"The CEO Must Be the Chief AI Officer"

you wake up, whatever problem you have in your life, why can't you solve it? And just like start there. I think the CEO needs to be the chief AI officer. Like it's not an engineering team thing. It's not like a pro- team thing. It's like you have to understand the balance of the technology, better than anyone. I think a good proxy for how to spend your time is what are things that only you can do the models cannot do. You have to sort of re-found the very concept of what the company self-identity is. Welcome back to another episode of The Light Cone. Today, we're joined by Pedro Franceschi, co-founder and CEO of Brex. Pedro started Brex in the YC Winter 17 batch and built it into one of the most important FinTech companies of the last decade. He's here today because Brex has gone deeper on AI than almost any enterprise company we know. And Pedro's own AI setup is so compelling that when he came to YC for lunch, it sent our entire team down a rabbit hole of building on their own. So Pedro, welcome to The Light Cone. Thanks for having me. Excited to be here. Thanks for changing all lives. Oh, God. That lunch. I'm like, I think the model company should be sponsoring me for the token consumption increase I generate. We supposedly generated on that lunch. That was the precursor to the brain, I guess. I was still working on G-Stack. I was still a 2013 Web 2.0 engineer who time traveled instantly to the AI tools of January 2026. I was, you know, probably half a million lines of rails code in. I created G-Stack because of that to help me make a software factory. And then after I met you, I realized everything is about freeing the claw. For example, I need you to say that. And then. And give it tokens. Yeah. Well, no, I mean. Let it rip. The craziest thing was realizing like what I had gotten wrong that I think actually most people in software are still getting it wrong is they've been treating the LLM like this very precious thing that's very expensive. Yeah. And so as a result, you have to literally put the agent inside a Foxconn factory. And it's like, like, can you imagine? Like, I mean, that's what the half a million lines of rails code was for me. It's like, no, no, I need to control what the LLM sees because it's about really, really like, I only want the context like from here. And let me write like all the if statements to make sure like, you know, like a Foxconn engineer, you're waking up at 6am. And you know, if you don't, you're going to get electro shocked. I mean, like, this terrible thing that you do to agents. Yeah. And they want to be like at the Esteline Institute. And that's what OpenClaw is. Exactly. Exactly. And it's funny because I feel like every single good AI product you've used is an agent loop with tools. That's it. Like, there's no, you try to sort of overengineer the harness and then do certain things by the end of the day. It's skills to us and a model. Like, there's not really much else. Maybe we start earlier because one of the things we'd love to kind of, you know, get down as a part of lore is like, how did you get so AI-pilled and like, all the way to the edge? Well, I'll tell you my encounter, LLM's, which was so, so I remember in the pandemic, there was someone, someone gave me an API access to GPT-3 and I was playing with it and I was like, okay, this is, this is really cool. This is, there's, there's something here that could be, could be special. But it was the kind of thing that was like, yeah, I feel like a research project, the kind of thing that Google used to release and you're like, you play with it for 10 minutes and you stop. A chat your PT came out and I think everybody was sort of interested in it. Where I think it got interesting was when you started to see reasoning models and of course tools, but, but I think everything also sort of a blip until December. And the way I describe it to my team is like, you know, electricity was invented in December and I think electricity was Opus 4.5 and ensure Opus models and, you know, open AI models got got better and better since then. But to me, that was the the tip of the spear where you could say, yes, like coding harnesses actually work and, you know, Cloud Code existed for probably a year before, but it wasn't that valuable yet. And I remember, you know, during the holiday break, I was, was playing with it and it was, was pretty shocking, probably some reaction that that everybody here had. And I think the question becomes, you know, if you sort of, if you think about, you know, you're, you're, you're sort of standing, you know, looking at 200 years of history and then you imagine you are, we're not in May, you're sort of five or six months after electricity was invented and most people are still playing with candles and, you know, questioning, you know, what can you do of candles and fire and, you know, like, yeah, exactly right. With all these lanterns and what can you do with it? And, and you know, the steam engine is like, I don't know, maybe like 20 years away still, but, you know, electricity already exists. That to me was the sort of the fundamental light behind it. And I would say, I think since then, Open Cloud was kind of a interesting sort of