# AI Agents Could Drive 10,000x More Web Activity | Former Twitter CEO

Markdown copy of https://realalpha.app/v/iRI6gTe1wGM/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

Parag Agrawal says agents could drive 5,000–10,000x more web usage, forcing search infrastructure and online business models to change.

- Show: Sourcery
- Hosts: [Molly O'Shea](https://realalpha.app/people/molly-oshea/)
- Guest: [Parag Agrawal](https://realalpha.app/builders/parag-agrawal/) is founder and CEO of Parallel Web Systems.
- Published: Oct 24, 2026
- Length: 59 min

## The best minutes

- [14:34](https://realalpha.app/v/iRI6gTe1wGM/?t=874) [GPU sourcing depends on location, timing and fit](https://realalpha.app/s/gpu-sourcing-depends-on-location-timing-and-fit/) (4 min)
- [47:25](https://realalpha.app/v/iRI6gTe1wGM/?t=2845) [Proactive agents will make search prioritize web changes](https://realalpha.app/s/proactive-agents-will-make-search-prioritize-web-changes/) (3 min)

## Also worth watching

- [11:42](https://realalpha.app/v/iRI6gTe1wGM/?t=702) [Web search may use one or two GPUs per ten inference GPUs](https://realalpha.app/s/web-search-may-use-one-or-two-gpus-per-ten-inference-gpus/) (3 min)

Skip the sponsor reads and the final eight minutes on ambition, partner choice and pressure.

## Summary

Parag Agrawal founded Parallel Web Systems, which builds APIs for AI agents to search and access the web. He spent 11 years at Twitter as an engineer, CTO and CEO; the YouTube description says Parallel has raised $230 million and is valued at $2 billion.

Agrawal argues that AI agents could multiply web usage by 5,000–10,000x from current levels, requiring search infrastructure built for higher volumes, lower latency and lower cost. He says agent use will also disrupt web business models based on human attention, and Parallel is developing a way to pay content owners when agents read their material. He expects the web to shift from people pulling information to agents proactively monitoring changes, while warning that security and access incentives must be addressed across the web. He also describes Parallel’s growth, its enterprise customers and the people and personal standards that shape his ambition.

### Minute by minute

- [11:42](https://realalpha.app/v/iRI6gTe1wGM/?t=702) **Inference will pull search demand.** Agrawal says he believed models would keep getting more capable and run for longer, while becoming smaller and cheaper at the same capability level. He expects agents to need substantial web search alongside model inference. His estimate is that for every 10 GPUs used for model inference, the world will use one to two for web search.
- [14:34](https://realalpha.app/v/iRI6gTe1wGM/?t=874) **GPU supply is a placement problem.** Agrawal says securing GPUs means finding the right type, quantity, physical location and delivery timing, not just a supplier with available chips. Some of Parallel’s GPUs need to be within five milliseconds of a large web index. He says the team has faced repeated acquisition scrambles, sometimes coordinating availability on iMessage, but has not turned away a customer for lack of GPUs.
- [19:30](https://realalpha.app/v/iRI6gTe1wGM/?t=1170) **Search expectations keep tightening.** Agrawal says Parallel benchmarks its growth against agent inference and sees 10x as a simple near-term growth measure. He says early customers such as Harvey and Notion pushed the company to meet high quality standards. Voice agents now require web search in 150–200 milliseconds, he says, compared with roughly 600–700 milliseconds for human search, so providers must improve quality, cost and latency together.
- [33:32](https://realalpha.app/v/iRI6gTe1wGM/?t=2012) **Cyber defense will stay adversarial.** Agrawal says agents will help attackers as well as defenders, and predicts consequential cyber incidents every few weeks will drive urgency. Drawing on Twitter’s crypto-spam and state-sponsored campaigns, he says abuse is difficult to eliminate because tactics keep changing. He sees a deeper risk in losing the ability to distinguish people from agents, and says defenses against prompt injection must operate at the web, model and app layers.
- [47:11](https://realalpha.app/v/iRI6gTe1wGM/?t=2831) **The web could shift from pull to push.** Agrawal expects agents to move the web from people searching for information to systems proactively monitoring changes and alerting users when action may be needed. He says the challenge is allocating compute across a trillion web pages, much of whose changing content is junk. As agent-generated material increases, he argues, search engines will need to prioritize useful signals.
- [51:20](https://realalpha.app/v/iRI6gTe1wGM/?t=3080) **Ambition grows through close peers.** Agrawal says his wife, Vinita, has pushed his ambition higher. He describes her as a practicing physician, an investor at Andreessen Horowitz and someone who has worked in startups, healthcare and biotech. He also credits competitive peer groups and high expectations with shaping his own standards.
- [55:12](https://realalpha.app/v/iRI6gTe1wGM/?t=3312) **Insecurity can drive better work.** Agrawal says he does not think pressure is necessary to perform and recalls sleeping well during the Twitter acquisition. He seeks jobs he would not have hired himself for, he says, because insecurity can make him humble, curious, hardworking and more willing to seek help. He sees internal expectations as stronger drivers than external ones.

