Scale AI CEO谈Meta的140亿美元交易、将Uber Eats做至800亿美元，以及前沿实验室下一步在构建什么

There's been a lot of talk these days about AI not delivering on the promise we hear, especially at enterprises.

These things take 6 to 12 months to get truly robust enough that an important process can be automated. Like with any of these major tech revolutions, headlines tell one story, and then on the ground, laying broadband means you need to dig up every single road in America to lay it. Someone's got to dig up the road or someone's got to run the undersea cable.

Is there anything you think people don't truly grasp or understand about where AI models are going to be in the next 2 to 3 years?

The general trend right now is going from models knowing things to models doing things. The next question becomes, “What can it do for me?” How does the agent make decisions for you?

Let's talk about scale and this whole world of AI that you're in. You essentially pioneered data labeling, training data, and creating evals for labs.

18 months ago, you would get a short story, and it would say, “Is this short story better than this short story?” Now you're at a point where one task is building an entire website by one of the world's best web developers, or explaining some very nuanced topic on cancer to a model. These tasks now take hours of time, and they require PhDs and professionals.

I've talked to a bunch of people who have worked with you over the years, and I heard a lot about just how high of a bar you set for new businesses.

From an entrepreneurship standpoint, it truly is about: What insight do I have? Why am I so lucky to have this insight? Why, in a world of a million entrepreneurs who are thinking, who are smart, and who are trying everything, am I in the position where I likely have an insight that others do not?

Today my guest is Jason Droge. Jason is the new CEO of Scale AI. This is the first interview that he's done since taking over for Alex Wang after the Meta deal. Alex now leads the super intelligence team at Meta. Prior to Scale, Jason co-founded a company with Travis Kalanick before he started Uber, worked at a couple startups. Most famously, Jason launched and led Uber Eats, which went from an idea that he and his team had to what is now a multi-billion dollar run rate business, and one that basically saved Uber during the pandemic when nobody was taking rides. This interview is following a theme that I've been following through a bunch of interviews, which is the evolution of how AI models actually get smarter. Along with scaling, compute, and improving the actual model code, much of the improvements we're seeing in ChatGPT, and Claude, and every frontier AI model is these labs hiring experts to fill in gaps in their knowledge and correcting their understanding of how things work, and basically showing them what good looks like in every domain that consumers are using models. Scale was the pioneer in this space. They created the category, and in our conversation, we talk about what is happening at Scale and just how this deal with Meta worked, what experts like doctors and software engineers are specifically doing to help models get smarter, how the whole market of data labeling and evals and data training has changed from when Scale entered the market to today, and also just how long will we need humans to keep helping AI get smarter. We also get into where Jason sees models going in the next few years because they have such a unique glimpse into the future. We also talk about a ton of really unique and really important product lessons from the course of Jason's career, including a bunch of advice on how to start a new business, both startups and within existing companies, and also a bunch of advice on hiring and leadership and so much more. A huge thank you to Alan Penn and Steven Chow for suggesting topics for this conversation. If you enjoy this podcast, don't forget to subscribe and follow it in your favorite podcasting app or YouTube. It helps tremendously. And if you become an annual subscriber of my newsletter, you get a year free of 15 incredible products, including Lovable, Replit, Bolt, n8n, Linear, Superhuman, Descript, WhisperFlow, Gamma, Perplexity, Warp, Granola, Magic Patterns, Raycast, ChatPRD, and Mobbin. Head on over to lennysnewsletter.com and click Product Pass. With that, I bring you Jason Droge.

Yeah, thanks for having me. Excited to be here.

As I was researching your background and preparing for this podcast, I learned a really interesting fact about you that I don't think a lot of people know. Travis Kalanick had a startup before Uber. It was called Scour, a peer-to-peer file-sharing app that I think got shut down. You were his co-founder. This was the early part of your career.

I'm guessing there are hours of stories we could talk about from that experience, so let me ask you this one question: What's a lesson from that experience that has stuck with you and that you've taken with you to future places where you've worked and built product?

1. Everything Is Negotiable

There are so many lessons. I'd like to pick one. I think the main lesson is that, in business and in startups, everything's negotiable.

We were 19 or 20 at the time. We built this search engine in a dorm room, and we were running it out of the dorm room. Our first URL was scour.cs.ucla.edu. These things weren't necessarily infractions at the time; we were just being practical. It was basically a project that we had started, so we built this search engine. People started using it, and we thought we would get in trouble, but it turned out the computer science department was excited about it, even though we had basically parked a domain on their servers.

We were using our own computers in the dorms to serve up this website and product. Then, when we got into financing, the financing process was fascinating. This is where the “everything's negotiable” lesson came from. Ron Burkle and Mike Ovitz were the initial investors in the business. We were in L.A., at UCLA, so we weren't quite wired into the entire Sand Hill Road scene.

As we were doing the deal, the terms kept changing on us. We thought you went and raised money and got a few million dollars at a $5 million valuation. This was back when that was actually a Series A valuation. Over the course of the deal, it was, “We're doing the deal. We're not doing the deal. Oh, you should give us 50% of the company. Oh, you should give us 75% of the company. If you want to sign the document today, this person's going to show up for breakfast, and if you don't sign today and give us 80% of the company, the person's not going to show up.”

The things we saw from day one of what can happen in business were completely wild. We thought there was a way to do things, and at a very young age, we realized there is no way to do things. There is just the way that you can negotiate your way through the world, which I actually think influenced Travis heavily and then me later, heavily, at Uber. If you can imagine it, it makes sense, and you can align incentives, then it can happen, but there is no way.

I think learning that at 19 or 20 years old was highly imprinting.

That is an amazing lesson. What happened to Scour? It got shut down, I think.

What happened there?

Well, yeah. So, basically, what Scour was is a multimedia search engine and then a peer-to-peer file-sharing network. But what it was used for was finding free content. At the time, the laws around this were pretty ambiguous because mixtapes were legal, but this was a hyper version of that.

We were eventually sued for $250 billion. I guess if you're going to experience something that's potentially as life-devastating as that, doing it when you're 21 or 22 is the time to do it. But it was this very cold splash of water about how the real world really works.

The MPAA and the RIAA were the ones who sued us—the entertainment industry, or the associations that represent the entertainment industry. Then they settled it for $1 million. So we're like, “Wait, you wanted $250 billion, and then you settled for $1 million?”

Of course, they were just trying to drive us into bankruptcy and drive us out of the market. These are established companies, so we're like, if these guys don't have a playbook to follow and they just make up numbers, then wow. How should we navigate the rest of our lives?

2. Scale Stays Independent

Let's talk about Scale and this whole world of AI that you're in. This is the first interview that you're doing since taking over as CEO at Scale. I'm honored to have you here to talk through this stuff.

Yeah.

This is also the first interview you're doing since the whole Meta deal, which was very complicated and confused a lot of people. I'm just curious to hear the current state of Scale and what people should know. For example, what's your relationship with Meta? What's your relationship with Alex? What is the current state of Scale?

Yeah. So, Scale is a fully independent company. The transaction was that Meta invested a little over $14 billion to get 49% of the company in nonvoting stock. They didn't take a new board seat. Alex fills the board seat, so the board is the same.

The governance is largely the same. There's no preferential access to anything that Meta has, and there's no preferential relationship. We've had a longstanding relationship with Meta on the data side of the business for a long time, and even on some business-development-related things, to maybe work on things in government together, et cetera.

Those things might get bigger just as we're closer now, but there's nothing that prevents us from doing things with other parties, and they have no access to anything that they wouldn't have had otherwise. All the privacy is still in place, and all the data security is still in place—that was there before.

In fact, only about 15 people went over in the transaction. Scale has about 1,100 employees or so now, and we have 2 major businesses. Each of those businesses has hundreds of millions in revenue, so we kind of have 2 unicorns inside the company today. That sustains. The business has grown every month since the deal happened, which I've read the reporting is not consistently reported. We haven't talked about it, right? So this is part of getting the word out.

