But the best way to tell as a founder of whether or not you're sitting on that border of delusion and ambition is whether or not your customers actually want what you're doing. What's maybe one cultural aspect of DoorDash that was present from the super early days?
I think the name of the game is to understand what is it that we're actually doing for the customer. It's almost like if you can run your business with metrics that only a customer would care about. A customer doesn't care about how many users you have or how much revenue you generate or whatever your ARR is.
Do you worry that there is a more AI native version of DoorDash out there?
All right, folks. Welcome to SPC. Welcome Tony to SPC.
All right, we have a short hour here, so I'm not going to take a lot of time to introduce Tony. I think he's pretty well known at this stage. So, I'm going to jump right in. Did you know that you wanted to be a founder? Like how could you have told, what were your earliest inklings of wanting to be a builder, a creator?
I don't know if there were inklings of my becoming a founder because I thought I was going to become an academic. So my parents brought me to the US when I was very young, when I was four. And you know, our family, I didn't know what this meant at the time, but our family came to the country with 250 bucks in the bank. As a result of that, I had to earn all of my toys. So one example that comes to mind, you know, maybe my first company, I didn't think of it this way until you asked the question, was a lawnmowing business when I was seven or eight. I was barely tall enough. I don't know if anyone has mowed lawns, but acres of grass, not like the San Francisco yards we're talking about. And the reason why I did that was because I really wanted to buy Nintendo and Duck Hunt. And so, I don't know if that's called tinkering, per se. That was probably the first quote-unquote entrepreneurial experience. And you know, something else that comes to mind for whatever reason when you ask that question is I had to move around a lot as a kid. My parents were always trying to find a better lot in life for our family. And for me that meant trying to find a better school. So I probably went to like four schools in my first, I don't know, six years of elementary school, maybe seven years, something like that. So, I was always kind of the new kid who had to figure out whatever, a new teacher system, a new friend network, a new logistic setup, etc. And so, I was always very good at figuring out the new thing. Those are probably the best examples I can think of that maybe were somewhat useful to me later on in life where you kind of have to figure it out. If I were to summarize those stories, it was figure it out. Figure it out on how you want to buy your own Nintendo and video games. Figure it out on how you're going to make it in school. But I had a very free childhood.
Yeah. And it's really interesting. So actually some similarities there. So, I also moved around a lot growing up. And I actually probably attended about 10 or 11 different schools all through high school.
And I think one of the things it makes you, kind of moving around a bunch and always being the new person, is be really adaptable, be really flexible and you kind of figure out how to fit in and it kind of makes sense, right? That you're very comfortable with the fast rate of change in Silicon Valley. So, here's kind of the way to morph this question. I think that sometimes it's only obvious when you connect the dots backwards as Steve Jobs said, right? Like what your life ends up becoming. So to jump around a bit here, what do you think the first 10 years of DoorDash, as you kind of now look back and you know it's kind of amazing to hear you talk both about the journey so far, but also what's coming? What do you think it uniquely positions DoorDash to do in like the AI world that we're kind of going towards, right? Like what are the set of skills, sort of the set of like almost happenstances that have happened that allow like what can Dash become?