## Transcript

Sourcery, hosted by Molly O'Shea, with Parag Agrawal.

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there. Where the podcast’s audio and the video part ways, some times are a minute or so off.

[0:03](https://realalpha.app/v/iRI6gTe1wGM/?t=3) Most people didn't believe me at the time. I was wrong. Why? Thousand access to low. 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? What pressure? You don't feel pressure? No. Parag, welcome to Sorcery. Thank you for having me. Well, we're in Index's office in San Francisco. It is gorgeous here. They're one of your investors.

[0:46](https://realalpha.app/v/iRI6gTe1wGM/?t=46) 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 are 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 on today. Congratulations.

[1:25](https://realalpha.app/v/iRI6gTe1wGM/?t=85) 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 browser-based. 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 thousand x more than humans ever have. And there were no agents. Agents didn't work.

[2:12](https://realalpha.app/v/iRI6gTe1wGM/?t=132) 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 lot 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 why thousand x is too low i think at the time i thought thousand x was like the crazy number to make up okay 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 thousand X's to lower number.

[3:03](https://realalpha.app/v/iRI6gTe1wGM/?t=183) If you just look at just some high level math right? A person, an average person does like order of 10 or less Google searches a day. I was checking how many I do. Several thousand now. My agents, the ones I track. Wow. Several thousand per day. All the agents that I use today. So I am now sitting at roughly 1000x, a little bit less than 1000, personally. Most of the world is not. Most of the world will in due time. It might take us a few years, but 1,000 is a low number. But is that efficient? Do you think that'll crawl back at a certain point because they'll just get smarter and know where to go?

[3:48](https://realalpha.app/v/iRI6gTe1wGM/?t=228) Yes, but it is efficient because the other part of the story is that you can't build old school search and do 1,000x of this. you think about like okay all the infrastructure that a Google might have built to serve whatever order of 10 billion searches a day like the best gift I have is 10 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 thousand x that infrastructure to do thousand x the searches right so part of what we do is built 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.

[4:46](https://realalpha.app/v/iRI6gTe1wGM/?t=286) Historically, search was done on CPUs, memory and CPU bound. uh now you get to architect and re-architect search for a price performance band that the world never needed until today of one of the things we were talking about before i recently interviewed g from versell 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 human is time constraint. 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 equal length 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

[6:23](https://realalpha.app/v/iRI6gTe1wGM/?t=383) estimation right now of where we are is agents are using the web 10 to 20% of humans, which is to say if my 1,000x number is right, we have from here at least 5,000x or 10,000x growth left. Just another way of saying, we are very early with agents. I think in the Bay Area and 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.

[7:12](https://realalpha.app/v/iRI6gTe1wGM/?t=432) That growth is faster than any growth ever seen. Like even ChatGPT, I think maybe, possibly. We'll see. but how do you expect that to grow and I guess where does the retention stay two things one I have to make a note We should call out Instinct2. Yes, I almost forgot. 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.

[8:07](https://realalpha.app/v/iRI6gTe1wGM/?t=487) 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. Yeah. 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 the my expectation is that 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 text this thing and you can do anything yeah that's the ux it's not complicated yeah now 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.

[9:32](https://realalpha.app/v/iRI6gTe1wGM/?t=572) 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 open claw maybe hit a wall like because that was really the first no i think open claw was right in all of the ways um yeah i think once someone got 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 the change i think the other factor we discount it's not just a one-time cause 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 not 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 open claw late last year and so you had to adopt alongside it but today as evidenced by instinct

[11:11](https://realalpha.app/v/iRI6gTe1wGM/?t=671) and Muse yeah 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 auto GPT 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, GP35 or whatever, like old models that weren't, like relative to today, they were like really bad models, relative to today.

