Mercor CEO on Why Application Layer Companies Have No Defensibility, The Model is the…

So there was an incident. All of the other parts are false, in that we obviously handled it very quickly. We were in touch with customers. We moved incredibly fast at engaging Mandiant and a bunch of other security consulting firms, and the company’s been crushing it ever since. We’ve expanded our relationships with all of the Frontier Labs and added 300,000,000 in net new ARR in the last sixty days. 300,000,000 in sixty days?

been pretty crazy, yeah. Keeping Where us were you when you found out about the hack and what did you do? Well, it was a Saturday, so I was in the office and I was talking with our engineering team. And I think the initial thing is of course, how are we communicating this to customers and trying to be very proactive about understanding exactly what happened, what was accessed, etcetera. And then how do we communicate this to the experts and just containing it, moving quickly on the comms.

And then from there, of course, making sure that we put in place all the right things so that it never happens again. There’s a

Well, it’s interesting because I feel like throughout the lifetime of the business, I have been through a lot of very stressful moments. That was definitely stressful, but it definitely wasn’t close to the most stressful one. Seriously. Yeah. I mean, there’s been times when I’m freaking out about making sure we get something right with a customer or whatever it is. But part of it is that there was this broad perception on Twitter that was much more exaggerated than what actually happened within the business. And so having a thorough understanding of what actually happened and having really strong relationships with customers gave us a lot of confidence that we would get through it, be on the other side even stronger.

And we used to have six values as a company, but we added a seventh value as security to make sure it’s very ingrained in the culture. But I think that, yeah, just that confidence that we know what’s going on and that there’s sort of this echo chamber on X that we need to hedge against a little bit. Do you pay attention to it? And do founders need to pay attention to it? Definitely. I mean, I think founders need to pay attention to it. Like we had an all hands with a company where we just laid out, here’s exactly what’s happening, here’s the trajectory of the business.

And I think that was very helpful to the entire team. But it was definitely annoying that there were all of these people saying things that didn’t actually happen and we couldn’t quite speak out against them too explicitly. Otherwise, there’s going to be the Twitter mob circling and all these recommendations from lawyers, etcetera.

That is exactly what happened. And like, I can even think of one person that’s very prominent who’s invested in multiple competitors and just like made this tweet about how all of our data was getting accessed by China when it was totally untrue. You mentioned adding security

I think so. I mean, we’re even seeing this on the customer side, where our customers obviously are very focused on how do we improve the model cyber defensive capabilities, so that we can have the best AI security engineer that is able to defend every enterprise from all of these attacks. Because in our incident, it was the attacker that used a swarm of coding agents to help get access to the system as is happening in a lot of these. And so I think there’s going to be an enormous boom in AI security engineering tools and various forms of defense that are able to help protect companies against all of the increasing waves of cyber incidents that are just getting started.

How does that actually work? When a normal attacker is trying to find vulnerabilities, they can only review so much code and go through a certain portion of it at a human speed bound by the amount of people in their team, versus when they’re using swarms of agents, they’re able to be very exhaustive in reviewing the entire code base, looking at the entire front end, all the different things that they’ve accessed. And so that has allowed a lot of these attackers to just move much more quickly.

And so we’ve been exploring various collaborations with customers and how we can strengthen their cyber defensive capabilities to hedge against exactly this type of attack as well. In terms of those various customers, true or false, you lost OpenAI and Meta’s

False. Our relationship with OpenAI is stronger than ever. Obviously, I can’t speak too much to specific customer relationships though. Can I push on Meta? Of course. I mean, I think that Meta, currently the relationship is still paused. Every other one of the Frontier Labs has grown their relationship with us since, and the company has been crushing it, but they’re the only one that is And it will be paused just because of the security? Well, there’s other things happening there. Like obviously I think that Meta’s a unique customer because of the scale acquisition.

And so naturally, they’re going to work with scale more, but I don’t wanna speak too much to the specifics of a customer.

shouldn’t speak too granularly to that, but yeah. Totally cool. Okay, so we have not I’ll speak to everything except customers.

a single offer to someone for Micro One.

Why does that come about? Because I read this article. So the reason for the article was that someone on our team sent an outbound to some people at Micro One saying that we were hiring a variety of people with these very high signing bonuses. I think one of them said $500,000 as a potential signing bonus. And they took first meetings, but we didn’t move forward with offers in anyone. And obviously, the way that gets framed to the press is, Oh, these are offers that are going out when there’s a giant distinction from one of our employees sending a message to one of their employees versus actually sending out a legal offer letter.

That one is false. I obviously can’t speak too much to other acquisition kind of stuff, so I’ll reserve any comments on future acquisition questions though. Would you sell for $30,000,000,000 No, I wouldn’t. I mean, ultimately, we’ve gotten a lot of acquisition interest, and we could walk away with like I could walk away with billions of dollars in cash. The thing is, that’s just not what motivates me. Like, I’m very motivated by how do we solve this incredibly important problem in the world of how humans fit into the economy.

I feel like we have the opportunity to build a legendary company in creating this new category of work, And our probability of executing on that vision wouldn’t be as high if we weren’t an independent company.

