, the founder of Home Depot, realized that same exact idea 30 or 40 years later, applied to a different industry. Constraints definitely breed creativity.

For us, because we had no money—or because I was so unsuccessful at raising money in the earliest years—we had to run these experiments. If a company has no ability to compete with its budget and other companies are outspending it with marketing dollars, for example, you only have one way to compete: You have to build a product that has better retention and better engagement. There’s no other way.

So what do you think when you see these giant seed rounds that we’re seeing now?

It’s impressive, is what I think. My encouragement to those founders is to actually find a problem worth solving first. But once you find the problem, go and solve it, because at the end of the day, that’s going to cover for whatever financial metrics they’re going to be solving for.

Everybody says, “I need more money, because if I have more money, I win.” One of my favorite historical anecdotes comes from the biography of the Wright brothers by David McCullough. Have you ever read it?

No.

Oh, I know you like to read history.

I do. I should probably check that one out.

Listen to the audiobook if you can. I know you’re busy. It’s great because human-powered flight was a centuries-old problem. People were trying to figure it out over and over again.

At the exact same time the Wright brothers were trying to do it, they had better-funded competitors and better brand names.

Sure.

I think it was Samuel Langley, who was backed by the Smithsonian. I think he had raised $500,000. This was a crazy amount of money at the time.

There’s a great line in the book where, essentially, the Wright brothers solved the centuries-old problem with the modest profits from their bicycle business. They tallied up how much it cost them, and it was about $1,500.

That’s incredible. I want to go back. You make some jokes that—I don’t know, maybe they were jokes, but I laughed when I heard you say this—you’re like, “I must be a really bad fundraiser.” What was this, like 1,000 days of hell? Can you talk about this? You’re trying to tell people, and it’s weird because the metrics in the business were all trending in a positive direction. They were. So explain what the hell was going on.

One of the hardest things you learn as a founder, and certainly as a CEO, is that you have to learn how to control your own psychology, because there are lots of things that are going to be out of your control and that won’t make sense to you.

This happened very early with DoorDash. I mentioned earlier that we had a difficult time raising the seed round, and then there was this difficult event with Stanford football in which we ran out of our money even faster. But we survived that. We raised the seed round, and our Series A and Series B were hot rounds. Somehow, I was able to raise money in less than a week in each of those instances.

But in the spring of 2016, a few things happened, and this was probably when I first started learning about the importance of dealing with your own psychology. It was actually the first time I took a vacation.

We started the company in January 2013, and this was January 2016—about 3 years or so. It was my first vacation: 5 days with my wife. We didn’t go on a honeymoon, so I promised her that we should make up for that. We went to Hawaii for 5 days.

We had actually received an inbound term sheet, so things were going pretty well. We were going to raise our Series C in the winter of 2015 or the spring of 2016. I specifically remember asking the investor, “Why don’t we just close this now? Why don’t we close this before the year ends?”

He said, “Don’t worry about it. You’ve never taken a vacation. Go take your honeymoon. Everything’s going to be fine. We’re good for it.”

I said, “Okay, all right. I’ll go with you on this one.” My intention and my style is usually to get things done quickly, but I said, “I’ll go with you on this one. I owe this to my wife.”

We went to Hawaii, had a great time, and came back in January. The public markets actually tanked. That was the first thing that happened. Companies at the time—I remember LinkedIn, when it was still an independent public company, or Salesforce—dropped 30% or 40% in value in a matter of a week or so.

All of a sudden, analysts and Twitter at the time—maybe it wasn’t as big as X is today—had all this commentary about how, “This is the beginning of the end. Finally, the bubble’s going to burst.”

That very quickly trickled into the private sector and private financings, where investors started backing out, including from the DoorDash Series C. This was the start, I would say, of 3 years during which DoorDash could raise very little money—a fraction of what our peers could raise—and during which we encountered several bouts of almost running out of cash.

But you're right, there was this tension internally. Here we are: the markets are going down. All of a sudden, the narrative for DoorDash was that this really hot company was now a company that could do no right. You could never make money; you couldn't beat all these competitors who were better funded. At the time, there was Uber and there was Amazon, either coming in, already in, or announcing more expansion. Even if you won, you were going to lose because this was a money-losing business, or a forever-money-losing business.

