{"id":938,"date":"2026-05-04T14:30:00","date_gmt":"2026-05-04T14:30:00","guid":{"rendered":"http:\/\/teknologi.news.eraenterprise.id\/?p=938"},"modified":"2026-05-04T14:30:00","modified_gmt":"2026-05-04T14:30:00","slug":"dara-khosrowshahi-on-replacing-uber-drivers-and-himself-with-ai","status":"publish","type":"post","link":"https:\/\/teknologi.news.eraenterprise.id\/?p=938","title":{"rendered":"Dara Khosrowshahi on replacing Uber drivers \u2014 and himself \u2014 with AI"},"content":{"rendered":"
\n

\"A

\n\t\t<\/figcaption><\/p><\/figure>\n

Today, I\u2019m talking with Uber CEO Dara Khosrowshahi. It\u2019s become something of an annual tradition to have Dara join us in the studio when he comes to New York for Uber\u2019s big GO-GET event every year, and it\u2019s always a lot of fun.<\/p>\n

The big news this year is that Dara is really starting to think about Uber as a much larger platform for travel \u2014 starting with the ability to book hotels in the Uber app, thanks to a partnership with Expedia. There\u2019s also new services, like being able to have coffee and snacks in your Uber when it arrives, and even personal shopping. Uber is going so far as to call this an everything app, so I wanted to see how far Dara thinks everything actually goes \u2014 and whether he\u2019s feeling pressure to own more of the user experience in a world where AI companies keep promising that their chatbots will book all the cars for you.<\/p>\n

\"\"\n

Verge<\/em> subscribers, don\u2019t forget you get exclusive access to ad-free\u00a0Decoder<\/em>\u00a0wherever you get your podcasts. Head here<\/a>. Not a subscriber? You can sign up here<\/a>. <\/p>\n<\/div>\n

I also wanted to know if those chatbots have created any opportunities for Uber. Last year Dara told me he was wide open to partnerships just to see if they were meaningful, but all the AI Uber integrations I\u2019ve seen so far have been pretty clunky, and far slower than just using the app myself. So we dug into what Dara is seeing there \u2014 and if there\u2019s any potential in the future.\u00a0<\/p>\n

I\u2019ve also been dying to talk to software CEOs about what AI is doing inside their companies, as AI coding tools and agentic systems upend software development. Just a couple of weeks ago, Uber\u2019s CTO said the company had already burned through its entire token budget for the year by the start of April, and Dara told me he was rethinking how fast the company would hire people as it spent more money on tokens. That\u2019s a big bet, and I wanted to know if Dara was rethinking how his software teams were structured as AI starts to muddle the relationship between product managers, designers, and engineers.<\/p>\n

Then we talked about Uber\u2019s increasingly large investments in autonomous cars \u2014 especially its big investment in Rivian, what kinds of milestones Dara is looking for as the technology evolves, and what happens to all of its drivers in a future where robots do all the work.<\/p>\n

And lastly, I asked Dara when he thinks AI will be ready to replace the CEO \u2014 it turns out there\u2019s already a rogue AI Dara inside Uber.\u00a0<\/p>\n

There\u2019s a lot in this one \u2014 Dara was as clear and candid as ever, and I think you\u2019ll like it.<\/p>\n

Okay: Uber CEO Dara Khosrowshahi. Here we go.<\/p>\n

This interview has been lightly edited for length and clarity.<\/em><\/p>\n

Dara Khosrowshahi, you are the CEO of Uber. Welcome back to <\/strong>Decoder<\/em><\/strong>.\u00a0<\/strong><\/p>\n

Thank you very much. It\u2019s good to be back.<\/p>\n

I\u2019m happy to have you; it\u2019s like a yearly tradition. You guys do your GO-GET event, you have a bunch of news, and then you come down to where we are.<\/strong><\/p>\n

Chock full of news for you. Yeah. Chock full of news.<\/p>\n

And we hang out together in person, which is my very favorite thing. So thank you for doing it. I want to talk about the news that you can now book hotels and other experiences in the Uber app, which is a big deal. But I always ask everybody the same two <\/strong>Decoder <\/em><\/strong>questions about how companies are structured in decision making, and I just want to do them as a little lightning round at the top.<\/strong><\/p>\n

So, <\/strong>last year on <\/strong>Decoder<\/em><\/strong><\/a>, I said, \u201cHow do you make decisions?\u201d And you gave me the Amazon answer. You said, \u201cOne-way doors and two-way doors.\u201d<\/strong><\/p>\n

Mmm-hmm.<\/p>\n

A lot of pressure on decision making. Lately, you\u2019re making big decisions. Even expanding the app is a big decision. Has your fundamental framework changed?<\/strong><\/p>\n

The fundamental framework has not changed. Now, I will tell you that I am pushing the company in something that we talked about, taking smart risks. The pattern that I keep seeing is that as companies get larger, they become more hesitant in terms of risk taking. It\u2019s more about playing safe. You\u2019re a public company, you have to hit your quarterly numbers, et cetera. And to some extent, as companies get larger, they get more resilient. They can actually make bigger mistakes. For us, we\u2019ve got almost $10 billion in cash flow. When I first joined, if we made a billion dollar mistake, it would be a disaster. It would put the company on its knees. And I\u2019m not saying that I want to make a billion dollar mistake, but the risks that we have to take in order to get the right return, in order to keep innovating in the world \u2013 for example, autonomous vehicles (AV), which I\u2019m sure we\u2019ll talk about \u2013 are getting bigger.<\/p>\n

We have to be willing to take those risks. And the patterning that I\u2019ve seen with a lot of companies is that as they get bigger, they get more conservative, and the way they operate gets more set in stone. You have more management layers, et cetera. We very much want to avoid that. And it\u2019s taking me really pushing that \u201cone-way door, two-way doors\u201d as one framework of looking at decisions, but then smart risk taking as well. We\u2019ve got to keep taking smart risks as a company. It means once in a while taking risks that in hindsight look dumb. But we\u2019ve gotta push the envelope, especially during this time when there\u2019s so much innovation going on.<\/p>\n

Risks, everyone wants to talk about it, but taking the blame for when things fail is like the other part of risk.<\/strong><\/p>\n

Yeah.<\/p>\n

It\u2019s the other side of the coin. Also empowering people to take the risk without that fear of failure is really important. How do you think about the stakes? How big of a risk is an individual software engineer at Uber allowed to take?<\/strong><\/p>\n

[laugh] I think if you can\u2019t identify the downside, don\u2019t take the risk. But if you can identify the downside, whether it\u2019s time that you\u2019re spending on a feature, compute that you are dedicating to a feature, or you\u2019ve gotta invest a certain amount of capital in building something, or going after expanding a new line of business in a country \u2013 we\u2019re launching Uber Eats in seven countries in Europe as well \u2013 then you can make the right calculus in terms of whether you should take the risk or not.\u00a0<\/p>\n

We want to learn from our mistakes. Some people talk about celebrating mistakes. I\u2019m not going to celebrate a mistake. But I do want to be able to make sure that I learn from a mistake so that the next decision I make can be incrementally better. That\u2019s usually the construct that we use. I think sometimes we overexamine our mistakes. We have meetings, we talk about it. We document the issues, what we did wrong, what could have gone better. I\u2019m honestly not a big fan of that. It\u2019s a big engineering thing, et cetera: \u201cUnderstand why you made a mistake, what you could have done better, and then move on with life. Let\u2019s go build the next thing.\u201d<\/p>\n

Put this into practice for me. What\u2019s a risk that came outside of your sphere of management control that worked out, and what\u2019s one that didn\u2019t?<\/strong><\/p>\n

One that absolutely worked out, that I was involved with \u2013 but it was the team that really pushed forward \u2013 was women riders and drivers preferred. There was some question as to the liquidity in a marketplace. One of the big things about Uber is that you push a button, you get a car in four to five minutes. There was a question as to whether or not we would have enough women drivers to introduce this feature for women riders. Because if you introduce a feature, it\u2019s not like \u201cwomen riders, women drivers preferred, and maybe you\u2019ll get one if you\u2019re lucky.\u201d That\u2019s not a good feature. So there was a real question as to the reliability of the marketplace to the extent that the vast majority of our drivers are men, in the US for example.<\/p>\n