I think to some extent. I believe there’s certainly going to be many more jobs in ten years than there are today. But there’s also going to be a lot of job displacement along the way. Amidst all of these layoffs, I think the most important question is understanding what jobs is AI able to do and what jobs is AI not able to do. And so we’re building a ton of initiatives such as the AI Productivity Index or APEX that are becoming the industry standard and answering that question of measuring across all the different popular job categories that people talk about ranging from consultants to investment bankers, to lawyers, to software engineers.

What are the actual tasks within those that AI can automate and what are the tasks that it can’t?

Definitely. Well, even another example on that is on APEX, the frontier model right now is at about 40%. And twelve months ago, the frontier model was one, which was scoring 1%. And so that’s been the progress of the last twelve months. And obviously we expect it to continue and be fairly significant. But I think that the key thing is that everyone underestimates the elasticity for demand and increased productivity in the economy. Ultimately, over the last two fifty years, we’ve increased productivity by 25x, equivalent to automating about 96 of someone’s job.

And during every technology revolution, ranging from the Agricultural Revolution to the Industrial Revolution to the Computer Revolution, people feared that there would be this enormous job displacement because of the lump of labor fallacy, where people assumed that there was a fixed amount of things that had to be done, and when we made people more productive, that would all of a sudden mean that there were fewer jobs. Yet two fifty years later, there’s more jobs than ever before, and it’s because we have no shortage of problems to solve as a society, right?

We still need to solve climate change and cure cancer and do all of these other new things, and so I buy that

Well, the thing I agree with you is about displacement. I agree there’s gonna be a very significant amount of displacement, but I also think that the economy is becoming much at creating new job categories and allocating new labor. Like a great example is what we do in that now we’re paying out over $3,000,000 a day in the fastest job category ever created in history. And I expect that’s going to continue growing exponentially from here. And I think that there’s going to be so many new job categories created across everything within AI, such as training agents for deployed engineering, building data centers, all the way to all of the problems that we otherwise wouldn’t have been able to address as a society.

Like how do we build solutions to climate change? How do we have more people working on rockets to explore space? Totally get you. You said 3,000,000 per day pay down? What is that in twelve months’ time? In twelve months’ time, that’s probably about triple that.

Maybe it’s quadruple that. We internal projections that are always much more aggressive than our external projections, but we almost doubled our projections last year. What new role will we have in five years that does not exist today? One of the largest things that people underestimate, both in the context of AI Labs as well as within the enterprise, is how significant of a job category it is gonna be to train agents. Like what we’re seeing is that all knowledge work is converging on training agents because it is structurally more efficient to do something once.

Instead of having a customer support representative that is redundantly responding to hundreds of tickets, they’re going to train an agent how to do that once. Instead of having a lawyer that is redundantly doing dozens of similar red lines on commercial contracts, they’re going to train an agent how to automate that. And even probably when you’re playing around with Claude, you see that there are so many repetitive workflows of how you prepare for a meeting or draft emails or whatever it is, where it’s just much more efficient for you to train the agent how to do that activity so that you can amortize that over the entire useful life cycle rather than doing it redundantly yourself.

And so I think that there’s going to be this enormous paradigm shift as agents enter the workforce and everyone begins to manage them. Can I ask you,

Well, I agree in part. I think that certainly the models need to have access to data to perform their jobs effectively. But the caveat is that they’ll be able to clean the data themselves fairly effectively as reasoning capabilities go up. The thing that humans will need to contribute to is all of the tacit knowledge within the organization that isn’t written down. Because I found that when I try to get agents to do all of these workflows throughout Mercor, there’s just an enormous amount of context that lives in people’s heads that the agents need to have access to perform effectively.

And so much of that is going to be the new job of employees of how do we codify all of this knowledge? How do we train agents so that they’re able to perform these tasks effectively across every function in the organization?

Why? Well, the reason is that if a model is able to, for example, read through every message written in Slack over the last six months, the model can presumably structure a table of here are all the different customer conversations that happened in the CRM, etcetera. And so I don’t expect humans to be doing like that type of stuff of how do we structure data, how do we classify it, etcetera. But I do think that humans will do the things that models inherently can’t do such as the test of knowledge.

When we look at the

Is that how it plays out? It’s interesting. We’re doing a ton of data collection in the physical world as well, especially across skilled domains, where you have electricians and mechanics and scientists strapping cameras to their head to record things. I think that there’s always going to be some degree of value in some of these niche vendors that are able to go really deep in a specific vertical. But what we’re finding is that there’s enormous value to aggregation and economies of scale, and that when we have this talent network of over 5,000,000 people that are able to refer their friends, it’s just so much easier for us to find the marginal doctor, because we have that enormous talent network that can refer us to their friends.

And even more importantly, that the kind of data shapes that we would build for a lawyer are often very similar to the kinds of data shapes that we would build for a doctor. And so all of the tooling that we build is very, very cross applicable, and that’s the way that most labs have been scaling out their data quite horizontally. And so for that reason, we are finding that the labs tend to prefer partnering with a very horizontally capable vendor that is able to flex across all of the different verticals and scale extremely quickly, rather than working with 100 different vendors that they have to train for the same data shape in 100 different domains.