We're excited to continue to build, deliver data, and do what we did before.

Okay. So the company today is independent—its own company. Alex, just to be clear, he works at Meta now. He's no longer at Scale.

Yeah, yeah, yeah. That's right. Excuse me. I should have talked about that more.

I think that's really interesting. So, basically, it was an investment, some people left to join Meta, the company continues, and you're running the ship.

3. Expert Data Trains Models

Let's talk about this whole space that you guys essentially pioneered. I don't know the best way to call it: data labeling, training data, creating evals for labs. You guys were at this before anyone even knew this was a thing.

I know even Scale pivoted into this market from other things. I think there was a bunch of stuff they tried with self-driving cars and all these things, and then it was like, “Oh, shit, AI labs need this data.”

One of the main stories I've been hearing—and I've had a bunch of CEOs from this space on the podcast—is that there's been this big shift from what Scale had pioneered and had been doing for a long time, which is generalist, low-cost labor training, to now, when labs mostly need experts: lawyers, doctors, and engineers doing training and writing evals.

I'm curious what you're seeing, how that's impacting you guys, where you think things are heading, and what people should know about this whole market of data training and data labeling.

Yeah, totally. I think the current positioning out there from competitors is just bogus. I'll explain what I mean by that in a second, but I think it's important to give 30 seconds on the history of Scale and what's been the thread going back to 2016.

Alex had this insight in the very early days that the important thing to models was data. I think he was 19 or 20 years old at the time as well, and he was like, “Okay, well, what business would I create around this?” The business that he created around it was labeling for autonomous vehicles, because if you label the data that they have, the cars do better.

That wave turned into the computer vision wave, and we have a relationship with the Department of Defense where we do labeling for them. That was in 2020. Then you move forward, and the models have gotten better over this period of time. As models get better, they need different types of data, so we've constantly been adapting to the type of data that models need to be successful.

Then the GenAI wave hit, and it went to the moon. As part of that, the industry is changing constantly, too. It is correct that when the models came out 2 or 3 years ago—we remember using them—they would hallucinate all the time and get basic answers wrong. They didn't know which poem was better, this poem or that poem. That was the state of labeling a couple of years ago.

Things have changed quickly, and we've changed with them. Now the state of things for everyone—and we've been at the forefront of all of this—is expert data labeling and more sophisticated tasks.

To give you a sense of what the task was 18 months ago, I've been here about 13 months, so I was interviewing and I remember seeing it. You would get a short story, and it would say, “Is this short story better than this short story?” Then you would edit it and say, “Yeah, it would be better if it was this,” and give some preference ranking to it. It was pretty basic 18 months ago.

You had the rise of some experts, but the models were so far behind that they needed even the basic stuff. Now you're at a point where one task is building an entire website by one of the world's best web developers, or it's explaining some very nuanced topic on cancer to a model. These tasks now take hours of time, and they require PhDs and professionals.

To give you a statistic to back this up, 80% of the people that we have in our expert network have a bachelor's degree or greater, which is very contrary to some of the positioning that's out there and some of the understanding of this industry. About 15% have a PhD or greater, and we have PhDs on the network earning significant amounts of money doing labeling and contributing their expertise to these models.

We've been doing expert data labeling ever since the models needed it. This game is about keeping in touch with the researchers, knowing what they need, and coming up with ideas internally. In some ways, we drove this because we were seeing that the models were not sufficient in more expert ways.

We would go to the model builders and say, “Hey, we noticed that this is a problem. If you would like to fix it, this cadre of experts can do that for you.” So, the counter-positioning that's out there is just what competitors say sometimes. It has nothing to do with reality.

Okay. That was extremely interesting. What I'm hearing is, yes, there has been a big shift to labs needing more expert folks involved in training, labeling, and writing evals. You guys are very aware of that and have evolved with it.

One of the allegations, I guess, in the market is that it's hard to find these experts. All these companies have their proprietary networks of experts and ways they find them. Is there anything you could share about how you guys go about that? That feels like the hardest part: finding these experts and keeping them from other companies.

They are hard to find. You have to have many tactics. As you would expect, there's not one way you do it. The largest way is that they refer each other, because when you're enjoying what you're doing and using your expertise to contribute to AI, which is pretty cool—if you're a PhD on this pretty specific topic and you're using a model and frustrated that it doesn't interact with you in the way that you want—this is a paid way to have an outlet for that and to make hundreds or thousands of dollars doing that.

And so a lot of times they refer each other. We also have campus programs where we'll literally go onto the campus and talk to professors and students, asking who would like to do this type of work. Then, of course, there's the more traditional scaled ways, like LinkedIn and places like that. But the best ones come from these grassroots and referral networks.

The only way you get that is by providing a great experience to these people, because they're doing it partly for money, but they're also doing it because they think that their contribution to the AI models is important and interesting. Many times, it solves a problem for them.

Something I've been seeing on Twitter just this week as I was preparing for this is a headline from The Information that came out. It mirrored something that Brendan from Workhorse said: Over time, the entire economy is going to move toward just reinforcement learning, and everyone's just training AI. That's basically the jobs that will be left.

Thoughts on that? Is that where you think things are going? Is there another perspective?

4. Agents Learn In Environments

Reinforcement learning is very important. I think this is a broader comment about the move to environments. There are these things called RL environments that effectively are sandboxes for AI agents to play in to accomplish a goal so that they can learn how to accomplish that goal. We've been doing this for over a year.

For example, you have a Salesforce instance. How does an AI agent navigate that instance? That instance has data that it needs to recognize, and it has configurations it needs to understand how to navigate. Salesforce is a highly configurable product. You're asking the agent to do a business process that needs very high reliability.

Then the agent needs to know, “Hey, if I can't accomplish what I'm going to accomplish, or if I think there's a low accuracy of what I'm about to accomplish, how do I pop it up to a human being for feedback so I can get guidance?” All of those things need to be trained, and there's no alchemy to it. You just have to put the AI agent in an environment that represents what a human being would be doing.

You can imagine that the number of environments in the world and the number of goals within each environment is enormous. The question—and the research that we have done over the past year to try to be a good partner to our model-builder customers—is how generalizable each individual task or environment is.

If you imagine the world of software systems, configurations, data types, sizes, user counts, and complexities, the permutations are endless. What you need is a strategy that allows a lab to collect data that is generalizable enough across a broad spectrum of use cases so that they don't have to collect 45 trillion combinations of what the agent should do in a particular situation.

Sometimes the work and the data are highly generalizable. By generalizable, I mean that you have a task accomplished in a simple way. The task might be, “Find the meeting on my calendar for my interview with Lenny.” The agent goes and looks through all of my calendar, and then it pops it out, right? That's a very simple example.

That needs to be generalizable to any calendar search, potentially, or any calendar action. The more generalizable it is, the more valuable the data is. Our job is to provide the most valuable data to model builders that accomplishes the goal of making the agents as useful as possible for their end users.

I love that you've been sharing these examples of specifically what these people are doing and the data you're providing to labs. Just to mirror back a few of the examples you've shared: One is an engineer building a website, sharing the code essentially with the model, and saying, “Here's how I would do it.”

In that example, is it just, “Here's the code,” or is it a recording of them building it? What is the data?

Yeah, it could be both.

Mm-hmm.

In some cases, it's just the website: Here's an example, and then they design it. In some cases, it needs to be annotated in such a way that it's like, “I made this decision for this reason,” or, “This decision for that reason,” or, “Here's how I would think about it.”

So it depends on what the model builders are trying to accomplish, and it can get quite nuanced in terms of what they're trying to train on.

Got it.