But because of our size and scale, we have been able to build liquidity in terms of women drivers. And now that women drivers can request women riders, we\u2019re looking to increase the number of women drivers as well. So you get this great flywheel. That\u2019s a risk that worked.<\/p>\n

We have built a taxi product twice. We tried it early on and we tried to build taxis the same way that we built peer-to-peer rideshare, which is kind of a one-on-one hail. And it failed, didn\u2019t work. Taxis didn\u2019t trust us. They didn\u2019t sign up. About six years later, Sachin Kansal, who\u2019s now our CPO \u2013 he used to build a taxi app \u2013 said, \u201cLet\u2019s try this again.\u201d And so, while it failed the first time, this time, we approached it differently.<\/p>\n

For example, with taxis, because we don\u2019t have the data inside of the taxi as to whether or not they have a rider in the car, what we did was a little bit different. We introduced blast dispatch. When you ask for an Uber and we want to hail a taxi, we will dispatch to 10 different taxis. And whoever says yes first accepts that ride. So we\u2019re able to get higher reliability and adjust the way that we\u2019ve built the product for taxis. Taxis are now one of our fastest growing products. That\u2019s also an example of making a mistake once, but then actually sometimes you have to try things again, even though it didn\u2019t work for the first time. A different flavor, a different approach. I\u2019m really glad that we took that shot on taxis.<\/p>\n

We\u2019re going to come back to risk, because you have a bunch of new products that seem risky.<\/strong><\/p>\n

I want to ask you the other <\/strong>Decoder <\/em><\/strong>question about structure. Last time you were here, I felt like I could have talked to you about the structure of Uber for the entire conversation. You had a wild answer that was very lengthy. I encourage people to go back to listen to that part of the conversation. But the short version is, you said, \u201cWe have a combination matrix and line of business structure.\u201d You have global leads for mobility and delivery and everything else is matrixed. And importantly, the thing that you had changed was you had made product a central function. You didn\u2019t have separate product teams for each.<\/strong><\/p>\n

And the ride business, obviously \u2013 I\u2019m guessing something has changed here because you have many new lines of business. You have an autonomy division. Quickly describe how Uber\u2019s structure has changed.<\/strong><\/p>\n

The only change in structure \u2013 because I do value stability \u2013 is that I now have a president and COO, Andrew Macdonald. Andrew ran our mobility global business. What we observed is that the platform that is mobility and delivery is coming together, and particularly users who use both mobility and delivery have been growing much, much faster than the individual use cases of mobility and delivery. And it was always my hypothesis \u2013 one of the visions that I had, coming to Uber, was that once we had the delivery business post-COVID grow so quickly and show that it has the potential of being just as big as a mobility business, we compete against mobility players and we compete against delivery pure play players. You could have a hypothesis, which is actually being a pure play could be an advantage.<\/p>\n

It\u2019s all Lyft. The only thing Lyft cares about, at least historically, was US rideshare, they\u2019re starting to expand internationally as well. Good for them. About time, you could argue. And the only thing DoorDash cares about, let\u2019s say, is food delivery. We\u2019re trying to do both. And it\u2019s hard as a company to do multiple things at once, to have skill sets in multiple business lines. To make up for that, we had a mobility team, a delivery team, and a bunch of common structures and services platforms. Where it came together was the technology platform. We started really pushing this idea of consumer-side platform, driver-side platform. To the extent we could get consumers to use both Rides and Eats, we had a hypothesis that we would retain them for longer. It turns out not only is the retention better, but they spend much more.<\/p>\n

Multi-platform consumers spent three times as much as single line consumers as well. We launched the Uber One membership, now almost to 50 million members, growing really, really quickly. They spend three times more, and they tend to be multi-platform versus single platform as well. And that we thought could be our secret sauce, which could differentiate us from the model line players and allow us to acquire more customers, bring them into the platform, get them to use more stuff, have better retention, et cetera. That sounded great, but the P&L often got in the way. It\u2019s every pixel \u2013 it sounds easy, let\u2019s use our mobility, let\u2019s cross-promote delivery as well, sounds easy, but that delivery pixel on the mobility app could be taking away from your mobility experience as well. And also could be costing mobility. It\u2019s P&L, I\u2019m sending a customer over to do something else. So, sometimes a P&L got in the way.<\/p>\n

I do a lot of stuff. I was pushing platform on the side here, in addition to everything else I do. I really wanted one member of our management team \u2013 and Andrew Macdonald\u2019s been here, he\u2019s one of the longest tenured employees, and one of the most capable team members that we have. I said, \u201cAndrew, it\u2019s time for you to move from running global mobility to actually becoming president and COO of the company, and thinking about the platform as a whole.\u201d It\u2019s been a big success, and it frees me up to work more directly with the product and tech teams. So, it\u2019s kind of a double benefit for me. But the platform is really starting to sing. The number of consumers using both Rides and Eats has increased six times in the past five years, and it\u2019s growing 50% faster than our general audience. It\u2019s definitely, definitely working, and I want to lean into it.<\/p>\n

It strikes me just as you\u2019re talking here that you\u2019re describing everything in terms of trade-offs. Even risk, you\u2019re describing in terms of trade-off.<\/strong><\/p>\n

Everything\u2019s a trade-off of life. Yeah.<\/p>\n

\u201cWe might use this compute instead of doing this other thing and putting a pixel on this screen might take a customer away from this line of business.\u201d So you\u2019ve installed the COO just to manage that trade-off more holistically?<\/strong><\/p>\n

Yeah. He negotiates the trade-offs on the ground, he\u2019s ultimately responsible for one number, if you want to call that, whether it\u2019s a customer happiness or that it\u2019s a P&L, and obviously, often you have to manage for all of the above.<\/p>\n

My joke on this show constantly is, \u201cIf you told me your org chart, I can tell you 80% of your problems.\u201d All companies are kind of the same and I can get to about 80% of the tension if you just tell me where all the executives are lined up and who controls what budget.<\/strong><\/p>\n

Kevin Scott at Microsoft as the CTO once was the person in charge of distributing the GPUs. And I was like, \u201cThat\u2019s all I need to know.\u201d I know almost everything about Microsoft at this moment in time. Now, it seems much more complicated for a variety of reasons, but at that moment I could just tell. It sounds like \u2013 and obviously the secret is in the last 20 percent \u2013 but it sounds like you\u2019ve installed an executive just to oversee the 20% of the prioritization and the trade-offs here.<\/strong><\/p>\n

It\u2019s the 20% of the prioritization of the trade-offs, but you could argue it\u2019s our most important 20%. It\u2019s a 20% that no one else has. And in one year, the 20% doesn\u2019t really matter, but when you compound it over five years, over 10 years, you get the results that we\u2019ve got, which is generally that we\u2019ve grown faster than our competitors, and we\u2019re able to be more profitable than our competitors. That\u2019s the power of the platform, and I really wanted to lean in. At some point it was getting up to a scale where it wasn\u2019t a part-time job, I needed someone really focused on the whole thing.<\/p>\n

So, the news here in that context feels like, \u201cWe\u2019re going to bet on the platform more.\u201d<\/strong><\/p>\n

We have bet on platform for the past five years. It\u2019s a vision that we\u2019ve always had, it\u2019s working, and when something works, you want to double down.<\/p>\n

I\u2019m going to be very reductive here though. The last time you were here, I described Uber as a magic button that made a Toyota Highlander appear in my life. Wherever I\u2019m in the world, almost statistically, a Toyota Highlander is going to arrive. That\u2019s great, and then it\u2019s going to move me around. And the jump from there to \u201cthe Toyota Highlander has food in it\u201d is reasonably small. \u201cWe\u2019re moving things around, we\u2019re a logistics business.\u201d The news here is you\u2019re also doing hotel booking in partnership with Expedia, you\u2019ve got shopping assistance, and now cars might have coffee in them.<\/strong><\/p>\n

We got a lot going on.<\/p>\n

This is far beyond logistics for a platform that was pretty much organized around logistics. Tell me about that in the context of risk and trade-offs and platform bet.<\/strong><\/p>\n