Do

of consolidation? I think there will. I think in most markets, when the markets are so frothy and anyone can get funding and run negative margins, of course, there’s gonna be this proliferation of companies that pop up. And when markets come back to earth and there are natural corrections, that’s when there’s periods of consolidation. And so we view having over $500,000,000 in cash and a super profitable business as a significant asset and allowing us to be prepared for when there is a market correction to make sure that we consolidate market share.

You’re profitable today? Very profitable. How long have you been profitable for? We’ve never really burnt cash. We burnt $500,000 after our seed round. And then from there, we’ve pretty much been profitable ever since. We have more cash than we’ve ever raised. And it’s just because the business has grown so quickly that we obviously try to redeploy capital as fast as we can to invest in growth, but yeah, the business has grown so fast that we haven’t been able to redeploy capital commensurate with that. Can I ask a Myth Buster

and what’s the revenue say? I can’t show the exact revenue number, but it’s, you know, dramatically higher than whatever’s been posted publicly.

So the revenue is between a 3040% gross margin, but the key distinction and why it’s not GMV, but it’s revenue, is that the experts are actually only one part of the broader value chain that we deliver to customers. So when a customer comes to us, they’re generally buying tasks where they would say, Hey, they’ll pay $1,000 for this task that delivers model improvement. And then we do the end to end process associated with how do we find the experts? How do we hire the experts? How do we build the platform that the experts work on so the experts can do the work?

How do we have our AI project manager manage the experts to automate all of the coordination of helping to produce this data? How do we have automated quality checks, etcetera, to produce the end product of the task that we’re delivering for our customer? And so that’s the large distinction of how we’re powered by a talent network in the same way that Uber is powered by a driver network, but that’s not the end product in the same way as some of those Marketplace businesses.

and it’s full stack. Exactly, very vertically integrated. Well, because so many parts of the downstream signal inform the upstream signal, right? Like we can use the quality checks on how high calibers each of the individual data points to understand exactly what are the types of experts that we should be onboarding to achieve the data that drives the most model improvement. And there’s oftentimes this very power law nature of data that drives model improvement, and that out of a dataset of 10,000 tasks, the top 2,000 tasks will create majority of the value.

And so it allows vendors that are extremely high quality to be super differentiated insofar as pricing power, because quality is the X factor that becomes dramatically more valuable than any other dimension. What tasks are super high value? Is it like the medical, the financial modeling style? It corresponds extremely closely to economic value. So think if you go through the top five domains that we serve, it would be software engineering, it would be finance, medicine, law, consulting, etcetera, and the super long horizon tasks within those.

And so I think we’re moving away from the paradigm of how do we get a investment banker to prepare a financial model, and moving towards the paradigm of how do we get a banker that can talk with five different colleagues and wait to hear back their responses and prepare an entire slide deck with a deliverable that includes the financial model, the analysis in a multi week long project. Those are the kinds of tasks that we need to be building to push the frontier of research and evaluation, so that those are the capabilities that people are able to use in the models in six to twelve months.

Can I ask, which segment are we underserved in?

Well, I would say maybe I’ll give it from Mercor’s perspective, and then I’ll give it from the lab’s perspective. We tend to be now so good at mobilizing experts that we’re able to access pretty much any domain. There’s always going to be some degree of like these niche pockets of oncologists or whatever it is that have a particular background, but generally we can fill those fairly quickly. And it’s more about people that actually are very acclimated to the frontier of AI, because it’s the people that both have the expertise in oncology, but also are power users of ChatGPT or Claude that are able to find where the model makes mistakes and help the model learn from those mistakes.

And so that’s from the Mercor perspective. From the perspective of the labs, it seems like it’s all encompassing. It’s just like the barrier to automating everything that you can do in say Google Workspace is how do we cover the full distribution of all of the context, I. E. Messages, Slacks, Slides, Excel sheets, and all the tasks, prompts, outputs that correspond to everything that you do in your job. And that applies to every individual and every domain throughout the economy.

And so there’s this enormous mobilization of hundreds of thousands and soon millions of people to build out the full distribution of everything that you could pass into Google Workspace and everything that you could want out on the other side in every job category throughout the economy.

Good question. Well, I’ll talk through the valuations of each, of the revenue of each. So our seed round Did Phantom not fly you in the chopper? That was Series A. So our seed round was in September 2023. We were at, call it, a million in revenue run rate, or just shy of that. And I initially didn’t wanna raise because I wanted to bootstrap the company, but Adarsh Hiremath’s condition on dropping out was that we needed to raise money. And so we met General Catalyst eight am on a Sunday morning.

They gave us a term sheet within thirty six hours for $2,300,000,000 at a $23,000,000 post money valuation. Pretty gross. So that was pretty reasonable insofar as price And this was was Max and him on? This was Max and Nico. And then at our Series A, the business didn’t grow that much from the seed to the Series A, but we found the market was a key differentiation. And we met Victor when we were at, call it, 1,500,000 in revenue run rate in May 2024. And Victor got super excited.