next step, which is, I think when we realized that, you know, the reality is, good AI products are agentic loops with tools. And we started doing that, you know, on product of Brex, but, but then on the personal side, I started spending a lot of time understanding, okay, what is at the frontier of using Open Cloud? And I think the insight was just, yeah, like, mark downs can take you really far, just like configuring and automating a lot of the things in your life. It's kind of funny. I remember I had this experience of like buying a movie ticket entirely in Open Cloud, using like a Brex card, it was provisioned through an API. And, and then I showed it to my team and they were like, oh, but like, you can go online and like book it in 10 seconds and I'm like, that's not the point. You're missing, you're completely missing the point. But anyway, and then I went obviously very deep into his rabbit hole and started spending a lot of time thinking how to change the fabric of the company and the way people the products and tell us about your personal Open Cloud journey. Oh, God. Before you came to the lunch, I had it like in store, but I was like, wait to scare to do anything with it. We were all scared. Yeah, don't get me wrong. Like, we deal with financial services data. We spend a lot of time thinking on how to be mindful of security and protection. But, and yet, I think people are a little bit more risk-averse than the technology probably requires them to be given where technology is. And when we started using Open Cloud personally, I started doing it a lot of my own personal setup. Basically, what I did, the V1 was, I'm going to give it read access to everything and just create like, a lot of tokens to my email, to Slack, and so everything to just literally not write. And I was kind of shocked how far it got me. And the next question that we spent time on Brexit was, okay, like, how do we actually get it to write into our systems? And everybody in the community and security team was, well, we cannot do that for all the reasons that we know. And then basically, where I spent, I don't know, probably four weeks of my time was, okay, let's solve the hardest problem, which is security. And we ended up realizing that the only way to actually do something about it was to do something in the network layer. And if we treat the agent like, you know, the agent has its own wills desires and, you know, they go to the Asalent Institute for agents. And, you know, they have to- >> Instead of Fox cons factory. >> They would try to do things at a network boundary that could not be the right ones. And we just had to actually just focus on that. So a lot of folks were, you know, and we saw in video in others, on Nimbleclaw, let's build these like open shell, four X that have controls over, you know, tools, the model calls, and the reality is, yeah, that you can do all that, but you can also just make an HTTP request wrong. So we focus on that layer and then we build a single crab trap, which we open sourced probably about two months ago, which is actually the way we use to secure agents at Braxton production. And the basic premise is you analyze, you, you HTTP proxy the entire network boundary of an agent. And the idea is, when a request goes through that becomes auditable. And you basically can use another agent to analyze the traffic and create a policy to let traffic go through or not. And surprisingly, or unsurprisingly, because these models are trained on, you know, hundreds of billions of web documents, HTTP traffic is actually, I would say probably the way the model's reason more so than anything else, because they just literally learn on the web. So the, the ability of the model to watch like a thousand requests and make sense of what's happening was way higher than we anticipated. So we actually build that, put that in production of Brex. And after you record a traffic of an agent operating for a day, you can build a pretty good policy that, you know, sets things that should be automatically approved. And for things that you, the agent isn't really sure, you can just use an LM as a judge. And the LM determines is this request something that should be approved or not based on the policy for what that agent should be doing. So for example, we have like a recruiting agent of Brex, Kojim. We have a policy for Jim. And you know, all the traffic goes to the same policy and 98% requests go through automatically, 2% use an LM. So we sort of got that problem solved to a degree that we got comfortable experimenting much more aggressively and sort of freeing the clock on the enterprise, which is really hard inside a capital one. So I would say if we found a way to experiment with these things and granted, we don't do the most aggressive things of this stuff yet. We don't use it on like, you know, customer data to a degree that we want one day to do. And there's boundaries to how we do it. I don't see any reason why a YC company shouldn't be at the bleeding edge of the stuff. Yeah, I mean, I think your intuition around like the proxy at the network level ended up being quite prescient. They go think a lot of the stuff that I'm seeing kind of around the open clause ecosystem at the moment, at least or