First I would say, these are different kinds of bets that we\u2019re making. Not all of them are going to succeed, and if they do, we\u2019re being too conservative. I expect some of this stuff not to work, but hopefully most of it will. One that I\u2019m quite confident that\u2019s going to work is actually travel and hotel bookings. In that Uber already is very highly used by the global traveler. We operate in more than 70 countries. Often what\u2019s the first thing that you do when you arrive in an airport in a city other than your home city? You open the Uber app. Part of what we announced is usually that the Uber app is the same app regardless of the context that you have. And if you think about it, when you open Uber at home, and we know you\u2019re in your home city, that should be a different experience than if you\u2019ve just landed in Paris and you open Uber.<\/p>\n

So, for example, we have what\u2019s called travel mode. You open up the app, and we first give you step-by-step instructions as to how to get to an Uber, and how long is the walk going to take? How long is the pickup? What are typical rides? We make it context aware, so to speak. And we give you highlights on what\u2019s going on in Paris. Lots of good stuff. Now, the sheer numbers that we\u2019ve got, which is that we have over 100 million riders taking rides to and from airports every single year. 100 million, that\u2019s a huge audience. We do 1.5 billion trips a year outside of your home city. We have the perfect audience and Uber\u2019s built for travel in terms of our being present all over the place. It\u2019s a perfect audience to start to build out the travel offerings.<\/p>\n

We started experimenting, actually, with trains in the UK, and it\u2019s worked out really well, it drives frequency, which is pretty cool. And now we announced a deal with Expedia, where now we offer hotel bookings through Uber. It\u2019s smooth, we have all your information, we\u2019ve got your context. And what\u2019s really cool is for Uber One members, they get 10% off every single Uber, every single hotel booking you get credits back to use, and then you get 20% off a rolling list of 10,000 hotels. We\u2019re making it really worth your while to book hotels on Uber.<\/p>\n

Tell me about the insight that led to that risk. Because I think about Uber and I\u2019m either, \u201cI just need to get somewhere so I\u2019m going to open the app,\u201d and the time sense of Uber is right now, I need something right now, or \u201cI\u2019m going to the airport tomorrow and I live in a reasonably remote area and I need to make sure the car\u2019s going to arrive.\u201d<\/strong><\/p>\n

Tomorrow is about as far out as I go. I never land at an airport and think, \u201cI need a hotel.\u201d Something bad has happened if that is the occurrence. The time horizon of needing hotels feels much longer than anything Uber has previously offered, at least in my experience.<\/strong><\/p>\n

So, that\u2019s a bet. You have to get people to think about Uber months or weeks before they need it.<\/strong><\/p>\n

What\u2019s the insight that said, \u201cWe can get people to do that\u201d?<\/strong><\/p>\n

It\u2019s a bet, and you just described an adjustment to your behavior, which is that Uber has always been about on-demand. One of the questions that we had is, \u201cCan we move from on-demand transportation to transportation by appointment, for example?\u201d The first step that we took was actually Uber Reserve probably three, four years ago. And if you remember, we used to have an old Reserve product, where you would reserve an Uber, but we would be hacking it in the backend. You wouldn\u2019t actually reserve an Uber. We would then call the Uber on-demand when we thought that it could get to you by that reservation time. It was okay, but it didn\u2019t get you the reliability that you needed, it wasn\u2019t a guaranteed reservation, so to speak.<\/p>\n

We took the signal, which is that some people were trying the product, but it wasn\u2019t that good, to be honest. We said, \u201cListen, what if we really up the reliability game?\u201d And we sent the dispatch to drivers in advance, we did some research. Drivers are like, \u201cI like knowing what my next day is going to be like.\u201d So, it was good for drivers. We were able to charge a premium, give it to the driver, essentially to up reliability, and we started building the habit of this as an on-demand service to \u201cActually, this is more than an on-demand service, and I\u2019m going to think about scheduling things in my life often having to do with travel.\u201d Now, what we\u2019re finding is actually some people are hacking Reserve, if you want to call it that, for reliability.<\/p>\n

If you\u2019re in Westchester County, in Armonk, and the liquidity for Uber is lower, you may not want to use on-demand for your commute, but you can use Reserve for your commute as well. What started as, \u201cLet\u2019s try this for travel,\u201d is now being used to hack reliability to some extent. That insight of Reserve building \u2013 and we\u2019ve been at it for four to five years, reliability is not perfect <\/em>perfect, but it\u2019s 99% now, and we\u2019re always working that trade-off between reliability and price, because we want the price premium to be as low as possible, but you can\u2019t lose too much reliability. That insight led us to believe that you actually can move from on-demand to scheduled, and the offerings, like the Uber One discounts, we think will hopefully, over a period of time, change behavior.<\/p>\n

So you\u2019ll actually come to Uber to reserve your booking in advance. We don\u2019t think this is going to be a last-minute thing. If you get to a city and you don\u2019t have a hotel, there is something wrong, maybe it\u2019ll be there on a cancellation basis, but we are trying to drive reservation behavior and we\u2019ve demonstrated previously then we can.<\/p>\n

Yeah. I feel like hardcore travelers who know to reserve an Uber, who are some of your best customers, they like price shopping hotels, and there\u2019s a lot of credit card points.<\/strong><\/p>\n

Yes, totally.<\/p>\n

My sister\u2019s a credit card points person. It\u2019s frankly a little terrifying, but she\u2019s really good at it. How are you going to compete with that? Because that\u2019s the customer. In my mind, the customer who knows how to book a hotel and Uber is also the person with five different credit cards trying to get the best deal, and they know that this portal is where they need to go at this time. How do you compete with them?<\/strong><\/p>\n

So, I actually had an earlier interview<\/a> with The Points Guy, and I asked them, \u201cWhat\u2019s the best credit card for travel?\u201d Because I was curious. Turns out Amex Platinum, according to The Points Guy, is the best credit card for travel.<\/p>\n

I don\u2019t believe you, because this worked out too well.<\/strong><\/p>\n

I\u2019m just saying. It was amazing. We have a great relationship with Amex, where you get benefits and free bookings on Ubers as well. There\u2019s actually a lot of layering that we\u2019re doing. If you\u2019ve got Delta Sky Miles, you can get Delta Sky Miles for booking on Uber. We have a relationship with Marriott Bonvoy. We\u2019ve got Travelers using Uber all the time. We\u2019ve got the Amex Platinum card, the best card for Travelers as well. We have the right elements coming together to get some percentage of our Uber One members to try the booking experience, and then we\u2019ll go from there. I do think that this would be a failure if it ends with hotel booking. One of the pieces of magic that Uber brings is it\u2019s actually the backend experience.<\/p>\n

One of my learnings when I was at Expedia was basically that after the booking, there weren\u2019t that many services that Expedia offered other than if something went wrong. You do everything you can to help the customer, but actually what we can do is connect all these logistical elements of your travel. So, obviously, your Uber to the airport, if you did your hotel booking, we already know where your Uber is, maybe we\u2019ll give you a discount to the hotel. I\u2019m hoping that as we build out travel, we can actually improve the in-market experience. I don\u2019t know about you, but why do I need to check into a hotel?<\/p>\n

What\u2019s the deal with that? I\u2019ve got my phone, and if you have a hotel booking, maybe you can walk into the hotel and we can give you all the information and you can just go up to your room, and maybe your app can act as a key, et cetera. There\u2019s a lot more that we want to do in terms of the in-market experience, and it\u2019s something that Uber is uniquely positioned to do because we\u2019re already in-market in almost every city that you\u2019re going to want to travel to.<\/p>\n

There are competitors in these markets. Expedia is an interesting partner because you used to be the CEO of Expedia. I assume you just made a phone call and said, \u201cHey, what\u2019s up? It\u2019s me.\u201d<\/strong><\/p>\n

I actually had to recuse myself from the process entirely. The idea, the strategy, \u201cLet\u2019s get deeper into travel,\u201d obviously I was involved with all that, but because of the conflict \u2013 I\u2019m still on the Expedia board \u2013 I had to recuse myself from the process. The team ran it, and I\u2019m like, \u201cGuys, what\u2019s going on?\u201d They\u2019re like, \u201cWe can\u2019t talk to you.\u201d Expedia won because of the great job that that team did, they got no help from me. I\u2019m sorry, Expedia.<\/p>\n