Initially, I refused to take a second meeting, but then he said, Oh, have you ever been in a helicopter? And so Peter took us on the helicopter flight and Benchmark really wanted to work with us. And so by the time that they gave us a term sheet, were at, call it, 2,500,000 in revenue and they gave us $250,000,000 post money valuation. Did that feel uncomfortable? Because that’s a big jump, $2.23 post to $2.50. So keep in mind, the time, this sounds crazy because we were at 2,500,000 in revenue, but I was projecting 50,000,000 in revenue run rate by the end of the year and 500,000,000 by the end of next year.

And so it felt like a bargain. Dude, you do know all founders project that day, we in the never the project.

Yeah, yeah, yeah. And then four months later, we met Felicis, and we never would make a slide deck or take investor meetings. And so Lisa sent us an email saying, Hey, we know your co founder Surya really likes Ferraris. Do you want to go racing Ferraris with us? Then I replied and I said, You caught my eye. Tell me more. And they said, We’ll meet at the airport in Hayward and go on Aydin’s private jet to Las Vegas to race Ferraris around F1 track. And so I was like, We’re available in three weeks on a Sunday.

And so we do this. We race Ferraris. We’re at $20,000,000 in revenue. They ask us what valuation do we think makes most sense? And I say one dollar to $2,000,000,000 So they give us a term sheet at a $2,000,000,000 valuation. And at the time, that’s 100 times revenue. And everyone thinks that that’s a high valuation. Meanwhile, was an incredible investment.

a high price. Well, so here’s the thing is, we’d been growing 50% month over month for the prior six months, and I think what none of them really realized was that it would continue for the subsequent twelve plus months. And so then that compounded more and more by September 2025, or say October, were at, call it, dollars 400,000,000 in revenue run rate. And then Felicis was like, We want to invest more. And so they gave us a term sheet at a $10,000,000,000 valuation. We didn’t really want to spend much time on a financing because the business was growing 50% month over month, and so we were very preoccupied.

And so that was about 25 times, and the business has almost 4x since then. So

If I had to choose any, the Series B priced in the most, the furthest ahead of our growth, or that or the Series A. Think it was probably the Series B, because both were 100 times revenue, but it’s very different to be 100 times revenue when you’re at $2,500,000 in revenue versus $20,000,000 in revenue. So that was probably the largest one. But obviously, both were great investments in hindsight. What is the next round then? We’ll see. Probably a much higher valuation. We’re getting a lot of offers at meaningfully higher valuations, but the company is fairly profitable, and so we’re taking our time to see who the right is.

I totally agree. I’ve never been on a warship before, but that sounds a lot of fun. There you go.

do you believe that? The reason I believe that is that the Application Layer companies’ businesses are not far removed from the Foundation Model companies’ businesses. Like it is not a far leap for Cloud Cowork to add capabilities across medical and legal. Obviously, they did it with software engineering and can do that across finance. And so I feel like building defensibility in the software layer on top of the models is going to be incredibly difficult. Whereas on the other side of things, in the infrastructure side, it feels like there are meaningful moats that are getting built.

Like we’re compounding enormous network effects in the business and a pretty significant data moat as we build out the inventory for our customers. Compute companies obviously are able to build moats through these very long R and D cycles. And so I think that there are going to be high margins that get achieved at the infrastructure layer and sort of sustainable, profitable businesses in a way that it’s less immediately clear at the application layer.

30%, zero impact on demand. It’s probably the same for us honestly. We have the demand to double overnight, we just don’t have the capacity. And so it’s mainly a question of how effectively can we scale to mobilize people to build out these environments much more quickly. Do you do pricing elasticity tests? Because if you can double price and double the business. We maybe can’t double prices. We could double capacity. We could probably increase prices by 30% without much of an impact, but the other thing you need to consider is that pricing is not merely a question of optimizing for the next six months.

It’s optimizing for a structure that wins the market over the next decade. For that reason, we’re very focused on how do we do what’s best for customers, how do we do what’s best for experts, and how do we build a sustainable business while we’re doing it, but make sure that we’re not leaving oxygen in the market because high margins invite competition.

Argue back. Maybe I would say two things. First is that I think over the last two years, everyone has increasingly realized that the model is the product. That we can build so many of these different abstractions of trying to stitch together API calls and having all this patchwork logic where people used to have all these drag and drop agent builders. And then they just realized that if we give the model the end goal and we train it to accomplish that end goal, it has outperformed every other solution in almost every case that we go after.

That bodes incredibly well for those that are training models end to end. The second thing to consider is that software layers are able to get recreated very quickly now. Like we’re building out an eval set that measures how effectively agents can build end to end SaaS applications, where 2025 was the year of how do you get a model to make a PR and a code base, and 2026 is the year of how do you get the model to clone Slack end to end. Those capabilities are going to exist in the models in the next twelve months.

That means very significant things for companies that are betting on software moats sustaining their businesses.

you’d build your own. I generally agree. I think that the caveat is when those companies have Network Effects, there’s probably a significant moat that isn’t being priced in fully. For example, Salesforce has tons of companies that are building integrations on top of their platform that creates this almost marketplace and network effect around it, or Slack has Slack Connect, right? And I think even Carta is another great example of this whole network effect of the people that use it and want to use the same platform across all of their companies.

