Today, I’m talking with Airbnb CEO Brian Chesky. This is Brian’s fourth time on the show, and he’s one of my favorite guests because he’s so clearly obsessed with things like company structure, design, and decision-making. You know, Decoder stuff.
We had Brian on the show last fall to talk about “founder mode,” a buzzy phrase inspired by a talk that Brian gave about his detail-oriented management style. As we were walking out of the studio, Brian told me he had some big news he was incredibly excited about but couldn’t tell me about yet. That news was a redesign of the Airbnb app with a striking new design language; new curated experiences in various cities, some led by celebrities and athletes; and a whole new services feature that lets you book things like private chefs and photographers.
You’ll hear Brian describe all this as a full-scale rethink of Airbnb, everything from how individual properties are stored in the company’s databases to how the actual company is structured, or changed, in order to get to where he wants to be five years from now.
Listen to Decoder, a show hosted by The Verge’s Nilay Patel about big ideas — and other problems. Subscribe here!
That would be a great episode of Decoder all on its own. But if you’ve been listening to the past few episodes, you know that I’m particularly interested in what happens to services like Airbnb, Uber, and DoorDash as new kinds of AI assistants and agents get more popular. Google just announced new agent features in Chrome and in various research prototypes, Microsoft is rapidly pushing on some of the core technologies to make agentic systems happen, and there are lots and lots of demos and test projects out there showing off what the next generation of automation might be able to accomplish.
But all of those things disintermediate service providers — after all, if you can just ask an AI assistant to bring you interesting vacation listings, get you a ride to the airport, or book a private chef, you might never actually open that beautiful new Airbnb app and see all the new things they’re trying to sell you to grow their business. So Brian and I talked about this quite a bit. This will be the next set of high-stakes negotiations in tech and business, and it’s clear he’s been thinking about it a lot.
It also wouldn’t be a Brian Chesky episode if I didn’t take the time to ask him about OpenAI and its CEO, Sam Altman — Brian is close friends with Sam, and he was a part of the drama that saw Sam fired and brought back to the company last year. He also introduced Sam and Jony Ive — an introduction that led to Jony taking over all design responsibility at OpenAI. So I did my best to see if Brian would reveal anything about what they’re all working on. You can tell me how well I did.
There is a lot going on in this one. At one point, Brian explains the difference between a product manager and a program manager by talking about architects and general contractors. It’s pure Decoder bait through and through.
Okay: Airbnb CEO Brian Chesky. Here we go.
This interview has been lightly edited for length and clarity.
Brian Chesky, you’re the cofounder and CEO of Airbnb. Welcome back to your fourth time on Decoder**.**
I love being here.
I’m very excited to talk to you. I’ve always enjoyed talking to you about management and running companies and strategy. You have very different ideas from a lot of the folks we talk to. It’s always interesting. And then there’s news. There’s big news that I want to talk about. You were last on the show back in October, and as we were walking out and you were getting on the elevator, you said, “I’ve got something really big. I can’t wait to come back and talk to you about it.” And that happened. You’ve launched Airbnb Experiences. You’ve launched all kinds of new services on Airbnb. Tell us what’s going on.
The story, just the short version, started 17 years ago when we hosted three guests that first weekend, and that really inspired the creation of this company. Something remarkable happened. These three strangers came into our home and we rented our space to them, but we also hung out with them all weekend. And as we’re waving them goodbye, I remember Joe [Gebbia] and I — we were roommates — were thinking there was a bigger idea here, but the bigger idea was not merely just renting your space. The bigger idea was what happens when strangers come together, and what if you could build this people-to-people marketplace where people could share not only their home but every part of their lives. And years later, once Airbnb took off, people asked me, “Well, what’s next for Airbnb? You’ve already monetized people’s biggest asset, their home. What’s next, their car?”
I started thinking to myself, I don’t believe the biggest asset in people’s lives is their home. It’s their time. There was a book written about Amazon called The Everything Store, but it probably should be parenthetically called “Everything in a Cardboard Box Store.” It’s not actually everything, and in fact, more and more of the economy is moving to services and eventually experiences, and we just thought this was an incredible opportunity for the company because when it comes to travel, more people stay in hotels than homes. One of the top reasons they like hotels is there’s a lot of services and comforts. We thought, “What if we could provide all the services in a hotel and more at a home?” And then we thought people travel to do things, but it’s hard to do really cool, authentic things even though people travel to have local travel experiences.
