Will Bryk. Exa.

Will Bryk:

The first advice I got when I told everyone that were building a search engine better than Google in 2021 was like, "What are you? Nuts?" I kind of laughed, but also was like, "Maybe I am nuts." Really try to avoid non-playable characters is the way of saying it, like people who just go through life as if they're on a stream and they have no agency over themselves. A month is an insane amount of time. Nothing should ever take a month to ship. It feels ridiculous that we can get to the moon and we split the atom. And yet if I want to understand what's going on at the border right now, it's really hard to. That to me is just an error in humanity. When I was a kid, I thought, we have a bunch of adults, like, why can't they figure out how to share information in a normal way? Why is everything politicized? I always thought this was ridiculous and I don't accept it. This is my mission if I choose to accept it.

Nakul Mandan:

If you've used an AI product that searches the web, there's a good chance Will Bryk's company was underneath it. Exa powers search for OpenAI, Cursor, Cognition, and a growing list of the best tech companies in the world. It's valued at $2.2 billion and it's quietly become one of the most important pieces of infrastructure in AI, and it's all happening fast. Exa went from about 10 people to 125 and tripled its valuation in under a year. The hard part of a company growing that fast is the person running it has to grow just as fast. Will is a first-time CEO doing exactly that in real time. What I find most interesting about our conversation today is Will's inner game. Will says instead of real life role models, he's inspired by fictional superheroes, people who decide who they're going to become and then do the work. “We’ll figure it out” from Mission Impossible, has apparently become a mantra at Exa. Today's conversation is a glimpse into Will's journey, how he's scaling himself, how he runs an AI native team, and what actually drives him. Knuckle up. Will, welcome to the show.

Will Bryk:

Thanks for having me.

Nakul Mandan:

Exa is building web search for AI agents, but Google already has a 25-year headstart on indexing the web. It is the best search engine by a long distance. Why do we need a new web search tool for AI? Yeah,

Will Bryk:

I mean the short of it is that Google was built for humans, and now there's this new kid on the block called agents, and they search in very different ways. And so Google is not optimized for this new world of agents searching the web, and Exa is built for this world. I have a list of 20 differences in humans and AI searching. But one example is humans are lazy, they type in a few keywords, then they get some links, they click on those links. This is how humans used to search. But agents, they're not lazy at all. They make very complex queries. They could instantly output a paragraph explaining exactly what the information they want. And they don't really want a list of links. They want a blob of data that exactly captures what they want. So another interesting thing is Google was optimized so that every link correctly is relevant to the query because you might click on the third one, you might click on the second one, but to an agent, it actually wants a very diverse amount of information because it's just gobbling it up all at once. So these are just a few of the examples in there, dozens.

Nakul Mandan:

But even if AI agents are searching differently, isn't this such an important piece of infrastructure for AI agents that OpenAI, Google, Anthropic would want to own it in-house? How can an independent company even exist against the onslaught of the big labs?

Will Bryk:

Yeah. So first of all, it's very, very hard to build a search engine. Very few organizations on the planet have ever done it. You have to crawl a massive amount of data, exabytes of data, and then you have to do a lot of research over years to, given a query, how do you filter down a trillion documents? So it's very hard. So only the very big companies, and company startups crazy enough to work on it for years can do this kind of thing. And then I think Exa's in a unique position where we are happy to be neutral and search infrastructure that powers any company. We serve 5,000 companies now. We don't need to be a platform that people come to. So I think the biggest reason why a company like Exa could exist is that these other big companies like a Google, they're not trying to enable Googles inside every product. They want you to go to Google and use it. There's an advantage for Exa that we don't need people to come to Exa. We want to enable every company, which is fundamentally different strategy from these companies.

Nakul Mandan:

Are there product decisions that you are making that fundamentally a Google or an OpenAI will not be able to make for this?

Will Bryk:

Yeah. For example, we expose the data of our index to all these companies so they could do downstream tests with it. Whereas Google or OpenAI, they'll more just do some sort of LLM summary of any data that they analyze.

Nakul Mandan:

Are there areas where those guys are still ahead of you, which you're working on?

Will Bryk:

Yeah, I mean, for example, Google has probably on the order of a trillion documents in their index. We're not there yet. It's very expensive. However, you don't actually need a trillion documents to be useful. In fact, a lot of customers don't want all the data on the web because a lot of it is pretty crappy. So you want just the highest quality information. But index scale is one thing that we are currently scaling up like crazy to get to hundreds of billions.

Nakul Mandan:

Is the goal a medium to long-term to remain a search API or you think beyond that you'll become part of the infrastructure for AI agents? What's the long vision here?

Will Bryk:

Yeah, it's less about search. Search is almost an implementation detail towards the information layer. We just want to give people high quality information. So for example, you could have perfect search, but if you're searching over information that's incorrect, you're a great search engine, but you're not a great information engine. So ultimately Exa will just be the information layer for any company that needs information data.

Nakul Mandan:

Is a key part of it that you are crawling real-time search? Because any lab has already indexed the world's information to feed it for training. So if I say X is good or Y is good, they've trained the model for a certain amount 'till a certain date. And so does Exa come in where it's real-time information or where does Exa come in where Anthropic or OpenAI might fail?

Will Bryk:

Yeah. Okay. So even if a lab downloads a whole internet and does pre-trains a model on it and then post-trains on that data, okay, so now you have a very compressed version of the internet. And so one problem, like you said, is it doesn't have real-time information, so it needs to use retrieval because the world's information is constantly changing. But also even things that are old, it can't memorize it. So it doesn't remember my birthday, even though that doesn't change. Yes, these LLMs, they're not going to remember Will's birthday. Let's say it was in some blog post somewhere, because it has to compress the whole internet. So they have to rely on retrieval for long-tail information that it's going to know Albert Einstein's birthday, but it's not going to know Will Bryk's birthday, and then real-time information. These are two categories.

Nakul Mandan:

As you build this, there's competition, but there's also the big labs, and again, they'll always want to own some of it, how do you think about defensibility and moat over time for Exa?

Will Bryk:

Yeah, moat's an interesting question because the honest answer is that execution speed is the most important thing and focus, at least in the early days. Then when you start making billions, tens of billions of dollars a year, okay, now you have a lot of people who want in on that. And then you do need some answers to moats. Maybe one abstract way of thinking about it is like, a company is kind of like a mix of intelligence or intelligences, physical resources, and maybe data, at least a software company. And if intelligence gets commoditized, everyone now can have access to fable times a thousand. And so all the intellectual capital you've accumulated is not as much of a differentiator. Then a lot becomes the physical resources and the information you have access to. So you can think of different types of moats. There's all sorts of different physical resources. Even feedback from customers is something in the physical world that you might have a lot of. So it's a very abstract way of thinking about moats. I still think having really smart people at the company is going to matter for years, maybe forever. It's hard to see how it plays out. But you could have moats in being just really smart people, which we have right now, in physical resources, which we are getting closer to, and then with information, which we also have a lot of.

Nakul Mandan:

Taking a giant step back, I mean, it is a fascinating world we are entering in in terms of we are going to go towards an internet where AI agents are doing the searching and I'm just talking to agents. You are at the forefront of it. Walk me through what is your vision of how the internet will look like in the future? Because the last 20 years of the internet has been built around Google and maybe some social networks like Facebook and Twitter, but we are entering a new age and you're at the forefront of it.