[12:28](https://realalpha.app/v/iRI6gTe1wGM/?t=748) 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 a different way almost nothing that is happening is diverging from what you could have predicted in a straight line three years ago yeah models will keep getting bigger and better. They will run for longer. Holding capability constant, models will get way 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 the big category would be coding agents first. I didn't predict the CLI form factor. Those were all bits of information in news to me along the way. But the macro thesis of agents are going to get more capable.

[14:07](https://realalpha.app/v/iRI6gTe1wGM/?t=847) 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 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

[14:27](https://realalpha.app/v/iRI6gTe1wGM/?t=867) **Parag Agrawal:** agents and for every 10 gpus the world spends on model inference will spend one to two on web search

[14:47](https://realalpha.app/v/iRI6gTe1wGM/?t=887) that's the business i'm in so are you aggregating gpus you've raised 230 million reported and announced i don't know but is are you aggregating gpus yeah we are not i won't use the word aggregating we are buying we're renting gpus uh to build the scale of search traffic that we are seeing and expecting uh 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 yeah in the last three months to be like oh shit we need gpus we don't have gpus can someone figure it out and like you're making a bunch of calls and like who are you calling sorry who are you calling 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?

[16:08](https://realalpha.app/v/iRI6gTe1wGM/?t=968) Physically, where is it? Sometimes we need GPUs within five milliseconds of where a large web index is sitting. Other GPUs are 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 tilts 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 maybe like there's too much more to it.

[17:23](https://realalpha.app/v/iRI6gTe1wGM/?t=1043) 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.

[17:36](https://realalpha.app/v/iRI6gTe1wGM/?t=1056) **Parag Agrawal:** The reality is that I wouldn't say the market is entirely efficient right now. Like a lot of it is on iMessage.

[17:44](https://realalpha.app/v/iRI6gTe1wGM/?t=1064) Really? 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 figuring 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 versell open ai anthropic granola and deep gram 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 Sorcery runs on Brex so I can spend time on building and not busy work. It's time to get Brex AF. Learn more at brex.com slash sorcery. That's b-r-e-x dot com slash s-o-u-r-c-e-r-y. Bye. From handshake to signature faster, Harvey, the AI platform built for legal work, is trusted by 80% of the AMLA 100 and and over 20% of the Fortune 500.

[19:12](https://realalpha.app/v/iRI6gTe1wGM/?t=1152) 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 zone frontier.com to learn more that's zone frontier.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 11 years you were cto then you became ceo i was also an engineer before that you were also an You've been 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, and now it's like, what do you have to achieve?

[20:17](https://realalpha.app/v/iRI6gTe1wGM/?t=1217) 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. 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. So what are that numbers? Like the easiest approximation is 10x. Think about like a 10x score 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 and making up a high number I have somehow managed to underestimate so now I can I feel like I can go to 10x feel a little bit uneasy that it feels crazy to grow 10x for too many years you know how long

[21:55](https://realalpha.app/v/iRI6gTe1wGM/?t=1315) 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. Okay. All of our insights, or at least my insights come from working with people doing really really interesting and unique things we started working in the beginning with all of the people pushing the limits of building agents so some of our customers like harvey notion they were one of the early people building agents for doing real work where the quality bar for that work was quite high and everyone in the adoption change was like, oh can we trust this agent to do our work? That was the question being asked. And we, by working with like really demanding customers or customers who had really demanding customers in turn, we ended up really pushing our systems capabilities on the quality dimension right the next phase i would say is like voice agents are starting to work and the interesting thing there is initially there was some acceptance for your voice agent

[23:24](https://realalpha.app/v/iRI6gTe1wGM/?t=1404) pausing a little bit unnaturally or being like let me look that up and come back to you like that was acceptable for like a three month period I think and then very quickly today if that happens to you at least the people who've interacted with multiple voice agents you'll be like that is ridiculous like that's a really bad product and so now you're seeing all these voice agents which are like super natural but also super fast So what does that mean for web search? Now you've got to be able to search the web in like 150 or 200 milliseconds. Like human web search was like a sub-second product, like 600, 700 millisecond product, roughly end-to-end.

[24:07](https://realalpha.app/v/iRI6gTe1wGM/?t=1447) Because we don't notice it getting any faster, right? For a voice agent, that latency constraint on web search is unacceptable. So now you have to push yourself on the quality dimension, on the cost dimension, and the latency dimension. And you've got to crush all three. And that's been something like we get pushed by sort of the frontier customer who is like, listen, the product experience that we can build today is going to be unacceptable very, very quickly. So what does every underlying infra piece need to solve in order to create the market expectation which changes so fast, right? Like, I think, I don't know, going back to web search, one of my, you should at some point try, even with the new model, turning off web search in ChatGPT once.