just agent ecosystem is essentially doing that. Like we're seeing that like with credentials, credential, brocreen, like agent vault is doing a lot of that. I think you mentioned the first version of CrapTrap included like credentials. vote. Why do you decide not to include that? I think it was just, let's just do one thing really well. And, and, and, and you know, at the end of the day, I think there's going to be a lot of solutions that do that. You could do credential broadcasting and other tools already. But the LMS of Judge was for us, the determining capability to say, do you trust this in production or not? In our security team of Brex, very rigorous and very good at what they do. For a long time, we're, well, you know, not really. To getting them to a, yes, we actually believe this is enough. Was a big unlock for us. And look, I always say this, like, we're not in the business of building HD Rocksies. We are in the business of being the bleeding edge of what he can do via. And to get to the bleeding edge required us to build this proxy. That's why we did it. Hopefully someone's going to build a YC company. Hopefully we're just going to build a better version and we're just going to go use it. But at the end of the day, that's the journey that took us to just sort of being at the bleeding edge in that way. And how much was was you like sort of pushing this forward and make how much resistance did you get internally and just how did you think? I mean, we got a I build. I think there was a lot of excitement about it. But the way I describe it, the option inside most companies is I think there's like sort of three tiers. There's tier number one, which is your token maxors, like your engineers that are pushing a bunch of code and typically living inside coding harnesses. And those are sort of well known. We know who those are. Then you have the sort of average engineer that is building a few things, but not not sort of token maxor to the same degree and probably I don't know, a tenth of the productivity. And then you have like the entire rest of the company. And the entire rest of the company typically is interacting with AI in what I call like Google search mode, right? Which is a chat bot with a few MCPs or a G Suite equivalent that you have a few tools from Google. But at the end of the day, it's really just like a search. And I think that where our thesis was, if you think about the value that AI creates for like a token maxor, for example, a lot of the value comes from the harness. And the thesis was how do you actually build an equivalent harness for other teams that are non technical. And our whole sort of thinking behind it was like, that's a lot of what open-cloth, you know, created, which is his ability that you can self-boot strap a lot of the capabilities of the agent by the way you edit your skills and mark downs and sort of set up the environment around the agent. And how far can we get this ability to 40 agents to self-boot strap capability without anyone actually going and coding by hand? So the analogy we use internally for, I would say the sort of the company-wide adoption of AI is we don't believe in the yes, like if people if you MCPs let them go. Because I think what people really want is, in my opinion, is really a way of saying, okay, this is actually a virtual employee, almost that has, you know, this on Slack, it has an email, I can actually invite it to me, can join a meeting, take notes, and you're trying to replicate that as much as possible. So how do you go with infrastructure to support that kind of use case? And I think the harness says, we'll look a little different and probably more like open-cloth than a coding model. Jared and I just did this this week for the first time where we installed Aqua Voice and then open Telegram with the claw, or actually we have it in Slack now. And then basically it was like me and Jared and like three engineers and are someone from the events team. And we're trying to put together, how do we put together 60 dinners with 20 people each of attendees from startup school? Nice. With 21 partners and visiting partners at YC. Sounds great, great problem. And then we just basically started talking about it and then I picked that up and then I pressed enter and then, you know, our claw just started doing it. None of us opened Cloud Code. Like it just sort of built a bunch of markdown, it did the analysis and yeah, people forget that Cloud Code is magic. It's just literally a harness around the same models we can use on an API, right? So I think that's the unlock of Embout a way. There's a few things that Cloud Code is doing that I think are really cool. Oh, there are amazing. And yet it's just, it's just a harness. And and and and and and the clock can use Cloud Code. Exactly. It's code. Yes, right. It really prefers to use code X these days. Exactly. Everything actually. I don't know why. Exactly. But ACP helps on that. So yeah, ACP is good. What do you think the adoption of token maxing hasn't really taken off? The thing that we found it very curious working with a lot of startups early on is a lot of founders are very shy about burning tokens. I think you really get to experience this when you really go all the way. Gary mentioned this point, which is token is expensive. And I think there are, you know, I'm in a fortunate position to be able to spend on tokens. But I would say I keep trying to picture myself. Imagine if I was like 14 or 12 when I started coding for real. And I had the technology we have now. I would be token maxing the cheapest way possible. And there are people doing that. You know, you look at the Chinese models, for example, like they're pretty decent. There's a huge hobbyist community where they build a gaming rig. But then they try to build like local alone. Yeah. And then that actually is like totally reasonable way to do it. 100%. 