Those were some of the headlines, or the themes behind the headlines. But at the same time, you look at the metrics on the inside and actually see everything going in the direction that you would hope for as an entrepreneur. You see repeatability from city A to city B to city C. You see unit economics improving. The reason the company wasn't profitable was because we were constantly launching new markets.

New markets require investment in the beginning because you're paying drivers to make sure that they can stay on the road even when you have no business. That's what was happening internally. That happened for about 3 years, though, where we were kind of stuck in one of these macro investment cycles, where the company could do no right. The sector was viewed as toxic, and that was probably the 3 years in which I certainly had to learn how to deal with my own psychology.

So, how were you doing that?

There was no one way. I think the first thing is that you have to make sure—you know, a lot of times, it's very easy to believe in your own beliefs. A place like DoorDash, which is very intellectually honest, asks, “What's actually real versus what maybe people are saying?”

We used to do this because we could all fit in 1 conference room. At the all-hands, I would show every metric in the company, including our cash balance, which was obviously going toward the x-axis. People were getting nervous, but they were asking a very good question: “Tony, I don't get it. The cash balance is coming down, but the business is going in the opposite direction. It's going up and to the right.” We weren't spending in a very organic way. We didn't even have a marketing team, let alone a marketing budget. We didn't have money.

Job number 1, to me, was to actually put the company in the best possible place by focusing on what we could control. Otherwise, I'm going to go crazy. I'm going to go crazy, and we're not going to put the company in the best position for success.

We got a group of maybe 20 or 25 people—the people who ran a lot of different important areas—and brought them under the tent. I basically said, “Look, we have to do the following: We've got to keep growing and keep taking share, we've got to get more profitable, and we can't run out of cash.” There's no “or” in any of these statements. It's an “and” function across all of them.

That was ultimately what I kept obsessing over, because if I obsessed over anything else—the markets, what people were writing about us, or another rejection from an investor—if I just obsessed over what wasn't in my control, I was certainly going to go nuts. That was part 1: focusing on what I could control.

Part 2, I think—and this is another lesson I learned during those years—is that it can be risky, but I actually think it's really important and undervalued to have genuine friends at work. This can't just be about financial success, commercial success, or some professional success on a résumé. There is this adventure, if you will, that we're on, this worthy, eternal mission, and at least we're going to die trying. Worst case, we're going to die trying.

Kind of like the Stanford football example. Yes, of course, we want DoorDash to make it, but what gets you through the next day isn't thinking about DoorDash as much as thinking, “I just want my teammate to be successful.” This willingness to think about someone else in addition to thinking about your own problems actually made it a bit easier to go through.

The final thing is going back to trying to build anything out of things that don't change. One thing I've been able to keep throughout the DoorDash chapter so far is my exercise routine. The routine itself has changed, but back then I was really into running marathons and things like that. Just keeping that up, having something that was a bit of a constant in my life while everything else was out of my control, extremely chaotic, and usually extremely negative—that helped. Part of the routine was also date nights with my wife.

There was no one thing to answer your question about how to manage my own psychology. Trust me, I didn't have everything together during every single period. But when I look back and ask, “What were the things that got me through it? What were the things that I kept trying to tell myself in my notebook about what to do every single day?” those were the things.

Yeah. You control what you control. I love this idea of having a mission bigger than yourself because you have this great line where it's like, at some point, willpower is going to give out. Your own personal willpower is going to give out, especially because you're doing this for over 1,000 days. How many rejections? How many nos were you getting from investors?

I stopped counting after 50, but it was over 100.

That's incredible. To this day, though, you don't look at the stock price. I heard—we have a mutual friend, Robbie Gupta. You went on his podcast and said, “I don't look at the stock price. I try to get everybody else in the company to pay attention to the business.” And he goes, “You had to remind me of our market cap because I don't know what it is.”

Yes. Yes, that's still the case today.

Yeah, I mean, back to the things that I can control: Usually, our finance team will remind me of the market cap during earnings calls and things like that. But sincerely speaking, it's not something I get to control, and it's also not what is fulfilling or motivating to me.