[25:04](https://realalpha.app/v/iRI6gTe1wGM/?t=1504) Okay. I think we all used ChatGPT for the first, I want to say, eight months, maybe nine. It did not have web search and we were just talking to a model. Try to do that today and see what happens. my bet is one 90% of things you ask it will disappoint you number two you will feel like a caveman for a cave world in using that product so it's only three and a half four years ago now but our implicit expectations of a product that looks the same have changed so dramatically and that is what sort of mind-boggling to me and I think that is essentially what drives all technical progress yeah is because it becomes table sticks right and now they might be getting subconscious you know they might have some consciousness to them and start judging us and attacking us I actually I don't know I'm more of an optimist on this okay I think they will actually get proactive. So they will know us. They will know our explicit and implicit needs and they will try to proactively do work. That's the phase, the next phase. I think you're trying to, you're starting to see glimmers of proactivity in instinct and muse and dot. Glimmers. I think they're still more reactive than proactive they're always on but they're not really proactive

[26:32](https://realalpha.app/v/iRI6gTe1wGM/?t=1592) i think the next era they will be hyper proactive which i'm really excited about i guess uh to peel back because this is a brand new economy and layer of the internet and what could i don't know the internet might not exist anymore maybe we'll just use agents this is what people are saying what are those components of the underlying infrastructure you're one of them what are like what makes up that base there are two ways to talk about this i think before going into the internet let's first talk about the the underlying infra components for building an agent you need a model which is inference you need perhaps a sandbox i think you need web search like we provide. You need an action space which includes a browser and a few other API hooks and you need access to a bunch of connectors to access your personal data or your company's data. You build these together and you can build an agent and then the rest of it is the all of the same stuff we've always had which is observability, feedback loop Oops, security. And then you have a cloud for agents, right? Now you're going to the web piece of the infralayer. You now need to do web search, but also you need to figure out to your point around the internet shutting down. I do think the primary consumer of the

[28:05](https://realalpha.app/v/iRI6gTe1wGM/?t=1685) internet is going to be an agent, which means the internet will get rearchitected. The first thing that's happening is agents are reading the web, which is what we do. What we haven't figured out is, okay, what is the new business model? Like previously, if you put your content out there on the web, you were either building a subscription business, which is like you put some content out there, a bunch of users come, some fraction of them subscribe for a monthly subscription to a newspaper. Other internet businesses were monetized by saying, I'm going to put ads and I will get paid based on human visitors seeing ads. Other business models were like either affiliate business models or you drive a transaction of some sort.

[28:54](https://realalpha.app/v/iRI6gTe1wGM/?t=1734) All of these business models have fundamentally relied on us random walking the web as humans, opening pages and taking the next action after the next action on it, some of which are monetizable actions for these participants. If agents are doing this, the buying action still remains. Every other action breaks. So you can't monetize attention anymore. So ads business model, perceived subscription business models, these things are not durable. So what replaces it? That's part of what gets me super, super excited about building Parallel. Because I think solving just technology problems, I think, is far less interesting than solving a technology and a business model problem.

[29:51](https://realalpha.app/v/iRI6gTe1wGM/?t=1791) So we have a highly opinionated point of view on how we keep the web open, on how we incentive align content owners with the AI app or AI agent so that the content owner isn't sitting there saying, I'm going to block all AI agents because I don't know how that becomes business for me. So we need to solve this problem for people, which is like, okay, give me an incentive to make my content available to AI agents. What is that? We've built a system which runs a model, not surprising, to compute how much to pay each content owner every time an agent reads their content. And we've been now doing a bunch of partnership with content owners and the bet we're making is that we'll be able to keep the web open as a result of our version of, we can call it adsense which is easy auto differential pricing actually provides content owner value for high quality content that's a really good point um i was interviewing usv union square ventures not too long ago and fred wilson was talking about how google the traditional business there is being cannibalized by ai llms agents it just is and so gemini is like the hail mary to pull it in and they're emphasizing that and they're taking advantage of that i'm curious to go deeper into

