a16z Podcast - Building Search for AI Agents with Exa CEO Will Bryk

Episode Date: June 6, 2026

Sarah Wang speaks with Exa cofounder and CEO Will Bryk about building search infrastructure for the AI era. The conversation covers Exa’s origins, why traditional search engines were not designed for AI agents, and how search changes when the user is no longer a human but an autonomous system. They discuss retrieval, agent workflows, coding agents, data access, and why search may become a foundational layer for the emerging agent economy. Along the way, Bryk shares his views on AI-native products, the future of information discovery, and why some of the most important problems in technology can ultimately be framed as search problems.   Resources: Find Will on X: https://x.com/WilliamBryk Find Sarah on X: https://x.com/sarahdingwang Stay Updated:Find a16z on YouTube: YouTubeFind a16z on XFind a16z on LinkedInListen to the a16z Show on SpotifyListen to the a16z Show on Apple PodcastsFollow our host: https://twitter.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Search is the gateway to the world's information. If you can make it perfect, then that has so many downstream positive implications for the world. You can kind of think of Google as being synonymous with search, right? It's one of the greatest technoplies of the last few decades. If you want to go really deep into some topic, Google fails. Most people want to understand the world. But they're getting fed information that's just like, you know, misleading in some way or straight up wrong. And if everyone had like information that was accurate, most reasonable people would be reasonable.

We have a family open club, Michael Claudeberg, and we wanted to give it web access. he was like, I recommend EXA. The world of agents' searching is just completely different from human searching. An agent doesn't just want 10 pieces of information. It wants everything. The XA, like, you could search something and then get not just like 10 results or 100 results, but 1,000 results or 10,000. How have we a team that, you know, has always been below 100 people,

have been able to build a search engine that's better than Google in all sorts of ways? Well, it's because for most of the Internet era, search was built for humans. But AI agents search differently. They need deeper context, more complete information, and the ability to navigate far more complex questions than a traditional search box was designed to answer. That shift is creating an entirely new set of challenges

around retrieval, knowledge discovery, and how information is organized online. Sarah Wang speaks with Exa co-founder and CEO Will Brick about search, AI agents, and the future of information retrieval. Welcome, Will, thank you for being here. Well, excited to be here. So I want to start with the origin. story, you've been interested in search for a long time. In fact, you and your co-founder, Jeff,

actually started building a mini search engine in college, which is not what I was doing in college. Can you say more about when you started getting interested in search and why you wanted to solve this problem? Yeah, yeah, sure. So I would say it's a life mission. So since I was a kid, I've cared about finding the highest quality knowledge, right? I was obsessed. And then in high school, I wanted to start a new type of news organization because I thought we're a civilization that got to the moon. We split the atom, and yet we can't understand what's going on at the border or in science news. Like, why can't we fully understand any topic? And then in college, I was roommates with Jeff,

and we were like, we could just build a better search using crowdsourcing the highest quality links. And we did build a pretty solid search. But then five years ago, so in 2021, that's when Transformer started to get really good. And it suddenly became possible to build a better search than Google. And that was a really important opportunity because search is the gateway to the world's information.

If you could improve search, if you can make it perfect, then that has so many downstream positive implications for the world across every industry, across every part of human life. So it just felt like this huge opportunity and no one was pursuing. And I was like, I'm going to devote my life to this because everything I care about is about information. And so, yeah, start XA and now it's gone. We actually made a lot of progress.

And we're a lot closer to that mission. There's still a huge amount of things, a huge amount to go. But yeah, it's been crazy to see how far we've come to achieving that mission that I've been thinking about for years. Maybe just to probe a little bit deeper, you can kind of think of Google as being synonymous with search, right? It's one of the greatest technoplies of the last few decades. And the idea that a startup could be better than Google at search is quite amazing. But how do you define perfect search?

And what do you see as the limitations of Google? And I'll throw in more recent events because obviously, you know, I.O. just took place last week. And they're very focused on AI mode and how they talk about the idea of information. agents and things like that. How do you think about beating old Google, if you will? And then there's, of course, new Google that's evolving. Yeah, I mean, Google was amazing and is amazing for what it's meant to do, which is like get quick answers to consumers. And now it's like increasingly longer answers. But really it's focusing on like, what do most of the billions of people in the world search for care about and like making sure they're really happy. And they do a great job of that. That's

what they're optimized for. That's why they're optimized for human clicks. It's like, you're really tired. You type in a few keywords that make no sense. And Google just magically understands what you're saying, that's magical because it has like billions of other people searching similar things. I'm excited by Google too, but like there are certain times when you want something deeper. And I was actually before starting XA writing a history book, I just got obsessed with history and I wanted to get to the bottom. What did it feel like to live in every period in history going back 5,000 years? I don't know if you ever talked about it.

I want to read this book. It's probably on hold right now. I guess it'd be really good. I would have finished it around now. But at some point I was like, okay, maybe I can build a search engine and then I could automate the building of writing of books. And I feel like that turned out to be true. But anyway, in writing that book, it became extremely obvious that if you want to go really deep into some topic, Google fails.

Like Google is great at service level information, which is great for most of the billions of consumers. But if you want to really understand what it was like to live in the Roman Empire, you know, in like 100 AD, it's actually quite hard. And like that information is scattered. It's everywhere. But it's like you need really, really good search, like really deep search to understand it. And so that was like the first realization that, or one of the first realizations that way, like, what if you could have like true perfect understanding of any topic? And so, yeah, okay, so Google has been changing their search engine a little bit. I would say I've seen many Google IOs now at this point at Exa.