100%. I have a friend that that has the exact same setup. He has his like little GPU farm in his house. And and first time when they're like, wow, heating is on here. It's really hot here. No, it's my GPUs. And it's like, great. Like, you know, power efficiency all the way through. It's funny because at Brax and we should talk about managing token cost and spend management for tokens, which is a topic we're spending much as high goes on now. I think that the cost part is one. But even if you take the cost part aside, you know, the first symptom is a lot more people should be complaining about the max plan limits. And you know, you see what's the percentage of Twitter that probably complains about it like 0.1%. So I think more probably still early. To me, there's this like the AI pill test in my opinion is whatever problem shows up in your life. Do you default to AI first or not? It's like, of course, mechanically, you can do it. But there's a point that it becomes like second nature. And then your whole like brain gets rewired and you cannot think in a different way. And there's the whole topic about AI dependency, human machine interaction. Yeah. There's all these things that we can we can talk about and put in the corner is still sort of surprises me how many people you go talk to about a problem. And I'm like, it's so cheap to intimately understand the bounds of this problem now. Like why haven't you done that yet? And come in with like a much more digested view on the problem. And I think the second thing is like, I think I think if you have the luxury of building a company now, the fabric of the company from day one can be built in such a different way that I think I if I were to sort of company today, I would say, okay, the premise is why can't it be just me? Like, and then you start from there. And your token consumption is probably going to be a lot higher than if you said, well, I'm going to have like three people or five people or seven people. And I think the fundamental constraint isn't as much in my opinion. Like, oh, like like AI as a cost savings, well, I'm going to be more efficient. I think the on lock is like the fabric of the company just looks very different when the boundaries become type systems interfaces agents talking to each other versus people. And I think people are still didn't fully grasp by, okay, what does it mean to build code with new agents like and the new technologies have I think that's like well understood, but how to live in a world where intelligence is on the tap and your default answer is let me actually solve this problem with AI first, even if you feel suboptimal. And then from there saying, okay, how do I actually make it optimal? Because I think for the majority of problems, there is a way to solve it. If AI that is probably better and your job is to figure that out, even if it's going to take you more time because that will compound. Why see startup school is back. We're hand selecting the most promising builders in the world and flying them out to San Francisco for July 25th and 26th to discuss the cutting edge of tech and startups. Apply now for your spot. When you started Brexit, I mean, like it's well known like you're like MVP, like had no web UI or something like all terminal like super scrap today would have because yeah, that's what I'm doing these HTML is assessing more like was it actually store the right approach to just have a really simple MVP and test that anymore. Which would you have like a way more fully featured. So I have this controversial view which maybe you all will disagree, which is like I actually think if I look into a pattern of companies that succeed, I think there's a really interesting pattern, which is minimal surface area. And the problem is with AI, I think you see like look at Stripe for example, Stripe other days was like literally an API. Brexit or other days know UI just like literally a terminal. You look at Airbnb is like the website was a form and the form was just like literally where you input it, what you needed and then someone somehow went there and figured out how to actually make the booking happen. Like doordash in the early days similar, right? Like it was just like little so the surface area was so small with the customer and so much of the band, the sort of the intelligence and the bandwidth of the founders were spent nailing this one single interaction pattern. And I think the risk with AI is that the agency behind choice goes away. So so you have this you have this this this this lack of discipline on what matters to solve. And I think