It's not like, “Here's a website,” and then it's great at doing websites. It's like, “Here's a website, here's why I made this decision, here's why I didn't make this decision,” or, “Here's a broken website and here's why it's broken,” if they're trying to accomplish a debugging tool for a website builder or something like that.

Another example you shared is a short story where it's like, “Here's one short story, here's another.” I imagine they're generated by a model, and then it's like, “Which is better?” And then, “How would you make it better?”

The other example you just shared is the Salesforce agent where it's like, “Hey, book a meeting with a prospect,” and then teach it how that happens. I love just how concrete these are, because it's like, okay, I get it—this is the stuff that these companies do.

Is there another, maybe one or two examples, just to give people a sense of what this data looks like?

5. AI Enters Healthcare

Absolutely. I can actually give you an example from our other side of the business. We have 2 sides of our business. On 1 side, we supply data to model builders; we sell the data. On the other side, we actually do solutions. We sell applications and services to healthcare systems, insurance systems, et cetera.

I actually think it would paint a more colorful picture if I gave you an example of one of those, because it involves data, but it also involves the use and manipulation of data for a very specific goal.

One example is that we work with a healthcare system. Health systems have lots of problems. This particular healthcare system has experts who see very rare cases on a regular basis. You go there only if no one else can figure out your problem, and there's a huge backlog.

There's a productivity element to this implementation, too. They want to be able to see more patients, provide better care, and prevent revisits because they want to give the accurate diagnosis on day 1 and determine what the treatment should be.

To do this today, without the help of AI, the doctor really needs to read 200 to 300 pages of documentation. It's rolled into 1 document, but in different formats. If you're a doctor, how are you going to read 200 or 300 pages of everything?

What they do is the best they can. They scan it, ask a nurse to look at it, or ask a more junior doctor to take a look at the case because they want to treat the patient well. Obviously, this is why they became a doctor. Then they go into the room, talk to the person, and make a diagnosis.

We basically built a tool that will read that document for them and point out the top 5 to 10 things they should take into consideration. One example is allergies that might not be obvious. We picked up on an allergy that a patient had that would not have been obvious from reading the document, and that allergy would have conflicted with the medication they were going to be prescribed.

The AI tool basically pulled out this correlation that would have even been hard for a human being to identify. To make this tool better and better, you get to a certain limit with off-the-shelf models, and the people inside this healthcare system have to do their own labeling.

We talk about labeling for model builders, but we're starting to see the labeling move into enterprises and governments because you can only get so far with off-the-shelf models plus RAG plus some fine-tuning based on recorded data.

One thing people often miss about these systems is that we assume, because you hear these numbers—like, “Oh, this bank ingests 200 petabytes of data a year,” or whatever fantastical number—that it's all useful. What we miss is: Is it the right data? Which of that data is useful to the models? Most of it is not useful.

Some of it is, but a lot of what we do when we're talking about knowledge work, when we're talking about making judgments, is human judgment based on synthesizing: How would this doctor in this case, or how would this banker in this case, make this decision? And how would they make the decision in the context of their overall enterprise?

That might be different from bank to bank and healthcare system to healthcare system because of the culture, the objectives, the incentives, et cetera. So we're getting to the point now where we see that digitizing judgment—human judgment and true subject-matter expertise—is becoming a bottleneck that we're unblocking for our customers.

That’s really interesting. It’s like the spectrum went from low-skill generalist labor to experts, and now to the specific expert at one company who needs to do this labeling.

Absolutely. There’s this broad narrative that we have 2 narratives: we have AGI, where everything is just going to become AGI, and then there are the skeptics who say, “Hey, this is all bunk. This is a bubble,” et cetera. My view is that most things are somewhere in between, and some of the extreme parts are probably correct.

The reality is that it’s very hard to get mission-critical use cases in agentic systems—where agents are talking to agents—to a level of accuracy that’s necessary to accomplish a goal. One of the main issues is that even a document that reads with the exact same words in Company A will have a different meaning and importance in Company B. So how do you have a system that knows that? All of this has to be built if you’re going to make good decisions.

This is a good segue to a question that’s always on people’s minds when they look at companies like yours and the other folks in this space: How long do we need people to be doing this? At what point will AI be smart enough to do it itself?

I know your incentives are to say we’ll never run out of people because that’s aligned with your growth. But how should we think about why we need people, I don’t know, in 10 years? How long do we need these experts telling AI things it doesn’t know?

6. Human Judgment Remains Essential

First off, the history of data labeling is a history of new beginnings. Autonomous vehicles, for example, do not need as much data labeling as they did in the past.

Scale is a company that believes data will always be important. At the point at which you don’t need external human data in models, I think we’ve gotten to a level of advancement in the world that is almost unfathomable, because you’re effectively saying that no new human skill and no new human knowledge is important enough to put into these models. That feels pretty far out there.

For a business like ours, we’re constantly looking at how to build operations that can find the new needs and then work with the contributor network—we call the experts contributors—to unearth that data and information. Sometimes it’s new people. Sometimes, within our existing base, we find that existing people have expertise we didn’t know about that maybe wasn’t useful to a model a year ago but is useful now.

This is a constant progression of getting more and more data into these models. Yes, we are financially incentivized to believe that humans will always be in the loop, but that’s not just a business belief; it’s a personal belief. These systems need to work for us, and if they work for us, then we’ll need to be in the loop on any of the decisions these systems make.

As to the broader point around labor, which comes up with the white-collar apocalypse and these things, I’m definitely on the more practical side of this—possibly just because of my nature, and possibly because I see what’s going on on the ground with customers where this transformation is supposedly going to happen in the next 1 to 2 years.

It might happen. The space is moving super fast, but I don’t think it’s going to happen in the next year. The idea that it happens in the next 2 years, I think, is very far-fetched, but nothing’s impossible here.

Long term, if you go back through the Pessimist Archive or whatever—these accounts that post, “The radio was invented,” and then, “Hey, all of this will be eliminated”—there will be change. But humans are very good at adapting, so I think what we’re underestimating in all of the doom and gloom is human adaptability. We, as a company, are highly adaptable, and I think the history of technology has shown that people are adaptable.

I really like that takeaway. I’m an optimist as well, so I’m always looking for reasons to be optimistic. I want to follow that thread, but before I get there, something very tactical I want to ask about is evals. They seem to be coming up a lot, especially with companies in your space. I’m still learning a lot about what this all is, especially in your market.

How much of what you or the experts are providing consists of evals versus other types of data?

Yeah, a lot of it is evals. Within enterprise customers and government customers, it’s mostly evals because somebody has to establish the benchmark for what good looks like. That’s the simple way to think about evals: What does good look like? Do you have a comprehensive set of evals so that the system knows what good looks like? It’s as simple as that.

So, in the case of the healthcare example you shared, essentially this doctor would be sitting there looking at all these reports, creating evals that are like, “This is what this should be discovering in this report, in this record.” Is that a way to think about it?

Yeah. That’s a very big part of it: What does good look like?

Awesome. Okay.

I have to reduce things down to simple terms.

Yeah. It’s interesting that you say “good” versus “correct.” Is that a specific term you like to use—good versus “This is the correct answer”?

I didn’t intentionally use that word, but these are probabilistic systems. I can get into some nuance here about the right types of problems that AI is good at solving.

If you have a human process that is only 10% or 20% accurate, AI is awesome. If you get to 50%, 60%, 70%, or 80% accuracy, you’re in the money. You’re in the green. Everybody’s happy. The system then has to know, “Hey, for the remainder, how do I make sure that humans are involved in the remainder of the decision-making?” But from a net-value-add standpoint, the humans are thrilled in that scenario.

If you have a human process or workflow that is 98% accurate and you expect an AI system to get you the remaining 2%, we’re not totally there yet. When I say, “What does good look like?” a lot of the processes and things that people are asking these systems to do—and asking us to build systems to do—involve making judgments on their behalf.