Every Google I'm like, okay, they say they're changing search. And they do, but they change it more to be valuable for the consumer-type use cases, which for me, it's like there are so many different use cases that go beyond that. There's like really deeply understanding the Roman Empire, but there's also like finding every competitor to your company. And right now, Google is just not going to do that. No matter how many changes they make, like you don't trust Google to find you literally every competitor of your company, whether it's in Europe or Asia. you don't use Google for recruiting, and this announcement doesn't change that. Like, you're not going to go, you say, hey, Google, I'm looking for machine learning engineers in San Francisco who have a background at startups because it's not built for that kind of thing.

So there is an opportunity to build like a new type of search engine that's meant for extremely like deep, complex queries that businesses really care about and agents really care about. Yeah, absolutely. And I mean, to the extent you talked about starting to build EGSA five years ago and then it feels like in the last five years, the world has completely changed, which is, actually, in my opinion, a great thing to happen to you and to Exa while you're building, because you can bring in this new technology versus trying to either fight it or have some sort of innovator's dilemma. Maybe share a little bit more on why you decided to build Exa from the ground up and what parts were the hardest or have been the hardest? And then how has it changed as the world of LLMs have changed? Building it from the ground up, basically we were like in 2021,

we could build a better searcher than Google. I don't care how long it takes. I guess we were young, high energy just ready to do anything, devote our lives to this. There was a thought experiment that really excited me, which is that, wait, I could totally build a better search than Google right now, and here's how I would do it. For every query, I would take all the trillion documents

on the web, and I would run GB3 over it. I would say, does this document match the query? Is this document match query? And then it would filter it down to the top 10 documents, and that would be better than Google. The problem is that would cost, like, you know, $10 billion per query. And so then it became an optimization problem.

But at least there was like an existence proof that's possible to build a better search on Google, and that was very inspiring for me. So then it was like, okay, like how do we optimize the hell out of that? And Transformers had gotten really good at the time, and Google wasn't really leading into it. And we just had this deep belief in the bitter lesson,

maybe more than Google, for search, which is that, like, if we could develop systems, like neural systems where you pour more data into it, it just gets better and better for the thing we're optimizing for, then you could actually just totally be better than Google. When we release this to the world in November 2022, like it was actually shocking. And Andre Carbethy retweeted it.

That's pretty popular on Twitter. It was like this new way to find information. It was the first time people were like, holy cow, it's possible. to find things beyond Google. And then, by the way, two weeks later, Chatsby Tea came out. So then people realized,

okay, there's another new way of finding information outside Google with LLMs. And that was very critical to us. So, like, we released our first search engine into the world in a year and a half after starting XA, November 22. Two weeks later, Chatsbyte came out.

Actually, to me, I was at NERPS and I saw the announcement. I played with it. I was like, this feels like J.B3, but a little bit better UI, and I went back looking at research favors. But to the world, I think it was the first time

they met this new creature, and it was just easy to use, which was a very big learning, which was like, okay, you make something easy to use. It's very important, obviously. Anyway, so then AI started really taking off, and then early 2023, people started asking us for API access to our search engine

that they had used based on that Twitter announcement in November 222. And that's when we realized, oh, wait, we could start serving this search engine to these, not quite agents because that wasn't the term of the time, just these AI products, these AI workflows. And they're going to want comprehensiveness. They're going to want to search in these more complex ways.

All the ways that we as nerds in 2025 one wanted to search, agents were very similar. So that was another interesting realization was like, I'm not an normal consumer. I want to get really deep to any topic, and so do agents. And so it's cool that we were building a search engine for ourselves.

It ended up being the exact same search engine for agents. They were very similar. This is a big paradigm shift. Even as investors, we're thinking, like, hey, it's not humans who are deciding the DevTil that wins. It's actually agents. Personal anecdote, I think I told you this one,

but we have a family open claw, Michael Claudeberg, and we wanted to give it web access, and he was like, I recommend EGSA. Before we invested, I was like, Sure. I'll go with whatever you recommend, right? And so it sounds like there's this nice dovetailing of how you were intending to build eggs in the first place to what agents want. But how do you think about what agents want, right? That's sort of the holy grail right now of, hey, I don't care

about database I'm using. My agent's going to select convex or super base. These are entire tailwinds that are making some of these companies. How do you optimize for that? Think about that. I've been thinking about this for a long time since there were the first agents, right? Like, we were the first search engine. Like, we were an early search engine, and the first AI products came to us because they were like, okay, they could be a search API.

And so I've been thinking about this for a long time. And yeah, I think the world of agents searching is just completely different from human searching. I guess you make the analogy of like agents to humans like humans to sloths. Like imagine we had a search engine. I'm picturing that Zootopia, D.C. It's like, imagine we had a great searcher for sloths

and then humans came around. They're not going to want to use that same searcher. And so you should think of agents has these, like, crazy creatures that have, like, infinite, like, time is meaningless for them. They just want to, like, make complex queries very fast and, like, analyze it really fast, and they want perfect output for their human users, right? So you want to build a searching for that. So how do you, what matters for that type of creature?

Well, lots of everything. So first of all, you need a search engine, like, handle complex queries, right? Like, you do not want that feature to have to simplify its complex need for its user into simple keyword phrases because you're just losing information. So you want somebody that could actually semantically handle complex queries, but also handle keywords because sometimes you just literally want, hey, like I have this complex chemical formula. Like I want that to be part of the document, right? So you want a tool that can handle both semantic queries,

keyword queries, really just like expose all the fundamental toggles to the agent. Because the agent, by the way, it has the patience to like, you know, make a domain filter here and a keyword filter there or like, searching this way, search in that way. So you want to like have a very controllable search engine. You know, like, with Google, like, you search something, and then you're like, no, way, that's not what I want. And then you try to change some keywords. And it's like, it's just missing it.

