Parallel Web Systems closing a $100 million Series B funding round last month, valuing that AI startup at $2 billion. You were at Twitter for 11 years. You were CTO, then you became CEO.
I was also an engineer before that.
We're building a collection of APIs that agents use to search and access the web. Parallel is an almost three-year-old company now. I was the lunatic walking around three years ago telling people agents will use the web a thousand times more than humans ever have. Most people didn't believe me at the time. I was wrong.
A thousand times more. My deck literally contained a section which is like, are we too early? Turns out we were not too early. We're building a parallel web for agents. We can transform the web even if you don't publish your website to be agent friendly. We'll do that for you.
In such a competitive career in every stage, how do you not break from the pressure?
You don't feel pressure? No.
Parag, welcome to Sourcery.
Well, we're in Index's office in San Francisco. It is gorgeous here. They're one of your investors. I'm not sure if you've ever gotten asked this question before, but what have you gotten done this week?
I'll tell you what I got done today.
We just shipped a really exciting integration with Browserbase.
We now power search in Browserbase. And if you think about every AI agent that people use, what are they doing? They're looking at your internal information. They're looking at web information. They're searching a bunch, a bunch more than us. Then they're going into a browser, logging in and doing something, buying something, writing something.
So that's what we got done today.
Congratulations. We were talking a lot before this and we were having a really great discussion. Every interview I've had recently is some sort of explosive growth of agents and it's going to be really fun to unpack all of that because from your point you were talking about read versus write, what you're doing versus Browserbase. So I guess to start, let's talk about that. Parallel. What is Parallel? How is it different? How has it evolved? And how are you learning as things are just changing so fast?
Yeah. Parallel is a three, almost three-year-old company now.
I was the lunatic walking around three years ago telling people agents will use the web a thousand times more than humans ever have.
And there were no agents. Agents didn't work. Most people didn't believe me at the time. Fortunately, some did, including Index. And at the time it felt like the science fiction hypothesis that there will be a lot of agents at work and in our personal lives and that they'll be doing like web searches for some reason. I believed it because I had obsessed with it for a long time. But today if you look at like what's happened in the last three years, like I was wrong, but on the other side...
A thousand times is too low. I think at the time I thought a thousand times was like the crazy number to make up.
To talk about, right?
And at the time I was like, my deck literally contained a section which is like, are we too early? Turns out we were not too early. Turns out today if you force me, a thousand times is too low a number. If you just look at just some high-level math, right, a person, an average person does like order of ten or less Google searches a day.
I was checking how many I do. Several thousand now. My agents, the ones I track.
Several thousand per day. All the agents that I use today. So I am now sitting at roughly a thousand times, a little bit less than a thousand personally. Most of the world is not... most of the world will in due time, might take us a few years, but a thousand is a low number.
But is that efficient? Do you think that'll crawl back after a certain point because they'll just get smarter and know where to go?
Yes. But it is efficient because the other part of the story is that you got to, you can't build old-school search and do a thousand times of this. You think about like, okay, all the infrastructure that a Google might have built to serve whatever order of ten billion searches a day. Like the best guess I have is ten billion a day is what Google will do. They have a lot of infrastructure to pull that off.
I don't think we get to build a thousand times that infrastructure to do a thousand times the searches, right? So part of what we do is build extremely efficient and optimized search, which again benefits from models getting better.
For us to be able to distill intelligence ranking into smaller and smaller models. For us to be able to use GPUs as the underpinning of search indexes. Historically search was done on CPUs, memory CPU bound. Now you get to architect and rearchitect search for a price-performance band that the world never needed until today.
One of the things we were talking about before, I recently interviewed G from Vercel and this is, you know, maybe this is in parallel, but AI agents for coding have reached 60% of traffic on their indices, what they see. You had a hot take.
Generally on the web, the statistic being shared these days in the last few months is that agent traffic is exceeding human traffic on the web. That's roughly the characterization. Now there are two things to remember. An agent for roughly doing the same work that a human does will likely do a lot more searches because it can, and a human is time constrained. An agent isn't really. It's more, and right now they're not even that dollar constrained because people are willing to spend on frontier. A human read on the web or a human search on the web to create equivalent value, it takes way more agent traffic. A lot of agent traffic is also people crawling for pre-training. And people, like when you look at, like if you're a website owner and you're looking at automated activity versus human activity, automated activity is ripping.
But my best estimation right now of where we are is agents are using the web 10 to 20% of humans. Which is to say if my thousand times number is right, we have from here at least 5,000 times or 10,000 times growth left. Just another way of saying we are very early with agents. I think in like, in the Bay Area, in Silicon Valley, we're all living in a little bit of a bubble.