The CEO of Expedia wasn\u2019t like, \u201cI\u2019ve got a board member breathing down my neck.\u201d<\/strong><\/p>\n

I had to recuse myself in those discussions. It was a little awkward, but it all worked out well.<\/p>\n

So, obviously Expedia would be a competitor, but they\u2019re your partner. There are other competitors: there are the hotel loyalty programs, Booking.com exists. They say the same sorts of things that you say. They\u2019ve <\/strong>been on the show<\/strong><\/a> saying literally the same sorts of things that you say.<\/strong><\/p>\n

Connected trip I think they talk about, right?<\/p>\n

All the time. Yeah. Why do you need a hotel? I think a lot of people like checking in the hotel, the free water especially is very useful when you arrive in a new hotel. That piece of the puzzle, where you\u2019re going to connect everybody\u2019s backend systems together and build one unified experience where the Uber app is the primary interface, I could abstract that away and say, \u201cWell, that\u2019s everything, that\u2019s what OpenAI would like to do. That\u2019s what Google would like to do.\u201d<\/strong><\/p>\n

Why is Uber going to win that fight?<\/strong><\/p>\n

It\u2019s a different question or service offering in terms of offering the availability of the service, but to the extent that you can actually deliver it in-market, it is truly different. OpenAI is an incredible company, they build a lot of cool things, but they don\u2019t live in the probabilistic real world that we live in. There\u2019s a Mike Tyson saying, I think<\/a>: \u201cEverything is theory until you get punched in the face.\u201d<\/p>\n

\u201cEveryone has a plan until you get punched in the face.\u201d<\/strong><\/p>\n

Yeah. \u201cEveryone has a plan.\u201d And we get punched in the face daily, which is drivers are canceling, riders are having issues, deliveries are late, et cetera. We already deal with this probabilistic world on the back end where things go wrong all the time, and it\u2019s one thing to try to chain all of these events together, and get the logistics right, but to adjust to real world traffic conditions, cancellations, road closures, all of that stuff, we do daily. We\u2019re much better equipped to actually fulfill this seamless, delightful end-to-end experience from planning to booking, making it incredibly easy, and then to delivery, the actual experience on the ground.<\/p>\n

Your partnership with Marriott, for example, Marriott wants those to be their customers. If you\u2019re the app that everyone\u2019s doing everything in, that relationship gets intermediated, is that a tension?<\/strong><\/p>\n

It\u2019s a tension. At the same time, it\u2019s a tension that everyone deals with. Marriott competes with Expedia, to some extent you could argue that they compete with us, although we\u2019re a much smaller player today in travel, maybe we\u2019ll get bigger. We work with Starbucks at Uber Eats, and of course they\u2019d rather have people come directly to their app, but the fact is that Uber Eats brings them a lot of incremental demand as well. So, this \u201ccoopetition\u201d theme is something that many, many players have been comfortable with for many, many years.<\/p>\n

Comfortable with for many, many years is in one context, right? Everybody has an app, and it doesn\u2019t really matter, you\u2019re all going to open the apps and maybe we can get you to open our app with a discount or a point system or something. Now you\u2019re in a world where you\u2019re going to open an app and maybe an agent\u2019s going to go off and do something for you. The idea of being the everything app in that context \u2013 Uber is describing this as a step to being in everything. It\u2019s in the press materials.<\/strong><\/p>\n

Brian Chesky <\/strong>was on the show<\/strong><\/a>. Airbnb is going to do concierge services for travel, and they\u2019re going to get way out of their lane, and maybe that\u2019s working, maybe it\u2019s not. I haven\u2019t talked to Brian in a minute about it. OpenAI wants to be in everything. X <\/strong>famously<\/strong><\/a> is already \u201cthe <\/strong>Everything App<\/strong><\/a>,\u201d as you know. We\u2019re all using X all day long for everything. Do you think the pressure on needing to be that interface is going up because of AI?<\/strong><\/p>\n

I think the pressure is going up to some extent, but I think AI is making it possible in a way that it wasn\u2019t possible previously. One is these models can adjust to real world conditions in a way that deterministic code can\u2019t. That\u2019s really cool. Whereas you had to build UI interfaces that were tight and relatively limited, AI is allowing for an interface that is unlimited, essentially. You can just tell the app what you want, and you can have agents then take that and break up that request and try to deliver it as best you can. AI is making it possible now. You can just build much faster. To go to smart risk, the cost of taking risks is going down. All of that is coming together in an opportunity set that I think a lot of companies recognize, including us, including Airbnb and the other companies, and it\u2019s going to be a race to many of these new markets, and we\u2019re confident.<\/p>\n

We\u2019ve raced before, we love competition. But this is another trillion dollar plus opportunity, and we\u2019ve done well with mobility, we\u2019ve done well with delivery. All of these businesses have been built organically. There\u2019s a builder mindset at Uber, and we\u2019re going to give it a shot, and so far the signal\u2019s pretty damn good.<\/p>\n

Last time you were on the show, we talked a lot about agents and accessing Uber as a service inside of an agentic workflow. I will tell you, I asked a lot of CEOs at that time this question, and everybody who had a physical product was like, \u201cWe\u2019ll be fine.\u201d And then it was Amazon, who has an interface to a bunch of dropshippers, that is filing the lawsuits.<\/strong><\/p>\n

They have a virtual product. Everybody who was in the world of atoms was like, \u201cGo ahead and try. Try to make another Uber, you just give it a shot, we\u2019ll be here when you\u2019re waiting.\u201d That was very much your attitude. What you said to me was, \u201cThe price of calling an Uber on ChatGPT should be zero until they prove it\u2019s valuable and then I\u2019ll figure out what the rate should be.\u201d It\u2019s been a year. Have you seen any meaningful uptake of calling Ubers from ChatGPT?<\/strong><\/p>\n

No. And it doesn\u2019t seem to be at this point a priority for a lot of the foundation model companies, whether it\u2019s ChatGPT or Gemini. I think they\u2019re experimenting with it, but I think the enterprise market is growing much faster than anyone thought that it was going to. There\u2019s been a pivot towards enterprise. Rightly so, based on the growth rates that we see, based on our internal usage of these foundation models. At this point, that part of the market hasn\u2019t developed, and the cool thing is, we\u2019re building some really cool products. You can scribble a shopping list, you can take a picture of food that looks really tasty, and we\u2019ll put together a shopping list for you. If you tell us what merchant you want to go shopping at, we\u2019ll put together the list for you and we\u2019ll get it delivered automatically.<\/p>\n

A lot of these experiences that I think people thought you\u2019d find on OpenAI, et cetera, you\u2019re actually going to find first on an Uber. I wouldn\u2019t be surprised if it\u2019s built over a period of time, but right now enterprise is coming first and you could argue rightly so.<\/p>\n

Uber is a favorite of agentic demos. You pop up all the time. I\u2019m just going to go down the list.<\/strong><\/p>\n

Is that right?<\/p>\n

Yeah.<\/strong><\/p>\n

Well, it\u2019s an everyday use case. It\u2019s great.<\/p>\n

Google and Samsung announced Gemini task integration on the newest Samsung phones, where the model will literally open the Uber app in the background in a virtual container, and click around it to get you a car. Have you seen any meaningful rides from that integration?<\/strong><\/p>\n

Not yet. Not yet, but we\u2019d be delighted to see it. We want to bring more experimentation, more opportunity for our drivers, it\u2019s just really small now. It doesn\u2019t mean it\u2019s not going to be big 10 years from now.<\/p>\n

We had a whole year of these things. Has Alexa sent you any meaningful rides?<\/strong><\/p>\n

No. Small. Very, very small.<\/p>\n

Okay. And I can keep going, but it seems like the answer \u2013<\/strong><\/p>\n

Have you used any of these products?<\/p>\n

I have to, I\u2019m required.<\/strong><\/p>\n

And how is it?<\/p>\n

I think they all have the same problem. Down the line, they\u2019re slower than me just doing it myself. Also, I\u2019m only ever calling a car from work to home or home to work or to the airport. The app is one tap away for all of those experiences.<\/strong><\/p>\n