Just as we would ask a human being, “Hey, what do you think we should do in this scenario?” what you’re looking for is the best recommendation or course of action given the current information.

To you, this is so obvious, and to people in your market, that I think a lot of people think about AI being trained on just, “Here’s a bunch of data. Check it out. Learn everything you can from all of human history and all of the written record.”

But what’s wild is that people are basically sitting around teaching AI things it doesn’t know and filling gaps. That’s how AI is getting smarter now. There’s no more real data for it to feed on. It’s just like, “Here’s what I don’t know.” Or, “Here’s what an expert found. You’re wrong. I’m going to teach you this.”

The fact that this scales, and that’s what’s keeping models improving, is so mind-boggling.

Yes. With any of these major tech revolutions, the headlines tell one story, and then on the ground, laying broadband means you need to dig up every single road in America to lay it. Someone has to dig up the road, or someone has to run the undersea cable. There’s always some operational chiseling going on in all of these industries.

If you think about how magical these models are, they’re remarkable. If you’ve been in technology long enough, it blows my mind even today that they get the punctuation right consistently. That sounds almost daft to say at this point in the market, but if you were to go back 3 years and think about that from a technological standpoint, a lot of things that we think are trivial now are very sophisticated.

The real answer is that it’s a combination of computational power, model improvement, and data, and all 3 are getting better at once.

Let’s follow that thread. You’ve been at Scale for a long time and, you said, CEO for 13 months. I feel like you see a lot more about where things are heading because you work with labs on things they haven’t even announced yet. You see more than most people, and I know there’s only so much you can share about what models and companies are doing.

Is there anything you think people don’t truly grasp or understand about where AI models are going to be in the next 2 or 3 years?

7. Models Start Doing Things

Look, there’s so much talk. I think it depends on how much X or news you consume. Our perspective is that the general trend right now is going from models knowing things to models doing things.

We're pushing the boundaries of knowledge. The benchmarks that we put out and that others put out are showing that the knowledge these models have is getting quite robust. The next question becomes, “Well, what can it do for me?”

As soon as you get into that world, the environments we were talking about start to come into play. How do you navigate a Salesforce instance? How do you navigate a healthcare system? How do you navigate even a weather app on your phone? How does the agent make decisions for you?

We're just getting into the beginning of that. It'll be very interesting to see how quickly that happens, and I think that's where a lot of the speculation has a wide variance. Because we're at the beginning of it, people take different trajectories on how that's going to improve.

If you take the most aggressive trajectory, it's, “Oh, it's actually going to be quite easy to train on these things,” and then it's just a change-management exercise in the economy. By the way, change-management exercises are not to be underestimated. There are still people in the world without an email address, so the adoption curve then becomes a human and policy issue, not a technological issue.

We're not there from a technology standpoint, but I do think in the next 2 to 3 years, if I take the bait and have to make a guess, the technology will get to a point where it will push change management and policymakers to say, “What do we do with this? Because it's getting pretty close.” That's probably 2 or 3 years away.

There's been a lot of talk these days about AI not delivering on the promise that we hear, especially at enterprises. There's this MIT study that showed there are all these pilots that people are excited about, and then they don't work, and companies aren't adopting these tools. There's data showing engineers are not actually as productive with tools like Cursor; it actually slows them down sometimes.

You work with a ton of companies implementing all kinds of AI. What are you seeing on the ground? What kind of gains are you seeing? Do you feel like it's overhyped or underhyped?

8. Enterprise AI Takes Time

There's a lot of hype out there. Our job is to actually build products that work, deliver value for our customers, and figure out where the rubber hits the road.

To get a sophisticated workflow—you know, MyHealthHer example is one. We do other sophisticated workflows, such as claims management for insurance companies. This is a financial decision that's happening, but it's an automatable process.

Basically, what happens is the POCs get to 60% or 70% of the way there, and the human mind goes, “Oh, the rest is no big deal.” But it's kind of like uptime in data centers, where every additional 9 is an order-of-magnitude investment in reliability, backups, and so forth. One 9 is basically a web server in a dorm room, like we had at UCLA. Five nines is a crazy high bar, but it just seems like a very small movement.

You have a similar dynamic going on here, where you have a bunch of people—one of the reasons why the POCs have failed is that there's a denominator effect. It's so easy to say, “I spun up a project. I spun up a project. I spun up a project.” It's really easy for people to try, so I don't necessarily know that the 95% number is accurate. I think it's a bit of clickbait in a way. It tells the right story, but it is a little bit hyperbolic.

If you take the efforts that happen in a company where they get a quality partner like us—or if they do it themselves with engineers who have worked with models before—and they put in the time, I'm talking about months, not minutes like you see in these videos, to actually get legal approval, policy approval, regulatory approval, change management, and an accuracy level that everybody's comfortable with, these things take 6 to 12 months to get truly robust enough that an important process can be automated.

I think that's where the hype is right: When you do it, the impact is, “Whoa. I never would have figured that out myself, and I'm one of the most educated doctors in the world,” as an example. But the time to get there is just longer than what people are selling.

It's such a good point that it's not only easy to try these things, it's just that everyone's doing it, so everyone's feeling FOMO. “I have to try these things. I have to try all these prototyping tools, Cursor, all these things,” just because everyone's doing it. Then you rush into it, and it doesn't actually work out.

Easy to learn, hard to master.

Mm.

That's my summary.

Yeah.

9. Customer Insight Beats Assumptions

Okay, let's move on from AI. This could be an endless discussion about AI, but you've got a lot more lessons to teach us. You've helped build Uber Eats. You've had a couple of startups in the past. We talked about Scour for a bit. I've talked to a bunch of people who have worked with you over the years, and I got a lot of really interesting insights into the things that you're extremely good at.

I'm just going to go through a bunch of these. One is your obsession with being close to customers and talking to customers.

Mm-hmm.

I love this topic because it's something everybody thinks they're great at, and they feel like they completely understand how important this is and why it's important. They all feel like, “I'm doing this. Don't worry about it. Everyone else is not doing this, but I am.”

Talk about what you think people might miss about how this looks when you're doing it well, and why this is so important.

I probably fall in the category of what you just described, which is maybe part of the hubris you need to start anything new. But I don't think it's a clean process. I think my process is that I'm constantly questioning every single thing that I'm hearing at the beginning of anything. I don't take what a customer says literally.

There's been a lot talked about on this topic from a product-management standpoint in terms of, “Don't do what they say; do what they mean, and look at their real problems,” and all these things. I think the way that I look at it that might be additive to the discussion is that I look at the underlying incentives of the customer.

The underlying incentives of customers are not always financial. Sometimes it's ego. Sometimes it's career growth. If you're selling enterprise software to someone, there's an executive sponsor, as an example. That person needs to trust that you're going to do a good job for them. How do you get them to jump with you on this big project?

That's part of the journey—not just the product, but what do they need to hear from us? What do we need to supply them? What do we need to do to actually unlock the opportunity to implement the product?

I think there's an incentives-alignment baseline. I'm a big believer that, clichéd as it is, “Show me the incentive and I'll show you the outcome.” I think that's absolutely true, even when customers will tell you things.

I'll give you an example. I've been out of the game for a while, so I can be open about it: Uber Eats. When we launched Uber Eats, I looked at the business. In terms of being close to the customer, we actually couldn't get a restaurateur to help us understand their unit economics. I knew nothing about this industry.

At Uber, my job was to figure out what other businesses we should get into, and so we looked at a billion businesses. Uber Eats—food delivery—was the one that we thought was most interesting, which turned out to be right, so good for us.

Very right.

We couldn't get a restaurateur to help us understand their unit economics. They'd say, “It'd be this percentage or that percentage,” or, “Why do you want to know?” Then we'd go to a different restaurateur, and they would explain it, but they were a little suspicious of why these Uber guys were talking to me about how much my ham costs.