I think most people's agents aren't doing the number of searches my agents are doing. But they will.
So Muse just came out with Jolly and I really like Jolly the character for Muse. We see dots coming out for OpenAI. We've been seeing a lot of these agents pop up for individuals, for consumers.
That growth is faster than any growth ever seen.
Like even chat, I think maybe possibly we'll see. But how do you expect that to grow and I guess where does retention stay?
Two things. One, I have to make a note. We should call out Instinct too.
I am. The reason is, listen, I'm on the Rebel Alliance here.
Which is there are all these big people.
And then there are us startup founders who are running at solving really hard problems and pushing the frontier of what's possible and innovating. So we should talk about Instinct.
They are growing to an amazing product. So yes, the category is amazing. It's interesting. I think it's growing really fast. I don't have any insider information around retention. I am, as you can imagine, it's a professional hazard that I use every one of these agents.
And I try to deeply understand what they can and cannot do and what problems we can solve for them.
We'll get into this category and how it changes our product strategy later. But I think my expectation is that with these agents, retention will be good. And it'll be good because for the first time the product onboarding feels natural, like not someone designing like a bunch of UX click targets and flows to get you to understand the product. The product is natural. All of these products are somewhat natural. And the question is can they deliver on the promise created by the UX. So the UX is you can touch this thing and you can do anything. That's the UX. It's not complicated.
Now retention is entirely a function of can it actually deliver on this very simple, high-level promise often enough to become a part of your life. Now it's going to happen, like a lot of other products struggle with the onboarding UX barrier to entry, but this new generation of products, that's just not a thing. And so I have a feeling that retention will get solved here.
Is that where you think OpenClaw maybe hit a wall? Like because that was really the first?
No, I think OpenClaw was right in all of the ways. Yeah, I think once someone got it, it's just not for the masses, right? It's just like most people don't have the ability to set it up and then to keep up with change. I think the other factor we discount, it's not just a one-time cough to set something up. Things are changing so fast that you have to adapt your instantiation of it alongside that change, and you have to adapt your use of it alongside new capabilities. So part of it might have been that the setup time because of the flexibility and you controlling it was far too much. Part of it is also that it was perhaps a little too early, right? Because in the beginning there was like one model which was really good, like the Opus model, and then as soon as you move to other models it wasn't as great. In the very beginning when people started using OpenClaw late last year, and so you had to adopt alongside it. But today, as evidenced by Instinct and Muse, you kind of don't have to know or think about it and the product will change underneath. But as the customer of it, you don't need to know what changed in the last month, you're like oblivious to the capability improving underneath.
I can't discount that this was really like just invented. So how did you, you mentioned this at the start of the interview, but like you've been obsessed with this for a while.
So what does that mean and how?
So I think I got obsessed with it now almost three and a half years ago. I was the, I don't know if people remember, there was like this moment of this Baby AGI and AutoGPT. Speaking of oftentimes developers in the open source world being ahead of everything else, people were building with like, I don't know, what is the model of the day, GPT-3.5 or whatever, like old models that weren't, like relative to today they were like really bad models relative to today. And people were trying all these things and I was trying all of these things to try and make things work, and they didn't work. But if you believed, you spent any time with the labs back then, you internalized how models will scale and how models will improve. Like in some sense, today it feels like everything changes every month. In some sense, I don't know if it's every month, it might be every week. But if you look at it in a different way, almost nothing that is happening is diverging from what you could have predicted in a straight line.
Models will keep getting bigger and better. They will run for longer. Holding capability constant, models will get way, way smaller and cheaper, and people will be able to productize around them to disrupt every category of work that can be done on a computer. These were very simple things that we believed and said three years ago, and we have only seen evidence week on week and month on month supporting these very few simple things.
And almost everything follows. So now you can say, yeah, I didn't predict that like the big category would be coding agents first, right? I didn't predict the CLI form factor, right? Those were all bits of information and news to me along the way.
But the macro thesis of agents are going to get more capable, you're going to do a bunch of inference, and for all the inference you do, you're going to have to do web search along it. I've now been on the record everywhere saying that like the one invariant I believe is we will use a lot of GPUs for model inference for agents, whether they be at work or consumer agents, and for every ten GPUs the world spends on model inference, we'll spend one to two on web search.
That's the business I'm in.
So are you aggregating GPUs? You've raised two hundred thirty million reported and announced. I don't know. But are you aggregating GPUs?
Yeah, we are. Not, I won't use the word aggregating.
We are buying, we're renting GPUs to build the scale of search traffic that we are seeing and expecting. And as I said, we grow alongside inference. So inference is growing, so we have to do a proportionate amount. I don't know if it's like already a, I don't think it's a core competency of ours, like we have been in multiple mad GPU scrambles.
