The Exchanges, every show

Q Um, my question to you on the back of that is, and it's a terrible question, you can chastise me for it. What are your single biggest takeaways from that experience that you took with you to Sierra, and what did you leave behind?

A It's such an interesting question. Of course, the, the scale of a, you know, two and then 10 and then hundred person enterprise software company is very different from, I think when I left Google, it was roughly a 150,000 people. Things that I've definitely brought with me, number one is a willingness to invest as far down the technology stack as you need in order to build the service and product that you want. Google, I think from the early days, famously built its own, if not data centers, cluster architectures, and they were the first really to use commodity hardware that required building novel distributed systems for, uh, serving and data storage and so on. And so we could see that language models and, uh, you know, as, as early as, you know, April of, Uh, 23, when we started the company, that agents were going to be a thing. This was before all anyone wanted to talk about was agents. And we realized, okay, this should be possible. It's not yet possible, but we're going to have to invent frameworks for building these things. Our own architecture is really from scratch. So actually our, our first, uh, founding head of research was the Princeton professor who literally wrote the paper on language model based agents, the react paper. And so we invented, we invented and, uh, went, you know, further down the stack than I think some companies at that point would have been willin…

AI assessment note: “Things that I've definitely brought with me, number one is a willingness to invest”

Q Is it a unique time because there is buyer pull like never before for this specific moment?

A There is effectively unbounded demand, I, I think, in two areas. One, we've talked about coding agents. The other is the space where we're the category leader. And so one of the reasons we've grown as quickly as we have is to meet that moment and meet that demand. We, um, we're now a hundred people here in Europe. We recently acquired a company in Japan, Opera Technologies. You and I were talking about this to hit the ground running there. And to have a team that can be attuned to the cultural nuances of, of Japan and, um, you know, the concept of omotenashi, which is like extreme hospitality, like that is what is expected in Japanese service, and that's what we intend to build there.

AI assessment note: “There is effectively unbounded demand, I, I think, in two areas.”

Q It's also not possible in everything, and it's not right in everything. How do you determine?

A You have to edit it. You have to have judgment for it, and I, I think You have to look at what is the thing that is not going to happen or won't happen as quickly without direct applied force from one of us, uh, or both of us. And, and so, you know, it's pointless to be in quote founder mode, you know, 17 layers in the details and something that doesn't matter. It matters a lot if it's Our next generation agent architecture, and there's something that we can add. And so we try to be selective about where we engage at, at that level, but, um, it's anything but kind of hands-off, uh, hands-off management. So I think it starts with the founders. Um, I think, uh, ambitious goals have a way of becoming self-fulfilling. You set out a goal, whether it's the quality of a product or a revenue number, it's like, well, what would have to be true in order to get there? Like, let's suspend disbelief and just imagine, like, what would have to be true to cover this much ground this quickly? Why can't we do that? Why, ok, why shouldn't we? Japan is an interesting example of that. Like, why, why can't we have a giant business in Japan this year and not next year? What would have to be true? Oh, we would have to have, like, 10 people on the ground. I was like, why don't we buy a company there? So you see how this stuff hangs together. So ambitious goals can take the form of, of a date. Sure. You…

AI assessment note: “You have to look at what is the thing that is not going to happen”

Q What was the most recent disagreement you and Brett had?

A Couple weeks ago, we were trying to figure out how to get something to move much faster in one space, and it's interesting, we, we basically always converge. It's like, we're, we're highly truth-seeking. It's like, what is, we have a funny expression, like, this is correct. Okay, what, what does that mean? From, from some objective truth-seeking perspective, like, this is the right way to do it. So we, we try to get to, okay, what is the correct solution? I was on one side. It was like, I think we need better kind of process and structure around this thing. Brett was on the side of people and maybe we need different leaders or a different leader in this space. Um, the answer is with most things like turned out to be some of both, right? Turned out to be some of both. But, um, I, I think, um, we, we started from, uh, No, it can't just be solved. It's like, no, it's not just people. And, you know, we pull on those threads and this wasn't think apart, think together so much as just kind of interrogating each other again with the goal of just getting to the right and best approach to something.

AI assessment note: “I think we need better kind of process and structure... Brett was on the side of people”

Q You sell, again, you sell to some of the biggest enterprises in the world. Um, I had a guest on the show the other day say you can't sell to enterprise without an FDE motion. Would you agree with that knowing all that you know now selling to 40 of the 50?

A I would like to think at least in the AI space, I would say rediscovered and borrowed this model from Palantir, and we came to it almost accidentally. So we started the company, And the first thing we did was reach out to people we trusted to understand what are the biggest unsolved problems that you were looking at and saw, oh, interesting service and support as a, a foothold into something much broader, helping support customers across the entire life cycle. We then enlisted, uh, uh, half a dozen design partners that we built the first version of our product and platform with and for. And these are, in the history of the company, legendary, legendary companies. Olokai, great flip-flops, you should buy them. SiriusXM, Sonos, Weight Watchers, and we built the first version of our platform with our engineers deeply embedded inside those companies. So much so that our founding engineer, Mihai, was actually an employee of Weight Watchers, including getting, like, It's performance review time, emails, and so on. And what we realized in that was no one has ever deployed an AI agent. No one has ever put AI in this way in front of their customers. And in order for us to build the best thing as quickly as we and our customers would like, being so close to the business, the mechanics of it, the people, their business model, That we understand it. I won't say as well as our customers, bu…

AI assessment note: “So no need for Ford deployed if you don't”

Q more and more advanced, Does that not mean the problem set for frontier models becomes more and more challenging? As you said, we've seen the progression of open so much that actually they can do the majority where it's like, I get it for like solving climate change, cancer treatments and materials, but actually like for the majority, like what percent of enterprise tasks can be done with open today?