What am I going to do on a daily basis knowing what the stock price is? Am I going to behave any differently? No, I'm not going to behave any differently. I'm probably still going to stick to my routine. I'm going to spend time with our teams that are trying to make sure they can hit the year. There's 1 group on the team that's trying to invent the future, and there are several of those teams. Then there are customers. That's how I spend my time.

I want to talk about that. I'd be remiss not to mention this because, again, I don't know why, but every time I hear you speak—and now, having this conversation with you personally—there's just so much Jeff Bezos stuff going on here.

I think you already know this, but there was a time in Amazon's history where Bezos talked about this. I don't know what the exact numbers were, but the stock price went from around $180 down to around $6. His whole point, I think he talks about this in his shareholder letters, was that the stock price dropped by 90% or whatever the number was, and he said, “Yeah, but I wasn't focused on the stock price. I was focused on the internal metrics of the business, and they were all getting better and better and constantly improving.”

He said, “I knew this was temporary. I will get out of this. I will survive because I'm going in the right direction.” What is this idea—you had this saying—that as an operator, you need 2 management systems? I think you just dropped a hint right there in what you said earlier.

If you're so lucky as an entrepreneur to one day find product-market fit, where you can organically grow and build a business that's self-sustaining and generates cash, you have the privilege of making a choice. That choice is: Do I keep doing what I'm doing, or do I keep expanding in service of our mission?

When you look at Amazon—and one of the reasons why I think Amazon is inspiring, as are a lot of these big tech companies—they tend to do 2 things at the same time. One is that they continue to build the core business, the business that got them to their place, both in terms of their place with customers and what they're known for, as well as their financial place, where they can invest from.

But they also do new things. They launch the next thing or the next thing, and they're trying to create the next thing. Those are 2 very different systems. One system is about making sure that you can constantly reinvent yourself. You're trying to build the next version of the product, to disrupt yourself, to build something that's 10 times better than what you have today, while you're also running the machine at the same time.

You're flying the airplane. It's a big airplane, you're carrying lots of passengers, and you're going to do a midair engine transplant. That's 1 type of system that you're constantly trying to build. Then there's the new stuff. It's not even an airplane; it's like a paper airplane. There are no passengers, nothing. You're in search of product-market fit all over again, and they require different ways to measure success.

They usually require different talent. They require a different amount of resourcing. They have vastly different timelines in terms of rate of progress, and they tend to have much larger error bounds in some of these newer areas.

It’s really hard to do because the more successful your big airplane is, the more paper airplanes you’re probably going to have. Some of those paper airplanes may be very expensive to build because you need more shots on goal to keep up this big business that you’re trying to move in service of your mission.

So the people scaling the businesses in DoorDash that are post-product-market fit, right? And then the inventors—do you separate these people?

We try to.

Okay, separate buildings? How extreme? No, no, no. I was just asking if you even take it to that extreme, like separating—

Separating them physically? How do you do this? Yeah, usually that happens, but I don’t know if that’s as important as needing very different goals, goal-setting systems, and incentive systems. That is probably more important than physically, necessarily, where they are per se.

DoorDash today also operates in more than 40 countries, so it’s tough to get every single person in exactly the same location. But it’s very important to separate how you actually track, manage, measure, and incentivize these projects. That’s more what I’m referring to.

Are you making these decisions about, “We’re going to allocate this amount of resources, this amount of time, and this amount of people to these experiments”? How do you actually structure this?

Yes and no. If I had to make every single decision, DoorDash would certainly move a lot slower than we would want to move. But certainly, I have to set the standards and the pace, if you will. That’s what I view a lot of my job to be.

Usually how it works is, first of all, anyone should be able to come up with an idea. It can’t somehow be that only the leaders come up with ideas. Usually, it’s the people closest to the problems who actually come up with the ideas.

If they can run an experiment—back to this process—that actually demonstrates some viability of success, that customers actually want this product, then it starts entering the phase where we can evaluate whether or not we should pursue it during our planning process. Through the planning process, we decide, “Okay, how many chips should we bet in project A versus B versus C?”