I think that the companies that have network effects will be able to, in some ways, generate more value because they can iterate 10 times faster while leveraging those network effects to create more value for their customers, and therefore build more valuable products, charge more money, etcetera, and increase revenue. The companies that don’t have network effects are going to struggle very significantly, because then there’s not really a defensible moat in the pure software associated with the products that they build. And so to me, that is the litmus test that determines whether this company is going to become worthless or whether this company is going to gain dramatic value from their ability to 10x product velocity.

You said we’re learning more and more that

So I agree with this in part, but the caveat I would give is I think it’s arguably more the forward deployed motion rather than the go to market. And forward deployed motion being the post sales, go to market being the presales, because ultimately, say you’re just really good at sales, and then you provide a SaaS product, and you have a savvy customer who’s spending $1,000,000 a year on the SaaS product, and they realize they could just like tell Claude to copy it and they’ll get the same exact thing, it feels very difficult to maintain your pricing power, even if you’re the best in the world at sales.

Whereas on the other hand, if you have a great forward deployed motion, where you’re going deep with a customer, you’re training the agents based on all of this tacit knowledge within the company so that it understands how to perform effectively, that feels incredibly differentiated and hard to recreate. And that’s also the reason that we see obviously the labs at OpenAI Anthropic investing so much in this for deployed motion. And so I think that the Sequoia article that services are the new software resonated a lot in that these software boats are whittling away, and it’s the ability to layer services on top of software to meet customer where they’re at and go the last mile that is creating stronger defensibility.

I think in a large way I do. You do? I think the key thing is that you need to make sure that they’re actually going to leverage AI. Like, I think there are a lot of companies that are just building services and not gaining a significant competitive advantage from AI and using that to that’s the thing you’ve got to be careful about, but I think it’s very rational.

I’ll give an example in the context of Mercor, which is that within this process of turning human time from the talent network into building these super rich environments that mirror everything that people could do in their jobs, there is a lot of human coordination of how do we answer people’s questions, how do we track the KPIs of the project and manage it effectively, how do we build the bespoke tooling for that project, and we have about 100 people, or call it 150 people, in our delivery organization that do that for Deployed Work of helping to go the last mile for the customer.

But now we have an AI project manager that just completed its first project managing that entire thing end to end, where it’s able to hire the experts, it’s able to answer their questions, it’s able to build the annotation tool using its coding tools within our platform and produce the end data type. And the experts all had a really good experience on the project reporting to the AI project manager that was running it. And so I think we’re seeing in real time that services are getting automated and that that is going to be this extraordinary transformation in the economy.

and why that is? Well, it’s a fascinating case study in Siobhan’s paradox, similar to what we were talking about in the context of making humans more efficient leading to more jobs, right? When we make models improve by 10x year over year, that has just been causing the total consumption of the models to go up and up and up as the cost per performance go down. I think insofar as how it’s going to develop is that this trend is going to continue very, very significantly before we start seeing any leveling off of token consumption within the enterprise.

Like right now, we’re spending more on tokens for our internal agents than we are on employee headcount. And I think most businesses are gonna look like that in But

Incredible. And so the way we manage it is that we have a variety of these key workflows throughout the company, where we have an AI project manager, as I was describing, that manages operations. We have our interview question agent where we’ve done over 5,000,000 interviews and asked all the questions in the interviews. We have our interview ranking, or the broader candidate ranking, where it helps to assess all of the candidates and figure out who we should be hiring. We have agents for accounting automation. We have agents fraud detection, etcetera.

And corresponding to each of these agents, we have an eval that tells us which model is best to use for this given use case, and what is the Pareto frontier of price performance for that specific use case. And that eval allows us to make the decisions around where should we be allocating our inference spend, what providers should we be using, etcetera. And I believe that over time, this is going to develop to look very similar across every Fortune 500, where they’ll need to have this system of record for evaluating and specifying agent behavior across every workflow in their business.

And they’re going to use that to commoditize the model layer because they want to enable perfect competition for the models having zero switching costs. And so we’ve been growing extremely quickly with the enterprise and helping them to populate the system of record and building out those evals for each of the use cases that they have throughout their business.

Yeah, so I think the key distinction is that I think the API layer will get commoditized. You can definitely build stickiness and workflows that people have on top of those APIs. Like for example, I have all of these routines running in Cloud Code, and I feel like it would probably be difficult, or I at least wouldn’t put in the time to move those routines over, and I have a bunch of similar things running in ChatGPT. So I think that there’s going to be various ways that people can build stickiness, but for pure API based products, where it’s like if we are just spending $10,000,000 a year on a specific workflow, obviously we’re gonna have an eval for that.