Will Bryk:

Yeah, I think the internet's going to change a lot. At the same time, I think a lot will be the same. So if you think about what the internet is on the most fundamental level, it's an information exchange network between people. So I find information from a blog post or from a YouTube video, someone created information and then I'm consuming that information. And even if you're buying plane tickets, it's kind of like an information exchange where an airline has availability for planes and then I want a plane ticket and now we do this exchange over the internet and now I have a ticket. So it's kind of like information exchange. And I don't think AI changes that. I think it accelerates it like crazy, where now everyone has their AI assistant. And so you imagine every human has AI assistance and they're exchanging information at a thousand X the rate. So they're doing the same thing, but more of it. Maybe the analogy is like, if you take a bazaar where people are exchanging goods, humans are exchanging goods, it's kind of crazy and raucous. But then you imagine New York Stock Exchange is an automated version of that and exchanges are happening way faster. They're doing the same thing, it's at a larger scale. That's kind of how I see the internet. Will it look different? I think a lot more of people's interactions with this information exchange network will be through their LLM, but I still think their internet is a place for creative expressions. People will still want to see all sorts of YouTube videos and types of content that people have created. So in a lot of ways it'll be similar just at a crazy rate speed.

Nakul Mandan:

In your view, will it be that we engage only with other humans and all informational access will be through agents? Meaning human to human on the internet. When we go on the internet, it'll only be for human interaction. Maybe the social networks or watching a video, entertainment maybe. But every time I need information, whether it's a booking available at a restaurant or some news, I'll only talk to LLMs.

Will Bryk:

Yes. I think often you'll be talking to an LLM for sure, or your LLM be talking to an LLM. Human bandwidth to the internet, it's too slow for the speed at which the internet has been moving. I do think though that a lot of people talk as if we're having trillions of new laborers or trillions of things on the internet. Really it's like, it's still billions of humans interacting. Agents are kind of like this, we think of them as real beings, but really they spawn and they spawn down. It's still software in the end of the day. They don't really exist the physical, but I don't know. It's still going to be humans interacting with humans, but you might be interacting with software that the human created in the form of an agent.

Nakul Mandan:

Your origin story with Exa is pretty fascinating given in your last job, you were literally writing a book on the history of civilization. And as part of it, you came to the conclusion that the quality of information in the world is the most important thing for civilization. And that kind of sparked or that was the genesis of Exa, right?

Will Bryk:

That was definitely one of the big motivators. I think I've been thinking about ideas surrounding Exa since even high school. I remember in high school I was like, "I'm going to start a new type of news organization" because I have had this frustration, and it continues to today, where it feels ridiculous that we can get to the moon and we split the atom. And yet if I want to understand what's going on at the border right now, it's really hard to. That to me is just an error in humanity. And when I was a kid, I thought this was like, "We have a bunch of adults. Why can't they figure out how to share information in a normal way? Why is everything politicized?" I always thought this was ridiculous. And I don't accept it. I think that's just an error in our coordination structure. Something's wrong here, right? And I think a lot comes back to our information tools. So if, for example, let's say there was some really well-funded news organization that didn't have bad incentives and just was fully devoted to the truth and they were super well-funded, I think that would have a huge impact to that in the world. So that was the thing I was imagining back then. And then over the years I've thought of different ideas. Me and my co-founder actually built a search engine in college together way before Exa because we were just frustrated that it's so hard to find high quality information. Again, that was like we were upset at the state of our information tools.

Nakul Mandan:

What kind of search engine was that?

Will Bryk:

So AI wasn't that good yet. So it was more like crowdsourcing for the best URLs. It's kind of like Reddit, but for everything, and really focused on high quality URLs. I still think that'd be really cool. And it's basically using the intelligence of humans with the right incentives, just organize, "Here are the best URLs on this topic if I'm trying to start a company or learn about physics or something." But then in 2021 when we started Exa, yeah, I was writing this book on the side and just really just became obsessed with history. I always thought you don't need to read fiction books because if you just read history books, it's insane what is happening. The deeper you go into any part of history, the crazier it is. It's just a fascinating story of humanity. I was writing this book because I felt like no book captured the excitement I had. And I thought maybe as a new... I was not an expert by any means, but I was just so excited. I was like, maybe I could capture this in a 300-page book. It had a really interesting structure. I was really excited about it. But as I was writing the book, I was doing a lot of research because I had a cover of 5,000 years really fast because I didn't want to write this book for my whole life. So I was doing research really, really fast, and it's just really hard to get deep into a topic using Google. So Google's great for service level investigations. "Okay, where is this Costco? What's the address?" Stuff like that. But if you want to find, "Give me every article on what they ate in medieval times across different regions of the world," you can't expect Google to do that. It's not meant for this deep research. And so yeah, just these were the ideas in the time. And then the critical reason we started Exa was the technology became possible to build a new type of search engine. This is the time when GPT3 had recently come out. So GPT3 was this magical technology where you could type in a whole paragraph and it would fully understand all the words in your paragraph. And then Google at the time was like, you can't type in a paragraph to Google, you still can't. It just wouldn't understand anything more than a few keywords. And so what if you could take the technology of GP-3 and put it into a search engine, you could have perfect search. That have a huge impact on our information ecosystem and hopefully solve some of these political problems that I noticed since high school.

Nakul Mandan:

So it's a pretty mission-driven start of the company. How much of that now is your private fuel or you're looking for that in every person who joins? Is this a recruiting filter almost for you?

Will Bryk:

So I've been doing this for five years. So no matter what happened to the company over the course of five years, I always thought this was a very noble mission and that drives me at the foundational layer. But definitely increasingly, there's this feeling of just like, it's almost like a video game. It's so fun to reach the next level. With every 10% increase in people, there's new problems, new emergent problems we have to solve. Now we're starting to have interactions with not hundreds of companies, but thousands of companies. It's just fascinating problems all over the place. Soon we'll have to talk to governments. I'm very excited for every next stage. And so that is an increasing percentage of my excitement.

Nakul Mandan:

But as a recruiting filter, is that a critical filter for you? If somebody is not bought into the mission and is joining Exa because it's a rocket ship, is that kind of a negative for you?

Will Bryk:

Okay. Yes. So people's motivation for joining a company is very important to me, and building the right culture is very important. Someone who wants to join just because the rocket ship, that's bad. But they don't necessarily have to believe in improving the information ecosystem. That doesn't have to be their primary driver to be like... That's not something I look for necessarily. I think if they do believe that it's awesome and I get all excited. But some people are really excited about Exa in various dimensions. So one is just like a lot of the engineers join Exa because they're super excited about exabyte scale systems. That's really cool. Or training models for retrieval, it's just a fact- ... scale systems. That's really cool. Or training models for retrieval, it's just a fascinating problem. That's okay too. I think definitely I look for passion and quirkiness and thinking differently. They don't necessarily have to believe in that mission. I think it's actually my job to convince them that this is the most important thing in the world. I have not done the best job, I think. I think it's something I think about a lot is I am so, so excited about this and I don't think I've done a great job of explaining it because it's actually, it's kind of hard. It's like perfect search, what does that really mean? For example, if I had enough time, I would make a video explaining what is the day in the life of someone with perfect search. What does the world look like? So there's more work to be done for me there.

