Invest Like the Best with Patrick O'Shaughnessy
00:18:20 · Patrick O'Shaughnessy, Dara Khosrowshahi
[00:18:20] S0: What's the most amazing thing you've seen someone do with the technology at Uber? [00:18:25] S2: Most amazing thing that I've seen is I'm seeing the impact show up in all kinds of random corners of the company. What we're seeing now is we've got devs in India who all of a sudden are driving 10x the code commits that they used to. They're using autonomous agents all over the place. So the uptake of the tools that we're seeing is not at all predictable in terms of who is racing ahead with this new technology versus what we typically saw. And so I think you're going to have very, very unpredictable patterns within the company. It's up to us to find it, and it's up to us to promote those individuals who are the rebels inside the company. We want the rebels to win here. [00:19:07] S0: One of the interesting questions today is what is the actual latent demand for intelligence? It's such a weird question. I'm curious if you've had an experience like me where at first you tell your team, burn tokens. Let's use as much intelligence as possible. Then there's a moment where you're like, Wow, this is expensive. Intelligence is Totally. How do you think about the trade off between, it seems like an unbridled good to have more intelligence, but right now it's quite costly?
00:19:30 · Dara Khosrowshahi
[00:19:30] S2: We're dealing with it now. Actually, it blew through our AI budget in a quarter for the whole year, essentially. And it is forcing us to adjust. We are going to meet our headcount increases because to the extent that my engineers are getting much more efficient, their throughput is increasing, they're becoming superhuman in terms of their output. There's a cost to that, and it's a significant cost. And at the same time, we're metering the increase in headcount. So my view right now is drive adoption, encourage usage across the company. We do believe in efficiency here. We've got over $10,000,000,000 in free cash flow, which is great, but it's on well over 10,000,000,000 trips a year. So we are not a high margin business. We want to really drive efficiency in order to lower prices for riders, to get more earnings to our earners, etcetera. So I do think it's a combination for us right now of encouraging adoption, but then driving efficiency. The way that I put it is, we're using the more expensive models to explore. Hey, let's try a new interaction here. And obviously, these frontier models, whether it's an OpenAI model or a cloud model, they really are terrific. And they're great to experiment against. Once we scale some of these experiences and interactions, we'll look to bring in more efficient models that are more efficient on a token basis or open source, again, lower costs. So exploration is go, go, go, but then efficiency and scaling is something that we're also talking about.
00:21:03 · Patrick O'Shaughnessy
[00:21:03] S0: So you're one of the few companies, many of them amazing companies with amazing leaders, that are in this AI meets the physical world where you've got a scaled brand and distribution, you're an aggregator of demand in many ways. And I have so many questions about this, but I guess the first one is just what that is like, how you think about that challenge, that problem, that opportunity as one of the small handful of companies that has the right to win in how that new technology links up to the physical world.
00:21:31 · Dara Khosrowshahi, Patrick O'Shaughnessy
[00:21:31] S2: I think the first is there's an enormous opportunity in just it's boring, but it's wonderful boring, which is just building larger models. So building our models in terms of our feed models or search models now are probably 10,000 times bigger than our older models. They're just able to take much more data about you and make predictions about what are you going to search for. Now we have universal search, for example. If you search for something, it's not just classified as rides or eats or grocery. We can show you everything that we have available across our services. So first, it's just like the tooling has gone better, the capabilities have gone better. And today, we can guess when you take an Uber ride, three quarters of the time, we can guess where you're going and it's just a one tap interaction. That can only improve based on these larger models. And as it relates to the real world, we can anticipate what's happening in the real world much more effectively now. And then the form factor with which you operate is going to change as well. And more and more, and it's going to take a while, you'll have drones getting your food to your home or you'll have AVs delivering you from place A to place B. So everything is changing about it. Some of it is immediate opportunity now, but we think there's a ton of opportunity going forward. [00:22:47] S0: What do you think has to be true for Uber to win the demand aggregator top
00:22:47 · Patrick O'Shaughnessy, Dara Khosrowshahi
[00:22:47] S0: What do you think has to be true for Uber to win the demand aggregator top [00:22:52] S2: slot in AV specifically? Supply. So we are a supply led company. The more drivers we have in the marketplace, the more restaurants that we're wiring up for you, a more latent demand that we have. And it's one of the lessons that I learned when it came from Expedia to Uber. At Expedia, we were really focused on demand. How many consumers can we get to the website, etcetera? And we were demand first company, and then we would build out hotel inventory and flight inventory in response to demand. At Uber, everything is upside down. One of our very, very significant growth opportunities right now is building at our surface, not just in big cities, but in the sparse markets in all of the suburbs of the world and smaller cities of the world, not the top 10 cities in a country, but the next 50 cities, the next 200 cities in the country. The first thing we do is actually go out and recruit drivers, recruit merchants and restaurants, groceries, couriers, and then just the demand shows up. I don't mean it's not that easy, but the biggest opportunity and challenge we have is making sure that we secure every single supply, every single car out there that's going to take you from point A to B, every single restaurant, every single retailer, every single grocer. If we do that, the demand will take care
00:24:09 · Patrick O'Shaughnessy
[00:24:09] S0: of itself. In AV specifically, what does that feel like? It feels obviously like we're just at the very early, early part of the S curve. I'm sure that if the answer is supply, you want to just be the demand aggregator with the most supply. So what does that feel like? What does the state of AVs feel like to you?