[31:29](https://realalpha.app/v/iRI6gTe1wGM/?t=1889) this monetization standpoint um in another interview i did at browser base with john colson of stripe he posed the question do we need bot detection anymore there's obviously good bots there's bad bots but if everything all the traffic becomes essentially bots these agents like what What is the point? Yeah. So I think today there's a lot of people doing a lot of bot detection. The word bot detection I think is a few different things. It is crawlers, it is actual agents working on behalf of a human or a business it is um malicious bots uh which are trying to do harm so i think we have to use that word now a little bit more carefully than we have in the past because historically people have mostly had negative connotations with that word in that like it's not serving anyone yeah today there's a category which is actually highly useful for you or for me like i would like my agents to inherit privileges i have so i think we should be careful i do think there is like a the bot detection market is doing well right now and it's doing well right now because we haven't figured out how to do incentive alignment and there's a separate kind of incentive alignment needed let's call it on the action side. So earlier I spoke about the incentive alignment

[33:08](https://realalpha.app/v/iRI6gTe1wGM/?t=1988) needed on the content side which is like how why do I let you read my content. The other side is what happens if I let your agent buy on my store and you're starting to see divergent strategies there. I'm of the belief that we need to figure out how to solve problems so that all agents not not just the ones built by a large lab, can both access all content and take all actions that their owner is allowed to, right? But we're not there right now. One of the biggest topics behind growth is cybersecurity and active threats. How do you, I know you said you're an optimist, but realistically, like how do you see the biggest concerns around this no i think the cyber security risks are real i think agents help both defenders and attackers problem always is that so the people who rush ahead to adopt agents for defending themselves, I think will be fine because attackers like to find the weakest link and there will always be a weakest link.

[34:34](https://realalpha.app/v/iRI6gTe1wGM/?t=2074) So broadly that's what I think. Now the security risk for something highly sensitive like the model weights of a frontier model are a whole different conversation yeah but by and large my intuition is that we will use coding agents aggressively to secure but along the way there's going to be a cyber event every few weeks of consequence which will drive action and urgency in getting ahead on defense yeah we're going to spend a bunch of influence on both attack and defense do you see any of this analogous to your time at twitter on consumer internet obviously that's a huge one of the largest social media platforms And so that creates a lot of vulnerabilities.

[35:31](https://realalpha.app/v/iRI6gTe1wGM/?t=2131) No, all of the, there are obvious parallels, right? Not to just keep calling out my company's name again and again. So at Twitter, right, like there was a real incentive to attack, whether it were crypto spams. Yeah. Those were popular for a period, right? Or whether it was some state-sponsored influence. campaign, which we used to disclose a lot publicly. Those things always happened. It required real energy, focus to control. The thing most people don't understand is it is almost impossible to take it to zero. You have to contain it because it is a knife fight and the tactics change. Anytime you control some vectors, another new vector will emerge then you'll control that. The other thing people don't understand is most of the bad stuff is stuff that doesn't show up on the surface. So if you saw an account and thought this is a bot, it's harmless because you could see that it's a bot. So the real problem is going to be in the margins when we lose our ability to differentiate either as people or as the technical people building these systems when you can't differentiate a human from a bot or an agent that's when you're like okay now I don't know what I'm up against and I think that is very similar today that we used to deal

[37:14](https://realalpha.app/v/iRI6gTe1wGM/?t=2234) with back then do you think bots should address themselves i have two contradictory feelings about this i do think good bots should declare themselves so that others can recognize them as good bots and treat them differently from bots that might be adversarial and again we keep using the word bots I would like to use the word agent. I really think bot has a more negative connotation while an agent is a more appropriate word. I think agents are perhaps the good bots mostly. So yeah I think they should declare themselves. I know there's like a bunch of protocols now being done around declaring identity or to the world at large. The core challenge there, why the other side is like this is a conceptual question. Should your agent have the same access as you? Should it inherit your access or should it have only a part of your access? Now I would really, it would be so much better if we could figure out a way to get rid of the bad bots but live in a world where we could enable our agents to have the access we have and grant that privilege. Today it's strange right like I can't get my instinct or muse to have the access I have on many websites because they got blocked there and they will get even more blocked if they started declaring themselves. And that is why I think

[38:49](https://realalpha.app/v/iRI6gTe1wGM/?t=2329) we're in this moment of transition where these products have these rough edges, partly because we haven't done incentive alignment. But if we started just declaring them everywhere, and they started declaring themselves everywhere, like we don't participate in that space, but we started doing that. A bunch of people today, because we haven't done the work to do business model alignment, are going to start blocking them yeah and so these products then will look worse and be less useful and that sucks so to me the existential problem is how do we drive alignment between agents and everything on the web do you think everyone will need alignment and safety on their team is that the next higher class yeah i think how do you not deal with the problem right like we if you think about us we are the web layer the web is the is a dangerous place like anyone can post anything malicious anywhere, right? A lot of that content will flow into an agent's context.