I feel like right now, and maybe I had this thought in 2021 or '22, that this decade you're going to have like two big battles in business. And they're almost like, you can call them battles, you can call them wars, you can call them explorations. You know, there's a handful of companies that are fighting for your attention. They want to be the super assistant that can solve any query that you may have or later on any task that you may have. And that's not the world in which DoorDash lives. DoorDash lives in the second world, which is we are trying to facilitate all of the physical atoms, not the digital bits, in the world to actually make something useful that hopefully will complement actually what's happening in the race for your attention because the two kind of systems kind of need to work together if you think about it. You know, no app ecosystem or operating system is that useful if it can't really solve everyday problems for you. I would argue it's why you find probably the device in your pocket to be the most valuable or most used device you may carry or have. And it's not just because it solves lots of fun digital problems in the background for you, but it can do things in the physical world for you. And so when I think about the physical world that, you know, probably certainly isn't getting the amount of attention as the digital world, there's lots of exciting things happening. You know, for example, if you think about the amount of things or stuff that people look for or buy today or could buy today, there's tens of millions of those items today. On DoorDash, however, there's only one of them, 1 million of those items. That means we can solve 5% of your needs expressed, you know, through searches, expressed through things you've actually bought, expressed through things that sit on shelves but are not digitized in any way, shape or form. Bring everything inside the city to you, I think, is the best way to grow the GDP of those cities and do something productive for your life and hopefully, you know, team up with all these people trying to compete to be your digital assistant. I think that's a really big mission in what we can do. We can talk about how we do that, whether it's bring inventory closer to you, building autonomous technologies. Those are all getting into the space of the how and the details. But I think there's a big kind of battle, if you will, of how do you actually facilitate all the atoms in the physical world in a way that actually solves real problems for real consumers and grows the GDP of the cities that house all of us. That's a big mission of where DoorDash is going.
Yeah, it's a compelling answer. As you were saying that, I was trying to do a mental checklist in my head about, okay, probably use 70 apps on my phone regularly and I was like, okay, probably 60 of them are just bits flying around the internet and maybe 10 of them interact with the real world. Maybe Amazon, Uber, DoorDash. And it's a pretty small list of apps that kind of like actually interact with the physical world that have a digital interface. And this is interesting because you have clearly created a company that believes it can tackle that divide, right? So, and there very few companies in America that I think have that belief relative to say if you go to Bangalore. So, S has a place in Bangalore. You see entrepreneurs there and they see logistical operational challenges and it's like they have it for breakfast. It's not a big deal.
Right. And so it's kind of always remarkable to me that when you come to America, there's a real trepidation, fear, or maybe almost lack of know-how. How did you build that within the company?
Yeah. Well, we didn't have any of it either to get started. I mean, but I guess the short answer of how any of you guys are probably building your own companies is you start by doing the work yourself. You know, for us, it was four of us who started the company, we just did deliveries every day. We did customer support every day. I still do customer support, you know, about 15 minutes every day and I still do deliveries every year. And that teaches you many things that teaches you things that there's no way that if you're just staring at, you know, I guess now if Claude is staring at your databases to try to just come up with whatever the right algorithm is, it doesn't teach you that, oh, actually, you know, if I try to assign the closest driver to a restaurant for an order, that may not be the best idea if the food's not ready. But you would know that almost immediately if you actually did deliveries yourself. And so there's no like secret sauce I think in any of the domains that each of you are pursuing. You have to be the expert and you can be the expert maybe not on day one but if you did it every single day for a decade I think you'll be pretty close.
You know, one of the things that is tricky about building for the real world is that, you know, margins are thinner and kind of like, you know, and sometimes when you talk to entrepreneurs who are kind of building in hardware or the real world, you realize what a great business model software is, right? Like if you have 75-80% growth, you have a lot of experience with that. Yeah, exactly. So 75-80% gross margins, you can a lot of up and still your business will be alive generally, right? It's much harder with like kind of real world businesses. So I guess how do you balance in some ways, how do you create a system within these companies that allows you to move fast while also realizing there is a cost to actually making mistakes that is much larger than your average software company and obviously there's a very famous example that you've talked about about the Stanford football game I think but how do you not let that experience intrinsically just slow you down so how have you found that balance at DoorDash?
Yeah, I think that, you know, this is probably the answer that would be true honestly in building great software too which is if you care a lot about craft, you should care a lot about efficiency. And I would argue every craft, every function has some efficient frontier. There's always some version of the optimization where you're trying to maximize quality subject to some constraint. The constraint could be time, it could be budget, it could be something else. But every function has that. And I think everyone who I've always admired, you know, in terms of their pursuit of their craft excellence always gets that. So there's no like one thing, Adita, to answer your question but it's almost like every project if you think about the unit of work at any company including your own probably boils down to some project that project boils down to some set of workflows and for each one of those projects and workflows there is probably this efficient frontier. I think even the fact that you institute that and then measure it and then hold people accountable to it, that ultimately becomes a system that reinforces in real practice, not just culturally, but actually operationally how you actually build a system that cares about quality and cares about efficiency.