Exactly. And it\u2019s pretty easy to use. One area that, for example, we are looking at is while the front end, the initial demand may come from any agent, I am going to want our pixels in front of you. For example, I\u2019m perfectly fine with OpenAI calling Uber, but then I want in that web interface and within the ChatGPT app, the Uber pixels and the Uber brand so that you know who is fulfilling that ride for you. We\u2019ll see how things turn out. If you\u2019re an Uber One member, you\u2019re going to want to use our product, especially for travel.<\/p>\n

Again, this is the fight that <\/strong>I\u2019ve seen coming<\/strong><\/a>, where getting people out of your app and just using Uber as a backend service, as a commodity against every other service, pure play or not, nobody\u2019s going to want this. But it seems like they\u2019ve all pivoted to enterprise so fast that that fight is delayed or maybe never coming.<\/strong><\/p>\n

I think it\u2019s delayed. It\u2019s going to happen because I think the size of the prize is too big. If you talk about history not repeating itself but rhyming, there\u2019s some of what I went through in my former job at Expedia. If you remember during those times, there was a big debate about metasearch. There were these metasearch players \u2013 Kayak, TripAdvisor, Trivago \u2013 that were amalgamating a bunch of travel content, and there was a point at which metasearch was quite powerful in terms of customer acquisition, et cetera. But as supply consolidated, really the value started accruing to the suppliers much more than the meta players and the travel business consolidated to Expedia, Booking.com, and Airbnb. There\u2019s more, but those are the three very, very big players. On the supply side, when you look at mobility, when you look at delivery, there\u2019s usually two or three players in every market.<\/p>\n

Even if you get that front end being particularly big, in a consolidated, let\u2019s say, supply marketplace and with our size and scale, multi-platform, all the countries that we operate in, I think we\u2019re going to be more than okay, in terms of the leverage and the negotiations that happen. I always try to push the negotiations to the backend \u2013 build a great experience, figure out the economic balance later \u2013 but sometimes you\u2019ve got to figure that stuff out upfront.<\/p>\n

This is a slight difference from the last time you were here, and I just noted the companies are all different \u2013 not Uber, but the AI companies \u2013 they\u2019re all in a slightly different posture than they were a year ago.<\/strong><\/p>\n

Yeah, totally.<\/p>\n

They\u2019re racing towards IPO, they are constantly calling code reds. Every week it\u2019s a code red at OpenAI.<\/strong><\/p>\n

It\u2019s a cool thing to do.<\/p>\n

Yeah. We\u2019ve had CEOs come on the show and say they\u2019ve called a code red, and I\u2019m like, \u201cDid you actually do it? And they\u2019re like, \u201cNo.\u201d They just wanted to say.<\/strong><\/p>\n

We definitely had our share of code reds. And there\u2019s a danger of code red fatigue in companies too because then it becomes meaningless. It\u2019s a real issue.<\/p>\n

OpenAI was a partner of yours, you\u2019ve obviously launched things with them, you\u2019ve used the products. As you broadly think about, \u201cWe\u2019re going to build AI services, we need a model provider,\u201d do they feel like a stable partner?<\/strong><\/p>\n

Yes, their products are excellent. For example, we\u2019ve used, I think, ChatGPT 5.5 for some of the cool stuff that we demoed today in terms of the shopping list or taking pictures, et cetera. Codex is something that a lot of our devs use. OpenAI has been a strong partner. Whatever drama that you see in the markets isn\u2019t showing up in terms of the quality of their product. They continue to be first-rate.<\/p>\n

The drama in the market is all encompassing. As you and I sit here today, Sam Altman and Elon Musk are <\/strong>in a courtroom arguing<\/strong><\/a> with each other.<\/strong><\/p>\n

Listen, it used to be Uber. When I was looking at joining the company, it reminds me of that, and we got through it. We got through it, and it\u2019s a great company now, and I think that it\u2019s an adjustment that every company has to go through. So many people are interested in how OpenAI does things because it\u2019s an important company in the world. They\u2019ll get through this.<\/p>\n

Do the model companies feel interchangeable in a way that has always seemed like a small danger here?<\/strong><\/p>\n

I think interchangeable is a little bit too strong a word. I do think that what Anthropic is building, Claude, it\u2019s spectacular. Our developers are using it all the time. Codex is definitely picking up use by our developers. What we do <\/em>do is we use some of the frontier models and some of the more advanced models to pilot, build demos if you want to build something quickly. And then what we do look to do is \u2013 it\u2019s much more than an API layer \u2013 we\u2019ve got a platform, Michelangelo, that has all the data feeds, and then essentially you\u2019re able to switch models. And early on when we\u2019re trying to explore something, we will use some of the more advanced models, but then once you get up to larger volumes, we will try to switch out either cheaper models or open source models to control the costs and the token costs on the backend.<\/p>\n

Interchangeable is too strong a word, but we definitely experiment with various ones, and at this point, nothing is hard coded into our systems. And frankly, we\u2019re going to make sure that none of them are hard coded into our systems.<\/p>\n

That seems like a hedge against the companies and their needs and also cost, right? The cost of tokens is still quite high.<\/strong><\/p>\n

Yeah. You never want to be overly dependent on one technology unless you\u2019re highly confident or it is very, very, very proprietary. Part of it is that all this stuff is so new. You and I were talking about Cursor last year. And Claude wasn\u2019t a thing, at least internally. Now Claude is really, really increasing at incredibly surprising rates internally. So, early on, as this market is developing, we want lots of experimentation, and we want to give our devs the freedom to try a bunch of stuff. I don\u2019t want this to be top down, \u201cthou shalt here or there.\u201d Of course, there\u2019s going to be optimization, but right now there\u2019s a lot of experimentation going on internally.<\/p>\n

Let me ask you about running a software company in 2026. This is the thing I was most excited to talk to you about. It is true, the last time you were here, we were talking broadly about AI, and had all these questions about agents and the big labs coming for you with their consumer chatbots. Maybe that\u2019s not happening yet. The thing that we did end up talking about just as you were walking out was, \u201cWe had GitHub Copilot, but all the engineers want to use Cursor.\u201d And now you\u2019re saying, \u201cAnd Cursor\u2019s around, but they\u2019re all using Claude Code. Or maybe they\u2019re using Codex.\u201d<\/strong><\/p>\n

The increase in Claude Code usage and sometimes the replacement of Cursor usage is fairly remarkable. We use both, they\u2019re both terrific products. And then there\u2019s a group that\u2019s using Codex. And they\u2019re all really good. And I\u2019d say the big change is with Cursor. It was coding and coding assist, so to speak, complete, but now these agents and agentic coding is something that is just blowing people away. It\u2019s very, very cool.<\/p>\n

And when you say \u201cblowing people away,\u201d I would say many of your peers have gone crazy. They have seen agentic coding, it\u2019s looked them in the eye, and they\u2019ve responded by losing their minds and saying that the entire structure of a company should change around this. I\u2019ll give you some examples. Meta is reportedly going to have teams where 50 people report to one manager; Jack Dorsey can\u2019t lay off enough people fast enough, and his goal, he said this out loud, he wants all 6,000 people agentically assisted to report to him at Block. I don\u2019t even know how you would do that. It\u2019s a show about org charts, and I read that, and I thought, well, our show\u2019s going to keep going for another decade. We\u2019re on the cusp of the weirdest org charts in history. Are you there? Are you saying, \u201cAgentic coding is going to fundamentally change how you construct a software company\u201d?<\/strong><\/p>\n

We have not gone and examined the fundamental org chart of the company yet. I\u2019m not saying it won\u2019t happen. We are pushing the company hard, and I\u2019ve got to push the company harder to go to first principles in terms of how you work, period. Our culture is like bottom-up, let people do a bunch of stuff, and listen. The engineers are using it, the debugging, all the cool stuff is happening, as it should. But what we saw is in sales. Salespeople now use agents to summarize information on a client that they\u2019re going to call to build out a really cool presentation. We\u2019re using agents and AI, I would describe, around the edges of how we work. That\u2019s one. And we\u2019re not thinking about, \u201cLet\u2019s think about the sales function from the bottom up,\u201d I\u2019d want to push a company to do that.<\/p>\n