A Well, I think if you look at what percent of enterprise tasks are completely automated today, it's a rounding error, right? It's very low. So is that, is that a model gap? Is that a diffusing the technology into the company? Is that an application layer gap? I think it's probably all of these, some combination of them. You're obviously correct that as the open weights models become more capable, the set of things they can do grows larger. The set of things where all else being equal, if they're much less expensive, that you would want to point a frontier model, uh, becomes smaller. But again, I think we're not imagining just how high the ceiling is, uh, in terms of demand for frontier intelligence, invention, discovery, building new products, building new services. I, I think it's hard to get your mind around when you have intelligence that can work around the clock and, uh, to invent, to build, to discover how you would use that and how much of you, how much of it you could use.

AI assessment note: “what percent of enterprise tasks are completely automated today, it's a rounding error”

Q Also one of the most humble leaders I've ever met. Um, so I absolutely agree there. Um, opening the US has lagged behind. We see Chinese models being unbelievably advanced and impressive. Do you agree that we have a challenging open ecosystem in the US and does that worry you?

A Part of the driver of the difference is probably the willingness of Chinese companies to do scale distillation of the frontier models, uh, from, from the labs. My impression is many of the models, uh, the open weights models coming from China are derived from training runs, uh, done in the US. I think if you have the US based Uh, labs and hyperscalers developing the frontier models. There's an obvious, you know, like, are they going to compete with themselves and drive, you know, price pressure on the frontier models by, you know, developing and releasing models, uh, open weights models that are of similar capability. You know, if, if I was running that business, that's not something I would do. So I think that's, if, if you can't build frontier models yourself, Okay, maybe the next best approach is to distill them and offer them up. I think that's probably the main driver of the difference.

AI assessment note: “I think that's probably the main driver of the difference.”

Q Also one of the most humble leaders I've ever met. Um, so I absolutely agree there. Um, opening the US has lagged behind. We see Chinese models being unbelievably advanced and impressive. Do you agree that we have a challenging open ecosystem in the US and does that worry you?

A Part of the driver of the difference is probably the willingness of Chinese companies to do scale distillation of the frontier models, uh, from, from the labs. My impression is many of the models, uh, the open weights models coming from China are derived from training runs, uh, done in the US. I think if you have the US based Uh, labs and hyperscalers developing the frontier models. There's an obvious, you know, like, are they going to compete with themselves and drive, you know, price pressure on the frontier models by, you know, developing and releasing models, uh, open weights models that are of similar capability. You know, if, if I was running that business, that's not something I would do. So I think that's, if, if you can't build frontier models yourself, Okay, maybe the next best approach is to distill them and offer them up. I think that's probably the main driver of the difference.

AI assessment note: “Part of the driver of the difference is probably the willingness of Chinese companies”

Q Can you expand on those? Again, two that I don't often get.

A Intensity. So I think it's, it's back to this great thing about giant market, giant market, hard thing about giant market, giant market, and others are in it too. I think there is an inevitability to companies interacting with their customers via really sophisticated agents that capture all that they know and all they can do on behalf of their customers and get the job done on their behalf that handle the complexity as opposed to Pointing you to websites and, and so where the, the conversation is the interface, right? I, I think there's an inevitability to that. And, uh, and therefore in order to win, in order to build the best company in this space, it is about pace. It is about winning. It is about, uh, building the best product. It is about, uh, being competitive and being intense about it. And, uh, knowing that, you know, we don't have the, the luxury of patience. There's no, Nothing, nothing written in the wind, right? That, that any particular company will be the company showing up in our fifth engagement, uh, and 500th engagement, you know, as intensely as we did our first, like you have to do that. And so I think there's also, I talk about the Venn diagram of who we hire for smart, nice, intense, and it's hard actually to get all of those three in a single person. When you, when you do, it's fantastic and you can feel it in the office. And Another way of translating int…

AI assessment note: “Intensity. So I think it's, it's back to this great thing about giant market”

Q It's also not possible in everything, and it's not right in everything. How do you determine?

A You have to edit it. You have to have judgment for it, and I, I think You have to look at what is the thing that is not going to happen or won't happen as quickly without direct applied force from one of us, uh, or both of us. And, and so, you know, it's pointless to be in quote founder mode, you know, 17 layers in the details and something that doesn't matter. It matters a lot if it's Our next generation agent architecture, and there's something that we can add. And so we try to be selective about where we engage at, at that level, but, um, it's anything but kind of hands-off, uh, hands-off management. So I think it starts with the founders. Um, I think, uh, ambitious goals have a way of becoming self-fulfilling. You set out a goal, whether it's the quality of a product or a revenue number, it's like, well, what would have to be true in order to get there? Like, let's suspend disbelief and just imagine, like, what would have to be true to cover this much ground this quickly? Why can't we do that? Why, ok, why shouldn't we? Japan is an interesting example of that. Like, why, why can't we have a giant business in Japan this year and not next year? What would have to be true? Oh, we would have to have, like, 10 people on the ground. I was like, why don't we buy a company there? So you see how this stuff hangs together. So ambitious goals can take the form of, of a date. Sure. You…

AI assessment note: “look at what is the thing that is not going to happen or won't happen”