So we wanted to bring back Airbnb Experiences but in a whole different way. So we’ve announced a few things. Number one is Airbnb Services, to make your stay more special. You can Airbnb a chef to come to your home. You can Airbnb a masseuse, a personal trainer, a photographer to take your photos. Then we relaunched Airbnb Experiences, bringing in some of the most interesting people in the world. You can do these really cool activities with them. And then for the third thing we said, “Well, we want to make it really easy.”
Our app was designed to do one thing, which is book a home. And so we had to completely reimagine our app to not only book a home but book a service and experience. Well, along the way, we basically rebuilt our technology stack, rebuilt an entire app to become a platform that could book almost anything. In the process, we also created a whole new design language. We’ve departed from this flat design that I think was popular about 10 years ago on the internet, to this really cool, robust, dimensional, vibrant interface. It’s kind of the beginning of a whole new company. And I think this is just the beginning of the next chapter for Airbnb.
I have a lot of questions for you about the design aspects of this. We’re going to come to that. It’s interesting you talk about monetizing people’s time and then the actual services in the app. I think you’re starting with chefs. Private chefs are an industry, and it’s not like I’m a pretty good cook so I will come to your house in the way that I might have a room in my apartment and I’ll let you rent that for a day or two.
There’s a little bit of a gap there. Are you expecting most of the services to be provided by professionals who use Airbnb for discovery, or are you expecting it eventually to just be regular people providing whatever things they want to do in their extra time?
I think it really depends. I think with services it’s primarily going to be people who are professionals, and we’re going to give them a platform. We vet everyone. We make sure they have licenses. We make sure they have certifications, but it’s good to remember, a huge part of the American economy is a service economy, and I think a lot of the people in services have fairly unsteady incomes. They have unsteady demand. It’s very much word of mouth. You don’t know who’s great. There’s not a system of trust. I think what we’ve created is a system of trust where we can vet everyone and make sure they’re really, really excellent; our brand stands for quality, and that’s really where we want to go. I think experiences are a little different, because we want really interesting people, but for many of the experiences the hosts have never done this before.
For example, I was just on a photo tour with a photographer who’s got a million followers on Instagram, but he’s not a tour guide. But he takes you around SoHo to look at all the cast-iron architecture and teaches you how to take photos. Now, he’s not a professional tour guide. He’s never done this before. So I think there is this opportunity to take people with a skill and monetize it. I think down the road there could be ways to take this to even more casual people as well.
I’m one of these people. This is a very common story in New York where it was cheaper to buy a house in the Catskills in 2016 than to buy an apartment in New York City. So I bought a house in the Catskills. We ran it as our own Airbnb for a minute and then the pandemic happened and we moved into that house by accident for two years. This is a very cliche story. I apologize to the audience. If you live in this city, you’ve heard this story a million times and everywhere else it sounds insane. But then we moved in and then we left and as we left I thought, “Well, now I have a kid. I’m not going to have this side hustle of running this Airbnb.”
We turned it over to a professional management company, and it just runs the Airbnb for us and it’s great. It takes a cut and it’s fine, and it seems to be going well. But there’s a part here that’s a lot of what Airbnb has become. The actual experience is people’s homes, but they’re managed by professional vendors because they do a good job of it. They’re consistent. They manage the platform on behalf of whoever owns the houses. Are you expecting that layer to emerge in the services category as well?
Hard to say, but I don’t anticipate it. There’s a couple of points there. One of the things and one of the reasons we launched what we launched is, and I could go in really interesting places with this conversation, when people think of Airbnb, most people think of homes and of empty homes, homes you get all to yourself. And that is most of what we do every single day. We have nearly 4 million people a night staying in homes. In the vast majority of the homes, the host isn’t there, and a large percentage of them are using third-party services to help them, not the majority but a bunch of them. I think that in the future, I want Airbnb to be a bit more of a real community where you’re actually connecting with the host, and with services. I don’t think these things get industrialized.