Nakul Mandan:

What I understood from preparing for this discussion was that you did hire for rebelliousness in the early days and first principles thinking. How much of that is then scalable? The first hundred, maybe they were all rebels of some form, but has it stopped scaling as you've gone from 10 people to 125 people in the head?

Will Bryk:

I wouldn't say it's rebels. I think there was some article that wrote that. It's not quite rebels because you can have a rebel without a cause, someone who just wants to be different. That's not good. You want as first principles thinkers, meaning people who just will really think about the problem and not care how it's been done before. We've always hired for that at Exa. We've always attracted people like that, because from the beginning it was like, "We're going to build a better search engine than Google." Only someone who could think from first principles could imagine that that's possible. And so we continue to really try to hire only people like that. Problem is you might run out of people like that. There's not infinite supply of people who are like that, but we try as best as we can.

Nakul Mandan:

How do you test for that first principle thinking? Because a lot of people look smart on paper. What is the recruiting process to assess this?

Will Bryk:

Yeah. I mean, so for example, on the engineering side, we have a couple of technical interviews. Those are reasonably standard. We've chosen for interesting, not just leak code type stuff. But then I think the big thing that helps us test for this is we do a work trial, meaning people come to the office and work with us.

Nakul Mandan:

For how long?

Will Bryk:

For one to two days. And that's critical because then you really get to see what someone's like. So you get to see first principal thinking because you give them a real problem. So you can't really see someone's first principle thinking if you give them some lead code style problem. You have to give them a real creative problem and see how they think through all the trade-offs and stuff like that. You could tell just from just seeing the way they think through the problem. Awesome thing about a work trial is you also get to see all these intangibles. Like, "How do they interact with people at lunch? Are they optimistic? Can you imagine working with them every day? Are they locked in during the day?" All these kinds of things you could test during a work trial.

Nakul Mandan:

I read somewhere, correct me if I'm wrong, that you try to, for people who are coming for the work trials, you try to have a meal with them. Is that kind of the intangibles?

Will Bryk:

It's starting to not scale, because we have multiple at the same time. I definitely walk with everyone or I try to, or me and my co-founder, that's very critical. We should continue doing that for literally as physically long as we can because we've gotten good enough for a 15-minute walk. We could see a lot in someone. We have all sorts of methods that we do and places we take them and just see how they react.

Nakul Mandan:

Double click on that. How have you gotten better? What are traits that you are now looking more for? And what are traits that you've stopped caring about that you may have overcared two years ago?

Will Bryk:

We really try to avoid non-playable characters is the way of saying it. People who just go through life as if they're on a stream and they have no agency over themselves. Someone should be like, they think about their life and think about the world and adjust accordingly and don't just follow. Like non-followers. And how do you test for that? Well, you could just literally just ask them about their life and their decisions and stuff like that, and you could kind of see. Also, I just tell people about where Exa is going and just see how they react to it. Do they get excited? Do I see a sparkle in their eye? Reasonably easy to see a sparkling eye. Sometimes people are jet-lagged because we have to fly them, so you have to adjust for that. And so all these subtleties you learn over many trials.

Nakul Mandan:

Are there traits you used to care a lot more about two, three, four years ago when you started the company that has just like you've gotten better at looking for something that now you think, "I was over-indexing, it doesn't matter"?

Will Bryk:

I'm sure there are a lot. One thing that just comes to mind when you're asking that is speed of talking. I happen to really like fast talkers just because I talk really fast.

Nakul Mandan:

Yeah. That is a surprise.

Will Bryk:

And I don't have time. I just want to move fast and make decisions fast and I get really bored. But I think one thing I've changed there is there are some people who speak slower, but every token that they say is really high value. So it's not actually the speed of words. It's valuable tokens per second is kind of the right metric. So I've adjusted there.

Nakul Mandan:

Is there a type of person who's clearly talented, but you have realized is not a good cultural fit for Exa?

Will Bryk:

Yeah. I mean, there are a lot of really experienced ... It depends when you are talented, but definitely there's a lot of experienced people and that could often be a negative thing. There are a lot of people who are like incredible pedigree, even have done amazing work in the past, but can they adjust to this new agentic world, whether it's the tools they use every day to build or just the way they think about retrieval? Sometimes people are locked in the past.

Nakul Mandan:

The war for talent is insane right now. Every company we've invested in feels it. I mean, every founder that I've ever talked to in the last few years feels it. It's always been intense, but it feels crazier now. And you're fighting against these labs who are growing super fast, have big missions. It's not that you're the only company with a big mission today. They're literally going for AGI and they haven't seen packages. How are you fighting for the top talent today?

Will Bryk:

I mean, the types of people that joined us in the early days were people who just don't want to join bigger companies. And I think that's still true today. Look, I mean these companies, they're building really cool stuff. No one could deny that, but they're really large now. And not everyone wants to join a 5,000 or a 10,000 person company. It's very different. You don't feel like you have agency over the output. Exa is still at a size where you could have a massive impact on the company. You could know everyone's name and it's just a really different experience. So the experience is very different. I think also just the upside could be insane at Exa. If you believe in the mission, whatever we're worth now, we could be worth a 100,000x that. And so that's a big deal. We shouldn't just have people who are interested in that, but it is true that you're not going to get a 100x return from one of these big companies anymore. So I think over time as Exa has become clearer that we could actually do this and these other companies have become really big, it's been easier to recruit.

Nakul Mandan:

As you've scaled, there's an early team, which is first principles. I know rebel is not the perfectly right word, but directionally that. As you've scaled to 125 people, have you gone towards a point where you now need some people who are execs and more process-oriented? As you are finding that balance, how are you ensuring that the culture and the early culture retains itself with more scale?

Will Bryk:

Yeah, for definitely hiring execs now, which is a whole different ... It's interesting to do that, and I've had to learn. I try to avoid process, kill it before it lays eggs. I'm very against process. Someone recently tried to introduce a process and I got really upset. No process. I think you could solve a lot of problems with culture, not process, because process is kind of a rule. So you add a rule here, and then you add another rule here, and then you add another rule here. And eventually you feel all corpo because you can't move and no one can do anything because they're constrained by all these rules. I would much rather get rid of all the rules and everything's fluid and just like, "Oh, we have this cultural thing where we do this, we do this." And I just don't like codifying things into process. But yeah, I mean, we are hiring people who have experience coordinating large groups of people, which is definitely a skill, obviously, that you could learn over time from other companies. And we try to hire from the best companies so that they've seen what great is, and then they bring that into the company. Definitely a different type of skill than a new grad who's just works crazy fast and just has no priors on what companies are like. They have amazing ability to think differently, but they don't know what great is necessarily. And so having this mix of people is very valuable at a company our stage.

Nakul Mandan:

Maybe double click on culture. You touched on that, culture versus process. How would you describe Exa's culture?

Will Bryk:

So culture, I could say it in words, but the best way to feel someone's culture is to go into their office and just experience it. And so it is pretty shocking. You come in and just so you hear laughter. Well, first of all, it's very bright in the office, which is beautiful. The sun comes in, which we deliberately chose the office. It would feel like an amazing experience. It's like this double-decker. It looks like one room, like a coliseum. You come in, there's just so much activity. It's just the energy is super crazy. So it's high intensity, people are having a lot of fun, they're laughing. Every company probably has kind people, but I think they're passionate kind. I don't know if there's a new type of word for that where it's like they want everyone else to succeed. Everyone's helping each other. So these are common things in other cultures, but it's really dialed up.