So what we did was order a bunch of food from these places. Then we got a restaurant supplier to give us a base catalog, and we matched up how much the ham weighed, how much the cheese weighed, how much the bread weighed, and how many pieces of lettuce were on there.

We tried to compose our own independent view of what the ingredient cost was versus what the labor cost was. Then we triangulated what our ground truth was, what restaurateurs were telling us, and what the site guys were telling us about restaurant economics.

If those things all overlapped, then we would say, “Okay, we have an insight about what to do here.”

How does this relate to Uber Eats? What we found as part of this is that a restaurant pays roughly 20% to 30% of every meal for ingredients. They pay roughly 20% to 30% for labor, and they pay roughly 10% for real estate and a bunch of other costs.

Anyway, so it kind of goes down the chain. But the important part is: What’s the value of incrementality? We came in and said, “We’re going to charge you 30% of the bill.” They were like, “Oh my God, is this Groupon all over again? This is way too high. Oh my gosh.” We explained the economics to them, and they were like, “Okay, we’ll give it a try, but this is way too high.”

They were right. The real number—the real clearing price—is 25%, but we weren’t that far off. When you go to find product-market fit or be close to customers, it’s a combination of asking, “What’s the most valuable thing?” In a restaurateur’s case, it’s, “Give me incremental demand.” If you were to take a restaurant location and triple demand based on the same labor, but you’re just scaling ingredients, you’ve got a 70% to 80% incremental gross-margin business.

Restaurateurs would hate when we would say this because it doesn’t work out exactly like that in reality. But because we had that insight, we had confidence that we could go to market with, “We need to charge you this so that the delivery fee can be that. And then if the delivery fee is that and we charge you this, then we think the consumers will adopt, and that’s what you need to get your incremental demand. And then we could pay the driver this.”

So you kind of fit this whole puzzle together without totally satisfying any individual’s 100% of their needs. What you’re satisfying is getting a clearing rate for them to participate in the market, in the case of a marketplace. So that’s one example.

Yeah, I love this example. You figure out how to help them with something they don’t even fully know themselves yet. You think through their goals for them, as if you were them, break down the economics, and then here’s the solution versus, “Hey, what can we do for you guys?”

Yeah. If you walked into a restaurant, they would tell you a bunch of things. They would say, “Labor scheduling is an issue.” They would say, “My rent is an issue.” They would say, “My ingredient prices are an issue. That’s 20% or 30%. If you could shave off 3% of that, that would be huge.”

You might then take that and go, “I’m going to go build a business that’s going to save you 10% on your ingredient costs.” But that doesn’t actually get into their head about what’s truly important day to day. That might be important for them on an annual basis, but on a daily basis, what are they doing? They’re looking at their numbers. They’re looking: Do people show up? How did I make money yesterday? Am I going to make money tomorrow?

The urgency, I think, is the biggest thing people miss when they’re building new products. You can build something that provides a lot of value, but if it’s not the top thing that the customer is thinking about in their busy days, then you’re just going to have a long road to a small TAM.

Hmm. This touches on just the theme I heard a lot about you: this idea of independent thinking and how much you value that. This feels like a really good example of that. Is there anything else along those lines of why this way of thinking is so critical?

I think a founder’s job—and I stretch that term because at Uber we had all the benefits of Uber, so I wasn’t really a founder. I just started the business there, but there are some elements of founding there—is that you’re looking for alpha in the market.

When we started our first company in 1997, it wasn’t that cool. It might have been cool in Silicon Valley, but it was definitely not cool in LA. Now it’s super cool to start a business. As a result, everyone’s trying everything. So how do you get alpha in that market? If your research is highly influenced by what the world is saying around you, you’re not going to have an independent insight. You kind of have to go off and do your own thing.

This is why, from an entrepreneurship standpoint, I have very strong feelings about what the approach to founding a company should be. It’s probably very particular to me, but it truly is about: What insight do I have? Why am I so lucky to have this insight? Why, in a world of a million entrepreneurs who are thinking, who are smart, who are trying everything, am I in the position where I likely have an insight that others do not? And then why am I the one to do it?

The answer might be that I’m in this narrow, far-flung place. The other answer might be that I’m inherently a contrarian personality type, so I’m constantly looking for the thing that’s true that people don’t believe is true, which sometimes works. But the second part of that is super important: Why do I want to work on this problem for 5 to 10 years?

People get this wrong all the time. They go and talk to a customer and say, “They have a problem. I’m going to go solve it.” It’s just not a great way to start a business. You really have to have this burning desire to constantly be questioning yourself.

The other thing about independent thinking is that you can’t fall in love with your ideas. I do not proclaim to be the world’s greatest independent thinker, for what it’s worth. This is what you’ve been told. Part of that is basically throwing away who you are, who you’ve been, and all your ideas for the mission that you’re on, which is trying to accomplish something for a customer.

This is great. I’m glad you went here. This touches on the other theme I heard often about you: just how high of a bar you set for new businesses. I think this advice is useful both for founders, as you said, and also for people starting companies within companies and new business lines. You’ve talked about this a bit already, but is there anything more there about how high that bar needs to be for it to likely work out when you’re starting something new?

Look, if you want to give yourself the best chance—and this isn’t always how it works—but if you’re in my position, 25-plus years into your career, I think there are 2 ways that companies end up working out.

The first way, which is probably the most important, quite frankly, is that the founder is just a force of nature over a long duration of time. You’re going to have to pivot. You have to have that energy to pivot, and you have to go years and years and years with it being hard. That’s probably the most important thing.

The second most important thing is that you can educate yourself on what good business models are, what bad business models are, what good markets are, and what bad markets are. Even if you’re this force of nature, if you’re going to go into a bad market with all your energy, you should at least know. Maybe ignorance is bliss because you just throw yourself into it and it kind of works out with time, but that’s not how I would operate.

The point is: marketplaces are good businesses. SaaS, at least historically—we’ll see how this changes—is a great business. Recurring-revenue businesses, sticky businesses, and network-effect businesses are good businesses. If you look at what the top VCs invest in, yes, there is a lot of portfolio building, but there are similarities in terms of the types of business models that they believe could be worth tens of billions of dollars. They have network effects. They have lock-in. They are more valuable at big scale than at low scale.

So if you just have a filtering mechanism on a new business, like I did at Uber, it doesn’t take that long to eliminate the bad ideas. Then, of what’s left, you can pick, “I’m very passionate about this,” even though it might have more problems than this other thing that, on paper, looks better. And then you have to have passion about it.

But I think people just miss a basic understanding of what businesses even have a chance of being worth $100 billion.

So you launched Uber Eats, and you figured out this was the place to bet. As an outsider, it feels obvious: Of course this is going to be a massive success. Of course food delivery is such a good idea. I know you looked at a ton of ideas in that process. Can you talk about what you explored and why you ended up picking Uber Eats?

I am definitely not the smartest person in the room when it comes to figuring these things out, so I keep a very wide aperture on ideas for as long as I can, until I’m thinking, “Okay, everything is coalescing.” There are a bunch of reasons why you have to keep an open aperture when considering ideas that might seem bad at the start, but you just keep digging to see if you’re right that they’re bad or whether you’re wrong.

As a general philosophical principle, I’ll start there. We looked at some crazy stuff. One day, I went walking around San Francisco and looked down Market Street, and there was a CVS, a 7-Eleven, a CVS, a Walgreens, and a 7-Eleven. I thought, “How many SKUs could possibly be inside one of these things that people want? And couldn’t you just put that into a van and hit the button on the van, and the van comes around and you get whatever convenience items you have?”

They’re convenience items, so why would that be a problem? We launched that in DC. We put 10 of these trucks on the road and put 250 SKUs in them. I mean, “Crickets” is an understatement of how bad it was.