It's not like, doesn't feel like very controllable, toggleable. You want the opposite for an agent because the agent is just going to keep searching until it gets to its outcome. And you don't want it to have to make like a thousand, like 10,000 keyword queries and still never get to its comprehensive information. You want it to like make a few queries and get comprehensive information. Anyway, so complex queries, toggleable. Also, like, comprehensive results. So this is a thing where it's like, you don't, an agent doesn't just want 10 results. or 10 pieces of information, it wants everything.

Because imagine you're an investor. You don't have to imagine that. If you're an investor and you're looking at biotech companies, you want complete information because you're making very important monetary decisions and you don't want to miss anything. You don't want to have a phomo. Like you're missing some critical startup that exists that actually reflects well or badly on the current one you're thinking about.

And so you want your agent to have complete information about every topic. So like with XA, like you could search something and then get not just like 10 results or 100 results, but a thousand results or 10,000 and increasingly agents are wanting this. You also want like lower latency because like agents search faster than humans. But at the same time you want higher latency

because certain applications don't care about latency at all. So I think another big thing with serving agents is like extreme customizability because like we're serving businesses, we're serving agents that are very different. Some want super low latency, somewhat super high latency or latency doesn't matter to them. And so it's just a whole,

It's hard to express how different. I have like a list of like 20 different ways, like humans and agents are different. And when you just build it from scratch for agents, you just make fundamentally different architectural decisions. So maybe just to go back to this point you made on model intelligence improving and how that's kind of changed the game in search as well. And I want to pose this thought to you that I'm sure you've heard before.

But given the fact that model intelligence is, you know, getting better, it sort of can almost make up for or do some of the heavy lifting in this user signal, right, that Google has collected over 20 years for page rank, et cetera. It can actually help get over that hump and do a pretty good job. And so my question to you, I guess, in that, is how do you think about this tradeoff of compute, latency, cost, right? There's all these tradeoffs that have to happen in terms of what you're actually using to, And to your point, you said it's like a big optimization exercise.

Like, how do you think about what to optimize? And I know you have different products, right? And so maybe your answer is like, well, depends on the product offering we're doing. But bring that into it as well. It's easier and harder to build a search engine for AI agents. Oh, interesting. Yeah, please, yes. I know that spark your interest.

Okay, so why is it easier? Well, first of all, like, this whole click, you know, Google has an insane amount of human click data. It just doesn't matter, like, that much for serving agents. I remember saying this, like, years ago, people thought that. was crazy. But it turns out to be right. Like human click data is great for humans, when you want to find results that humans click on, which is obvious, right? Like, so if you get a huge amount of human click data, you could train on that and now, you know, Google can understand what you mean even when you

don't even know what you mean. However, agents like just don't, they don't benefit that much from click. I mean, maybe a little bit in terms of like the ranking signal. It's also, it's valuable for agents to know what humans think is valuable. That's a very minor thing. So it's interesting that like all that click data that Google is accumulated, it just doesn't really match. matter for agents. And so it's a whole new ballgame. So there's not much of an advantage there. There's also other things like, yeah, like Google probably had, you know, hundreds of people working on re-ranking. Whereas, you know, because that was a complex thing before LLMs. But now with LMs, you could have a Reranker that you just call an LLM and like, you have one engineer work on it.

Now, obviously, we have more than one engineer working on rerankers at this point because you don't want to just call an LM. You want to train your own models and make them, you know, faster, higher quality. But you don't need a team of hundreds. You could do it with like a couple people. So like, like, how are, how have we a team that, you know, has always been below 100 people have been able to build a search engine that's better than Google in all sorts of ways? Well, it's because like LLMs unlock, the technology unlocks like new types of techniques. And then also like, because serving agents, like you don't need all the click data and like, yeah, data that Google has been collecting. So those are some ways why it's easier to build a

search engine and why we've been able to build a fantastic search engine with a small team. I think a small crack team. But it's also harder. Why is it harder? Well, it's because like the, the, requirements for a search engine now are getting more and more intense. And so, like, you know, I wake up at the morning and we have like, you know, customers love us, but we still have customers being like, wait, why can't you be perfect at this search? Or what about this search? And like, they're constantly pushing us towards the edge because we're starting business use cases

that have like deeper and deeper needs, like, you know, billion dollar investments around the line. Like, this has to be perfect. And that's great because it's pushing us towards perfect search. And so, like, basically, like, if traditional surgeons had like 99.9.9% quality or a liability. Like these these new search ends for agents need 99.99.99.99.99.99.99.99. And this is a great thing because it's just pushing us towards that dream of perfect search I've always been dreaming about. It's very similar to like LLMs. Like, you know, Opus 4.6 comes out.

Everyone's really happy. This is working out 99.9% of use cases. And then when the next one version comes out, everyone wants that thing, right? Because the extra nines of quality are so important in this new agentic economy. And so you have a similar thing for search. So it's both easier to build a search engine fast, but really hard to build a perfect search engine. Yeah. You know, it's interesting. When we were doing our diligence in the space, which, you know, has taken place over the last few years now, some people came to us with the opinion that search is just getting commoditized. And I think we look at that and, you know, if you go out there and search for information we know is out there public, et cetera,

you can't find the answers to everything still. So the fact that it's commoditized, you know, you'd have to kind of divide up, oh, what type of search is commoditized? So maybe I'll ask you that question. And then what is the type of search that's still really hard? and what is the key to unlocking that? Is it data partnerships? Is it a technical breakthrough? How do you think about the edges, as you mentioned,

in terms of perfecting or pushing forward the frontier of search? Yeah, I guess what does commoditize mean? It means like over time will the thing, will it just not matter which tool you use? Like, they're all kind of the same. I would argue that the LLMs are going to get commoditized or are getting commoditized faster than searches.