Customer service is another example where we\u2019ve got agents who generally follow policies. There\u2019s a policy, if you\u2019re an Uber One member and your order is delayed by 20 minutes, we\u2019re going to give you $15 back because you\u2019re a loyal customer, et cetera. That\u2019s a policy that\u2019s in place, and there are agents that are following those policies, et cetera.<\/p>\n

Human agents you mean. Your current agents.<\/strong><\/p>\n

Human agents. And we then said, \u201cLet\u2019s build virtual agents to follow those policies,\u201d and it turns out that actually our policies on a global basis, the documentation is complete crap, to use a technical term. What happens is, an agent, a human agent, I\u2019ll be sitting next to you and be like, \u201cWhat does this policy mean? It\u2019s kind of unclear.\u201d And you coach me and then I figure it out. Humans are quite flexible. When we had AI agents go through these policies, they just went nuts. One approach was, \u201cLet\u2019s rebuild all the documentation and policies the right way, and then let\u2019s have the agents work based on these policies.\u201d<\/p>\n

But why do we put those policies together in the first place? It was to get to goals and outcomes based on standardized ways. I don\u2019t want to go bankrupt, but I want to keep you, the Uber One member, happy. And so we made a policy to approximate the optimal outcome for the population. But now I can just tell the agent what that outcome is. I want actually to be fair to a person, I want Uber One members to be happy, I don\u2019t want to go bankrupt, et cetera. The approach that we\u2019re taking now within customer service is to throw away the policies, describe to the agent what you\u2019re trying to accomplish, and then let the agents go and obviously train them on good interactions, bad interactions, and give them feedback, et cetera.<\/p>\n

Wait, just a foundational philosophical question.<\/strong><\/p>\n

Yeah.<\/p>\n

Why trust computers to make those determinations and not people?<\/strong><\/p>\n

Because the model can learn based on the population of everything that is happening, versus an individual human just learning based on the experience that he or she is having that day, and models are easier to track and tune than humans are to train.<\/p>\n

Okay. So, this is a scale answer. It can see all the data.<\/strong><\/p>\n

Yeah.<\/p>\n

So, you can just describe a generalized outcome.<\/strong><\/p>\n

You can retrain based on that data and you have perfect visibility into the actions, reactions. The retraining output you don\u2019t have perfect visibility into, but you can iterate around that. It does demand a different approach and it\u2019s a little bit back to what you and I were talking about, which is a smart risk. It\u2019s a riskier approach. We got to throw stuff out and just completely rebuild in a different way. And I\u2019m really glad. In this case, it wasn\u2019t me who pushed the customer ops team to throw everything out. They were frustrated with the results that they were seeing early on, like, \u201cWe have to be able to do better, we\u2019re going to try this out.\u201d The signal looks really promising, but I can\u2019t tell you if it\u2019s actually going to work in the end.<\/p>\n

That kind of dynamic customer response \u2013 in terms of pricing, people are making it illegal in this country to do dynamic pricing in that way because it feels unfair.<\/strong><\/p>\n

Yeah. That is actually an issue. What we don\u2019t want to do is have different outcomes based on targeting you versus another person versus another person. But you can have different outcomes because there were circumstances that were different. So, for you, if your food was 15 minutes late, another person\u2019s food was 45 minutes late, and you\u2019re both Uber One members, you could actually have different outcomes because the circumstances are different. It\u2019s not based on targeting or optimizing based on targeting, it\u2019s optimizing based on context.<\/p>\n

That\u2019s really interesting, and it strikes me that we could probably do another whole hour on, \u201cWe wrote a bunch of rules for humans, and now we have to write a system prompt that isn\u2019t the rules.\u201d<\/strong><\/p>\n

It\u2019s actually the outcomes that you\u2019re trying to get at. Yeah.<\/p>\n

Yeah. That\u2019s another ad.<\/strong><\/p>\n

We\u2019ll see if it works.\u00a0<\/p>\n

You\u2019re going to come back next year, I\u2019m going to ask you if it worked.<\/strong><\/p>\n

But let me ask you just more at the base level. When I think about software companies generally, the creative tension of any software group is you have a PM, you have a designer, you\u2019ve got some engineers. They all want to be in charge. They all think they are going to do it right. And they all need to work together. If you can get that right, it\u2019s magic. It feels like with the power of vibe coding, everyone is going to try to do everyone else\u2019s job, and no one\u2019s going to be good at it, and it\u2019s all a mess. I can see it happening all over the place already.<\/strong><\/p>\n

Totally.\u00a0<\/p>\n

Are you rethinking that basic triad inside of Uber?<\/strong><\/p>\n

It depends on the kind of project that you\u2019re working on. There are some larger projects where you need design, you need proper planning, et cetera. But we are having some product team members, whereas previously, if there were some simple bugs in the code or very, very simple features, they would have to then prioritize it with their engineers, et cetera. Now, they\u2019re just going in and they are vibe coding, and an engineer is going to review the code, but essentially the product person is going directly into the code base, so to speak, or going directly with an agent into the code base. I do think for simpler problems, smaller problems, the dynamics are going to change. We\u2019re going to try it out, we\u2019re going to see what happens.<\/p>\n

When you look at a company like Meta, which seems to just be in the midst of endless rolling layoffs, they\u2019re saying it\u2019s because AI is making everybody more productive, it might be because they\u2019re just freeing up capital expenditures (CapEx) to go spend on whatever they\u2019re spending CapEx on, to whatever end that Meta is going to do AI. Super intelligence, I\u2019m told. Are you in that same spot where you\u2019re like, \u201cWe\u2019re getting more productive, I need fewer people\u201d?<\/strong><\/p>\n

No. My view is if an engineer is going to be 50% or 200% more productive, I want more engineers. The list of ideas in terms of what we want to build so outscales our throughput at this point, that generally we are looking to add more engineers to our employee base. Now, there is a trade-off, and we are dealing with the trade-off right now as we speak. I don\u2019t know if you saw it, but our CTO was talking to a reporter, and made a comment, which is true, that we have blown through our AI token and infrastructure budget for the whole year in about three to four months. And it was a big thing when that happened, but it happened. And the trade-off is going to be headcount. We are budgeting differently. Previously, you would have a headcount budget or plan, doesn\u2019t mean it would actually happen, but as a plan going in, you would have an infra budget.<\/p>\n

Now there\u2019s an active trade-off going on between the two, and to the extent that we have overages in terms of token spend or infra spend \u2013 theoretically those overages are products that are being built and are productivity that\u2019s being added to our engineers \u2013 we\u2019re going to hire less aggressively, so to speak. That is a live trade-off. How far it\u2019s going to go, I don\u2019t know at this point.<\/p>\n

Are you all the way at, \u201cI\u2019m spending so much on tokens that it\u2019s costing me more than hiring one junior engineer\u201d?<\/strong><\/p>\n

We are spending a lot on tokens. I haven\u2019t done the math yet, but it\u2019s significant. But the throughput is really accelerating. At this point, it\u2019s something that needs to be managed, and I do think it\u2019s just taking different muscles. The way that we\u2019re managing budgets, especially on tech, is fundamentally different from how we did three, four years ago.<\/p>\n

All right. One last AI question, then I want to talk to you about autonomy. Which is also AI, but in a very different way.<\/strong><\/p>\n

Physical world AI. Yeah.<\/p>\n

You were on <\/strong>Diary of a CEO<\/strong><\/a>, and you said the employees at Uber have created an AI version of Dara to practice presenting and pitching to. Is that real and how close are we to AI replacing the CEO?<\/strong><\/p>\n

It is real, I have not witnessed the Dara AI, but it is real. People have done it. Honestly, I don\u2019t know how good it is. It\u2019s clearly not as good as the real thing. Come on, how is that possible?<\/p>\n

Decoder <\/em><\/strong>listeners, every time we do an AI episode, they say the AI should replace the CEO. It is a reflexive comment we get.<\/strong><\/p>\n

I\u2019m not there yet. I think that the AI powered CEO is going to be better than the AI CEO. I think there\u2019s a magic in terms of teaming up humans with AI and with agents, and based on what I see, that is a superior product than pure play AI or pure play human.<\/p>\n

You should recuse yourself from this. You have a deep conflict of interest here.<\/strong><\/p>\n

