Invest Like the Best with Patrick O'Shaughnessy
· Patrick O'Shaughnessy
[00:00:00] S0: Meaning they have to have contact with reality in some sense? A thing a customer ends up seeing or touching.
· Brian Chesky
[00:00:00] S1: Yes. An engineer would call it, they have to be technical. In other words, every engineer needs to code. Even the managers need to code. And whatever the version of that is for your field, you need to code. So if you're a lawyer, you're just not managing people, you have to actually read the case law and you have to get involved. And that makes sense, right? You can't just be these managers where you're kind of people's therapist and you're just doing meetings, you're doing one-on-ones. People who have lots of recurring one-on-ones are not going to survive. Because what they're doing is like that, oh, you come with me with whatever problem is, I'm here to help you, like a mentor or professor. That kind of leadership style is not going to work. You need to have context. And I think a lot of people will survive this age of AI. The two types of people that will not survive the age of AI are two types of people. Pure people managers who think it's all about just leadership. No, it's about content and leadership. I hear about design leaders where the heads of design, they don't actually manage the design. Jony Ive manages the design – he designs, and he leads people. A design leader who only manages the people, that's crazy to me. The way Frank Lloyd managed his design team is through the work. I say you manage people through the work. You don't manage the people, you manage the work. Otherwise, what are you doing? I mean, maybe a couple times a year you should have a check-in, have a heart-to-heart, ask them about their family, like build relationships. You should do that. You should have relationships. That's not a day-to-day thing. You're not their therapist. You're managing people through the work. So two types of people will not make the shift to AI or pure people managers and people that are rigid and don't want to change and evolve. As long as you've got a growth mindset, I think the tools are going to be very easy. There's an economic incentive for the tools to be so easy everyone can figure them out. So I don't think the AI tools are going to be complicated. Right now, there's a lot of command line. I think Clawdbot and Cowork are not the most intuitive to an average person, but I think economic incentives will be for this to become incredibly intuitive. I think AI is really an enterprise thing right now.
· Brian Chesky, Patrick O'Shaughnessy
If you take out ChatGPT and people screwing around like image generation, it's pretty much just an enterprise phenomenon. I'm on the board of Y Combinator. 175 companies last batch, 159 were enterprise. There are no consumer companies. Because there's no consumer companies, the incentive to make the interfaces really simple is not there, because it's their job to figure out the interface. The next wave of AI is going to be consumer AI. Consumer AI is going to be the big prize. Can you think of a consumer AI company? Maybe OpenAI, but most of their energy is now going to Codex. Google, yeah, with Gemini, but most of it's still going to search. They don't want to cannibalize their business. I think that's where it's all going to go. I'm trying to start thinking about AI as a consumer play and how do we make tools so simple that everyone here can figure out how to use them. [00:00:00] S0: It's surprising that so few have tackled consumer. So many of the biggest companies in history are consumer companies. You built one of the great consumer companies of the last generation of companies. What advice would you give to people that do want to build consumer businesses using this technology? If the next batch of YC startups were to invert because of your advice, what would you tell them? How would you guide them? I'm so curious why you think there aren't more. It's surprising.