If you want to get a chef to come to your home, you’re still going to get a chef. It’s going to be a real person. If you’re going to go on an experience with somebody, it’s still going to be a real person. I think we, the company, can provide a lot of that platform layer, but I do think that most of this is going to be peer-to-peer, person-to-person. I also think that if I were to zoom out for a second, I think we’re in a really, really interesting time in the world with Silicon Valley and tech. I think I heard the average Gen Zer is spending four hours a day on social media. I think AI is an incredibly exciting tool. Probably the most powerful tool developed in our lifetime or many lifetimes. Maybe the way to think about AI is as an accelerator. It’s an accelerator of the path we’re probably already on, and the path we’re already on is people spending a lot of time on devices, a lot of time living in a digital world, a lot of time consuming content.
I remember more than 15 years ago, 20 years ago, there was a thing called social networking. And it’s funny, that term doesn’t really exist anymore because around 2012, your friends became your followers and social networking became social media, and so then connecting became performing and the relationships became kind of parasocial. I’m not saying this is a bad thing, but what’s clear is that there’s now a void, and there’s a void in people’s lives, which is people living in the real world, making real connections with real people, having real experiences, real memories, and this is where I’d like to take the company. I really want Airbnb to start to feel like more of a social network in the real world. We’ve made these experiences really social. I think it can be the platform to meet one another, to connect, and just to build this entire ecosystem around people, their passions, their skills, their time.
So you’re starting with 10 services. I think chefs are the first and the rest are, as you announced at the event keynote, you basically pointed at hotels. Here’s all the stuff hotels do, and then there’s some other stuff like photographers. How’d you pick those 10?
We basically just did a bunch of surveys with our guests and asked them, “What kind of services do you want to use at your Airbnb?” There were a few around food: chefs, prepared meals, and catering. We noticed that people were booking entire homes. The homes come with really big kitchens. Not everyone wants to cook, and so the kitchen is often not used. So, what if somebody could come make food for you?
Photography was a very, very popular request on Airbnb, because we have a network of thousands of professional photographers. We photograph all these really wonderful homes on Airbnb that look really well when photographed. So, with a network of thousands of professional photographers, we thought, “What if we allowed that network to take photos on your trip?” We noticed one of the most shared types of photos or even videos on Instagram and TikTok were of travel. Travel photos, travel experiences, but people struggle to take really good photos.
One of the problems if you’re traveling with your family is you can’t really take a family photo unless one of you is not in it, unless you give your camera to somebody else. Then we thought about nails, makeup, and hair. Why would we do those? Well, a lot of people travel for special occasions, like weddings or other events, and so a lot of people need these services, and it can be really difficult. Let’s just imagine you live in New York and you’re going to Chicago and you need to get all these services. How would you find them? So this was kind of where we started. I think eventually, who knows, there could be literally hundreds of services that we could offer. The real question is could Airbnb one day go beyond travel? Could you use Airbnb to find services in your own city? And I don’t see why that’s not possible down the road.
We just had Dara Khosrowshahi from Uber on the show. He was talking about a similar shift for that app, where I think of Uber as the button that just brings you a Toyota Camry anywhere in the world, which was very difficult to pull off. It’s a simple thing to say. It’s very hard to do.
Now, Uber’s moving toward wanting you to use the app every day — it wants you to schedule rides and have this ongoing relationship with this platform as opposed to “I need a Toyota Camry, I’m going to push this button.” It sounds like you’re making a similar move, right? You might use Airbnb a couple of times a year now as you travel. You want people to use it every day.
Ultimately, what we really want to do is just be useful in people’s lives and be able to solve problems better than anyone else. If we can do that, we want to do that. Right now people use us to book one thing once or twice a year — a home. It turns out though that we’ve done 90 percent of the work to be able to go into a hundred other businesses. Not to say it’s not a lot of work to build those businesses, but from a platform standpoint, we’ve built this reputation system. We have these really robust profiles. We have 200 million verified identities. We handle more than $90 billion flowing through the platform every year. We’ve got one of the best design application teams in the world to make this product ostensible.