Nakul Mandan:

Is it intense? Is it 996? How would you characterize in today's AI, Silicon Valley times?

Will Bryk:

I don't require people to work at certain times, but people do work extremely hard because they are excited by it. So one thing that's really important to me is I make sure everyone works on whatever they're most excited by. And luckily, we are in a business where in any dimension of improvement, we make more revenue, because information is so essential, everyone needs it. And so it happens to be the case that people are just working on the things that are most exciting to them. And when you do that, it doesn't even feel like work, so people work really hard. I'm not telling them what to do. Having to enforce a rule like 996 is a symptom of a problem that you are not matching the things you need at the company to the right people.

Nakul Mandan:

Could you double click on the process part of it though? You specifically said you're anti-process. How does that scale over time?

Will Bryk:

I guess I'm learning that.

Nakul Mandan:

Fair enough.

Will Bryk:

And I would say some of the places we can improve at the company are probably due to this we don't want to ... Not having process sounds great and it has benefits, but they're also negatives where there are coordination problems. And so we're figuring that out as we scale very rapidly, all these systems. But yeah, I mean I'm anti-process, but we basically make mistakes and then we quickly have reviews on what happened and then try to ... Now, we have this cultural learning. It's kind of like a tribe, right?

Nakul Mandan:

That could be a process in some ways, reviewing.

Will Bryk:

Yeah, we do retros on any problem. We constantly talk to each other. People are talking a lot at Exa. They're exchanging information. So we're in around a hundred people. When we're 300 people, then this might not work anymore. I would, as much as possible, I'll try to have as minimal process as possible. Probably at some point we'll need real process. And of course we do have some processes, being like the founder mode type thing. If you don't just constantly think about the culture, it could become siloed and rules everywhere. And you leave for a week, you come back and it just looks like a different company. So I'm constantly observing what's going wrong and trying to remove bottlenecks. And processes are an example of a bottleneck.

Nakul Mandan:

We talked about the team and the recruiting and the culture a little bit. Let's shift gears to how you actually run the company. Starting with you, maybe how do you spend your week these days? How much is spent on product, vision, engineering, sales, communication with the team, recruiting?

Will Bryk:

Obviously, change over the years, five years. The first couple years I was just doing engineering and research, which is very fun. Then there's a spectrum from that to all people stuff, and there's a spectrum and it's increasing. I mean now, I don't do any engineering and research anymore, but I do still at a high level. It's basically become more and more abstract. I stopped coding, but I still did architectural decisions and now it's more high level architectural decisions. I'm still involved in those. Yeah, engineering and research direction, sales, talking to a lot of customers, marketing, I think a lot about that. And then just high level what's happening at the company.

Nakul Mandan:

If you had to break it up into percentages, do you have a sense today? The Exa of today, how does it get well?

Will Bryk:

30% just the operation of the company. 30% go-to-market marketing. 30% engineering, research, vision, reviewing of work.

Nakul Mandan:

10% just catch all?

Will Bryk:

I can't do math.

Nakul Mandan:

Yeah. I was very curious that you talked about how you believe bitter lesson is how you approach search. And just maybe just give color to our audience on what bitter lesson refers to. But I was curious if you've extended that to how you run your company.

Will Bryk:

I think bitter lessons like the idea that you could try all sorts of heuristics to solve some problem in software, but if you just dump data and compute onto something, it'll get better. Those systems always went out. I feel like the bitter lesson, it doesn't really apply to humans as much. You can't just dump humans onto a problem and it just works better. For example, you do not want to over-hire. We could hire hundreds of people right now. That would not be good, because humans are not like GPUs. They're not all the same and they don't perfectly work. They don't perfectly coordinate. So I would say a bitter lesson does not apply to humans. Maybe that's why humans will not win out in the end because we're not like an ant colony that perfectly works with each other.

Nakul Mandan:

Is part of the bitter lesson not applying is that there are more talented people and less talented people, not every human is fungible, but every unit of compute is fungible and is an additional equal force?

Will Bryk:

Let's say imagine a world where a bitter lesson worked for humans, meaning you just dump more humans onto the problem. Or you could try this organizational structure of that one, but really it's all about just dumping humans on the problem. You would need automaton humans who just work perfectly in coordination. Sometimes when I imagine humanoid robots in the future in a factory, they're tossing a tool and they perfectly use it, like that's just not how humans work. Humans are way more inefficient. It doesn't quite apply. Yeah.

Nakul Mandan:

Partly I'm trying to understand as to the fascinating thing about Exa is it's growing super fast. It's become an important part of the infrastructure and is continuing to be. And you're a first-time CEO who's learning as you scale. And you've obviously done a great job. That's why you are here in terms of the number of customers and the explosion and usage that Exa is seeing. But I'm also curious as to how you are evolving yourself as a leader and a manager. Have there been managerial philosophies or thoughts that you've evolved over the last few years where you used to think this way and now you've completely flipped in how you run the company?

Will Bryk:

Well, definitely how I run the company has changed a lot. Again, I'm doing a lot more people management, people coordination issues, less just obviously engineering. I've definitely had to adapt. Being a leader for five people is very different being a leader for a hundred people. A lot more of my job is aspiring. So I've had to learn how to inspire, how to talk in front of a crowd. I don't know about its philosophies. It's more just like I'm picking up skills as I go and I realize, "Okay, wait, the way I used to operate, the way I used to spend my time, it worked when it was 30 people, but now it's 60 people and it doesn't work anymore." So I'm learning as you go as fast as I can. Definitely meeting amazing people, whether it's our investor who's just ridiculously good at sales, let's say, or a new engineer who's ridiculously good at architecting things, and I just take that. I'm trying to be the best possible leader, but I think I have no ego, basically. I just take things and then just I'm trying to form the best possible leader. So I'm constantly asking people for feedback. One thing I do is I constantly be like, "Hey, just give me critical feedback." And one thing I've learned is people never give you critical feedback, because they're worried or something, and I try to force it out of people. And so, one thing I've tried is you have to ... I will not stop. Often I go on walks or we're a walking culture. Like I say, "You have to give me critical feedback or else we won't stop walking." And then they'll say something. But like I said, I'm trying to constantly learn. Yeah, so I've had to change a lot.

Nakul Mandan:

You also mentioned how there are different stages of the company that you've already experienced. The company was one way when you were 10 people, then it changed, and now you're 125 people or such. Could you talk about the two or three different- 25 people or such.

Will Bryk:

Yeah.

Nakul Mandan:

Could you talk about the two or three different stages that have happened, and what broke, and what then changed in you?

Will Bryk:

A lot of it, by the way, has to do with physical office structure. So for example, when you're four people, you fit in a room. And then a big stage that happens is when you can't fit in a room anymore. You just physically can't fit in one room. Well, now there's an information... Because a company's like a bunch of people sharing information. And when you don't fit in a room, now you have to... How do you make sure everyone knows what's going on? The next physical infrastructure thing that changes is you can't see each other. Even if you can't fit in... You might have one office space that's like one giant thing, so you can see each other, but then at some point the office gets big enough where you can't all see everyone. So now, not only you don't know what everyone's up to, but also it's just harder to exchange information. It's really about information exchange. Information flow is probably the biggest thing as the company scales. There's also so much going on at the company now that I don't want everyone to know everything that's going on. So you physically can't fit it into your context finder or your memory, and it's just not efficient. So you have to make sure that only the information that someone needs goes to them. And that's a really fascinating problem, and I think AI can totally help with this. Everyone uses AI to be way more efficient on an individual level. I don't think we've seen AI really accelerate productivity on an organizational level. And that's a really interesting concept. For example, basically, we would love to have a bot where I have an AI assistant and it tells me all the stuff I need to know. So as CEO, the stuff I need to know is a lot more high level. Sometimes it should go low level. For an engineer it might be like this thing, for a go-to-market person it might be this thing, but it's like some subset of all the information at the company. Right now we use Slack, and there's just like so many things going on. So we're trying to figure out this.