And the reason is because you don't need to run like mythos over every cell on your Excel sheet when you're trying to find competitors or something. Like there's a lot, like most of knowledge work does not require the smartest model. You act like, you know, like just an open source model that's big enough. You know, and now

the infrastructure for running them is very good. Like, it's pretty cheap. Like you can just reuse open source models for most of knowledge work. Not that the crazy smart LMs don't, they do have a super amount of value in terms of like inventing new science and math and and certain

cases you want to find any bugs in your code or something like that. But like increasingly, like I would say like if you think of knowledge work, like all the different tasks you might do as like concentric circles of difficulty, a huge amount of that service area is covered by like off-the-shelf models you get right now. Yeah, very fair. Yeah, right. So, but like on the other hand, like search, like when you are trying to, you know, enrich every cell in your Excel sheet with competitors or people you're trying to recruit, then like every extra nine of quality and search really matters. Basically, I'd argue that a lot of knowledge work is actually search problem, not only an intelligent problem. And so, yeah, I mean, what are some examples where search is not

good? I do think company and people search is the most, like, easy to see and just most value to people. Like, every company in the world has to search over companies to sell to, or almost every company in the world has to search for companies to sell to and people to hire. You could ask yourself if, like, finding companies to sell to or finding people to hire is a solved problem. I think every company would say, no, it's not. That's why people are constantly switching tools, like trying out new tools, is because we just don't have comprehensive information over all the people or companies we want. So that's a really good example.

That's something X is leaning very deeply into, like go to market intelligence. Because we care a lot about it. It's very exciting. It's also very useful to use internally. We have companies to sell to and we have people to hire. It's been great to dog to their own thing. It gives us some advantage. But yeah, that's an example.

I think you just want comprehensive. like you just want all the people that could be connected to you that are relevant. And by the way, it's a very beautiful thing. Yeah. You didn't ask this, but I think one beautiful thing about search is that a lot of important problems in the world are actually search problems, like, dressed up in a different way. Say more. What's the example? Okay, political polarization.

I would argue that's a search problem because there are people out there who want to, everyone wants to understand the world, or most people want to understand the world, but they're getting fed information that's just, like, you know, misleading in some way or straight up wrong. And if everyone had like information that was accurate and like controllable and like comprehensive, I think most reasonable people would be reasonable. And I think because our information environment is so chaotic, it's so polarized, it's causing reasonable people to be unreasonable. By way, I'm part of it's like I'm sure I have incorrect beliefs on all sorts of political things because my information is not perfect. It's something I really want to solve. Loneliness is a search problem. We,

No one thinks it's lonely. Same more, yeah. A lot of people are feeling lonely in modern society. Well, it's because they're not finding, not finding people to hang out with or to be in a relationship with, right? And so, like, yeah, lonely is a search problem, and especially in a city life,

like it's hard to find other people with similar interests or you might bond with. And this is like, you know, with a perfect search engine, with whatever information people are willing to share, it would help to find other people. So, for example, I have a lot of crazy ideas about flying cars.

I really want flying cars. I hate cars on the road. I would love to go to a group of people talking about flying cars. I'm sure I'm great friends of those people. I can't just be like, find me all the flying car enthusiasts in San Francisco. I would love to at some people to do that. Okay. And this gets in your earlier part of your question is like, what allows for this differentiation?

So it's a long answer to how to search not a commodity. You start to see that search is like a bigger thing that people think of search as like in 1990s. You see a text box. You type in a few keywords. You get a few things. Like that was search. No. Like search is way broader.

Search is coordinating the human species around anything we're trying to do. How does search become less and less commodity? Well, it's always about really good retrieval and really good data. So if you do perfectly on both those things, you have all the data and you have all the best possible retrieval, that is perfect search. And so it is like both accumulating all the web's data, you know, accumulating data that's not on the web, and then training extremely powerful models to search over it. Those have always been the two, like index and retrieval have always been the two. to pushes at Exa on the engineering side since five years ago. Can you say more about the data element?

And, you know, obviously, no need to share any secret sauce there. But it does feel like the web is getting increasingly closed, to some extent, parts of the web, right? But there's a lot of fear out there from data providers on, oh, we don't want to be stack overflowed, right? Especially if your business model is around impressions you serve to humans visiting your site. I think the fear is the greatest among those businesses. business models. How do you think about just sort of that interplay with data providers and making sure you can get to the path of perfect search, but work with these data providers that are maybe becoming more closed? Yeah. So I want to get to the ideal world where you have perfect

search and then everyone's intensivized to create amazing content, even more so than before. I actually think there's an opportunity here and I'll talk about how to create a system where like content providers are making more revenue because they are participating in this massive agented economy. Yeah. Right? So, like, I think from first principles, you know,

the first principle's idea here is that the agendasic economy is going to be huge. Like, basically, like, we're all going to have agents. So those agents are going to be participating. You can imagine it's like agented economy and cyberspace. We're basically they're doing like commerce. They're like reading information. It's like everything we do on the internet,

they're doing, but like a thousand times bigger, right? So there's a massive amount of value in this agentic economy, meaning money. And so if there's so much value, like, Instead of, like, you know, hundreds of billions of dollars going into one company, what if, you know, $50 billion a year went to one company and the other $150 billion went to all the content providers, right?

Like, there's ways to distribute the value in this new agenda economy that are more favorable towards providers of content. And, like, this is kind of how, and by the way, it won't be $200, but it'll be a trillion dollars a year of value. So there's, if we could figure this out well, like, there's an opportunity for everyone to just, like, do amazingly. I loved all the use cases you talked about. There's go to market, right?