So I paid a lot of attention to Amazon, and Amazon in the late ’90s was a bookseller, as you recall. I’m not sure Jeff Bezos had the ambition at that moment — maybe he did — but he certainly went to these adjacencies, and the adjacencies were CDs and DVDs. Then he went to electronics to play them and then he went to toys and then the rest is history. I think there was this opportunity for us to be much more than a marketplace for vacation rentals and homes, and I think at the highest level what I want us to build is a community. Not just a marketplace, but a global community where you can literally travel anywhere, get anything you need for traveling, live anywhere, get anything you need in the real world, and essentially belong and connect with people anywhere.
So travel, live, and belong. I think that’s where we’re going to go. I think it’s probably a five-year journey to get there. I don’t want to say we’ve done most of the work from a technology application standpoint, but we’ve rebuilt the technology and rebuilt the application from the ground up to make it extensible enough to offer really anything.
You launched riffs on some of these ideas before Experiences had been around. I think you had virtual experiences in the pandemic. You started testing experiences I think in 2014. What gives you the confidence that you’re going to pull it off this time?
It’s one of those things where sometimes if something doesn’t work the first time you ask, “Was it a bad idea or was it just…” There’s this great saying by Marc Andreessen, who was one of our early investors. He said, “There’s no ideas, just ideas that are too early.” And he basically made the comment that almost every idea that filled in the .com is now a popular app. Webvan is now basically Instacart and / or DoorDash. I’ve always believed there was a consumption from physical goods to services to eventually experiences and experiences at the top of the pyramid. I think that with social media, people want to share experiences. Social media influencers want to make extra money. How does a social media person, how does an influencer get paid right now? They build a huge audience and they do essentially paid promotions, or they try to parlay into creating a product.
One of the big things we notice is a lot of these really influential people, they don’t want to just broadcast and monetize attention. We think we can monetize their experience. And so we started seeing that. We thought the timing was right. Post-pandemic, people are looking for things to do. Social media is a great distribution channel. We have a huge audience, we have the capability to pull this off and people love experiences on Airbnb. They just didn’t really know about them.
So we’ve completely reimagined the product from the ground up. I think the big difference this time is we’re not going after traditional tour operators. We’re trying to find some of the most interesting people in culture around the world, like Olympians to do workouts with. It turns out this could be fairly scalable. We can get Michelin chefs to do cooking classes with you. So many people, I think in the future, are going to be offering experiences or going on experiences, and we just zoom out and say, “What are people going to do in the future?”
It’s pretty obvious we’re going to use more devices, and these devices are going to get more powerful and we’re going to be able to live in these digital worlds. That’s obvious. The question is, well, what else are we going to do? I think we’re going to use these devices to live in the physical world, and what jobs will AI not replace? I think that AI is not going to replace all these jobs that are people-to-people oriented, that are rooted in connection, rooted in skill, rooted in having an experience. So I think many times you want to either bet on a trend, or almost bet on the opposite of the trend, which is to say bet on the gap that a trend makes. If you’re betting on AI and the world being digitized, you also want to bet on this gap in the world, this huge void. People are going to need things to do, they’re going to need ways to make money. So I think this could be a whole new economy that could emerge.
You’re talking about curating the experiences that exist. The initialist is pretty fun. There’s a Patrick Mahomes experience, there’s something called the Otaku Hottie experience.
[Laughs] With Megan Thee Stallion.
I’m definitely signing up for that one. That’s a lot of input into the system. The benefit of traditional Airbnb is people put up their houses, you can get reviews, you’ve already built the system for that marketplace. People show up at houses and the house is not a variable. The house can’t have a bad day. Maybe it’s dirty and that’ll tank your reviews, but Megan Thee Stallion can have a bad day. There’s a variability to that experience. How do you defend against that?
Part of that is why we’re doing so much quality vetting. And I think in general, and I’ve talked about this a little bit in other conversations we’ve had, we have this philosophy and I learned this philosophy during Y Combinator. Paul Graham had this philosophy. He said, “Do things that don’t scale.” He said, “It’s better to have a hundred people love you than a million people sort of like you.” The way you grow something is you focus on just getting a hundred people to love you and maybe it means you do things by hand that seem completely unscalable. And then what you do is once you’ve figured it out, we might call this product market fit, then you use technology and the industrialized part of your brain to figure out how to create systems and software to scale it. We decided to do something similar with Experiences. We decided to try to build it out by hand.