Nakul Mandan:

Exa is also probably one of the most AI native companies on the planet. How do you guys use AI operationally? So maybe starting what you just hinted. Sounds like you use an AI assistant?

Will Bryk:

Mm-hmm.

Nakul Mandan:

And it gives you exactly, on a daily basis or an hourly basis, this is the information that Will needs to consume right now?

Will Bryk:

I do have this, but it just doesn't. It doesn't work. But why? Because it's missing context. And everyone's realizing the same problem, that there's so much context. If people are talking with each other and in the physical world it's not being recorded... And if you miss that, then your summary is just going to be off. If it doesn't have eyes, if your AI is only in the textual world, then you're missing someone's emotion, and maybe someone's not having a good day. All these things matter. So we are very AI native, but there are huge limitations. And we are not yet in a world where I could just get a summary and that's all I need to know. I still need to constantly walk around the office, talk to people, get a feel. I actually will look at people's faces and be like... Just see, is something wrong? I overhear conversations. I'm constantly just walking around and getting a feel for the day.

Nakul Mandan:

It's probably slightly reassuring to hear that even an Exa-like company has not solved everything with AI. But everybody kind of knows how engineering has been AI-fied, but let's talk about some of the other departments. How do your sales and marketing departments use AI today?

Will Bryk:

Okay. So talk about Exa being very AI native. Yes, we have Exa.ai. We also want AI Exa, which is kind of like an initiative to make Exa super-AI-pilled. And one thing we do is we have, anyone at the company can get access to... Can run evals, for example. Whether you're an engineer or a go-to-market person or ops person, you talk to the assistant that we've created and it just has access to a bunch of our systems. Or hey, I want to know, has this customer's usage gone up? We could just do that. Everything's just very easy to interact with. So that's been a huge boon. Yeah.

Nakul Mandan:

But are salespeople fully using all kinds of AI tools? Are they building their own AI tools? Are your salespeople on Claude Code and just building their own automations for outreach or faster responses to customers or prospects? Or ultimately sales are sales and they use some AI tools but it's not as much as if you were a salesperson?

Will Bryk:

No. People are definitely building their own tools, and everyone is a builder. Sales is a lot of human-to-human interactions and relationships, and that is a majority of the job. Of course we have go-to-market engineers now. They're trying to help as much as possible, but ultimately it's a lot of relationships. Yeah.

Nakul Mandan:

I'm assuming almost all of your code is AI-written now.

Will Bryk:

That's right. Yeah.

Nakul Mandan:

What is the skill that you're looking for in engineers now in this new world?

Will Bryk:

The type of person who's very good in this AI world is just very different. And everyone can adapt, I think. But someone who can think in very parallel ways. Now people are doing like 20 things at the same time. The people who can do that and function in that world are just ridiculously productive. So it kind of rewards the ADD-type engineer. I would also say it rewards people who are obsessed with the final outcome. Maybe you could divide engineers into those who love the tools themselves and those who love the final outcome and have very good judgment about what is good product. They'll just keep telling their agents until it works whereas people obsessed with the tools and the code, and the syntax obviously doesn't matter as much anymore.

Nakul Mandan:

Do you feel like... There's always been this concept of the 10X engineer. In the AI world, what's the delta between the best engineer versus a solid AI-enabled engineer?

Will Bryk:

I would actually argue there are like 100X and 1,000X engineers. And the reason I'm saying that is because there are some engineers who will do something that, even if you had a 100 or a 1,000, they would never be able to do it, which is a little different from the AI tooling. But those engineers who are able to think those thoughts can now just implement it. So AI tools make people extremely productive at doing their things. So you can imagine someone... Yes, you could have someone do what other people would do, but 10X faster. But I think the more interesting thing is someone who would do things that the other person would never be able to do. So I'm very excited about those people who can just think thoughts that would never be though of by 100 average engineers.

Nakul Mandan:

What about the engineering leader? How has the role of an engineering leader evolved in the AI world?

Will Bryk:

I think because everyone has so much agency, an engineer leader is a little bit more hands-off. At least at Exa, we have a very decentralized organizational structure where the best ideas come from anywhere. And so I think engineering leader is very much like a coordinator, an unblocker as opposed to a top-down, like, "Here's what you should do." Just because the best ideas are coming from everywhere. Everyone's talking to genius AIs. Imagine if you had an army of interns and each intern had access to Terence Taos. You don't want to tell those army of interns what to do. You want the interns to talk their Terence Taos and then come up with the ideas. So I do think an engineering leader is more a coordinator now.

Nakul Mandan:

Are there roles at Exa today, because of the pace at which this AI is moving and enabling individuals, that didn't exist? Maybe one that you already hinted at is a GTM engineer. Are there roles in a department that are performing this role? Do you have almost like every department has its own engineer?

Will Bryk:

I don't think every department needs an engineer. I think everyone could basically act like their own engineer, so increasingly you might not need that. I would say there are interesting roles now that are happening at the intersection of roles that used to be separated. For example, design, front engineering and then, I don't know, deployments can all be one person. Right?

Nakul Mandan:

Mm-hmm.

Will Bryk:

Basically, people can do a huge amount of things now across different departments. We don't have departments, but across different parts of the company that they couldn't do before. The generalist really wins in this world, basically. You don't want specific people.

Nakul Mandan:

When I talk to people, a lot of CEOs are telling me, even in the AI orgs, like I was talking to Arvind at Glean recently. And many of the AI native orgs where they're saying, "Even within the organizations, there's the top 5% of fully AI native people who are just supercharging themselves with all kinds of automation." These are, like you, pretty AI native orgs, but they're still laggards.

Will Bryk:

Yeah. No, I would say everyone is using AI tools that... There's no one who's not. There are people who use it better or more and then there are people who use it less, but everyone understands that it's just way better to use it.

Nakul Mandan:

What's the hardest mental shift that people have to make if they're coming up in traditional orgs, they've been working for 15 years?

Will Bryk:

At Exa the hardest thing people need to transition to is the speed of shipping. And so in traditional orgs or bigger orgs you might take a long time to ship something. Maybe a month. A month is an insane amount of time. Nothing should ever take a month to ship at Exa. And so we try to teach this culture really fast. The first week, you've got to ship something important and then post about it. And that's really critical because it teaches people, okay, this is the standard. And you see people shipping things like nonstop, multiple times a day, and that's expectation. I think people will rapidly adjust to that. We only hire people who can do that kind of thing but, yeah, some people have to adjust.

Nakul Mandan:

How do you even manage today's pace of shipping? You're talking about people shipping multiple times a day.

Will Bryk:

Yeah.

Nakul Mandan:

How is the organization coming together on the engineering side as a cohesive unit attacking the goals if everybody's shipping almost every day?