There's also the, you know, I'll put it in the loneliness bucket, but finding people you can connect with. One that you didn't talk about that I just want to pause on really quickly is coding. Can you share more about why WebSearch makes coding agents more powerful for this use case in particular and why EGSA is such a perfect fit for the coding use case? Yeah, for sure. So every agent at some point, agents are like humans. Like at any point when you're coding,

or when humans used to code, you would have to look up information, right? Because you want, like, the most recent technical documentation or you might look at a blog for inspiration. And so agents are very similar. In particular, they really want the freshest information so that every line of code they write does not have some critical error.

And, you know, especially with coding agents, like the stakes are so high that, again, every extra nine of quality matters. So these coding agents are very intelligent right now. But in terms of retrieval quality, they've been in the dark ages or the dark ages of like the early 2000s in terms of search quality,

but like search can be way better over coding it. You know, we're talking technical documentation, we're talking SDKs, just like perfect, perfect retrieval over the sinks. That was our goal. And so we're not yet perfect, but we're extremely good at search over any sort of coding material. And so, yeah, so, you know, when, you know,

a bunch of coding agents try us, like cognition, for example, we've talked about, tried us, like we powered Devin now, and they've just found when they tested it that it just makes Devin way better, way more accurate, make way fewer mistakes, which really matters. And this will just continue. By the way, you can also think about coding agents as just like, like every agent is going to want to do everything.

And so it's not just searching over code. It's also at some point searching over the world's information, just being up to date with the news because like if the coding agent will become your agent. It's like I think coding agents and just general agents are going to like merge. Yeah. And it's an interesting trend. But yeah. Right.

No, I mean, it's clearly codex is not for developers, right? It's sort of like the Everything app to that point. I want to bring in mostly because it's in the, it's all over Twitter right now. And I think it might be an interesting tied eggs. And that's sort of this topic of token maxing, tokenomics, just the fact that, you know, Uber is talking about how they're spending too much. Service now hit their budget for the year already. You know, I think Microsoft talked about pulling their cloud code licenses. Tokenpocalypse.

Tokenpocalypse. Yeah, okay, there we go. Yeah, exactly. A better word for it. But we've talked about this before, but in terms of how you think about search actually making token consumption more effective and efficient, can you share more perspective on that and to the extent you can, like, what are results that you're seeing on that front? Yeah, sure. So, like, retrieval can help solve the token apocalypse because, like, we should not be using gigantic models for every test. We should be using, like, and people are starting to, realize this. Like you should use a family of models of different sizes. The big model decides what to do and it dishes out of commands. So the small models. And those small models can be way more accurate and reliable if they're using retrieval. So retrieval helps small models act like big models. It's kind of in a cheap way. And so we do save our customers a huge amount of tokens because they can use smaller models and use retrieval. They could also, we have all, we care a lot about this. So we have like, we've put a lot of research effort into how to, you know, extract only the most

relevant information from documents so that these models can just, like, not have to consume too much tokens because, like, a lot of, you know, any sort of input tokens can dramatically increase spend. So we could, like, we could save, like, 20x on cost for customers compared to other providers by, like, being very efficient in, like, what information does the, from the web, does the agent actually see? But yeah, in general, smaller models using retrieval is much more efficient. And, like, Andre Carpathy had a tweet about, I keep mentioning Andre Carverthley on Twitter. I know, we all do. Yeah, he's great.

he had a tweet about this a couple, I think a couple years ago where it was like the trend is towards like smaller raw intelligence modules using tools. And that trend will, like that's an important trend. Because like you have,

you know, you have a limited, like the cost of the model is determined by the number of weights. That, you know, determines the cost of inference. And, and if those weights are,

if you're wasting those weights on like all sorts of information about the world, like the capital of France or, you know, this random blog that you read, like you're just wasting tokens. Sorry,

you're wasting a weights. Those weights should be focused only on, like, intelligent processing. And you could probably get to models that are like one billion, even less than a billion parameters that are extremely hyper-intelligent and completely unknowledgeable. It's like Einstein, like, who never saw the world. That's kind of like the way to think about it. And then it uses tools that are very cheap and efficient.

And that's a much more efficient world that will help solve this, like, this compute shortage that is affecting everybody. Yeah. Do you think this is more of like a hot take moment, but do you think that reality will start to take place second half of 2026? Because right now we're in this phase where everyone's playing with the biggest best model that just arrived. And to your point, probably overspending. So when do you think this reality kind of sets in? I mean, the reality is definitely starting.

I don't have the exact. It's like trends are everywhere. Yeah, yeah. When does it become noticeable? Yeah, I would say by end of 2026, it's very noticeable. Wow. Okay. Kind of a hot take, actually.

So you talked about doing just the research that you're doing. And I think EGSA, you sort of famously structured as more of a research lab, honestly, than what people are calling application layer companies, infrastructure companies, right? You know, this is sort of a research lab focused on search. And coming out of that, one of the things that we were most excited about, frankly, is just the exciting cutting-edge work that you guys are doing. One of those things was actually search as it pertains to RL, and there's a lot of efficiencies there, et cetera.

I wanted to just flag that because it was an interesting finding. I think you used Tinker, so shout out to thinking machines. But say more about what you guys are finding there and also just sort of what research directions you guys think about as being important. Yeah, at a high level, like a lot of the big ideas in training LLMs apply equally well to trading the search models. So, for example, like, we do pre-trading of embedding models. We do post-training of embedding models.

We do RL on, like, search tools, right? So, like, a lot of these things that are working in. LMs work in retrieval, too, which is kind of interesting. And you don't hear a lot of people talking about. So, yeah, in that RL blog post, we just basically try. Like, a lot of people RL on a search tool, but we haven't seen as much many studies, like, testing different search tools that you are on. Right.