We wanted to build out, get some of the biggest icons in the world, get some of the most interesting people in the world to get the network going, to show people what’s possible. And what we want to do is use software and community to scale this. I think it’s going to be a much more curated, hands-on scaling process than the original core business of homes. But it’s almost like the difference between Amazon and eBay, where Amazon did the hard work of building out fulfillment centers, and eBay didn’t, but ultimately, the best experience wins, and I do think with software and technology and community, we can do this.
So what I would imagine going forward is we recruit most of hosts, they come on the platform, it’s very hand curated, and then what we’re building are tools that will be very much assisted by AI, and we’ll get more and more automated to be able to do this, and we’ll get more of the communities reaching out to us to provide more experiences. I mean, for example, like Megan Thee Stallion, Patrick Mahomes, and a lot of other celebrities have reached out to us because of Experiences. But the other thing, and one of the reasons we want to get these celebrities in Airbnb is a lot of people say, “Well, if Megan Thee Stallion could do this, I would want to do this.”
It might be significantly lesser-known people, but it’s something to aspire to. Now, to the point that people can have bad days, I guess that’s what makes it real and authentic. These aren’t cookie-cutter experiences. This is real life, but I think there’s something wonderful about it. I think Airbnb is ultimately not a SKEU, it’s not a standardized product. People are living, they’re breathing, they have good days, they have bad days, but I think it’s really about authenticity, and I think that that connection is what makes it so exciting. That variability is what makes life so rich.
You have a version now, especially as you expand into delivering more and more services of what I’ve started calling the DoorDash problem, where the app is beautiful, now, you’ve invested a lot into the app. I want to talk about the decisions to do that. You want people to use your tool and all of the agentic AI executives who come on this show are like, “You’re just going to have Alexa book you an Airbnb,” and that they’re going to cut you out and this is the dream. You’re just going to say, “I want a sandwich,” and they’re going to go ping the DoorDash API, or they’re literally in some cases going to click around DoorDash’s website on your behalf and DoorDash gets none of the customer relationship.
You have a version of this problem now, right? I’m going to Toronto, get me an Airbnb. Some agent’s going to show up and now you’ve expanded the surface area of the problem. I need a chef. I’m going to go click on the Airbnb website. Have you thought about whether you’re going to work with those agentic AI systems or block them or build your own? Because that seems like the platform change that’s coming that no one has really worked out the business of yet.
I totally agree. First of all, let’s zoom out and ask how we think the future’s going to look. There’s this AI maximalist view that there’s going to be like one or two AI models and one or two applications that rule them all and you use this one app and this one model for everything in the world. If you take that to its logical conclusion, you also start to go to this place where almost one company rules everything, and I think there’s numerous problems with the AI maximalist view that it’s one company to rule them all. One problem with it is, I don’t know if everyone wants one company to have total power and primacy, but the other is just one company is not going to build the entire future. This entire future is going to be built by millions of people in thousands or even millions of companies.
There’s an alternative view, which is to say that AI can democratize the world. It’s almost like when technology stagnates the world consolidates, and when there’s this campaign explosion of technology that could actually create a lot more startups. I think that’s another alternative. I do think that every company is going to have to be an AI company or risk disintermediation. The models that are being developed we have access to as well. I think there’s a couple of things that are going to play out here. Number one, I think Airbnb will in and of itself be an AI application. We’re hiring really great people. I think we have one of the best software design teams in the world. We have great application layer design, and I think we can broaden and broaden our app. That’s partly what we’re trying to do.
The more companies become a platform, the more it’s the reason to go directly to that company. I think service experiences are just the beginning of things we can do on Airbnb. Also, I think Airbnb is a community, so you want to be able to connect with the guests and hosts. Our messaging platform is really important. The sense of trust is really, really critical. So number one, I think Airbnb is going to be like a concierge for your traveling, for your life, and maybe beyond. We’re going to try to be as broad as possible. The second thing is, I think these AI applications, these native AI companies (take OpenAI), are going to have software development kits. I think they’re going to have SDKs and just like Apple created the app store, but Apple didn’t build every app. Very few of the most popular apps are Apple native apps. Why isn’t Apple able to make the most popular apps? Because it’s just so much for one company to do, to make hardware, to make an operating system, and to make apps.