Will Bryk:

Yeah. One thing we put a lot of effort into is automations. With our search engines we have constantly doing automated evals. And I've heard of bigger search engines like Google. They have to go through. A human has to look through the final thing. And we try as much as possible to just have AI automation so you feel very comfortable shipping something. And as long as it passes automations, you're good. And so we have really smart intelligences testing all parts of the product before a new thing ships. So that really helps.

Nakul Mandan:

Maybe let's shift gears towards how you are scaling as a CEO. The company has 10Xed in the last year or so. What's the part that you've had to change in your leadership style just in the last year, from 10 to 25 people?

Will Bryk:

Honestly, I'm thinking the biggest thing is I've had to learn how to speak in front of crowds. I would say that's one of the biggest skills I've learned, is every Monday I give kind of like a speech in front of the whole team. And now it's getting quite big. And I was scared of that kind of thing when I was younger but now I love it. And so I think there's an example of just, I was forced to do this every week, and what used to be really scary is now something I really enjoy. I've had to learn how to make decisions really fast, and so... And this is, I think, an awesome skill that has carried to the rest of my life. Like, "Oh, should I go to that restaurant or that restaurant?" Boom. It's done. I just don't.

Nakul Mandan:

Yeah.

Will Bryk:

It's really fast. I've had to learn how to communicate extremely clearly, because one thing I was surprised by when we got to the 20, 30 person size was if you're at all unclear in, like, "What exactly do we do here?" It'll just trickle down to the company. I'll have serious problems. So I try to be extremely clear in my messaging. Certainly when I send a message to the whole company, I edit that for like an hour.

Nakul Mandan:

Because as a CEO, like every word matters. Like everybody is hanging on to me.

Will Bryk:

That's also a bit of surprising thing. I've had to learn to watch, a little bit, what I say in terms of-

Nakul Mandan:

Yeah.

Will Bryk:

I might say something like, "Oh. Lunch should be at this time." But I'm joking.

Nakul Mandan:

Yeah.

Will Bryk:

And then suddenly it'll be at that time.

Nakul Mandan:

Yeah.

Will Bryk:

And so I love brainstorming but sometimes when I'm brainstorming, people interpret it as commands. They're not. So I have to be extremely explicit. Like, "FYI, this is just complete brainstorming. Feel free to just completely contradict me." And that's been helpful.

Nakul Mandan:

But you've also, maybe on the human side, have become a manager of managers. Right?

Will Bryk:

Yeah.

Nakul Mandan:

So one goes from an IC to a manager to a manager of managers. How has that evolution hit you? Or not really much different?

Will Bryk:

It's been sad to be out of the details of everything. So when we were five people, 10 people I could be in the details of literally everything. I used to pick out what chairs we would have in the office, and I loved that. I love being in details. And so there's been this thing where I have to increasingly not be in the details of things and only be in details of things that are problematic or the biggest bottleneck for the company. As I manage people who manage, I have to trust that they're going to get into the details. Because that's kind of been a thing to learn; how do you manage people and make sure that they bring that intensity and detail-orientedness to what they're doing? And so I haven't learned how to do that well.

Nakul Mandan:

You came up as an engineer. Have you already made the mental shift from being an engineer to a CEO now?

Will Bryk:

I like to think that they're not that different. What is an engineer? They're trying to solve problems. I'm not a software engineer anymore, but I am an engineer. I'm engineering the company. I'm trying to solve problems constantly. Every day is just solving problems. Even sales. Trying to have a customer like you and solve a problem for them. The way I said it is clear. It's engineering. So if you abstract engineering, I'm still an engineer. But yeah, it is a lot of people, like engineering people in the culture and processes and stuff like that.

Nakul Mandan:

Are there decisions or situations where your engineering instinct kicks in naturally and you've had to correct yourself and say, "Hey, because of the human aspect of leadership, I need to think differently?" Are there instincts you've curbed?

Will Bryk:

It could sometimes be overly-logical, saying exactly what is true. And sometimes you have to... It's more of an emotional argument. For example, if I'm inspiring someone, or the team, I don't know exactly how big we'll be or how many people will be using Exa, but I might choose a number. Like, "Hey, like a billion people will be using us in two years." I don't actually know that so my logical brain is like, "Well, it could depend on this and..." When you're giving a speech, you just say, "A billion people," because it's roughly right. And does the details matter? So that's the kind of thing I've had to learn; go a little bit less logical and more emotional. And that's an example.

Nakul Mandan:

You're learning fast and the company's growing super fast. Who are your advisors or even CEO friends who are your sounding boards on decisions or even just personal growth as a leader?

Will Bryk:

My co-founder is super critical.

Nakul Mandan:

Yeah.

Will Bryk:

We've known each other for 10 years now. We were roommates at college. We can like read each other's thoughts. I could look at him and know what he's thinking. So that's been critical. I have no idea how we'd get here. We would not have got here without having each other. Yeah. We're like one organism, which I love. And then we have advisors, and I learn from... We have an amazing board who, I don't know what other boards are like because this is my first company, but this board is amazing. And we have advisors who are amazing. I don't have some crazy network where I go to retreats or something. I don't really do that. One interesting thing is, a lot of my role models, they're not real. They're fictional. I have been obsessed with superheroes since I was a kid, and I continue to be obsessed with them. And I think they're actually like... It sounds ridiculous, but they're critical to how I make moral decisions, for example. I look up to them like I do, like, to my older brother. I also have an older brother who's been a role model.

Nakul Mandan:

Is there advice that you've gotten along the journey, now you look back and you feel like, "Man, that was completely wrong advice as I scaled?"

Will Bryk:

Definitely the first advice I got when I told everyone we're building a search engine better than Google in 2021 was like, "What? Are you nuts?"

Nakul Mandan:

Yeah.

Will Bryk:

I kind of laughed but also was like, "Maybe I am nuts." So there was definitely, I wasn't sure this made sense for a couple years. I had drive. Like, "Okay. Maybe it's possible." I think maybe like a year and a half in we made a retrieval breakthrough that really was like, "Oh my god. You can be way better than Google. It is possible to do all this crazy stuff." And then there have been other... I don't know. Yeah. There's been so many, just advice that I've gotten that's been wrong. No one knows the company remotely close to you. And if someone knows it better than you, that's a problem. Something's wrong. You're obsessed with it. You're living it every day. There's certain types of advice, if it's been done before like building a go-to-market org, is not... We don't have to reinvent the wheel. So of course, for that kind of thing I could totally take advice from our CRO or something. But for the future of AI search, no one's going to really know that except for us.

Nakul Mandan:

You talked about how you've had to learn about hiring execs. And now that you look back, you have a ton of advice, even for the young CEOs who come to you sometimes. Could you talk about, what were you bad at when hiring execs that you've completely grown out of?

Will Bryk:

Yeah, yeah. I think it took only like a month or two to get good or get better. I'm sure there's people who are way better, and I'm still learning. There were a few things that were just very obvious like, for example, I would do Zoom calls instead of meeting in person. We were looking for various execs. In person is just way better. And these are important hires. It's worth it for you to spend a lot of time going to wherever they are. Even if they're in Australia, you've got to go there. So you've got to go meet them. One other interesting thing is it's really important to show that you care about them and really want them. Maybe at first I was more about, like- ... Know that you care about them and really want them. Maybe at first I was more about having them sell me on themselves, but really you have to sell yourself, you have to sell them, you have to tell them that you want them because if you have good execs you're trying to hire, they have infinite optionality. And so they have to really trust that they're going to have huge impacts at the organization, that you're going to care about them. And so I think the right way, the right balance is, "We're awesome and you're awesome. We really want you to join us." It's like you're leaning in really heavily. You're not acting too cool.