· Brian Chesky
[00:00:00] S1: Here are the three or four reasons I think it's happening. Number one, I think a lot of people, when ChatGPT came out were afraid. They were afraid ChatGPT was going to kill their business. I think a lot of investors didn't want to invest in something where they thought ChatGPT was going to kill it. Number two, the business model is tricky. There is no consumer business model for AI that I've seen yet. For example, ChatGPT, there's three ways it can monetize. Subscriptions – unfortunately, they're probably going to hit a local maximum percentage of users subscribing because Claude and Gemini are giving it away for free. Ads, again, they're hitting a local maximum because Claude and Gemini are not going to do ads. Then e-commerce, they shut down the third-party apps. And the inference costs are expensive enough that they're burning a lot of money. I think the first thing is you need to have a business model around consumer AI. You can't just be in the business of information, because people are not trained to pay for information. The second problem is distribution needs to mature. Now again, the top three apps in the app store are AI, so it does prove that if you have something revolutionary, you'll find your way to the top. The third thing is, while I think Silicon Valley, we like to describe ourselves as rebels, I think, and maybe it's a little bit like the always-on nature of Twitter, I think it's very trend-based and vibe-based. I think there's a little bit of everyone kind of does what everyone kind of does, if that makes sense. And I think the trend is enterprise. The other thing is with Y Combinator, we tell entrepreneurs, get other YC startups to be your first users. It's a really great distribution strategy. The way we got our first users is you do things that don't scale. Now, what happened was that has been taken to the logical extension that everyone just keeps making enterprise companies. Maybe finally, the reason people aren't doing consumer companies is they're just harder. They're more hits-driven. The prize is bigger, but the risk is higher. It's more all or nothing. You can pick a narrow vertical, build a good medium to large-sized business that's pretty straightforward. You can get other YC companies to adopt it, and you can grow into larger and larger enterprise. It is a much more hits-driven business.
· Brian Chesky, Patrick O'Shaughnessy
You have to be good at a lot more things. You generally have to be better at design, marketing, culture, press. It's not purely technology and sales, which is what I'd say enterprise is. You make a product. It can be a technology-based product. Oftentimes, in enterprise, the person using the product is not the person buying the product, so sales becomes really important, but you can start sales really small with some companies. It's really hard to figure out how do you start small consumer? Who do you start with on the street? I think these are the reasons why, but it might also just be that everyone's doing enterprise because everyone's doing enterprise, and it's a trend. My prediction is that we're living in the age of enterprise AI, and I think in the next 12 to 24 months, you're going to see the beginning of a consumer AI renaissance. Almost every app on my home screen has not changed fundamentally since AI, including Airbnb. I think that's going to change in two years. [00:00:00] S0: On your most recent earnings call, you referenced this thing – Project Hawaii, I think it was called – which is the on-the-ground direct application of the soon-to-be-defined founder AI mode or whatever. Maybe just describe what that is and how you as an incumbent consumer company are trying to make yourself into one of these AI companies, and sort of what's possible with a project like this.
· Brian Chesky
[00:00:00] S1: This is such a great story. I had this conundrum. We had 7,000 people, we were down to 5,000. I wanted to basically take the magic of the founding of the company in the early stages, when we were in this little apartment working together, and I wanted to work with a team of 10 people, and I wanted to have that team work on one problem. So we ended up focusing on one problem, we said, let's put a team together to focus on improving the guest experience, improving conversion rate. Conversion rate is basically search to book, the conversion rate of people typing in the location and dates and booking. There's a funnel, and you can basically measure the funnel, and you can do A-B test. But more than A-B testing, I wanted to do something where the north star was let's make the experience better and increase conversion. We're going to start with a better user experience. So we put together a team, I think it was like 10 or 12 people. Designers, engineers, a couple of product people, data scientists. It was mostly just a pure software team. We treated it like a little startup. We said, we're going to just focus, and we're going to do a system of crawl, walk, run, then fly. Crawl, fix the bugs, fix the problems of conversion rate, just the easy things. Once you build a little bit more confidence, walk. Start to develop features, Really start to reframe the journey. Then run. Rethink the entire flow. Big features. Fly. Completely reinvent yourself. Everything's going to be measured. Everything is going to be operationalized. The team's results were phenomenal. They ended up delivering the equivalent of, I think in year one, $200 or $300 million in revenue for the company. The following year, it was like $400 or $500 million. Now, we're at a run rate of over 600 basis points, 600 basis points of $13, $14 billion. You get the sense of how much of a lever that is. It was just one team. It grew into dozens and dozens of people, maybe 50 or 60 people, but it was this really lean team. We took that team, and we said, what if we apply it to the next problem, pricing? Same people, another team, totally different team, same model, then another model, then another model. I really tried to work with a team. Initially, I would meet with them every week. Then it became every other week and then every month.