When the iPhone came out, all the apps except YouTube were native. Now all the apps I use, other than iMessage, are not made by Apple, except for maybe the calculator because I don’t really care to download my own calculator. This is probably where the world is going, that there are going to be companies that develop devices, there are going to be companies that develop operating systems, but I don’t know if there are going to be single apps just like with the App Store because every app is going to want to have its own interface. Every app is going to want to have its own kind of culture, and so this is my theory for where it goes, but there is a maximalist view that it’s all consolidated to one or two companies.
That maximalist view is, I think, best expressed by the companies that are promising agents. OpenAI is one of them. I know you have a relationship with OpenAI. I want to talk about the work you might be doing there, but they’ve built some agents and some prototypes of agents. There are other companies that have built even jankier prototypes of agents that at the beginning were just using testing software. They weren’t even using AI. There’s stuff like Model Context Protocol that Anthropic is doing, which sort of creates API layers for agents, right?
All of that basically implies I’m going to talk to my computer and the computer’s going to go do stuff for me. The next version of Siri, which is now delayed, the promise was you would talk to Siri and it would use the apps on your phone for you. I don’t know if that’s maximalist to “one or two companies will control everything,” but it is maximalist to “there’s a platform change coming and natural language will be the interface.”
You’ll mostly communicate with your computer by talking to it, and then it’ll just do stuff. And that’s the thing that disintermediates your interface. It disintermediates your customer relationship, and I don’t know why you would participate in it. I’ve asked this of all the companies that provide services, when you watch the Alexa demo and it’s like, “I got you a sandwich,” it’s like why would any of the delivery services disintermediate their customer relationship in that way.
Well, yeah, there’s a bunch of things here. One is it’s not clear to me that voice is the best way to do everything. It’s not even clear that voice is the best way to do most things. Let’s zoom out for a second. Just like I don’t think a chatbot interface was the best interface for most tasks, hence your iPhone. You don’t want to text the weather, you don’t want to text the calculator. You want a specific interface. I think a lot of the future is going to be more visual. I think the amount of bandwidth you can communicate through verbalizing words is very, very limited compared to seeing something, and hearing is very, very low bandwidth. So it’s great for certain things, but it is very, very limited in being able to do other things. I mean, get me an Airbnb. Well, what does that Airbnb look like?
What does it feel like? It gets very limited very, very quickly. And additionally, you’re right. These companies are going to have to want to participate in the platform, and I don’t think companies just want to be data layers, and so these platforms or these new interfaces are only as good as the companies that participate, and the companies will only participate if they can have a relationship with their own customer. So we’re going to have to figure out this new world. It’s going to be, I think, really, really interesting. I think the future’s going to be multimodal. Voice will be critical to it, but I think it’s going to be much more than voice. There will be some things that will be voice only, but I think there’s going to be things that go well beyond voice, because it’s hard to receive information from an audio standpoint to do most tasks. I mean, you can get only so far with it.
I’ve asked people on both sides of that debate how they think it might be resolved. They all have a similar answer, which is, well, we have to convince everyone to participate, and then the specifics go to, well, maybe we’ll just pay them more money than they would’ve otherwise gotten, right? It’ll be worth your while to be a data layer here. We’ll just pay you a transaction fee on top of what you might otherwise get. And then other people have a version of what you’re saying, which is actually what I want you to do is just open my interface inside of the agent and then I’ll have a customer relationship, and I have no idea how any of this will play out. Have you had these conversations? Have you talked to the various agentic companies and said, “Here’s what I actually want”?
I mean, one of the things I’ve talked to numerous companies, including Sam [Altman], about is there has to be some type of software development kit, an SDK, and it would be great for us to be able to think about this together and figure out is there a win-win? That’s the big question. Is there a win-win? And there probably is. It is so early that no one really knows. This is the very beginning, but ultimately, and this is what I told Sam, the right solution will be whatever’s best for the customer. Whatever’s best for the customer will win because they’ll ultimately vote. And so you’ve got to imagine what’s going to create the best experience. But I think my instinct is you will have very few devices, you’ll have very few operating systems, but you’ll have more apps. And I think that’s kind of the way computing has always been. That’s my instinct of where it goes.