In the last three months to be like, oh, we need GPUs. We don't have GPUs. Can someone figure it out? And like you're making a bunch of calls and like...
Everyone. Everyone. Like every hyperscaler, every neocloud.
What's like the typical response back? Because this comes up in pretty much every conversation now. It's like, how do you get, I guess, the allocation of GPUs?
No, I think it's like, I think at least for us, I think it's true for many people, it's like, okay, which GPU is it?
Physically, where is it? Sometimes we need GPUs within five milliseconds of where a large web index is sitting.
Other GPUs, we're sometimes more flexible with multiple GPU workloads. It's not just like, okay, do you have GPUs for me? It's like, do you have the right GPU in the right scale in the right physical location for me available at the right time? And that is what makes the problem much harder. I think so far we have managed to solve every problem we have encountered. We've not had to turn away a customer because I didn't have a GPU, right? But we've had a bunch of GPU acquisition fire drills.
I mean, that's a really interesting line to go down. It's like, what is the sourcing process like for GPUs? Like, what are the main questions that you're asking? What do you need?
Yeah. So I think you mostly, so I think different workloads are different, but for bulk of our workload, it's like how many GPUs, which types are you open to, when do you need them, and where must they be? Like, there's not much more to it. I think for certain large training clusters, there's a little bit more to it.
And then for us, like, are you willing to commit enough duration to that physical thing? Yeah, but that's, it's not complicated. The reality is that I wouldn't say the market is entirely efficient right now. Like a lot of it is on iMessage, right?
Yeah.
So, so yeah, but it works. The market works.
Well, that's crazy. So you're doing deals on iMessage for GPUs?
Not exactly. They're filling out availability on iMessage, and then like the finance person then calls the finance person, and the paperwork gets done.
Oh my god.
This episode is brought to you by Brex, my favorite. You become what you spend on. And I refuse to spend my time on work that shouldn't exist. Expense reports, receipt chasing, and manual closes. The companies building what's next from Vercel, OpenAI, Anthropic, Granola, and Deepgram have all made the same call. They all run on Brex. Brex is the intelligent finance platform that combines cards, expenses, and banking into a single stack with Agentic Finance, built-in AI agents that handle expenses automatically, enforce policy before spend happens, and close your books in minutes. That's why Sourcery runs on Brex so I can spend time on building and not busy work. It's time to get Brexed. Learn more at brex.com/sourcery. That's brex.com/sourcery.
From handshake to signature faster. Harvey, the AI platform built for legal work, is trusted by 80% of the Am Law 100 and over 20% of the Fortune 500. Learn more today at harvey.ai. That's harvey.ai.
AI needs more than chips. It needs power, land, and infrastructure. Zone develops next-generation data center campuses, partnering with AI companies, site developers, and technology leaders to bring compute online, faster, and at scale. Zone is building the foundation of the AI frontier. Visit zonefrontier.com to learn more. That's zonefrontier.com to learn more.
I want to go back to the topic of scale because we were chatting about historical scale and what that used to look like and what you would expect. So you were at Twitter for eleven years. You were CTO, then you became CEO.
I was also an engineer before that.
You were also an engineer before that. You've done the whole thing. You also have a PhD. I've done all this research. But I want to talk about that. So like, can you set expectations now? Like what would scale meaningfully look like today if like back then it was a triple, triple, double, double? Now it's like, what do you have to achieve?
I don't know how somebody else could do it for our business. Turns out the answer is relatively easy.
As I've said, like we just build alongside inference. So if you think about inference growing year on year, to us that is the anchor. Like our right goal is to grow faster than inference for agents, for knowledge work, or consumer agents grow. So if we grow alongside or faster, I'm feeling pretty good, right? And that's to us the benchmark. Now, what is the right inference benchmark? That's a question I don't know long-term the answer to, but I think in the short term it's clear that GPUs that are coming online this year or next year are going to do inference, and so then they're going to do search alongside it. So the short term there is very clear, but our benchmark is just growth of inference.
I think like the easiest approximation is ten x. You have to think about like a ten x go as the very simple round number. If you think about today, like if you asked me last year, I would have probably said a lower number.
But the pattern here is even when I have thought I was being provocative in making up a high number, I have somehow managed to underestimate. So now I can, I feel like I can go to ten x. I feel a little bit uneasy that like it feels crazy to grow ten x for too many years. You know, how long can that happen? Maybe I'm wrong. Again,
I'm curious because you get to see so many different market insights from internally, externally, all around. What are some surprises that you've come across and what are the signals you're looking for?
So I think it might be useful to tell people about the kinds of customers we end up working with.