And so we simply RLed on SER, so like, Google Rapping versus XA, and found that, you know, RLing on XA does way better. better. Like it both like uses fewer calls, so it's more efficient, and then it's like higher performance. And this makes sense because again, like Exa was designed for agents to use. And so like it just it just allows agents to make more complex queries. Like it's really like capture what they actually want as opposed to having like to compress what they want into like shorter phrases that are more for traditional search engines. So that that was a cool blockpost to explore and think it was really helpful for that. In general, our research is like the bitter lesson. So just like

scaling laws in lots of different directions, some of the ones I mentioned. Post-training, pre-training, RL. We've been pretty under the radar. I don't think people don't realize how much research we're doing. We don't publish all of it, obviously. We don't publish much of it.

But there is a lot to go in search, and I don't think people realize that. And I think we realize it because we've just been obsessed with it, one, crazy enough. Like, that's just what required. But then also, I think the biggest thing is actually because of our business model and who we serve.

We've just been pulled. pushed in all these crazy directions. Because we're not serving two billion consumers who are kind of all the same, like two billion humans. We're serving, you know,

now, you know, over 5,000 businesses that are pushing us in all these crazy directions. Like, just like every day, it's like, why can't this be higher quality over companies or people?

Like, why can't this be faster? Like, why can't, why can't the information extraction be even higher quality? And so, like, just we're being pushed in all these crazy ways. And that's why we're exploring

all these research directions. Like, research always follows need. So we have insane amounts of needs at Exa to do better. And so that's why we do all this for crazy research. Yeah. No, that makes sense. So I guess kind of tied to this is I wanted to ask you about how you think about benchmarks and hill climbing. And I'll say this sort of tongue and cheek, but we've noticed that, especially maybe among folks in your space, but also in other spaces, right, there seems to be like benchmark maxing or whatever you want to call it. And, you know, of course, to no surprise, everyone is always at the top of their own benchmarks.

Yeah, yeah. That's how you know. It's not something's wrong. That's how you know. Exactly. So, and obviously you have this relentless pursuit of ground truth. And also, you know, self-improvement, right? I think, like, you're the first to admit, hey, here are the areas we could be better on.

We're trying to improve on a, like, continuous basis. But how do you internally think about what benchmarks matter? What is ground truth for you guys in terms of like, hey, we're actually better on this front, but worse on this front? Yeah, yeah. No, 100% the evals have been benchmaxed in retrieval. There aren't too many evals in retrievals. So that's one problem. There's aren't that many standard third-party retrieval evals, and they've been like benchmarks.

And they're not really actually good representations of agentic search, like what agents actually need. And so it is a problem in the industry where, like, customers can't really know what is true, which is sad. It also demonstrates the need for a really good search engine to distinguish what is true. It's just another example. But, yeah, I mean, the ground truth for customers is their own AB tests. And so when we do a, like, right, like literally they are testing us versus other providers on their use case. And if they have enough data to do AP test, that's the best. If they have, often they make their own e-vow.

So sophisticated customers will make their own evils. super-sophisticated customers will run A-B tests, and then, like, customers who just want something might just, like, look at evals that are published online. But certainly for the sophisticated customers, when they test us, it becomes a lot clearer who's, like, on the top. But, yeah, we want to improve this ecosystem. We want to, like, we want to be, like, the research lab

that, like, helps improve the ecosystem and, like, publishes things. And even if we're not at the top, we want to show. And so you'll see more coming out there. So you predicted that agenic search will be a bigger business than Google search by the 2030s. Say more about that. What trends are you seeing that we do believe this?

Yeah, I mean, just you could get this from basically, like, estimating the number of searches. So the number of LLM calls and the percentage of those LM calls that require search, then the cost of the search, and then you just like play it out. And the trend has actually been pretty clear. And we've have, you know, pitch decks from like years ago where we kind of predict where things we go, you know,

maybe we're off by a quarter here or there, but it's like pretty... you know, it follows a trend. And so, yeah, if you follow that trend, even conservatively, you get to a massive TAM for agentic search in the 20s. I mean, even before, like, late 2020s and then early 2030s. Basically, like, the number, it's hard to express, like, how many searches will come from agents, right?

Like, humans on average make, you know, a couple searches a day. But agents, when everyone has a personal assistant and every single software tool you use is going to be checking its work with retrieval. Totally. Like, the number of searches is going to, to be, we say thousands because that is like understandable and grokbold people, but really it's going to be millions.

Yeah. It's just going to be like the world to be filled with search in a way that the world is filled with electricity. It's like it's a fundamental infrastructure that powers everything. Like information, I think last stuff, search is like perfect information. It's a world be filled with like the highest quality information. And so yeah, I mean, there is a lot of, when the world is filled with something, it's usually a large tam.

And yeah, if you just play out the numbers, we think it will be bigger than Google ads in 2030. Not that ads won't be a huge part of the world, too. Like, ads are important for commerce, and that might be a percentage of the agenic search economy, too. But, yeah, it's just the numbers here are insane. And it really, it comes down to a belief, like, do you think LLMs will eat the world? We'll eat all software.

And, like, we have always believed that. Yeah. It seems like very true. It seems even more true every month. So, yeah. What do you think, and there's some debate on this, on the LLM side, on the training side, too, but what do you think is the bottleneck?

today versus, let's say, three years from now. I feel like five years is too far to predict. Is it, I think you've said in the past, it's no longer intelligence in terms of bottlenecking search. Like, is it data, accessible data? How do you think about how that evolves? Yeah, I mean, well, initially the bottleneck is going to be actually the infrastructure,

which is kind of interesting. No one realizes, but like if you actually get, for example, 10x, 100x, X, 1,000 X, more surges than Google, the infrastructure to handle that is insanely large. It just hasn't been built yet. So we're really excited to explore all sorts of cool new vector databases that have like super high throughput, for example, things like that. Yeah.