I think increasingly more of these AI companies are going to have to choose to be either the language layer, the foundational layer, and that’s where a lot of them will go, or they can vertically integrate. But if a company vertically integrates like Apple, you can’t vertically integrate and be wide because there’s too many things to do. Imagine Apple trying to build the device, the operating system, the Airbnb app, and handle customer service and do this and do that and build the community and handle all the money and deal with trust and safety. So there’s just a lot of jobs to be done in society and every company has to bring its core skill set.
One of the things I think we’re great at is interface and interface design and the connection of the online world with the offline world. And so, ultimately, the best product will win, the best solution will win. Part of what we’re trying to do is broaden our offering as much as possible, mostly for the customer and mostly not for strategic considerations because you have to align your interests with what the customer wants, but this is exactly where I think we could go.
The other thing I think about when you think about the service providers in the context of the agentic AI is I’ve seen a lot of demos where someone points a phone at a dishwasher and says, “My dishwasher is broken. Get me somebody to fix it.” And then the data provider is like a Thumbtack or an Angie’s List and it says, “I booked someone for you.”
Now, I’ve booked repair people on these services, and the problem is the individual repair people use that for discovery, but they don’t use it to actually run their back office. They’re not actually scheduling there. They would prefer you not to transact with them there because they have to pay fees. There’s a whole other side of it where you can tell the database that something happened, but the actual human being might not actually ever show up. And you’ve got to close that gap across all of the verticals that you’re now in. And one of the ways you close that gap is to just take it over and say, “We’re going to run your back office too.” Are you all the way there?
We’re going to get pretty vertically integrated so that we’re building the tools for these service providers. We’re building the tools for these hosts, and I think this maybe goes to a broader point, which is that most customers when they look at Airbnb see an app with five tabs, and they see an interface. It’s kind of similar to Amazon. When you see Amazon as a customer, all you see is the website, and then you see the cardboard box showing up. It turns out most of what we call Amazon — at least Amazon retail, not AWS — is not the website. It’s the fulfillment center. It’s everything that’s powering the website and fulfilling everything. I think the truth is that’s what’s going to be Airbnb.
You can almost think of Airbnb as three things. It’s the app that customers see, the guest app. There is this whole app the hosts use, which is probably even more robust than the guest app because that’s an app people use every day. And then there’s almost this third Airbnb, which is the biggest of all, which is the system that powers everything that makes all this possible. How do you make sure that when somebody wants to get a haircut, you have the tools to make sure that somebody can manage their business on Airbnb? But the bigger challenge is not even that. It’s how many people in New York City need to get a haircut every night, and what kind of price point do they want, and who’s vetting them and how do we make sure they show up? What happens when they don’t show up and what happens when they’re late?
There’s a thousand contingencies, and the question is how do you design a system elegantly [enough] to be able to solve all these different problems? And so there’s just going to be so much to do. And I think that’s what makes it so interesting, and one of the reasons it’s hard to fully disintermediate something like this is it’s the real world. If you think “what will AI automate?” It’s going to automate a lot of digital content. I think robotics and autonomy are going to automate a lot of repetitive tasks. I think the service and experience economy — I mean, who knows in 10 years, 20 years what isn’t automated and what isn’t done by humanoids — but certainly in the next 10 years, I think that’s a lot of where the human-centric economy goes, where people are doing physical things in the real world.
I want to ask about the decision to do this. You and I have talked about decision-making a lot in the past. We talked about founder mode, which was a great conversation. You did a big story with Steven Levy at Wired, which is great. I recommend people go read it, and it basically sounds like you decided to do this, right? You took a lot of notes, you wandered around your house, you decided you’re going to do this. You had a meeting, and you said, “We’re doing this.” That’s a big decision. Did your team push back on you? Did you just roll over them? How did that work?
No. When I say I decided to do it, I guess the better way to say it is I decided one weekend to write a vision of this that then became a multi-month conversation with the team. And so it actually happened in the wake of the OpenAI situation.
Yeah, I was looking at the timing and I was thinking, “This is all happening at the same time.”