We couldn't get an order to save our lives. What we realized was that we hadn't really done the research on what convenience stores really were. If you didn't have cigarettes, beer, Slurpees, and these things, you didn't bring people in to sell all the other things. We didn't know anything about retail. We were clueless.

So that's one idea. We looked at grocery, but honestly, the unit economics terrified me—the pick-packing and everything like that. I think Instacart did a remarkably good job of getting the unit economics to a good spot, and it was probably the hardest operational problem you could tackle.

We did generalized delivery—point-to-point delivery, what's now, I forget what Uber's product is called, but Uber Direct, I think it's called—where you have something that needs to go point to point in a city. That was kind of a flop from the beginning because the truth is, consumers don't really have this need. Businesses sort of have this need, and in 2014, when we were doing this, no one had this need.

We tried 15 versions of all these things before we eventually just said, "Okay, the food delivery thing is popping off on all signals, and we can make the unit economics work. People seem to want it." It's a super cool problem because we can enable independent restaurants with all these tools and allow them to compete with the big guys. We can take the real estate out of the equation, so you can have a real estate location that's non-prime, but if you have prime food, then you get to compete. We were like, "Oh, this is a very interesting problem, and we can really help local economies."

This ended up being, if I remember correctly, what basically saved Uber during COVID. Lyft didn't have something like this. How big is this business at this point? Is there anything you can share about how important this turned out to be for Uber?

Of course. We launched it in December of 2015 in Toronto, and within 2 hours, we had done $20,000 worth of sales. It was crazy how quickly we saw that it was the right idea. The unit economics were good.

I was at Uber for about 6 years, but it took us about a year and a half to figure this out. Four and a half years later, it was about $20 billion, so it was $0 to $20 billion in 4½ years, which is pretty good. Uber was very good at scaling things.

It was a competitive market. Others did well. We beat a lot of people, and some people beat us. Now I think it's pushing $80 billion, and that's been for another 4½ years since I left. I think COVID turned it from $20 billion—I left right before COVID, total coincidence—to $50 billion in about a year. The ridesharing went this way, and food delivery just went to Pluto.

What luck. Well done.

Luck is part of the game. That's the other thing that's important to realize. Luck is part of the game, so do not begrudge people for luck. This industry is hard. All these things we're doing are really, really hard. Luck is just part of the game.

One of your colleagues, Steven Chao, who I am an investor in his new company—he worked with you at Uber Eats for a long time—told me to ask you about the McDonald's story. I imagine that was a big milestone, a big moment for you guys. Why did you decide to put McDonald's on Uber Eats? There's apparently a story about how you won that deal.

It was kind of interesting. This just goes to maybe where ignorance leads you to accidentally find the right answer.

We had launched Uber Eats, and Uber had a global footprint. We were the only food delivery network with a global footprint, excluding China. Everything at Uber needed to be launched globally. That was a very big part of the culture, et cetera. It was a lot of work, and you can spread yourself too thin and cause other problems. But in this way, it was good.

My vision was, "Okay, let's help the little guy compete with all these chains." They have these systematized food systems, and food is what makes a city amazing. No one talks about the chain restaurant that they visited in Paris. They talk about the local place that they found, and that's who we want to be part of.

McDonald's actually approached us and said, "Hey, we'd love to do food delivery with you." I said, "No." They were like, "Hold on a second. We have 80 million consumers a day. You don't want to do this together?" I said, "It's not really our vibe right now."

I pushed them off for 4 or 5 months, until my team was like, "You're insane. These people are going to put marketing behind it. They really want to do this. They want to lean in." Because of that, I think—it's hard to correlate these things—we ended up with this exclusive relationship with them and got an insane number of customers.

Chains at this point actually weren't really on food delivery networks because everybody was so worried about the unit economics. They're so sensitive to basket size. My approach was like, "Eh, figure it out." That's a very Uber culture thing: "Okay, the basket's $17. It's our job to make that work. Reduce the radius on the delivery, figure out the economics, maybe mark up some of the food in some places. There's always a way to figure it out."

We did it, and 3 months later, the business started hockey-sticking again at a different level. My team was like, "Dude, you were so stubborn on this point." I think it actually ended up being a net benefit because we got a great deal with them.

The fact that you pushed them off helped you get a better deal is what I'm hearing. That's amazing.

Yeah, I think that's the story he would be referencing. The onboarding of it was crazy because we basically went global with them in about 6 months, and at this point, the business was less than 2 years old.

Activating an 80-year-old company that expects processes to be in place, while we had 2 of our office managers in New York managing it, was just mayhem.

I'm still sad In-N-Out is not on any of these apps.

Yeah, me too.

I remember someone was hacking it. There were all these ways people found a way around it, and they were like, "No, no, okay, you're Postmates. We know. We're not going to give you any food."

Yes. Love In-N-Out.

You've touched on this idea of gross margins and margins, and how obsessed you are with this. I wanted to spend a little time on it. I've heard that you're obsessed with understanding gross margins before going in on anything. Most founders have no idea what they're doing here. What have you learned about what people should be paying attention to, what they might be forgetting when they think about the feasibility of a business?

Look, it's one filter among many filters. There are certainly businesses that have low gross margins that are great businesses—Costco, Walmart, et cetera. Amazon talks about this all the time: There are companies that increase prices, and there are companies that lower prices.

By and large, high gross margins combined with healthy churn curves are a very healthy sign for the business. Think about it: If I were to sell you something and I can't mark it up a lot, how much value am I adding beyond what's in my hand? If I'm not adding that much value, then what am I in the business of doing? I'm in the business of adding value.

It's not quite that simple, right? This is just a litmus test. When someone comes to me, especially with a new business—we dealt with this at Uber, and I've dealt with it everywhere—someone comes up with an idea and says, "We can get into this business, and I think we can charge this, and it will get us to a 40% gross margin."

My next question is, "Start at a 60% gross margin. Why does that not work?" They go, "Oh, well, the customer..." Immediately, you short-circuit to what the real problem is. "Oh, the customer has an alternative." "Okay, who's the alternative?" "It's some offshoring company." "What's their gross margin?" "We don't know." You go find out, and it's 20%. They've been around for a long time, and they have scaled operations.

You're like, "Okay, your gross margin's going to go from 40% to 20% quicker than you think, and you're going to be in a world of hurt unless you do something to differentiate." So I take gross margin as just a very coarse instrument, not a perfect instrument, to think about whether I'm adding enough value and whether I'm differentiated.

It's not perfect, but it's a very quick short-circuit filter to see whether someone pitching you an idea has thought through this dynamic. If the response is, "Yeah, gross margin's super low right now, but here's the dynamic I'm going after," then you're like, "Oh, okay."

Sometimes it's like, "We'll just make it up with volume, and then the gross margin will go negative for a while," and you're like, "Wait, wait, wait."

“This doesn’t work.”

So what I love about this is that this is just a lens into: Is my idea good enough if, structurally, I can keep a high gross margin? Is there a reason why people in this space haven’t been able to have a higher margin?

Yeah, exactly. It’s meant to disqualify ideas for larger companies, and everybody has ideas. It’s a way to cut through: Do you understand the machine that’s going to need to be in place in 2 or 3 years? You might have a 70% gross margin now, but the next question is, why can’t someone else do this?

And if you have an answer like, “Well, they can now, but they can’t in 2 years if we run really fast,” okay, we might have something. If they can now and they’ll be able to in 2 years, you’re going to have margin compression.

Along these lines, I was just listening to, I think, the a16z podcast. Alex Rampell was sharing this story about Costco: As you said, their strategy is actually to keep margins very, very low because all their revenue comes from their membership. They have something like 50 million members charged, paying 100 bucks a month, and that’s their entire business. They don’t plan to, and they don’t want to, make money off the products.