Moving on to perhaps just kind of like, you know, tackling obviously what everyone here is working on directly, indirectly, obviously, you know, various parts of the AI stack. Maybe to think first about AI at the application layer. I think one of the tricky parts is that it's weird, right? Like you would think that we would have moved on from a basic kind of like text IRC interface to AI which kind of came about in November 2022 with ChatGPT but it's actually pretty, I mean if you use Claude Code in terminal mode it's literally exactly that right now. So I guess what I'm trying to do, you have thoughts on where does this, how does this UI evolve because I don't actually think we know yet as an industry like, you know, face mode, text mode, they all seem quite basic. So, I'm wondering if you are seeing something within DoorDash that you might have an idea to share.
I definitely agree with you that we're all trying to figure it out and that everyone knows that, you know, whatever we have today is probably the worst version of what we're going to have in the future, right? And so, I think in some ways that's, you know, hopefully should bring a sense of optimism. I mean these tools have already brought tremendous productivity gains and you know frankly ideas, new ideas, etc. I think the thing that sometimes gets missed though in the search for some of these, and I think this will be the answer or at least the approach that we're thinking about at DoorDash, is you got to figure out what the end-to-end job ultimately for the customer is. And I think sometimes it's easy to lose sight of what that is. You know, for DoorDash, well, the end job is to bring you something. Whether it's a t-shirt, a water bottle, a hamburger, whatever it might be. And our job's not finished until that whole sentence gets completed. Some of that happens in a digital interface. Some of that can happen through photos and video, maybe voice, maybe search, maybe an assistant that does a complicated query for you. But some of it has to get finished ultimately offline and physically. Actually, many things have to happen offline, physically in order for that one sentence to be completed. And I think that's why it's hard. It's because the end-to-end job for any user depending on what the job is may actually require a different interface and I think that's what we're all trying to figure out. We're kind of solving for different use cases here, different use cases there. On the flip side, you know, a lot has already taken off on just some of these chat interfaces. I mean, the usage, I think, speaks just by probably the usage in this room. And so that's where we're at right now. And I think, you know, the main advice I give our teams is what's the ultimate end-to-end job. And then from there, let's figure out what the interfaces ought to be. And I think right now I think every company is in some exploration of that.
As a prominent company that is obviously integrating AI, how do you make sense of the model landscape? Are you just happy that there are multiple model providers that are kind of like essentially trying to win your business? Do you also feel as though that you need to be moving down the stack yourself to own more components around the context, the memory, the long horizon planning? Like what is your kind of like chess play there?
It goes back to a bit the previous question where I think the name of the game is to understand what is it that we're actually doing for the customer instead of starting from what is the thing I must protect for my business. It's almost like if you can run your business with metrics that only a customer would care about. You know, for example, a customer doesn't care about how many users you have or how much revenue you generate or whatever your ARR is or your weekly growth rate. I actually think that that will give better insight as to potentially what to actually work on in terms of producing a roadmap. And so for us, the first thing is you got to play with all the tools and understand what are the actual true capabilities versus what their promise is. And I would argue that's different by different types of areas. In coding, like you're saying, I totally agree, I think the progress of reality versus promise is getting very quickly closed. In some other areas where you're trying to solve an end-to-end workflow that may have multiple systems that need to be touched that requires human judgment that does not have a verifiable good response, that's not a great place right now in terms of for AI agents where the gap between reality and the future potential is still fairly sizable. I think at least the way that we've thought about it is how are we going to make the most sense of what the end-to-end job is? How do we make the most sense of what are the tools today? What can they be used for? And then how do we actually marry those two things to deliver something unique. It's not rocket science to answer your question. It's also not starting though from a place of okay here is the big strategy and therefore these are the things that we must own because I think candidly I would say that the field is moving so quickly that it's very very hard to know exactly certainly within the next 5 years where the technology is going to go and how far in terms of what it can do will be.