And every time a new model comes out, we’re gonna benchmark that and understand exactly how we should be hot swapping between models and distilling models. Why does the API layer get commoditized? Because the switching costs are zero. Like when the switching costs are zero, there’s a new frontier model every two months. That means that we very quickly are going to swap them out, right? And ultimately the decisions that we make boil down to the score on the eval corresponding to that workflow. And so it’s very easy to compare model to model one for one in a perfectly hot swappable way, which is almost the definition of a commodity.

that is in twenty four months’ time? For Salesforce? Yeah. I don’t know about twenty four months’ time, but I would bet that in five years, the average enterprise spends more on compute than headcount. The reason for that is that the models are just becoming so capable that it seems like there is just enormous ROI to being able to have models do something for 100 ks a year, that is going to continue compounding at an exponential rate in a way that human intelligence is not going to.

And so humans will still play an important role at the things models can’t do, but I expect that cost of inference, cost of compute will exceed that. The reason that that’s so interesting to me is that having an eval for your specific workflow, like say we take the case of Salesforce, having an eval for how good is a specific model at code generation and their use case is often a 10x lever on the price performance of that model. Because they can distill the model, they can have an open source model that is performing as well, if not better, for a dramatically lower cost.

And so as we see this enormous shift towards compute and significant inference spend across every workflow in the enterprise, they’re going to need to have evals that act as a source of truth for whether those workflows are being done correctly and whether they’re using the right models to accomplish that.

That’s exactly the problem, right? Is that we used to have this paradigm of all of the academic benchmarks that were totally disconnected from the outcomes that enterprises actually care about, where people were building everything ranging from GPQA for PhD level reasoning to IMO for Olympiad Math to Humanities Last Exam for this long tail of academic problems no one really cares about. And now they’re focused on how do we get the model to do this end to end workflow coordinating with multiple colleagues for a financial model or a slide deck like we were discussing.

How do we get the model to build an entire SaaS application end to end? That’s why there’s this enormous build out in pushing the frontier of evaluation as a critical research problem for the next frontier of model development.

I think both are true. Like there’s going to be many orders of magnitude more demand in five years than there are today, maybe four or five orders of magnitude more demand. But there’s also going to be increased competition with people just distilling and having fine tuned open source models that accomplish their workflows. Ultimately, I think OpenAI and Anthropic are incredible investments, and it seems like they’re starting to be consensus around that in a way that there wasn’t just a couple of years ago. But at the same time, I think that majority of inference in five years is going to be using a open source or custom fine tuned or distilled model, not using a Frontier model.

Okay, interesting. You said they’re obviously incredible

wrong. I could definitely see one of them being a $10,000,000,000,000 company, maybe even significantly higher. It feels like the opportunity associated with being the frontier model is so large. It will just eat up so much of the other demand within the economy, because that also means when you have the frontier model, you can use that as a teacher model to distill your own models, to have the best small models, etcetera. So I would guess at least one of them is worth more than $10,000,000,000,000

and glib. I think it’s not a crazy idea. NVIDIA is obviously a phenomenal business that will continue to execute super well. The only caveat is that it feels like we’re starting to move towards a multi chip future, where obviously Cerebras is executing well. I’m good friends with the etch guys. Most of the labs are building in house chips. And so I would guess that in five years, it doesn’t feel like NVIDIA has quite the same monopoly, but that’s okay, because even if they only have 30% or 40% market share in the largest market in the world by far, that is the world’s most valuable company.

Maybe to some extent, I definitely worry about how do we smooth out the benefits to society? Like how do we ensure that every enterprise and every individual is able to reap the full benefits of AI rather than just like a handful of people in San Francisco? But ultimately, I also think that there is some natural dynamic associated with capital allocation where it is going to be more valuable to give the compute to an Anthropic, where they have the marginal demand and can use that right away versus a less successful company that might not be able to create the most value with that.

So I think that it’s probably good from a capital allocation and efficiency standpoint, so long as we are able to manage the societal implications of increasing inequality.

Well, believe this very strongly. I actually wrote this essay as a research paper when I was a freshman in college. It was one of the few productive things I did in college. And essentially, the thesis of this was that the largest positive externality in the economy is jobs, right? People talk about all of these economic theory of how we have negative externalities like carbon, or like smoking, or whatever it is, we should tax those. But on one hand, the largest positive externality is jobs. Yet on the other hand, the way that most economies structurally collect income is by disincentivizing jobs, both on the income tax side by taxing the individuals, as well as on the payroll tax side of taxing the companies.

And as we move towards a world where there’s increased job displacement, increased uncertainty around how many jobs are there going to be, especially for the bottom half of Americans, I think that this is going to become extremely problematic. And so I would suggest that we move towards a paradigm where we instead focus on taxes of things that aren’t necessarily going to have a negative impact on incentives in the economy. Like one great example is capital gains, where I’m going to invest money in assets regardless. And so if there’s higher capital gains tax, it’s not like I’m just gonna not invest, right?

And so I think that taxing capital gains, especially short term capital gains, which I think is probably not as beneficial for the economy as long term capital gains, would probably be structurally much better off than taxing income.