Yeah, that’s right. They’re playing a slightly different game. I’m not an expert on Costco. I’ve spent some time with the company, but they use price as a way to get to scale, right? They’re basically saying, “If we discount”—same with Walmart—“we’ll get so much volume that we’ll just take the air out of the room for all of our competition.”

Then the question is, okay, so if you have a low gross margin today, in 2 or 3 years, once you land one of these centers in a market, why won’t your margins get eroded? The answer is, because we will have already absorbed all of the demand.

You try to go to 8% versus 10% gross margin, which I roughly think is what their gross margin is, and that’s going to be a really hard business if you already have a habit with a customer. They’ve already built their weekly trips around you, you already have relationships with suppliers, and you already have general managers who know how to stock inventory. That’s not a straightforward exercise.

So they’re first to scale, and then good luck competing with them.

Hmm. Okay, just a couple more questions. One is, there’s this term that I’ve heard that you often say and believe in: this idea of not losing as a precursor to winning. Talk about that.

Yes, yes. Tech is a culture where portfolios are built by investors, and a lot of the narrative is controlled by investors, frankly. Founders obviously participate, but this idea that you should just go for it is consensus: Just go for it. Who cares?

Well, I don’t know. If it’s my life and I only have one moment to take a shot, I might not want to just go for it. I might want to think for a little bit. I think the best entrepreneurs—I have no data to back this up, but these are my friends, this is my friend group—and the best business owners look at the risk profile of the decisions that they’re making, and they try to make asymmetrically positive decisions all along the way.

Oftentimes, I feel like we forget about the risk of a decision. There’s more to unpack there, because I actually think taking highly risky decisions and then having them work out is a weird cultural thing, too. How do you train people to do that? It’s very hard to take high-risk decisions and be right enough, because it creates a lot of volatility.

But it goes back to my comment about the most important thing in founders, which is just this ability to persevere. Survival is part of the game, and most people just give up before they can get their timing right, before they get the right insight from the customer, and before they get the right product in the market.

Life can change quickly in tech. You can go from being a dog to being a hero in a very short period of time, but you’re on this very, very long journey, and you have to survive for that condition to be met.

So then the question is, when you’re in a hype cycle—which I would argue we are right now—everyone wants to go for it, then go for it more, then go for it more, and go for it more. You don’t realize, “Guys, all of our customers are going to be around in 5 years. They just want us to solve their problems. We have to be around to solve their problems for them.”

Survival is a precursor to that, so let’s not put ourselves in a position that could potentially compromise the enterprise along the way. It doesn’t mean don’t take risks, but think about how you calculate them.

I love how clear it is that this lesson, and many of the lessons along these lines, have come from failure and things not working out and things breaking, which is the best character lesson.

Yeah. Have you ever gotten on the other side of a high-risk, high-reward decision?

Mm.

It is so painful because you are just cooked. You are done. Often, there’s no way out.

Is there a story along those lines that comes to mind, or an example of that?

Well, this is where it all knits together with why I try to be so thoughtful. I think you can spend a little bit of time thinking up front to save yourself a lot of pain downstream.

I had this business—not worth detailing it—but after the bubble burst in 2001, I thought, “I’m going to self-fund a business. I’m going to build a profitable business. I want to prove that I can do this.” We had started Scour, which had all the things we talked about.

What I did was, I was a golfer, and frankly, there was nothing to do in tech. So I started selling used golf clubs on the internet, and I was making real money.

I might have learned more from this business than any other because I started on eBay, and I was 22. I didn’t really understand that my margins would come down because anyone could do this. But I was one of the first ones to do it, so I was making a ton of money.

Then I built this business, and I failed to recognize that I had a lot of hubris. I was like, “Oh, if I could just buy all the used golf clubs in America, I could be the market maker for prices.” Don’t people do that?

I love this ambition. That’s great.

Yeah, and it’s just madness to actually think about the practicality of that. I just didn’t spend the time thinking, and then I ended up in this business. The business was profitable. It got to a couple million in revenue and paid me a dividend for a while, but it was painful the entire way.

I love the spectrum of experiences you’ve had. You’ve sold golf clubs. You’re helping achieve AGI, you could say. There’s also a whole part of your career we haven’t talked about, where you built Tasers, body cams, drones, and all these things. You also did peer-to-peer file sharing before anyone else.

The final topic I want to spend a little time on, based on this experience, is hiring and building teams—something that I know you have a really strong take on, and that I’ve been hearing a lot about on this podcast recently. It’s this idea that it’s more important to build the right team than to find the most optimal top talent. Talk about that, and why that’s so interesting and important.

10. Hire Curious Problem Solvers

As of late, I’ve developed a more nuanced view of this, which is: For certain roles, you absolutely need the right experience in this current market. You see this with researchers, right? The market is moving so fast that you don’t have time to train people up, so you actually have to find people who either have the right relationships with the customers you want to reach or who might not check other boxes but are awesome at that.

They might not check the classic boxes that I think you’re referencing—they’re a problem solver, they can grow with the company, they have a high trajectory, et cetera. I would say that’s 5% of the roles in a company, but it’s very important whenever speed to market is important.

For interviewing, I just interview for 3 things, and I have to interview across all kinds of expertise, which is hard. I can’t be an expert in everything. So I reduce it down to just 3 things: Are you a curious problem solver, and can you articulate that verbally? Can you work across people? Are you humble enough to do that? And are you a good leader?

If you just do those 3 things, I think you have a pretty high chance of success, at least in an organization that I’m running. The world is changing, so you do need people who are adaptable. All the experience is not necessarily one-to-one relevant.

To your point about working across teams, this actually came up at Uber Eats. I’m not sure if this was mentioned to you by that group, but whenever I would hire people, I was trying to compose almost like an organism of strengths and then minimize the conflicts.

That management team, for the most part—outside of some of the operations side—was the same management team from day one, when we had nothing, to $20 billion. I just believed that the team knowing each other's strengths and weaknesses and being able to compensate for each other was more important than the classic advice you get around, like, “Well, that person hasn't seen this much scale.” And you're like, “Well, yeah, but can they learn it? I learned it.”

So you do have to believe in people a little bit, which is my job, not necessarily their job. These are people systems. They're not straightforward, rules-based things you can apply.

I especially love this advice because there's all this talk about what skills will matter in this world of AI doing all our jobs, and it feels like these 3 buckets are maybe the same thing: Are they good at solving problems? Are they good leaders? Can they collaborate well with other people?

Yeah. I don't think that the core of the rise of humanity will change. I think that these things are pretty core to how humans have been successful for a long time.

Speaking of that, I'm going to take us to a recurring segment on this podcast that I call AI Corner, where I ask folks this question: What's some way that you've found a use for AI in your day-to-day life or in your work that makes you more effective, helps you get more done, or helps you get better stuff done?

Honestly, when I came into Scale, my history was in consumer, and I'd done some application-level stuff with government. This space is moving so quickly. AI is my tutor. I use it as a tutor.

As these new concepts come up, I have a lot of people in the company who can educate me on the nuances of the technical aspects of the data and the products, but they only have so much time. There are new concepts coming up all the time, and I need to stay on top of them. It might sound crazy, but a large percentage of my job is not dealing with the engineering issues related to AI. I'm managing an organization, but I love understanding it.

It's one of the most enjoyable, rewarding parts of my job to learn from all these AI researchers, but they don't always have the time to do it. So I use it as a tutor. I turn on voice mode and talk to it on my way into work. I think that's probably the most impactful thing that I use it for that's also relevant to this topic.

I do exactly the same thing, especially when I'm preparing for these podcasts: “What exactly is this?”

I think about when you say this. I did an interview with the founders of Perplexity a few years ago, asking about how they worked at Perplexity, and the founder said that before they'd ask a question of anyone on the team, they had to ask AI first. I was just like, “That's crazy.” But now it's so obvious.

Back then, I was like, “That's an insane new way of working. I've never heard of this before.” It was just a sign of how ahead of the curve they were.