Nakul Mandan:

What actually drives you is a question that I would love to understand. You're clearly an intense person who's working insanely hard. Could you take a step back for us and where does the drive actually come from?

Will Bryk:

I don't think I've really taken a break ever. I worked extremely hard in elementary school and in middle school, high school. My friends would come over to our house and I would not meet them because I feel I just wanted to work and stuff. I mean, I grew up in a hardworking family. My parents were really hardworking. I just wanted to do something big always. So I've been journaling for a long time. Looking back at my journals, I'll always see, okay, what if I was like this crazy thing? I could be like Iron Man or it's changed over the years. Albert Einstein was inspiring. I wanted to be like those types of people. I never really cared so much about my personal life, I think is maybe one interesting thing. I just don't care. I don't care about going to the park and enjoying myself. I just never cared. I really just want to build something amazing. So that's always been the drive. And so obviously to build something amazing, you have to work really hard. I also just ... I get excited by things. Exa is the perfect storm of exciting thing that really matters for the world. And that wasn't always the case. When we started, it wasn't clear we were going to build something awesome. It was a bet. And again, a lot of the inspiration was mission-oriented. It was like, okay, this is important for the world, so it's a noble mission. Even if it fails, it'll be valuable. And there were a lot of hard things, especially in the beginning days where it was like, "Okay, does this even make sense?" Rolling in my bed at night. "There's so many other things to do. Does this make sense?" But the mission really motivated me. I think once it started going, my motivation, once it started working, my motivation started to change. So certainly when we started hiring people, okay, now lives are on the line. People who are really smart, who could have done anything are now relying on this company. So when it was 5 people, 10 people, you kind of feel that, but the bigger it gets, the more that ... Now we have over 100 people who could have done lots of other things and they're doing this, they have families. I feel that pressure and I want this to be the best time of their lives. That's a pressure and it's motivating just to know that the tribe is relying on me and the company. So mission, it's just a noble cause. The people. And now the third thing is I actually think we could do it. I fully believe that we could be the information layer for the world in a way that no one's ever done before. Perfect search. Infrastructure that never existed before will exist because of Exa. One thing that's interesting, I have never had even a thought about burnout in the past couple of years. Zero. It's just not a thought in my head. I could do this for decades until it's solved. I don't think it'll take decades. Luckily, I think it won't take that long. I just feel this incredible responsibility. So I think competition is very exciting. So that also adds a little bit of a game aspect to it. I loved sports when I was a kid and still loves sports. I don't play that often, but it's just more fun when you're competing. For example, if you're running in a track, you could run fast, but when someone starts coming up behind you, you run faster. It's kind of like that vibe.

Nakul Mandan:

It does sound like the desire to go big seems like you were saying earlier that it was innate pretty early and then you found your mission and then it has compounded to a competitive thing that you're having fun building and now it does seem like it's compounded to a different level.

Will Bryk:

Yeah.

Nakul Mandan:

You also said superheroes are a huge part of your mental models and psyche. Where did that come from?

Will Bryk:

The first superhero movie I remember watching is Spider-Man 2002 with my dad. I was six, and that was life-changing. I mean, that was like holy cow. He was a nerdy kid. I was a nerdy kid, and he just became the superhero who changed the world. And I love that he was so humble about it. He just helped the world and he had to give up the thing he loved the most, Mary Jane. The last line was like, "This is my gift, my curse." I'm getting shivers when I say it. That's such a beautiful message. And so a lot of superheroes have that exact same ethos where it's they have something they need to do for the world and they don't take recognition for it. They often have to hide their identity to save their loved ones. But that's pretty cool that they're very self-sacrificing. So I've always looked up to those types of people. So Spider-Man was the first one. Then the next Spider-Man came out in 2004. That was pretty good too. And then what was it? Iron Man, I think 2008, that was a big one because I was like, okay, this is a superhero, self-made through technology. That was very inspiring. The Dark Knight series was the next big one. Became obsessed. I know every line. I love that one because, I mean, you said you like Batman? So I don't have to tell you. And then I would say the next era of superheroes for me was the Mission Impossible series. Now it's like I live and breathe that. I watch that all the time. Every weekend I'm watching clips from it. I listen to the music. I listen to this song called Firefight every day in the morning to wake up. And then also We'll Figure It Out is another song from Final Reckoning, the last movie. And it is very real for me. I don't meditate, but I imagine this is the same types of things people do, like their rituals, their religion. It has that similar vibe for me.

Nakul Mandan:

Why have real life superheroes like let's say Elon Musk or I don't know, political figures or Albert Einstein not have the same impact? Because you did mention Albert Einstein. Is it the hero's journey? Is it ...

Will Bryk:

No, I think it's actually because real life people are a lot messier. And I mean, I have immense respect for all the people you mentioned, but Spider-Man, Toby McGuire is so pure in his desires for the world. And at least a lot of the famous people in the world are just not like that. Because to get famous, often you have to be a maniac in some dimension.

Nakul Mandan:

Yeah.

Will Bryk:

And that's awesome. And they help the world. But I'm sure they exist, these people, but it's very rare to be a true Batman. By the way, why you don't know about them? Because they don't claim the light. You don't even know who Batman is. So it's almost like an oxymoron for there to be a famous real superhero because if you're a real superhero, you're anonymous. So I think that's maybe part of it. Obviously it's cinema. So cinema is going to be epic. The soundtrack is perfectly timed. If I watch the Batman movies without Hans Zimmer's soundtrack, I wouldn't be as inspired. So it's fictional. It's fantasy.

Nakul Mandan:

You mentioned you listen to the Mission Impossible soundtrack literally every morning.

Will Bryk:

Yeah.

Nakul Mandan:

And what's that ritual about? You walk towards office with that?

Will Bryk:

No, I do it in the home. I guess sometimes I do it in the Uber. Yeah. The idea is, well, what I like about the Mission Impossible movies is that they have this really high stakes mission, and they're constantly encountering problems that seem insurmountable, and yet they surmount them. And I just think that's a great analogy for a startup. We have this crazy high stakes mission and literally every day there are ... It used to be 5 problems a day, now there are like 20 problems a day. And you just have to solve them. There's no option. For me, I just, okay, there's a problem, got to solve it. You can't just let it fester. You have to solve it immediately. You do that over and over again for years, and then you just get this confidence that like, "Oh, we'll solve it." So I do say there's this meme at Exa now, "We'll figure it out." It's completely taken from the Mission Impossible movies. They say that a lot. "We'll figure it out." And it's a beautiful phrase. I don't know why it's not used more. It's very optimistic. People see we have these problems and then we solve them and then you forget about them. And I often like to remind people where we came from at the company because they're like, "Oh wait, we did that." If the search got so much better in three months, then of course in three months it can be so much better. So for new hires, we do an onboarding, I do a mission onboarding with them. I show them ... Actually, I made a video of the past of Exa, the past five years because I always start, I say, "I'm going to say a lot of crazy things about where we're going. For you to believe me, you have to see where we come from. So here's the video where we came from." And then they see that like, wow, this company went from nothing to this. We're still nothing, but look how far we're going to go. And then because they saw how far we came, they believe me.