So that's like an interesting bottle. Like in the same way there's compute bottlenecks, there'll be like infrastructure bottlenecks. I mean, it'll be solved at some point. Yes. And then like other bottlenecks are like data bottlenecks. Like so agents are increasingly going to want to ask questions about the world and that data might not be on the web. It might not even be recorded anywhere. And so like, you know, there will be this trend towards how can we like accumulate all the world's data.

Like literally unearthed the world's data. I'm really excited by this because the world is filled with information and it's not all recorded. You know, the history of humanity, the history of humanity is the world's been, you know, when we're hunters and gatherers, there's also information. Then they started writing things down. And like the amount of information the world has been like skyrocketing. There's still so much information that's not recorded. You know, you go from all the way from the first clay tablet.

Oh, really the first like paintings on on caves would be arguably the first time things were written down. And then like, and then, you know, clay tablets and, you know, now I've in newspapers. and obviously now we have like the digital age, but there's so much information in your head, like satellite images, like that are not just like in the world's soup of information that we could search over. Yeah.

And like to fully understand the world, fully understand how props, crop yields are going to, you know, affect, you know, some company, like, or like, what are people thinking about the world or how do we unite the world? Like that requires like understanding the world at a deeper level.

So I think the bottleneck will be data in a lot of ways. And then once you have, the problem is once you have all this data, now the bottleneck will be retrieval. Right. Right, right, right. Imagine, it's just a crazy idea. Like, imagine, you know, we're, as we expand as a species,

like, we're thinking, like, very far in the future. Like, think about, like, or in the solar system, like, there's so many things going on in the world. There's so much data being accumulated. The retrieval over that is very expensive. And so, like, they're actually really fundamental, they're interesting fundamental questions here. Like, right now the web is, like,

what's called a trillion pages that matter? What happens if the web were a thousand times bigger, like a quadrillion pages, meaning, like, everyone started uploading data. Well, then, like, any search algorithm that works over a trillion pages no longer, like it might work over a quadrillion pages, but it might be a thousand times more expensive.

And that's not practical, because if anything, we want search to get cheaper, not more expensive. So what kind of search algorithms to work over a quadrillion pages? These are fascinating questions that, like, I don't know if I've ever heard anyone else talk about. Well, actually, I was going to say, at least the,

thinking about when we're living in, you know, on Mars or whatnot, there's probably one other founder that has thought about that extensively, Elon. And, you know, I listened to your, when you went on the latent space podcast last year, one of my favorite podcasts. And you talked about actually working at SpaceX. How was that experience? And like, are there elements of Elon's leadership style that you've taken as, you know, CEO of your own company? Yeah. Well, first of all, the internship was magical. So, like, for example, I saw the first landing on the barge. Wow. And like, like, just outside mission control. Like, I'm getting, uh, tingles just thinking about it. But yeah, like, just, just, just, seeing that, like, people coming together to do something magical. Like, that was very inspiring and made a big impact on me. And, yeah, in terms of my leadership style, yeah, I like to think that I've incorporated some of what I think are the best aspects of Elon. So, you know, for example, like, he's very detail-oriented.

Like, he gets into details of everything. And, like, for better or worse, like, I do that too at XA, like, everything from, like, the algorithms, like, the vector DB algorithm to, like, the office space and making sure, like, every part of the office is, like, really just, like, inspires and excites. Yeah. And like, like,

shows our passion. And that goes across everything, from our marketing to the engineering to, uh, to, to go to market. And, um,

so that, that's exciting. Obviously it doesn't scale. So one thing I've been learning as we scale the company is like, what details can I choose to go into? And so it's, I like the ego metaphor of like, uh,

you know, I'm an eagle flying above the company. And then when I see some detail that I think it should be important to, to, to fix, like I go dive down and go into it and then come back up. Um, so I think Elon has,

some of those properties. Obviously, Elon is like all in every day for, what is it, like, decades. Yeah. I've only been doing for five years, but I intend to do it for decades. One last thing that he does really well, which is like he's very good at memetic, like names and like inspiring through like memetic things. Like, you know, like you realize that SpaceX's mission is not like improve rockets to get to or, yeah, like, like, oh, make rocket travel really fast, like good. It's like, it's like make humanity interplanetary. Like that's a really good memetic thing. So I think a lot the names of like projects and like when I when I you know I do a team stand up every Monday in front of the whole company now we have like this like double floor office where people now are

surrounding on the top floor it's really cool it's like a stadium and I give like a speech and I and I like to like simplify what we're doing into like what is the core what is the memetic core and like come up with a cool name and that inspires by the way a name is really important because it's when you have a company of a certain size people are constantly communicating in ways you're not you're not part of those conversations and like the name like really like uh grounds the mission of that project. Yeah, absolutely. This is really important.

Like, I could think for like a day of just about a name. So I just totally divergent, but on the topic of names, one, I love goat. Talk a little bit about the symbolism of that. And then, I mean, besides, you know, greatest of all time, obviously. But, and then also EGSA. You shared before what EGSA means, but talk more about why you name the company EGSA. Yeah, sure. Okay, so the goat thing is basically.

there was a coffee shop that opened up next door, and so many people, so like the mayor of SF kept posting about it, and so a lot of people go, it's really cool coffee shop, hedge coffee. And so many people were walking by, we're like, we got to, like, have them stop and look at what XA is.

So we have some posters about XA on the entrance to our office, but we were like, how do we get them to stop? And so I was like, just put a goat. I don't know why I thought this, but just put a goat there. And I actually, some part of me wanted a real goat. I don't know how we'd maintain the goat.