The thing is when investors are taking very high risks, it’s generally in an aggregated way in a portfolio. And so you would tax the gains on the portfolio overall. And so even if you have a portfolio of like, and I know that you don’t like to hear the capital gains tax, Harry, but

I agree that the main thing you need to be careful about is if people would move to other geographies, because obviously that creates problems. But I think capital gains is one option. Think But I’m so sorry to be addicted,

I agree. I think that there needs to be sensitivity analyses associated with how does the increased amount of taxation cause people to just leave and reduce overall government revenue. But I think that another way of going about it is also taxing consumption of items that probably aren’t the best. Like, it’s crazy to me that instead of taxing carbon, we tax the bottom half of Americans. But why don’t we tax carbon, right? That’s like a very clear negative externality in the economy, at least in The US, that’s not taxed.

And so I feel like there is a lot of low hanging fruit with respect to things that we could tax without damaging incentives in a perverse way or causing people to flee the country that would be far better than taxing the bottom half of Americans. And the other thing is that it’s only 3% of government revenue. The fact that it’s only 3% feels like it’s a very easy decision for policymakers to make in the grand scheme of the impact that it would have on people. Would you tax prediction marketplaces?

It’s gambling. I probably would. There’s probably some value of having good prediction marketplaces for allowing people to have effective predictions of the future and hedge things within their lives and investment portfolios, but it’s likely okay to tax. The thing on that point around taxing the bottom 50% is Jeff Bezos retweeted me, which I was ecstatic about. It’s pretty cool. It was pretty great, yeah. Who’s the coolest person you’ve met? I really like Jensen, and I really like Satya. I mean, so many incredible people. Obviously, Dario and Sam are incredible.

But if I had to choose one person I mean, Jensen’s so cool. Right? Like the jacket, his style, he’s always on point. So I would say Jensen is probably one of the coolest.

that gives me confidence that I can do that too. Actually, will say this one thing, which is that I remember when I went to Georgetown, I didn’t get into Harvard, I was like, Wow, the people at Harvard are probably dramatically smarter than me. And I went to this nonprofit called Prod, where it was a bunch of kids from Harvard and MIT that were all building startups. And they’re very smart, don’t get me wrong, but I do think that most of us have this very equalizing feeling.

That majority of people that accomplish extraordinary things, when you spend more time with them, you realize that they’re sort of just a normal person to a significant not all of them, but most of them to a significant extent. And I think that makes you feel like like when I saw Ethan Thornton from Mock raising $70,000,000 as a 19 year old, I’m like, Wait, Ethan is like a chill guy and a good friend, and maybe I could do something like that one day. It just gives you the sense of being able to accomplish so much more.

Does that matter? I think that it’s going to be difficult to change because there’s just so many strong network effects around talent, right? When we have the best talent, even I knew so many brilliant French researchers that go to work at OpenAI Anthropic and DeepMade, right? Because when we have the best talent at those labs, that’s where they all aggregate, and then that compounds to them having more capital, more compute, more impact, etcetera. And so I expect that trend and to be one of the largest, not only economic, but geopolitical advantages that The US has.

So if you’re Europe

Do we just accept that? I would accept that, yeah. I think that maybe it’s worth having some post training capabilities, because there is gonna be value to distillation in some of the work that happens after foundation models are built, and there’s definitely gonna be some value in applications, but I don’t know if I would lean aggressively into how do we compete head to head with Anthropic. Do you buy the

wherever that is? Maybe in some cases. There is value in localization, and I’ll give an example, which is that often oftentimes labs will come to us and say, they need their models, not just to be good at American law, but also to be good at British law or good at French law or whatever the jurisdiction is in the world. I think that that is going to be an important last mile in making the models useful in whatever jurisdiction that they’re operating in. That said, the labs are just going to hire 10,000 people in France to teach the models how to be better at French law.

I don’t think that there’s so much that others are gonna be able to do to stop that because the transfer learning capabilities from all of the other domains that they’re focusing on are just so powerful.

It’s definitely one of the things that’s most top of mind, in particular because markets for people founding companies are so hot, where we’ve had three employees that have founded companies worth in excess of $100,000,000 I saw your Tweets

you do the

Yeah, exactly. And we’re a very young company, right? And I think that it’s difficult for a variety of reasons. A lot of people probably don’t have a full understanding of just how hard it is to build a company, as you know well, Harry, and how low the probability of success is and how fortunate we were and how lucky we got along the way. And so I think that that’s definitely one of the large challenges. Even like there was someone I was hiring the other day and he had an offer for $20,000,000 in cash per year from TBD.

And like that’s the kind of stuff we run into on a regular basis. TBD? Meta’s super intelligence group. 20,000,000 in cash? Per year. Or it’s in stock, but liquid. That’s hard to compete against. It’s hard to compete against, yeah. Does that change? Does that just continue to escalate? It’ll probably continue to escalate for a smaller group of people. But I also suspect that as more people gain knowledge of how these labs operate and what the capabilities of how to train a frontier model, that means that there’s going to be more supply in the market for people that have that skill set and thus a little bit more reasonable pricing.

And so I expect there to be some craziness that continues, but hopefully sort of the ninety ninth percentile at least within the market will bounce itself out. What is the hardest role to hire for today? Researchers. Just because of supply? Because of supply and demand. It’s just this market where there’s 10 times more demand than there is supply, and that makes it very difficult. We’ve been building out an incredibly strong research team, like Edward Hu, the first author on Laura, who was previously at OpenAI is working with us and a bunch of other top researchers, but the market is definitely getting very hot.