Yeah. I mean, I love that answer, particularly the part about everybody should just be using these technologies. And I often am just like, you know, there's still a little bit of an inertia where you go in, you're like, I'm going to have to learn how to use this thing. I'm going to suck at it in the beginning. But these tools are just like, honestly, all of us are capable of using these tools.
By the way, one of the things, one of the notes I have here is that Tony is customer obsessed, not competitor focused. And it's pretty amazing. You've kind of brought up like having a customer-centric point of view multiple times and it's clear that that's how you kind of like frame a lot of your kind of thinking about both what to build and what DoorDash should be doing. It's pretty amazing.
You know generally when there are these huge platform shifts is when both old and new incumbents tend to get displaced right or at least they're under the most threat right so what do you, I mean do you worry that there is a more AI native version of DoorDash out there and if so like what would that look like? You don't have to give out the ideas but like I'm curious like what is your, what would be your, like what is a more AI native version of kind of an interface to the physical world. What would that look like?
You know, I always go back to what is the end-to-end problem that we're trying to solve and what are the different component parts, right? One of the challenges we found in building DoorDash is it's not just an app. It's not just a logistics algorithm. It's not just a lot of customer service issues that we have to resolve within usually seconds or minutes. It's all of these things coordinated together in order to just do something simple like bring you lunch or bring you dinner. And the question I always ask is where can a new technology, to your point, and it doesn't have to be AI even though I'm sure that's getting more than 100% of the air time these days, whether it's autonomous technologies, things that are physical like hardware, but regardless of the technology, regardless of the specific implementation details, like what is this thing actually capable of in terms of which part of the stack can it actually do better? Because in truth, I don't view any digital solution that's going to replace an end-to-end system that has physical parts if it can't also do the physical part, right? But maybe there's parts of the digital experience in which it can, certainly for instance you can write much better queries on chat assistants with greater intelligence that right now the DoorDash app can't do. Okay, that might be an area in which we look into, right? But that alone would be insufficient to actually bring the customer what they want, which is maybe a lunch or dinner or the latest pair of shoes or lipstick that they forgot because they're on a work trip or something like that. And you kind of have to do all of those things. And so, the way I think about it is like, well, where are we most vulnerable or where do we have the most opportunity? Is the same version of the same question, the other side of the question where it's actually going to help. Again, if I looked at every metric at the company from a customer's perspective, not whatever the output metrics might be in terms of what we report to Wall Street, well, very quickly you're going to know. And then I'll just evaluate the new technology against each one of these areas to see where can it do better and can it do better? We're going to be the fastest or the greatest adopter of that tool.
Are you doing anything interesting in terms of accumulating kind of a data moat that you know perhaps in the coming 2-3 years would allow you to essentially build your own kind of like RL models or your own whatever the version is of kind of doing post-training, whatever the current regime is, kind of like in three or four years are you doing anything special in terms of readying yourself for that?
Yeah. So, one of the things if you think about what we've had to do in terms of building DoorDash is we basically have to digitize the physical world. And I think while LLMs can obviously train on all of the internet, if the data is not available on the internet, that would be very difficult to perform that task. You know, data such as where can we find the last parking space near this building in San Francisco? Data such as how many Coke cans are available to buy at the Safeway on Fourth Street nearby here. Or where can I find the cheapest pair of shoes in size 10 and a half for my upcoming 5K race this weekend? Well, those answers don't exist on the internet because those answers don't even exist sometimes in the physical world in a digitized format. And so if you think back about what one of the things that DoorDash is building, we're building lots of inventory systems effectively where we're trying to catalog information for the first time that's never been available on the internet. You know, this is why we have to use our own sometimes mapping systems, our own inventory systems. It's not because we don't trust third party systems. It's because third party systems either don't exist or the information is very incomplete. And so for us, that information can be very useful. And in the way I always think about how you build technology is how do you actually grow the pie so that you can get more participants into the ecosystem. And this is a perfect example where we're building lots of this information and we probably will be a very large partner or participant where we will team up with some of these companies building these foundational models to actually build very useful models of information that doesn't exist that it would be impossible to train on because they don't have it.