Some projects aren’t all starting at the same time. Some projects are older; some projects just got born. It’s almost like an internal venture system, if you will, where it’s stage-gated. There is no, “Oh, you get all the money up front.” You have to earn your right to the next stage, and that’s going to be based on how well you’re solving that customer problem.

Where did you get that idea from, this internal stage-gating—essentially treating it as internal venture capital?

A lot of it came from DoorDash’s own history, where DoorDash kind of worked this way, right? Maybe some of it wasn’t in the exact formulation we wanted, but that’s how DoorDash was born. You started with little resources, or not a lot, and as we got progressively more successful, or discovered more product-market fit, we were given more resources.

To me, when I think about the things that we built that most solved customer problems, it tended to be when we were most resource-constrained. I do feel like that’s important to know, because the most important thing, again, when you’re starting something is: Do you really have something, or are you just believing that you have something?

You don’t get to make that call as the inventor. It’s the customers that you’re inventing for who ultimately are going to tell you whether or not they’re going to buy it. That’s the most important thing. We’re trying to make sure that we can actually create something that’s 10 times better than the status quo.

If we can do that, of course we’ll keep scaling. Now, some projects cost more money to start, but that’s just the nature of the problem. We’re still, relative to its size, giving it a small amount of budget to begin with.

Are you also learning from your peers? The reason I ask is because we just did one of the episodes I’m most proud of so far for this new show—the one we did with Tobi Lütke. I was really excited to talk to Tobi because I’m constantly asking world-class founders, “Who are you learning from?”

Tobi is your favorite founder’s favorite founder, and his name kept coming up over and over again. That was the conversation where you ask a question and you cannot predict what’s going to come out of his mouth next, because he has all these uncorrelated ideas, which makes for a fascinating conversation.

Who are the people that you’ve either built relationships with or studied in your peer group that you’re also learning from and taking ideas from?

You’re right, Tobi is absolutely great. First of all, some of the peers I have are just people that I grew up with. One of the benefits—which we didn’t get into—of Y Combinator, besides just being a forcing function for testing our commitment to the project, was actually the peer group.

We never got into that part, where, if you think about the 2010s, the companies that grew out of Y Combinator that we grew alongside—maybe we were in slightly different batches, or not exactly in the same one—but whether it was Airbnb, Stripe, or Coinbase, we all kind of grew up in the same era, if you will.

As a result, we got to know each other through different events and venues and things like this. Trading notes with one another certainly was important, and we’ve all had our share of challenges and triumphs.

Then, looking at companies that are ahead of us, in my not-day job—in my other job—I play a small role at Meta, where I serve on the board. I’m learning from founders like Mark, who certainly have built companies that are at a different level of scale versus where DoorDash is at.

Let’s stay on Mark for a second, because we were talking before we recorded. I got to spend some time with him and had a few conversations with him, and I came away even more impressed than I thought I would, given the fact that, for his age, he doesn’t really have a peer. I actually told him that.

I said, “I wish you did more podcasts and talked about how you built a company that no one else your age is even remotely close to.” What are some things that you’ve learned from observing him, being on the board?

I think the first thing that impresses me a lot about Mark is his willingness to always learn new things. I think one of the traps, if you will, of success—or fighting your own psychology—is not just the challenging parts when things aren’t going well, but also when things go well and maybe you’ve actually achieved some milestone.

One of the trappings of success is actually wanting to hold on to it. What you see in someone like Mark—and, I would argue, the team at Meta—is this willingness to reinvent themselves, betting early, for example, on building a different platform in the case of virtual reality and augmented reality.

Obviously, they’re going all in on AI. Those things take a ton of courage. There’s not a lot of data early on in either a platform shift or a new technology’s arrival to know whether or not you’re on the right track all the time. But you’ve got to place the bets before you can see the success.

I think that willingness to learn a new domain where you’re the rookie, where you’re going to stumble, where you’re going to get criticized and misunderstood, and where you don’t know the answer—which is the opposite, if you will, of the successes that they came from in terms of the previous businesses they’ve created—is really impressive.

That willingness to always be the beginner, to always go in the arena and sweat and bleed and toil and struggle, is really impressive.

You both share a love of jiu-jitsu. Have you found anything from your jiu-jitsu practice that you’ve brought back to your day job?