people tend to believe that I can just experiment a lot of things and that's absolutely true. But but that doesn't preclude you from actually choosing what matters. I always tell people like I think if you don't if you can't minimize your surface area and and solve the problem with a very clear set of boundaries, you haven't found a problem to solve. solve. And I think that's an, and you can of course find how to compress the problem into a smaller surface area using AI. And that's really valuable. But I don't think you're used as an excuse to not do that, which I think is why I can just build so many other things. But you know, I always tell this to people like intelligence is compression. So when someone comes to pitch me an idea in the company, I'm like, it has to fit in an applicant. Like, where I just fit in an applicant. What's the, where it's an applicant? And then someone comes with this and I'm like, I don't know where you buy an applicant, but the ones in my house are not as nice. How about the set before it then even I'd like actually a lot of the pivot advice I give founders during the badge comes from you talking about how you found the Brexit idea. And if I like the approximate view I remember is that you thought about it as like two week cycles and like you're either in like exploration or exploitation mode, you know, like trying a bunch of things, but then you want to like hone down. Like would you still use that pattern now? I think one of the most one of the hardest things of building a company is talking to customers and not just having the conversation, but how to extract the sort of unspoken signal from these conversations. And I think to me, the can AI solve this lens, like like whatever problem shows up in your life, can AI go solve that? And you think about like building a successful company. Like why can't you prompt your way into that? And the reasons are simple is because there's signal that the models were in trade off. And the signal is when you go talk to person and they tell you about a problem they have, they're not going to tell you they're not going to give you the answer shaped into a prompt that you can put into an LM and that LM is going to go and output the product that's going to win and be a billion dollar company. They're going to tell you a very sort of local optimum answer based on their worldviews and their constraints and the way they see it things. And I think a lot of the job is the job now is to have the wisdom to choose what you want. Because before the wisdom was not just to choose with to choose and know how to execute it. The execution is out. Right. The execution is gone and the model is going to do that better. The wisdom to choose is still I think the missing bottleneck. And to me that all comes from which signals are not in the models. So say like pre AI, you had personal bandwidth to explore like three ideas in parallel. You're saying like now in AI world, you'd still do three in parallel. Would you like 30 and let the models try and the way I probably approach it is like let's pick a broader universe of things to sort of early initial exploration. But to me, the lens is okay, why can't AI solve it and like which signals on in the model. And I think the signal was typically the customer. And then and then when you go talk to the customer, I think I wouldn't paralyze that probably. I would be, okay, let's try to get into head space of this person. And and and I think there's like it's so easy and we did a lot of exploration with like synthetic customers and building customer models and things like that. And and those are really valuable once you know a lot about the customer. But when you don't know enough yet, I think there's this like very basic thing, which is even a brex, for example, like one of the hardest things for us as a company was we initially sold to founders, we were founders, we knew about ourselves, we knew about our problems. And then as the company got bigger, we were selling to finance teams and finance teams are different. So so building that mental model of like what's the value system? Like of course, you can eventually make the model represent that and and have that worldview. But but there's there's an intangible that I think is is where a lot of the alpha still comes from. And I think to me is like the I think a good proxy for how to spend your time is what are things that only you can do? And even in the company of one, what are things that only you can do the models cannot do? And that to me is like one of them. I think that's so on point. I think a lot of founders like you that successfully navigated pivot have this loop basically there's this the just book others in mind from psychology. That has to do with people that have very good emotional connection with people are able to simulate what the other person is thinking and what the others and others theory mind. Exactly. And I think the founders that get that and have the empathy to figure out what the customer is not verbalizing is the what is make the I think Gary says is make the implicit explicit 100% of what are all those desires 100% and the very subtle signs a lot of time because they're murmurs. As founders go through them and figure