[39:41](https://realalpha.app/v/iRI6gTe1wGM/?t=2381) So you can do injection attacks. Now, to the credit of model builders, in the last two years, models have made very real strides on not being very vulnerable. But I think this whole world will rely on defense in many many layers so almost every layer has to obsess about this problem and you can't say oh models will solve it we have to really focus on it to solve it at our layer before we get into the models then models have to really focus on it then the app builders who are building with models and with us have to really focus on it at their layer so we have to deal with adversaries at every layer and that is actually what gives you true defense and still the occasional thing will leak out. How good is AI at real professional work?

[40:26](https://realalpha.app/v/iRI6gTe1wGM/?t=2426) Merkur's AI Productivity Index, or APEX, tracks it with every major Frontier model release. Built with partners like Harvey and Cognition, see where the Frontier actually stands at merkur.com slash APEX. Today's episode is sponsored by VCX by Fundrise, the public ticker for private tech, allowing investors of all sizes to invest in venture capital. Learn more at getvcx.com. Some of you may not have heard this yet, but our sponsor Public just launched something called Generated Assets, and it brings AI into investing in a way I've honestly never seen before. Here's how it works. You type in an idea like AI-powered supply chain companies with positive free cash flow or defense tech companies growing revenue over 25% year over year. Public's AI then dispatches a swarm of agents that scan every single U.S. stock, evaluates them, and instantly builds a custom index around your thesis. What really stands out is how clearly it explains why each stock is included. And before you invest, you can even backtest your idea against the S&P 500, so you're making decisions with real context, not just guessing. And beyond generated assets, Public lets you invest in stocks, bonds, options, crypto, all in one place. They'll even give you an uncapped 1% match when you transfer your investments over

[41:33](https://realalpha.app/v/iRI6gTe1wGM/?t=2493) from another platform. If you want to build a portfolio that actually reflects your thesis, visit public.com slash sorcery paid for by public investing full disclosures in the description founders scale faster on deal set up payroll for any country in minutes hire anyone anywhere get visas handled fast and get back to building visit deal.com slash sorcery that's d-e-e-l.com slash sorcery how did you recruit for your team especially as this is moving so fast so you're you have global offices now yes um how big is the team and how have you like recruited what are the main roles the team is about 120 people now i think we started the year close to 30 so this has been a year of explosive growth a lot of that growth has been because we had no gtm at the beginning of the year we were pure tech play now the last three months six months we've built a pretty sizable gtm team and that's been done because we're seeing all of these almost every business that buys inference needs to buy web search so we're finding really large like whatever the top 100 500 companies in the world starting to really spend on tokens and starting to really buy us alongside those tokens to make them more effective so a lot of our growth

[42:54](https://realalpha.app/v/iRI6gTe1wGM/?t=2574) is the sales team which how do they know to buy you it alongside the tokens they run into problems so one of the common things people run into is people start using in A model. That model perhaps bundles itself with a web search. Okay. And they start building an application to that. Now, sometimes it's not good enough. And the team realizes that it's the web search component, which is causing it, despite the model being great, it's causing it not to be capable enough to solve the problem they're solving. which is perhaps how we started first working with some of the AI legal, AI finance, AI science kind of customers.

[43:32](https://realalpha.app/v/iRI6gTe1wGM/?t=2612) Right. The second point at which they really need us is they decide to use multiple models. So we're like, okay, we want to switch across multiple proprietary models on a dime or we want to start adopting open-weight models. And at that point, having the highest quality web search, which works with every model, actually enables model independence for you to have flexibility to build your application. So that's the second reason people buy us. The third reason is when people have really scaled use cases and now start optimizing for measuring ROI, optimizing for end-to-end costs. because our search is so high signal and token efficient, the models do half the number of searches, use up far fewer tokens, end-to-end are cheaper, faster.

[44:21](https://realalpha.app/v/iRI6gTe1wGM/?t=2661) That's the third reason people buy us, which is like, I'm spending too much on tokens. It works. It's useful. Can you help us be more efficient? So it's not just routing. You also help with token optimization. For a lot of categories of knowledge work, where a lot of work happens on the web, We help a lot in token efficiency. Wow. So when we talked to Gia Vercel, we were talking about the shift of GTM. It was SEO, then it was GEO, and now it's AEO. So, like... I think GEO and AEO are the same. A little bit, but not really. Okay. Because the agents now... More than chat. Yeah. GEO was like chat.