Jiu-jitsu is a fascinating activity. It’s like some version of physical chess. It’s a great way to think about it. It’s almost like an exercise where there are so many opposites that you have to hold at the same time.

The best jiu-jitsu athletes can be extremely firm and strong, yet at the same time extremely relaxed. They’re very capable of being intentional with their game plan, but then give up and release their agenda within a nanosecond if they see that they’re losing their position.

I think the willingness to be so flexible is certainly something that I’m trying to teach both myself in my personal life and also bring back to DoorDash.

I think the other thing is just, no different, frankly, from any craft: the willingness to get 1% better every day. In a particular position, any particular flexibility exercise, or simply improving your balance in order to hold a position, very small things ultimately compound. When you look at the elite athletes—not someone like myself, but the elite jiu-jitsu practitioners who win world championships or medals at events—they all have that. When you actually talk to them about their craft, it’s the tiny details. It’s the edges of a move.

It’s not some silver bullet that they’re looking for in a match or something like that. In fact, these matches at the most competitive levels are decided sometimes not even by points. They’re decided by what are called advantages, and that is one thing that I think is just a great reminder: you always have to be trying to master that craft.

I have to ask you how AI is changing the way you guys are running a business. You have this great line. You said, “I think some of the technical advances, like AI, have given people new ways to run companies.”

Yes.

How are you using it? What are you doing? How’s it affecting your work?

It changes by the month. This is a question that, if we were to talk in the future, I’m not sure it would actually be the same answer. One of the first things I would say is that I think about some of the systems that we’ve architected here, about how you can learn from doing things that don’t scale all the way to shipping, especially with something like coding.

Right now, I think where the agents are, they’re still good at what I call functional tasks—for example, coding. But outside of coding and looking at cross-functional areas, they’re not quite there yet for a lot of reasons. Within something like coding, anyone, frankly, from any function, can come up with an idea, run the prototype, run the experimentation and the analysis, and then actually ship to a small group of people all by themselves.

That is very impressive, and it collapses the amount of activity required, or speeds up the learning loop you can have in any scientific process inside your company that touches code. That’s very cool.

Second, what are LLMs good at that humans are not good at, or are less good at? They can have almost infinite memory and infinite context, and search across any sort of file. So then the question becomes: how do you actually feed it the right information? If you can feed it the right information, it probably can do a lot better than humans can at the same activity.

I think those are 2 areas in which, whether it’s speeding up your learning processes or actually improving the same activities that are effectively done manually right now, they can be done with not just higher efficiency, but also greater effectiveness.

Would there be any benefit for you in partnering with one of the big model companies, with all the physical data that you guys are collecting, or would you keep that proprietary?

Most of the information needed to run DoorDash as a great service consists of things that we use for ourselves. The reason why we use them for ourselves is because it’s not just that simple—like, we just give away information and then somehow someone’s going to be able to do something positive with it. You also have to take the action that the data suggests.

For example, if the data says something is missing in an order, or the Dasher is at the wrong location and cannot find the customer, those are all pieces of information. Some corresponding action has to take place in order to actually solve the end-to-end job: to get the missing item or to find the customer, wherever the Dasher is.

A lot of DoorDash is—sure, we have a lot of information, but we have to do something productive because it’s the end-to-end job that ultimately we get judged on with customers. If we can partner with anyone, frankly, in order to solve the end-to-end problem better, of course we would do that. But I think it’s very hard sometimes to just give away something if there’s no ability to correspond that with action that will ultimately solve some customer’s problem.

Think about what a wild ride you’re on. You start the company and your competitors are literally using fax machines, and now we’re in the age of AI. You have this great arc in 13 years.

Yeah, and I love this quote, and we’ll end here. You have this great quote: “There’s just no better way to be an expert than to just do the work. You might be surprised at how quickly you get to become the expert.”

Yeah, that’s beautiful. Thank you very much for the time, Tony. I think we just scratched the surface. I think there are a lot of things that you said today that I haven’t heard anywhere else. I’d love it if you just came back on every few months, every year, whenever you want.

Sure. That was fun. Thanks for making time. Thanks, David.