Of course I do. I\u2019m hoping the board sees it that way as well. Maybe the board is planning this and I had no idea.<\/p>\n

That would be, in keeping with the Uber story, that would be there.<\/strong><\/p>\n

Exactly. How is AI changing our board processes? I\u2019ve got to think about that one.<\/p>\n

Oh, trust, I get those pitches. You don\u2019t want anything to do with those. Let\u2019s talk about robots, actual robots, actual AI in the world.<\/strong><\/p>\n

Yes.<\/p>\n

Uber has made a bunch of big investments in robotaxis. I want to start with Rivian. It\u2019s over a billion dollars, I think it\u2019s $1.2 billion in total commitment to Rivian over some number of years. That\u2019s a partnership announced in March, that you\u2019re going to buy up to 50,000 fully autonomous R2 robotaxis by 2031, but it\u2019s also called an investment. And I\u2019m just doing the math \u2013 at the price of the R2 platform, you\u2019re just buying a bunch of cars. Is buying a bunch of cars an investment or are you actually getting equity in Rivian?<\/strong><\/p>\n

We actually invested in Rivian equity, and we\u2019ve invested in a number of our partners. Usually we will invest in our partners like Lucid, in WeRide, in Avride, for example. It is an investment and it\u2019s a vehicle commitment as well. It\u2019s both. It\u2019s based on deliverables, obviously. They\u2019ve got to deliver, and based on everything that we\u2019ve seen from RJ and team, putting together a first-class AI team, we\u2019re confident that they can deliver on those R2s.<\/p>\n

Yeah. The deliverables are very vague, I\u2019m just going to read you the press release. \u201cUber will invest up to $1.25 billion in Rivian through 2031 subject to,\u201d and I quote, \u201cthe achievement of certain autonomous milestones by specific dates.\u201d<\/strong><\/p>\n

Well, they are very specific contractually, and they\u2019re fairly fuzzy as far as what you know.<\/p>\n

I put this into five different AI systems and no one can tell me what they are. What are the autonomous milestones?<\/strong><\/p>\n

I could tell you, but then I\u2019d have to kill you.<\/p>\n

The reason I\u2019m asking is that I desperately want to know the specifics.<\/strong><\/p>\n

I\u2019m looking at this industry in total, and I will tell you that we\u2019ve thrown out whatever autonomous milestones we used to have, the level system that everyone used to talk about. That\u2019s all gone. No one cares about this anymore. No one\u2019s like, \u201cWe shouldn\u2019t do level four.\u201d We\u2019re doing it. And I can\u2019t quite tell you what milestone an autonomy platform has to hit before I can say \u201cThis is a robotaxi.\u201d<\/strong><\/p>\n

I\u2019ll give you examples of milestones, not specific to Rivian. Usually there\u2019s a milestone, for example, if you release in-market with a vehicle operator. Usually one other milestone may be if you take the vehicle operator out, you can only take the vehicle operator out to the extent that you complete a safety case that we put together along with an autonomy provider, then another deliverable might be delivering a certain number of cars that are NVO-capable, that have a redundancy at a certain bill of materials as well, at a certain cost. Those are examples of deliverables that have to do with either capability or economics, because ultimately this is about going to market with a product that\u2019s proving to be a very, very popular product.<\/p>\n

Your big partner in the past was Waymo.<\/strong><\/p>\n

Yes.<\/p>\n

Waymo has gotten there in many cases to some of the kinds of milestones you\u2019re describing.<\/strong><\/p>\n

Yeah, sure.<\/p>\n

You\u2019re obviously diversifying away from Waymo, you\u2019ve got the Rivian deal, you mentioned Lucid, you\u2019re going to buy at least 35,000 Lucid vehicles designed exclusively for use as part of Uber\u2019s robotaxis.<\/strong><\/p>\n

Yeah, and a partnership with Nuro.<\/p>\n

And a partnership with Nuro, which is the platform there.<\/strong><\/p>\n

Yes.<\/p>\n

Overall, you\u2019re going to commit some $10 billion to autonomous efforts. You launched Uber Autonomous Solutions, which feels like a bet on this is happening, but we don\u2019t know who\u2019s going to win. You\u2019re diversifying.<\/strong><\/p>\n

It\u2019s a little bit different from that in that we believe that it is going to happen, and we believe that just like there isn\u2019t going to be one foundation model to rule them all, there isn\u2019t going to be one physical world foundation model to rule them all. And all the evidence that we see is, yes, Waymo has passed the finish line, they\u2019re the leader, they are in many ways inspiration for many, many companies in this industry. They\u2019re a great partner of ours in Atlanta and Austin. There are many other companies that are getting to the finish line. WeRide, for example, or Pony.ai or Baidu \u2013 these are Chinese companies \u2013 are already at the finish line, and we are in-market, for example, with WeRide in the Middle East. There are players like Nuro, Waabi, Avride or Wayve, all of whom are accelerating to the finish line.<\/p>\n

And if anything, the speed of getting to the finish line is accelerating. One, model capabilities are much, much better now. It used to be deterministic code that you had to slog through, now obviously it\u2019s learning AI models. SIM capability is much better so that data will go much further in terms of model training. And what we\u2019re trying to do with AV Solutions is build out the whole necessary ecosystem around these companies so that they can focus on what they do best, which is training these models to get them to superhuman safety. We can help them get there, for example, with data collection, and we can both then get to market as quickly as possible.<\/p>\n

It\u2019s not, I would say, a diversification bet. It\u2019s a bet that there are going to be many players. And as a platform, we\u2019ve always been supply-led. The way to grow our platform is to build out supply, whether that\u2019s more drivers or more restaurants or more hotels. As we build out liquidity of supply, demand shows up, and just like we want every safe human driver on the platform, we want every safe robot driver on the platform, whether that\u2019s a Waymo driver or a Nuro driver or an Avride or a WeRide. It\u2019s a bet that we\u2019re making, which is that there won\u2019t be one physical AI model to rule them all.<\/p>\n

There\u2019s some real confidence in this bet. I\u2019ve talked to a lot of rideshare CEOs over the years, a lot of autonomy CEOs over the years, and it\u2019s always been 10 years away.<\/strong><\/p>\n

Yeah, yeah.<\/p>\n

The confidence I\u2019m hearing from you is, \u201cThis is happening. We\u2019re spending a lot of money to get there faster.\u201d<\/strong><\/p>\n

All the evidence we see is that it\u2019s happening. Waymo has shown the way. A lot of Waymo engineers now are working in other companies. For example, the Chinese players have shown the way, and you\u2019ve seen it, the speed of foundation model development, whether it\u2019s digital foundation models or physical foundation models. Nvidia is betting on this as well. These are big bets made by capable companies and we think we\u2019re on the right track here.<\/p>\n

In the context of our conversation, I\u2019m going to bring up the trade-off.<\/strong><\/p>\n

Sure.<\/p>\n

By saying it\u2019s going to be more real, you no longer get to kick the can on, \u201cWe\u2019re not going to have drivers in the cars,\u201d which famously got Travis Kalanick in a lot of trouble by saying, \u201cI want to get the driver out of the car,\u201d long, long ago.<\/strong><\/p>\n

Yeah.<\/p>\n

Because autonomy was so far away, we just didn\u2019t have to solve this problem. You have been on podcasts recently saying, \u201cThis problem is here. I don\u2019t know what\u2019s going to happen to 9.5 million Uber drivers when autonomy comes.\u201d You literally said, \u201cI don\u2019t know,\u201d to Steven Bartlett.<\/strong><\/p>\n

Well, if you don\u2019t know, you should say it. Now, here\u2019s what I know. 10 years from now, I am 90% certain that we\u2019re going to have more drivers on our overall platform than we do today. I don\u2019t know if that\u2019s going to be true in San Francisco, but with the way that the business is growing, and the capability of building these cars at the right bill of materials in all the markets that we operated, not just the high cost markets, we\u2019re going to have plenty of drivers, and we also are actively looking to build out more use cases for drivers that are more complex. One of the announcements that we made was about a personal shopper. It was Courier, people started hacking Courier, asking Uber Couriers to go shop for them, so we decided to productize that as well. That\u2019s a very, very complicated interaction.<\/p>\n