out the inside that oh it's this really a thing. But how do you know when to poke for it? And the problem with relying on models and right now which is I'm still very optimistic that there's still a lot of job through founders. Definitely is you don't even know what the right incantation or set of prompts to ask the model because they're there you don't even know what to ask. Exactly. There's like another meta layer. Yes, it's the whole like you long thing of like you know which question is the universe the answer for kind of and of course these these are generalities right but but I think what I've seen is you have to remember that elements are not magic. Like elements are trained on a very specific course of information optimizing for very specific set of benchmarks and outcomes. And I think the biggest bitfall of elements is you have no sense of how much training data the model has seen for the exact thing that you're asking it. So imagine imagine if like every time you ask an all-time question it gave you like yeah like I the sampling frequency of this in my DS had was I don't know X and on this other answer was 0.0001 X. You would trust is very different right the distribution is so different. Oh I would pay for that. That's a great start up idea. Exactly. So we need to do a model that does that. Yeah. I would pay for it. Yeah. Well because it's fascinating because then like anything that's out of distribution you just go and like fill that in. Exactly. I mean actually as an applications engineer on top of the LLMS that's actually a huge like blind spot. And that's a more core and a lot of the other data companies are doing like a lot of the jobs for them is to say well where what are the mind spots for LLMS and it's funny like I think a lot of the data labeling companies right now trying to understand the bitfalls and the models. But the problem is in order to do that you have to be an expert to know what the gaps are in the answers. But the problem is the founder when you're looking for an idea is you know nothing about it. So there is a there is a there is a curse of knowledge and a curse of not even knowing what the balance of knowledge is which I think can can can make you believe that you understand something that you actually can do an art of model action to understand. Can I confess something weird about like after creating G brain now I do use AI in a different way where now that I have a retrieval system that is actually usable if I have a problem or question about anything like for instance I was trying to work on a really really like the last humanized prompt. And you know a lot of that stuff probably isn't in distribution yet. Yeah. There's a whole Wikipedia article about like you know characteristics of AI writing. But you know now I can just go tell it like go spend a day like deep research literally every single paper article like read everything put it into my Git repo and then I'll be able to retrieve it and summarize it into something that actually is usable. And so the sort of like filling in 100% the things that are out of distributions like I can sort of like pack it with whatever context. And it's like you can do that with anything is like if you're interested in you know running a restaurant literally you could have you could go and read like 500 books about like every the top 500 books about what it's like to run a restaurant. And you would have like the compendium of all information about it. Yeah. And I think a lot of what like for example like one of the things that we do a Brax now is building this customer world model is a similar idea where we're trying to get every single touch point that the customer has of us like literally like what how many times they click a button to the dashboard all the way to where they tell someone on an email or would they say on the phone or they say on a call and and just that in months all data. Okay what should this customer need next from us? What should this customer be thinking about like what are the issues that they will face with having faced and again is just a distribution problem. This is actually an answer to one of the questions which is like will there be jobs or whatever is like as long as there are limits on RAM. Mm-hmm. Actually like there will be. Mm-hmm. So I don't know. I mean, I think it's kind of an interesting one, right? I think so. Literally you can't have a model that has enough parameters that could like have everything that you could possibly need in distribution. Like there aren't enough atoms in the universe, right? It's like a modeling problem. I think we forget that the world models in which the models are trying like there is something that the designers of the models influence the way the model actually behaves in the end. So you know one of the things that we spend a lot of time thinking is like how to make elements work for people to look very different from us. How to make elements work for like the average finance person in the US that for talking about an answer and you know the model defaults to like AI cap ex as the final as a default category for like like for example, that's a really funny example. Like I was playing with the AI for Connie categorization and like the