[44:57](https://realalpha.app/v/iRI6gTe1wGM/?t=2697) Exactly. So it's like Vercel has optimized itself for the agents coming to its website so the agents can read the website and know exactly what products they offer how they can use it and so because of that Vercel is like one of the top uh uses for building out your products and that kind of coding and all that environment no i think it's it's interesting you bring that up three years ago when we put up our first website up when we were still kind of it didn't say anything we wouldn't tell we had a machine version until this date if you go to our website you'll see like the human version and then there's a toggle to transition into the machine version really we've had it as far as i know you're the first website to take on this metaphor because we're building a parallel web for agents yeah and so we have always carried this metaphor which now the world is coming around to it's like a cute thing to do because the truth is the presentation layer almost doesn't matter because we can we can transform the web even if you don't publish your website to be agent friendly we'll do that for you like we can we can understand what the human version is and how to help it become agent friendly oh wow um

[46:12](https://realalpha.app/v/iRI6gTe1wGM/?t=2772) we don't do a or geo but i think your your website will become consumable by agents with us but i do think like we for example even think of our onboarding and our documentation and every part of the developer or customer journey with the primary assumption that it's a human with a coding agent or a developer with a coding agent rather than a developer. And so you'll see a bunch of on our website, you'll see like, put this prompt into your agent and then it'll exactly know how to use parallel to solve whatever problem you're trying to solve. Like one claim just like you just need the human needs to just authorize slash add a credit card slash accept terms and the rest of it the agent can rip. So really rethinking how you design every experience to be agent first is I think pretty fundamental for sale like I know is very agent built so people who live that world have a advantage we have like probably like 10 or 15 minutes left uh but I'm curious I mean you're doing a lot of press like you're out there you're used to getting asked a bunch of questions like what are the questions you want to get asked I spend a fair amount of time thinking about today proactive agents and always-on agents, how that changes the

[48:04](https://realalpha.app/v/iRI6gTe1wGM/?t=2884) compute world and how that changes the web. So the question I think there to me is and we've now built a bunch of systems which some people are starting to use around this notion of a proactive web or I've been out there saying the web will go from pull to push. So far your primary, as humans, our primary way of searching the web was just us going and searching on Google. For the rare few you use Google alerts, it didn't really work. People tried to build like what Google now and a bunch of these things which would send you notification. Nothing really worked on the web in terms of being proactive I think with agents and the cost of intelligence going down it is going to work and so now if you imagine an interesting term news implicitly understanding your world and your life and being able to set up essentially really really high signal high precision high recall alerts on everything that changes on the entire web to know when your agent needs to act on something on your behalf. So something changes in the world or something might be happening, what do you need to do as a result or does that change your plans or does that change your actions? How we allocate compute in that world is a problem I am currently obsessed with

[49:36](https://realalpha.app/v/iRI6gTe1wGM/?t=2976) because the web is really large. There's a trillion pages out there. A bunch of much change is changing every day. Most of it is junk. No matter how cheap you make compute,

[49:46](https://realalpha.app/v/iRI6gTe1wGM/?t=2986) **Parag Agrawal:** we have not enough compute to compute everything. So now we need to be really smart at prioritizing our compute over everything that is changing on the web so that more compute can be spent

[50:03](https://realalpha.app/v/iRI6gTe1wGM/?t=3003) on things which are actually actionable rather than junk changes. The rate of change of the web is also going to dramatically increase, right? You're starting to see that, like people are publishing with agents, like you can like, the AEO, SEO business is like churning out agent-generated content on the web, which just means the rate of change on the web is increasing, which means the job of a search engine is to pull the signal from the noise as potentially more noise gets created. How do you think about data centers in space? I wish I knew more about it. I hope they work. Have you gone down any of these rabbit holes in this compute?