It\u2019s a random store, take a picture of the goods, \u201cThis is what I want.\u201d We\u2019re building out much more complex use cases for humans to migrate onto as more of the work is being automated. 20 years from now, I don\u2019t know what that\u2019s going to look like, because then you really start increasing capabilities. I think these are big societal questions. It\u2019s going to be true of white-collar workers, and it\u2019s going to be true of certain kinds of blue-collar work as well. CEOs should talk about this, not in a way to scare people, but we should also be honest about it. I\u2019ve never seen a wave of technology that has such a direct impact on how companies work and how people have worked with the accelerated pace that I\u2019m seeing today. Doesn\u2019t mean that society can\u2019t adjust, but the pace of change here, it\u2019s pretty remarkable.<\/p>\n

One of my theories about the extremely negative polling on AI is that it\u2019s fundamentally an enterprise technology. You\u2019ve described this even in this conversation, the frontier models, those companies are moving to enterprise use cases, you at Uber are using them in enterprise context, and there are not great consumer products in front of people.<\/strong><\/p>\n

Not yet. Yeah.<\/p>\n

I haven\u2019t seen them. Maybe they\u2019re coming.<\/strong><\/p>\n

We\u2019re trying to do that and it\u2019s these moments of surprise and delight where you can talk to your Uber to get an Uber, lots of complex situations, you can transcript a shopping list, take a picture \u2013<\/p>\n

Sure. But I don\u2019t think that stuff is going to change <\/strong>the overall polling on<\/strong><\/a>, \u201cThis is a threat that will take my job away.\u201d<\/strong><\/p>\n

Yeah. Listen, if it\u2019s your job, I think you\u2019re right. Yeah.<\/p>\n

This dynamic of everybody showing up saying the jobs are going away, and mostly because it\u2019s so good at writing code, this is a weird disconnected dynamic for regular people. Uber needs customers, you need people with money to want to ride around. How do you see that economy playing around?<\/strong><\/p>\n

Right now talk is louder than what we see in the market. The economy remains robust, the consumer remains robust, we don\u2019t see white collar people out of work at this point. I just don\u2019t see it in-market. Now, the fear that you see might be a leading indicator of what\u2019s to come, but at this point, I see no signal in our actual business that it\u2019s having an impact on consumers at large.<\/p>\n

What do you ascribe the extremely negative polling around AI to?<\/strong><\/p>\n

I do think that it\u2019s some fearmongering from the press. They love the drama. Are you part of the press or no? A little bit, okay.<\/p>\n

Yeah. Can I have this level of influence?<\/strong><\/p>\n

Can I point at you here?<\/p>\n

You can point at me all you want.<\/strong><\/p>\n

But listen, it\u2019s a conversation that people are constantly having, it\u2019s a dramatic conversation. And I do think machines replacing humans has been a theme for eons. What you do see in manufacturing, for example, with automation is that machines complement humans, and then there are other capabilities that humans always adjust to. It\u2019s just things are moving so fast now that I think the fear is out there. I\u2019ve got 14-year-old twin boys and two other older kids. My 14-year-old kid is like, \u201cDad, why should I study? I\u2019m not going to be able to have a job.\u201d And I was just blown away. My 14-year old is asking me. Now maybe he doesn\u2019t want to study.\u00a0<\/p>\n

This does feel like the main thing 14-year-olds say.<\/strong><\/p>\n

Yeah, exactly. So, it\u2019s in the ether, you see signals, there are some companies, like you mentioned, who are acting on it. We\u2019ll see what happens in the next two years. But I don\u2019t see how it\u2019s going to reverse. Once we get more data, maybe the reality will be less dramatic than someone makes it out to be, and then we\u2019ll see. We\u2019ll do our best.<\/p>\n

Yeah. I would love for it to be real that it\u2019s the press. The media history is not at a moment of intense strength right now. It is contracting.<\/strong><\/p>\n

Yeah, but there have been some. I do think that the media is incentivized sometimes to overdramatize these things. Could be real, maybe it\u2019s not. I do think that there is a reality in it. The question is, how quickly is change going to happen, and will society be able to adjust fast enough?<\/p>\n

Look, I get all my news from X the Everything App, which assures me on the daily that AGI is just around the corner.<\/strong><\/p>\n

I want to ask you the question I ask every time I talk to you. I always take an Uber to come see you, it\u2019s just my little tradition, and the drivers always have the same question. So, I have the same question every year.<\/strong><\/p>\n

Sure, sure.<\/p>\n

And then this time I actually got a very detailed follow-up question to ask you.<\/strong><\/p>\n

Oh, cool. All right.<\/p>\n

The drivers all want to know: How are they going to get paid more?<\/strong><\/p>\n

They are going to get paid more by some of the newer jobs that we\u2019re giving them. Shopping, for example, on a per-hour basis can pay more. But I do think that driver pay is based on what market rate pay is, essentially. The local pay goes up and down based on the spot cost of labor in a particular market. The way that drivers are going to get paid more is the cost of labor generally goes up, or it goes down. Right now, the cost of labor is fairly steady, and driver pay has been fairly steady. Nationwide, it\u2019s probably $32, $33 per utilized hour. Here in New York City, it\u2019s over $50 per utilized hour. Drivers are making decent money. Of course, they\u2019re going to want to make more money.<\/p>\n

They all want to make more money.<\/strong><\/p>\n

Of course.<\/p>\n

Do you think autonomy changes that rate?<\/strong><\/p>\n

I don\u2019t think it will significantly. I think that drivers are probably going to take longer trips. When we see autonomous inventory coming into a market, we slow down driver recruitment because we want the drivers who are in-market making as much. At this point, in markets like Atlanta, like Austin, where we have a significant autonomy presence, because we\u2019ve reduced recruiting, driver pay is actually up. And I\u2019m hoping that we can continue those trends for a long time.<\/p>\n

I\u2019m glad you brought up utilized hours because this is my very detailed follow-up. It\u2019s actually good because you brought up all the keywords of this question.<\/strong><\/p>\n

Sure.<\/p>\n

So, you mentioned Westchester, I live in Westchester. The drivers in Westchester are allowed to drive into New York City, but they\u2019re not allowed to pick up in New York City and drive back to Westchester. So they literally lose one utilized hour. I\u2019ve been directly requested that you go and lobby the city and state so that they can go home with a utilized hour instead of an empty ride.<\/strong><\/p>\n

We have already been lobbying. Some of these regulations have unintended consequences. New York is unfortunately one of the most highly regulated markets out there. A significant amount of your fare goes to the city, et cetera. I think Ubers are too expensive here, and I think regulation sometimes goes over the top. It\u2019s something that I will absolutely take to the powers that be.<\/p>\n

The powers that be in <\/strong>this <\/em><\/strong>city is Zohran Mamdani. Have you met with Zohran Mamdani?<\/strong><\/p>\n

I have seen him speak, I have not met him one-on-one yet, but I look forward to that dialogue.<\/p>\n

Well, here\u2019s my tips. One, say you love New York City, he loves it when you say you love New York City.<\/strong><\/p>\n

Cool, cool. I do.<\/p>\n

Two, tell him the drivers want the return trips from both the airports and the city.<\/strong><\/p>\n

I will absolutely relay that to him. Maybe he listens to your podcast, you never know.<\/p>\n

We know some people. The same thing, I can\u2019t tell you. I can\u2019t tell you what the milestones are.<\/strong><\/p>\n

Dara, this is always a pleasure, thank you so much for coming.<\/strong><\/p>\n

Thank you. I really appreciate it.<\/p>","protected":false},"excerpt":{"rendered":"

Today, I\u2019m talking with Uber CEO Dara Khosrowshahi. It\u2019s become something of an annual tradition to have Dara join us<\/p>\n","protected":false},"author":1,"featured_media":939,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"colormag_page_container_layout":"default_layout","colormag_page_sidebar_layout":"default_layout","footnotes":""},"categories":[1],"tags":[],"class_list":["post-938","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=\/wp\/v2\/posts\/938","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=938"}],"version-history":[{"count":0,"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=\/wp\/v2\/posts\/938\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=\/wp\/v2\/media\/939"}],"wp:attachment":[{"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=938"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=938"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/teknologi.news.eraenterprise.id\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=938"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}