How much does it cost to hire a high quality AI researcher? Oftentimes it would be in the tens of millions of stock per year. For the really good people, yeah. I remember when researchers weren’t paid very much. This was

Is it harder than ever to run the company? I don’t think so. Like to give a frame of reference, we were 40 people and 50,000,000 in revenue run rate last year, at the start of last year. Since then, we’ve seven or 8x headcount and we’ve increased the broader scale of the business by 25x, 30x. It’s definitely been very stressful to keep up with the growth along the way. But I think that now we have the supporting functions, like we have HR, or not really HR, but we have finance and we have legal and we’re building out HR.

And that brings some sense of stability where I don’t have to deal with all of these little escalations. And I’m able to just spend my time focusing on building great products, research, and time with customers. And that I think has made it easier, significantly easier to run the business. I get in a lot of shit for everything I

agree with me? The caveat I’ll give is that I think it’s really important. We definitely had challenges in scaling culture when we went from 40 people to 400 people. How does that show up? Well, it’s so many things, ranging from making sure that we keep a really high talent bar, to making sure that people are bought into the mission of the company, to even the tactical things of making sure that managers are communicating to their team about their performance review and how they’re doing so that they’re never surprised by a performance review.

And when we have a young team with a lot of first time managers, that just creates culture challenges of people that aren’t used to giving feedback and maintaining all of the values and commitment to the mission of the team. And so I think that to some extent, I agree. And I think that some of the big tech companies probably go too far on empowering HR, but I also think that it’s important, and one of the large lessons we’ve had over the last eighteen months or so is that it’s critical to really get these foundations in place as you scale headcount, otherwise it creates problems.

Culture challenges.

So the reason it’s not true is that we’ve never mandated hours at the company. And obviously I work extremely hard. Adarsh works extremely hard. We work from when we wake up until we sleep pretty much all the time, aside from maybe working out, but I’m still thinking about work during that time. Most of our leadership team of course does as well, but at the same time, majority of my leadership team has kids, and we want them to be able to go home and see their families and all of that.

And so I think that it’s some combination of knowing that building a legendary company requires immense dedication to the mission of the business, while also recognizing we need to ensure that it’s a sustainable environment for the best people in the world to do their life’s work. Are you ready for a quick fire round? Of course. Would you like to go public? Definitely. When? In the next few years, I think that all legendary companies eventually go public. And so it’s an important part of the journey and maturing and having a much larger company than we have today, but it’s not something we’re rushing to do this year or next year, in part because we dropped out of college less than three years ago at this point.

And it’s still a very young business where we want to make sure that we properly actualize everything that we’re working on, on the enterprise side especially, before going public. Don’t laugh. Do you ever lie in bed at night and just go like, Wow, pretty wild. I’m always pinching myself, and I feel extremely grateful for the team and Adarsh and Surya and how all of them made it possible because I could have imagined a 100 things that would have gone differently and we’d be in a totally different circumstance.

What have you changed your mind on in the last twelve months? What have I changed my mind on? I used to have some questions around whether the Foundation Model Labs would be the largest businesses in the world because of the exact things you asked about in the context of how much those models are going to be able to maintain pricing power amidst a competitive environment. But I think that as we’ve seen the sheer revenue ramp of these businesses, I’ve gained immense conviction that they will be the most valuable companies in the world.

I really admire Jeff Bezos. I think he’s so disciplined about the culture of Amazon. That’s one of the things that’s always stuck with me. Everyone there just understands the values and is steering in the same direction as such a strategic business leader. I’ve never met him, but I’ve always wanted to. Which competitor do you most respect and why? Good question. I admire that Edwin from Surge has done a really good job in staying super close to research, and it’s something that we’ve obviously been doing a lot of as well.

But I think that’s probably one of largest things that differentiates both us and Surge is our ability to train models, to hire some of the best researchers in the world, and I admire them for execution on that front. What percent of data providers

competitors. About half. Half? Yeah. I’d love that. What would you most like to change about your role today? I would say that there’s a decent amount of HR things that get escalated to me. And so we’re looking for a really strong head of people that is able to handle a lot of this. Final one for you, dude. What’s the kindest thing

One that really stuck with me is I remember, and I’ll probably attribute this to the entire prod community, namely especially a couple of people like Rob Walkin, Ben Spector, and Richard D’Angelo. But Prabh was this nonprofit that got started at MIT and Harvard, and I was sort of a blow in because I didn’t get into those schools, but I went to Georgetown. And for the first year of the business, they would meet with us every week. Ben became a big customer. Richard would give us tons of money just to float working capital, and Rob gave incredibly valuable advice.

And they had nothing in it for them. They took no equity. I tried to give them equity, and they wouldn’t accept it. And Mercor wouldn’t exist if it weren’t for any of those individuals, I would say. I think that that is something that I’ll always be grateful for for the rest of my life. Dude, I have to say, I