So the OpenAI thing, like Sam was fired from OpenAI on a Friday before Thanksgiving. I was pretty involved in that situation, more as just a helpful friend. From Friday to Tuesday, my parents and my sister and her husband were in town, and then they eventually left my house for Thanksgiving weekend to go to my brother-in-law’s family’s house. I had this weekend by myself with all this pent-up energy, and that’s when I basically just poured all these ideas down.
Now, these were things I was thinking about for a long time. It was basically, what if you could Airbnb the world? What if you could have Airbnb for everything? And I basically started saying, “Well, what would everything be?” And I wrote down a list of things, services, experiences. It was really three ideas. Idea number one was Airbnb is going to become a platform where you could go from short-term rentals — vacation rentals — to kind of everything you’d need to travel and live, kind of like Amazon went from books to everything.
The second idea was Airbnb, to the point of AI, was going to become an agentic app. It was going to become the ultimate concierge for traveling and living, and we’d become the ultimate agent. By the way, if you think the future of AI is agents, what are the most common agents in the world? Travel agents, customer service agents. That’s what we do. So we know a lot about that. The third was, and maybe most importantly, we were going to go from a marketplace to a community and put people at the center. So I wrote this out. It was like thousands of thousands of words. I tried to distill it, distill it, distill it finally to these three basic ideas. I shared it with an executive team, I think on a Monday morning, and I think the team was both enthusiastic and had a lot of questions.
Basically when I communicate, and now this goes to organizational stuff, I try to communicate in concentric circles. Some founders and CEOs just do things and just tell a few people and no one knows. That’s probably the worst thing because you’re not bringing people along. Some people have an idea and they email the entire company. I think that also is problematic because you don’t want to tell somebody and their manager at the same time. Because then people go to their manager and their manager’s not bought in. They’re like, “I don’t know what we’re doing. I’m not sure.” And people aren’t really bought into it, and then everything’s half-baked.
So what I did is I brought in my executive team, which was maybe 10 or 12 people. We beaded the idea up, I refined it, refined it. Then I kind of went to the next consensus circle of 20, 30 more people, and I just kept widening the aperture. There weren’t really a lot of edits from the original vision. It was very clear that this was inevitably where Airbnb was going to go. It was going to be a community where you could travel and live anywhere. AI was going to be the center. People’s profiles were going to be at the center.
Then we just started working on it, and we were actually transforming the company before everyone’s eyes. We basically rewrote the technology stack, rewrote the app, and it was great because we had to do it anyway to update our core business. So, we basically rebuilt the entire app. It worked out insofar as it actually advanced our core business, made our core business better, made our core business stronger, but we were able to turn all of our components into primitives that were extensible. So now it wasn’t a page for a home. It was a page for anything, if that makes sense. But we put on a new technology stack, and then on the page that was an anything page, the homes performed better, because we built it in a much better way.
That’s super interesting. So you’re abstracting the core of the platform and now you can sell basically anything?
You can sell and do almost anything. And so this gets to the point, which is to say … it’s an oversimplification to say there’s going to be these broad AI companies and there’s all these companies that are narrow verticals. Well, we’re going to be an AI company too, because it would be like saying we’re an electricity company or we’re an internet company. I think there’s AI-native companies, companies that were founded on the premise of AI, but even that’s not novel anymore. I mean, it’s basically every single startup in Y Combinator, and I’m on the board of YC, I see a lot of companies. Maybe 500 or a thousand companies come through YC every year now, and every one of them is an AI company. Just like every company 10 years ago was a mobile app, but companies weren’t native.
[Content truncated due to length...]
From The Verge via this RSS feed
iPhone repair kit (2022)
Russian Governor Igor Ivanovich Kobzev shares an image of the strike’s aftermath in the Irkutsk region. | Image: Telegram
The Pixel 10 phones are likely to look an awful lot like last year’s Pixel 9 models.
20 iPhones: great for filming horror movies and protecting your modesty.
Laurene Powell Jobs (left) said that Ive’s design process is a “wondrous thing to behold.” | Image: Getty Images
The student-built Purdubik’s Cube set a new Guinness World Record by solving a Rubik’s Cube in just 0.103 seconds. | Photo: Matthew Patrohay / Purdue University
A screenshot from Marvel Cosmic Invasion.