Yeah. I think number 2 would be taking the internal documents and asking, “What's the most important thing in this document?”

Hmm.

I'm shocked. Then I'll read it and double-check, but I'm shocked at how good it is at pulling things out. There's so much in organizations that's like, “I don't know what you want me to say, and I don't know what I need to know.” We each have our own agendas, and so there's this matching problem.

Then there's this huge broadcast problem where, of all the information you might want to receive, what's actually important to you? I use it a lot for that, too.

Amazing. That's a really good tip. I use it for legal documents, just like, “What should I know about what they're trying to do here for me or against me?”

Yeah.

Jason, is there anything else you wanted to share or leave listeners with? Maybe double down on a point before we get to our very exciting lightning round?

Yeah, absolutely. I think the really important thing—the reason why I'm doing this, the reason why I want to spend time here, outside of wanting to be on the show for a while and being a longtime listener—is that there's a lot of amazing work going on at Scale. The teams are working super hard, and we're delivering a ton of value for our customers.

The public narrative has not represented the work that the people here are doing and the work that our customers are doing with what we're doing for them. I just think that deserves the respect and reward that all those people are putting in, and we'd like people to know that.

I appreciate you saying all that. With that, we've reached our very exciting lightning round. I've got 5 questions for you. You ready?

Yeah, let's go for it.

What are 2 or 3 books that you find yourself recommending most to other people?

Something that's going to sound interesting: The Selfish Gene is one of my favorite books.

Love that book. I don't know if anyone's ever mentioned it. It was one of the most influential books for me, too. Sorry, keep going.

Yes. The Selfish Gene. The Road Less Traveled I've read more than once. It's just a classic human psychology book. Then, in business, I think Good to Great. It's not the read that you're going to be most excited to enjoy on a vacation, but it's pretty much right. I think we should take advice from people who have analyzed these business problems before, because not a lot has changed, but we keep acting like everything's changed.

What's crazy about that book—you look at all the companies they talk about. I haven't read it in a while, but I believe the whole book is about companies that last, or maybe that's the other book. Anyway, all the companies that they talk about, I don't know if they're still around. It's so hard for a business to last a long, long time.

I would also recommend Thinking, Fast and Slow, right?

Mm-hmm.

Yes.

Thinking, Fast and Slow? Yeah.

Thinking, Fast and Slow.

Yeah.

Excuse me.

Yeah, yeah.

Sorry. It's been like a decade since I read it. But just in terms of the point that human biases are very important to understand.

What's really crazy to me about that book, and Kahneman in general, is that someone asked him, “How has your life been impacted by learning all these biases humans have?” He's like, “Not much. I have the same biases. Knowing them doesn't really help me avoid them.”

See, I find myself checking myself. Whenever I get super-convinced about something now, I'll be like, “Okay, what is the list of things that I'm inclined to do—”

Wow.

—to try to catch myself, because I think we're most inclined to make these bad decisions impulsively, which is what I think the book is largely about. I mean, it's a long book.

So long. Oh my God. It feels like that's where AI could help us in the future, just like, “Hey, Jason, are you sure this isn't a framing effect or whatever?”

Yes.

Okay. Next question. Do you have a favorite recent movie or TV show that you've really enjoyed?

Most of the movies I watch are with my kids, so I wish I had something deep and profound.

No, kids' content is also a very acceptable answer. You can go with that one.

Yeah, the Formula movie I thought was really good. It's a classic action movie. I don't think it informs anything in AI or business, but it's good to check out from the craziness of tech once in a while.

Is there a product you recently discovered that you really love? It could be an app, clothing, a kitchen gadget, anything along those lines.

VO-3, right? It's not totally new, but when I was in high school, I wanted to be a screenwriter. I actually grew up in the Bay Area, and everybody was an engineer, but I wanted to be a screenwriter.

I went back and got the first page of one of my old scripts—which weren't good scripts, but I got the first page. I took a picture of the script, fed it to VO-3, and said, “Make this scene.” It got it right.

Wow.

I was shocked. I was absolutely shocked that you could just take a picture of a script. So now I'm thinking about that for family videos: How do I use these tools for family videos? Some of the Grok tools now, with live image generation and making images more active, are really interesting.

I think they need one more step of iteration, but I think those are going to be really emotionally life-changing for people. Just a little bit of movement in an image from a grandparent or a relative, or whatever you haven't seen in a while, really does make a big emotional impact on you.

I love that when you play with these tools, you can probably think, “Oh, here are the people who helped train this thing.”

Here's the people that helped them fix a problem that it had.

Yeah.

Oh, man. You just unlocked a whole new business unit right there. Yeah. Okay, two more questions. One is, do you have a favorite life motto that you often think about or find useful in work or in life?

Yeah. “The end is never the end.” That’s my favorite internal saying. It goes to the comments before about survival being a precursor to thriving. You’ve got to survive before you thrive, which your brain tells you.

Along these entrepreneurial journeys, I think this is most applicable. This is the hardest journey anyone can go on. If you go on this journey for 5 years, you are mentally harder than 99.9% of the population. People don’t understand the Chinese water torture of having self-doubt and having things go wrong.

More tactically, you get this when you’re working out. In a day, you think, “Oh, I’m too tired. I need to stop.” But the truth is, you can keep going, and the world’s going to keep spinning. So I find in the moments where it’s the hardest, or you have this hard decision that seems impassable, and your body is having this visceral reaction of, “This is not a—this is impassable,” just remind yourself that you’re going to wake up tomorrow. This isn’t the end. There’s another end somewhere.

I find that unlocks me to be like, “Okay, there might not be a perfect solution. There might be an imperfect solution, but it’s a solution, so let’s just keep going.”

Final question: You helped create Uber Eats. I imagine you’re still a power user of Uber Eats. Do you have a favorite restaurant on Uber Eats that maybe people should know about, or that you order from most?

I order a shocking amount of McDonald’s, actually, despite my original story.

It’s sort of the family treat in the house. I would say that’s probably the top thing that we order.

Oh, man. I’m worried for your health, but I love it. I haven’t had McDonald’s in so long. Maybe I should give it another try.

More practically, we’ll order Mixed Greens or Tender Greens or something like that on a day-to-day basis. But I think the more notable, surprising thing is that despite my initial aversion to working with a global chain, it’s a good treat once in a while. You just shouldn’t have it all the time.

Jason, this was incredible. I really appreciate you making time for this. I’m really honored to be the first chat you’ve had since taking over at Scale. Where can folks find you online if they want to reach out, learn more about what you’re doing, maybe look for or join Scale? Where do you want to point people to, and how can listeners be useful to you?

Yeah, absolutely. I’m @jdroge—J-D-R-O-G-E—on X. That’s probably the easiest way to follow me, keep up with things, and shoot me a DM if you like. I think that’s how you would keep in touch. And sorry, what was your other question?

If you’re hiring, where should people go check it out? And then also just—

Yes.

Okay.

Yeah, absolutely. Just go to scale.com, go to our careers page, and we have 250 open roles. To the point about being in business and growing, we’re hiring a ton of people. Our applications business, our data business, is growing. Our applications and services business is growing like crazy, and we’re going to need a lot of people to help us on that journey.

And you guys just signed some insanely large contracts with the government, I was reading.

Two $100 million contracts.

$100 million contracts?

$100 million, yeah. We didn’t sign just one. We signed 2 in 1 month.

Oh, Jesus.

So, yes, our federal business is doing well. Our enterprise business is doing well. Our international government business is doing well. There’s a lot of demand out there.

Some salespeople are getting some great commissions. Good job, them.

Yeah.

Jason, thank you so much for being here.

Yeah, thank you. I’m honored to be a guest here. I’m super excited to be with you, especially so early in the journey—or at least my journey here leading Scale.

Appreciate it. Thanks for coming.