[50:51](https://realalpha.app/v/iRI6gTe1wGM/?t=3051) I haven't. I haven't gone down the data centers in space rabbit hole. I do think we need compute. Yeah. I have gone on a little bit of the how sovereign will party. Participate in the data center. And that is fascinating to me about how that will play out. Well, one of our partners is Brex, and they're all about performance and spending so that are moving faster. I take this topic in the personal way. So I do think performance is a matter of who you spend your time around. Some people say it's your closest five people. Some people think it's like your mentor, people who inspire you. For your career, tremendous career very interesting i'm so curious who is that person or those people for you my wife is her name is vinita i was an engineer at twitter when we got to know each other so if you think of the part of my career where i've done well that people know about it's mostly her she is a superwoman she's hyper ambitious and I was ambitious but I think she has driven my ambition even higher like I have always aspired to be the top X in the world at anything I do. That's like the only way like I was driven to do that for math, I was driven to do that for physics, I was driven to do that for

[52:27](https://realalpha.app/v/iRI6gTe1wGM/?t=3147) computer science research and I think I got there for all of those. But I don't think I would have thought I was capable of this until she pushed me and then I was like yeah I think if you care enough about something I think you can go become top 10 in the world at anything. Now the top one is a whole different field but I think like for me it's like the the standard setter is my wife. She is a she's a physician, she practices still, she is an investor at Andreessen Horowitz, she's done a she's done a startup she does healthcare she does biotech she works really hard and is an amazing mom and wife that's amazing so you guys have a nice little competitive dynamic yes yes yes I think I am privileged that when I go to a lot of things like people use me to access her really people will come to me to figure out how to get me to forward an email to her oh my gosh that's amazing well that's a really great call out I wasn't expecting that why do you think partner selection or who you choose as a partner is so important probably the most important decision in life and like you as you said right like you your peer group like I was fortunate growing up to have like a really crazy weird outlier of a peer group like all

[54:12](https://realalpha.app/v/iRI6gTe1wGM/?t=3252) of my friends were essentially kids of nuclear scientists in India and everyone was like okay I'm gonna be the best at whatever we try to be and so there was like this ambition slash competitive expectation all around and so if you surround yourself and then I went to college and my college was the most competitive and I went to grad school that's the most competitive so you surround yourself in that and like you kind of surprise yourself by being around high expectations so when you like this is what now everyone's talking about right like you build a team if you have high talent density it's a superpower if If you don't hire outliers, you don't get outliers if you hire only outliers.

[55:02](https://realalpha.app/v/iRI6gTe1wGM/?t=3302) You get only outliers. There's a lot of self-selection that happens. I've been fortunate to go through that journey and I got started in life with that kind of peer group. And then you kind of won't ever let yourself not be in that peer group. And that percolates into your partner selection, which then further drives things. So it's like kind of a self-fulfilling prophecy of sorts. in such a competitive career in every stage how do you not break from the pressure what pressure you don't feel pressure no no why why would i a lot of people do yeah i think it's kind of a made-up thing right i really believe that like i think i through the twitter acquisition i used to track my sleep score really and i was sleeping fine i think if you know what you must do you do it you can you don't have to feel pressure like the world can come at you but if you know what you're capable of and you have a slightly long-term view like everything is a week or a month so you don't have to feel pressure you have to perform every day which is a completely different thing so now the core question you have to ask is there are moments where you must feel pressure to perform so i really like feeling insecure is my word which is i have i've always

[57:00](https://realalpha.app/v/iRI6gTe1wGM/?t=3420) told people that I only want jobs which I wouldn't hire myself for. So I want somehow to feel that someone's thrown me into a challenge that on paper I'm not supposed to be able to accomplish. One, that drives you to be slightly humble. Two, it makes you a little bit paranoid. it makes you want to learn things and makes you curious and it makes you work really hard and it makes you seek help. If you feel secure in what you're doing, it doesn't drive any of the most important behaviors in my experience. So pressure is a different word because it suggests that you are trying to meet somebody else's expectations instead of your own. And I think internal expectations are way way way better drivers of success than external ones.

[58:03](https://realalpha.app/v/iRI6gTe1wGM/?t=3483) And at least for me, I would say other than my wife, my expectations have always be higher than the world's expectations off of me. Even from like true believers like who believed in me what a fascinating turn i did not expect that this is like we i will say thank you so much okay for such a fantastic conversation um it was really awesome to learn more about parallel and then your personal life too if only we started with this your philosophy and like how you think about things is really yeah that's fun yeah well thank you so much frog amazing cool hey it's molly if you enjoy our interviews check out our newsletter sorcery.bc where we deliver a once a week top deals and tech headlines email and also go deeper on our podcast interviews subscribe to sorcery today and don't forget to subscribe to the podcast on youtube spotify apple or wherever you listen link in description to sign up