I think the ROI would still be valuable. But instead, we bought, like, a nice, like, fake goat. Yeah, I love that goat. And we put like the swag on him. So it's like it's a beautiful. Yeah. And and people stop by. So like we're actually like tons of people stop by. Like wow. Even on Saturday and Sunday, which is like when the coffee shop is very popular, like people come by and I go and I say hi to them. I talk to them. It's actually very beautiful. We're actually like an important stop for for kids on their walk to school. Like you know when you're a kid you had like to stop. So you like move the ice cream store. Like the goat is a stop and they're always taking pictures. Very cute. Starting the recruiting funnel early.

but it's interesting how much those things matter. Yeah, absolutely. We've hired people because of the goat. Anyway, XA, I think is a great name. I love XA. One value of XA is like, it's a great prefix. So like XA anything, XA data, XA this, XA that. But XA means 10 to 18th,

which is in contrast to Google, which is 10 to the 100th. And the idea there was like, you know, we're kind of overwhelmed with information. Even though you could technically find lots of things on the web on Google, It doesn't mean it's like organized and you get the highest quality knowledge. So like in here it's like 10 to 18th is like you're extracting the most important information.

It's still very large amount of information, but it's not like overwhelming. That was one of the ideas. Great name. We love the name as well. And the goats. My kids love the goats. So I wanted to, this topic feels a little bit over talked about, but it's been in the news more recently. So I'm going to bring it up, which is I don't know if you saw Will Minetus wrote a

an article recently called Grind Slop. And it's almost this backlash of, and in fact, I pulled a quote from it. He said something about, we're witnessing a phenomenon that masquerades is disciplined, but represents perhaps the most extravagant squandering of economic surplus in our civilization's history. So extreme words, right?

And he talked about sort of all of these articles that are glorifying the grind over what they're doing, you know, all of these things that, you know, maybe are more relevant to work-life balance, et cetera. So I guess I'll start with saying one of the things that I was actually most impressed and excited by when I visited Exa at, God, I think it was like 8 p.m. was that there were a ton of people in the office. And it wasn't just that they were there for FaceTime,

because obviously that's not what you and Jeff care about. But they were excited about what they were working on. And I think really importantly, they were excited by who they were doing it with. And so just say more about, like, how you think about that culture of going all in and going after what you guys are, you know, sort of the mission that you just talked about. And, like, what do you think the important things to balance are in culture? Yeah, yeah. I mean, this is critical to me.

And it should be critical to any company that wants to do great things. Like, people need to be super excited about what they're doing. And I have a ton of fun, right? So, like, you might have also heard at APM, like, people are just laughing. Yeah. Oh, yeah. We laugh a lot at Exa. And it's almost too loud sometimes

and it's like, distracting. So we have to have like headphones for everybody. But, but yeah, like people are, it's very important people are having fun and that's necessary for doing the best work of your life. And I would say like having fun just like, just good vibes. But also it's very important that people work on the projects

that are most exciting to them at any point. And so like people, especially now with like, you know, all these AI tools, like anyone, especially on the engineering side or even on the market side or off side, everyone can work on whatever they want. So for example, like we had someone who built the vector database. who was like, hey, I want to go train models. I was like, okay, just go do it.

And like, he didn't really have experience in training models, but I was like, you're really smart. You'll learn it in like two weeks. And then you'll just do amazing work. And he's done amazing work. So, right? So I make sure everyone's working on exactly what they want to work on. And luckily in search, it's like, at least in our space,

like in any direction we improve things, it will be good for the business. So it's like, okay. And it actually happens to turn out somehow that like what people want to work on and what we need, like perfectly aligns. It's like a magical thing. I don't know how that's possible. But yeah, so everyone's working on exactly what they want to work on.

also working on multiple projects because like AI tools will enable you be like productive in parallel. And so it's just and then what we're doing is just super cool. Like, you know, there aren't, like it's hard to find maybe these days like really exciting like humongous projects and missions that other people aren't doing. And this one happens to be like organizing world's information, making perfect search. That's very exciting and important for the world, but also like really hard problem. And engineers are really excited about that. And on the go to market side, it's like, we're then selling this search to the whole world, and we're selling to some of the coolest companies,

and it's very exciting to them. They get to talk to all these hot companies and give them search that really helps their products. It's just an exciting space, but it's very important to me. So I hope we maintain that forever. Yeah. No, absolutely. Well, maybe just to end on one last question,

and I'll say, first of all, Exa is hiring. And you told me once that you actually still interview every single person at it. Exa or who comes to, you know, interview for Exa. And so I'm curious what, and you've, you've attracted some of the most incredible talent, junior, senior, everything in between, very high slope, very experienced. And I'm curious what it is that you look for when you do that last final interview and, you know, how you've been able to get these incredible people to come to Exa. Yeah. I mean, people might not like this word, but I look for passion. It's like someone

just like a fire in their eye. Like, do they really care about what we're doing? Or some aspect of doing. I really want to scale this thing or I really want to sell this to everyone. Because why is that important? It was important five years ago, but I think it's especially important now because like the world goes to who is most passionate and who's most agentic because like you can now do anything. Like you could be whatever you want. Like, you know, like with agentic tools, you could literally like do anything. And what matters most is like how much do you care about the end result? And then like your judgment and everything. And so like I I look for the fire on the eye because if you have that fire, you could literally do anything.

It's actually crazy how meritocratic the world is becoming because of these agentic tools filling in the gaps. Yeah, absolutely. Yeah. Well, thank you so much, Will. This was great. Such a fun conversation. Appreciate you and very excited to be partners. Awesome.

Thanks for listening to this episode of the A16Z podcast. If you like this episode, be sure to like, comment, subscribe, leave us a rating or review and share it with your friends and family. For more episodes, go to YouTube, Apple Podcast. podcast and Spotify. Follow us on X, A16Z, and subscribe to our substack at A16Z.com. Thanks again for listening, and I'll see you in the next episode. As a reminder, the content here is for informational purposes only. Should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund.

Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see a16Z.com forward slash disclosures.