Nakul Mandan:

Last superhero question, otherwise we can spend the rest of the evening here. Favorite superhero?

Will Bryk:

He's not a superhero technically, but Ethan Hunt.

Nakul Mandan:

Ethan Hunt.

Will Bryk:

It used to be Bruce Wayne, but not anymore.

Nakul Mandan:

Really? Ethan Hunt.

Will Bryk:

I mean, you know Ethan Hunt?

Nakul Mandan:

I know Ethan Hunt, but I would not have said ... My answer is the opposite, so we can debate that later over a drink. Segue this back into you as a CEO and then again that you've gone through, where does the superhero thing show up for you on a day-to-day or a weekly basis as you run the company? Is it more in the back of your mind, the motivation? Where does it show up?

Will Bryk:

I think it's like a guidance. It's like, what would this person do? How would they act? If I'm facing a hard decision, whether it's a people problem, I always try to like, what would Ethan Hunt do? He would never get emotional. He really cares about people. He deeply cares about people. He deeply cares about the mission. He'll make hard decisions sometimes. They're actually perfect role models. Just simulating them in my head can tell me often what the right thing to do is. They would be great CEOs.

Nakul Mandan:

What's a part of being a leader or a CEO that's been lonelier than expected?

Will Bryk:

Weirdly, as we get bigger, people are a little more ... When I sit down at the lunch table, people are a little more, I don't know, cognizant of what they're saying or something like that, which is sad because I'm just a guy. And it's just like you become a little more distant from the people at the company, though I'm extremely friendly. I walk around, I have inside jokes with literally everyone. So I try to counter that, but I think as the company gets bigger, that is just necessary or happens. So it does get a little lonelier. Yeah.

Nakul Mandan:

Is there ever doubt in the journey? Is there a nagging thought that keeps coming back every other night at 2:00 AM as you keep growing yourself also?

Will Bryk:

So 2:00 AM is I think where my best thoughts happen.

Nakul Mandan:

Okay, great.

Will Bryk:

Mostly because I know still people are up at the company, but I know most of the company is sleeping. You have to imagine, I come into the office around 9:30, and the second I come in, it's just chaos. There's so many things going on. There's so many problems. There's so many amazing things. Slack is going crazy. It's just this monstrosity of energy. So many things are going on, so many threads. We have so many priorities at same time, which I know is also an oxymoron. You can't have so many problems. We do. And then it's just a jumble and then it goes till literally ... Then I'll go home at maybe midnight or something, and then I'll shower or whatever. And then I get to process. And then my brain starts to synthesize everything that happened. And then it's like, oh wait, this is the most important. It's seeing, I think, the most important patterns. And then I'll send a bunch of messages at 2:00 AM to people, "Okay, we got to do this, we got to do that. We got to change that." And then I'll fall asleep. And I usually collapse. I'll just throw my computer aside and just collapse and then I do it over and over again. And the weekends are slower, so I get to think a lot more there. And a lot of important strategic decisions happen over the weekend. But that's kind of the day, which I just love.

Nakul Mandan:

My question partly was about more, does self-doubt ever creep in or you don't know self-doubt?

Will Bryk:

I have pretty clear problems that I need to fix. For example, I over-promise. It's something I have to learn. I over-promise. "Oh yeah, I'll get back to you. I'll do that." And I've had to learn I can't do that, but I still do it. Here's an example promise. So I don't know about self-doubt, but I like self-criticism. Oh, I'm constantly like, "Oh goddammit." And I have strong emotional responses for things that are pretty trivial. "How did I miss that?" And I'll think about it really hard, but because it's so emotional for me, I think I often don't do it again or don't do it in the same way. And so I am in a constant state of self-critique.

Nakul Mandan:

Is there managerial feedback that you've gotten about yourself, whether through advisors, CEO coaches, or board members that has been hard to digest?

Will Bryk:

I don't really think it's hard to digest. Again, I welcome feedback. But I mean, yes, I've definitely had board members that are like, "We've seen this a thousand times, you're doing this wrong." And I' m like, "Oh, got it." It's usually not hard to digest. I know they're right. It's just like, ugh, I don't want to do that. I really liked the way I was doing my job and now I have to change. But that's just necessary. For example, I really didn't like recruiting at first. I just wanted to code. And then I was like, "You, Jeff, you do the recruiting." But then I started to do it because I had to. And then I started to get some wins, hire people. So you get a lot of dopamine. And now I love recruiting. I can do it all day. But now I can't just do recruiting all day. I have to do other things. And so I've had to force myself to change, but then I end up liking it if I get good at it.

Nakul Mandan:

You mentioned that burnout is not a thing for you, and I understand that, but at the same time, there's a speed and intensity at which you're running. How are you managing your own psyche and health or you're young, you're running on adrenaline and it doesn't-

Will Bryk:

I really don't think much about this. I'm trying to optimize my sleep. I'm very sensitive to sound, light, and temperature. And it's actually impossible in San Francisco to find a place that both has AC and is not loud because all the AC is in the modern building. This is the most annoying thing about San Francisco and I just can't solve the problem, so I just deal with it. And I think I'm sensitive to all these things probably because I am very stressed. There's a lot of things going on and I'm constantly thinking. So if some sound happens, if some firetruck goes off, I'll wake up which is really annoying. So I think sleep is probably my biggest problem, but everything else is fine. And I would just say one thing about burnout. If you actually think about what burnout is, it's like it's wheels turning, but not moving. Because we're progressing, we're progressing fast and the company keeps changing, I don't feel burnout. I think if, for example, we were working so hard and we weren't progressing, then maybe burnout would start. But because we've just been smoothly progressing over the course of a couple of years, I don't feel it.

Nakul Mandan:

Okay. Next up is our quick fire round. You ready?

Will Bryk:

Yeah.

Nakul Mandan:

Red flag you never ignored in the candidate when recruiting

Will Bryk:

When someone is confidently wrong.

Nakul Mandan:

Most overrated thing about being a young founder?

Will Bryk:

You don't get much credit for it. People just want, "Does the product work?"

Nakul Mandan:

Most underrated thing?

Will Bryk:

You have no real obligation, so I can just work nonstop.

Nakul Mandan:

A search query that Exa does better than Google's?

Will Bryk:

Really any list of entities, so lists of companies or people.

Nakul Mandan:

A search query that Google still does better than Exa that irritates the fuck out of you?

Will Bryk:

I was going to say flight tickets, but it doesn't really irritate me.

Nakul Mandan:

Okay. The book on civilization, finished or abandoned?

Will Bryk:

Delayed.

Nakul Mandan:

The bitter lesson applied to something people should take more seriously?

Will Bryk:

We could have all sorts of heuristics to solving problems in the world, but I think a lot of it just comes down to more and more knowledge. So you throw more and more knowledge at the world, I think a lot of problems get solved.

Nakul Mandan:

So Ethan Hunt always figures out how to get to Mission Impossible. What are you, Will Bryk, trying to figure out as a leader for Exa's Mission Impossible to come true?

Will Bryk:

I guess perfect alignment within the company on what is most important and why this matters. This is my mission if I choose to accept it.

Nakul Mandan:

All right. Well, this was great. Thank you for having me.