first example of like an example of an expect is just like writing pros and an example is like AI cap ex and I'm like, Oh, why is it AI cap ex? The first example comes up with because the people that are building the models fucking only think about AI cap ex, right? So there are things like that that I think is like kind of interesting to think about that the mental models of the model, I think are out of the box are more biased than we may give them credit for. I mean speaking of AI cap ex like earlier you're saying that you know we're so early still. I don't know the funniest thing about AI to me is how often I find myself thinking crypto maxims. Yes. This is the worst the models will ever be. Yes. My favorite now is telling people who hate AI coding like have fun coding at one X speed. Exactly. Exactly. I was telling a friend about you know, how how how to be you know long inference that based on the thesis that there's going to be a lot more inference people think. people were expecting a lot of him for instance, we just look at public markets and, you know, semi supply chain, all that. - People are saying like 10,000 X. - Yeah, but the underwriting, which is kind of funny, is like, I think there's one image. 2,500 dots, each dot is a 3.2 million people on the planet. And basically, you know, 84% of the world never use the eye. 16% have used at least once a free chatbot. Then 0.3%, which is, I guess, six or seven squares. Pay 20 bucks a month for AI. And one box out of the 2,500 actually use agents in whatever capacity. So that's the arguments to be long inference. And I think this is just starting out. And I think a funny thing on this is, I think the, you know, it will be the biggest expense in a company, like easily, right? And it asks, there's a lot of margin and tokens right now, but people always want to be the bleeding edge, but even token costs decrease by 10X. They're going to have 10X more usage, so it will be a still large cost. And we're spending a lot of time thinking how to help companies actually manage token spend on brex. We ended up building our internal version of this. We call it magpie, where the idea is you can effectively, you know, every dollar of token spend in the company, you can attribute to product we have to customers and internal tool that we used to serve, or an internal employee, and understand model usage, et cetera. And we're now figuring out how to build analytics on, what are we trying to do with the tokens? To start to get a sense of ROI. But anyway, it's a fascinating topic that I think has a lot of early work compared to what it will be one day. - Can you share any of the data that you've gotten from brex about just like what tokens spend just like in the economy? - It's increasing. (laughing) - No, look, I think two things are surprising. One is, I think to your point earlier, on how do we look at token maxing? I do think there's such a thing as how much cost boundaries you create internally dictate token consumption obviously. But to me, I think what's the most fascinating is, when you look into the sort of 10-hour rate is written now, and maybe you include New York, tons of token consumption, and you could probably argue, and we see in the data also a faster revenue growth. I think what's really interesting is the gap between anyone in these two 10-hour radius and everything else. And this is like not small companies, you look into like very large companies of very large budgets, and that could be token maxing. And the economic thing for them to do would be to token max, and they spend like, I don't know, 10,000 a month. And you're like, you should probably spend 10 times more, or 20 times more, or 100 times more. That's still surprising. And I think the reason is again, sort of similar to the point at the beginning, like we did this exercise two and a half years ago where I sat down with a lot of the engineering product leaders in the company, and we had this question, which is if it started Brexit again in 2024, the answer would be even more different now, what would we do differently? And turns out like everything. And we started going down this route, and it's like, it's kind of maddening because we're like, okay, we have this completely old way of even thinking about the fabric of the company, and we build the product, and we build our processes internally. The first best answer is, yes, we should have started now. Second best answer is like, let's go do something about it, and change what we do things, right? And I think a lot of our approach in terms of adopting I has also been, you know, how do you pause and say, okay, like there is a discontinuity in the, not just in how we solve the problem, but on what the definition of the problem actually even is, and sort of take a step back and rethink it. And you know, like there's, there's like millions of examples of that, but you know, one example, which is kind of funny, is, you know, we're redesigning our KYC process. Like whenever we onboard a customer, we have to do all these checks to KYC the customer. And KYC historically is something that you can automate, like 80% of it, 20% is manual. And of course, the original impetus for anyone is that it's really an agent that does it. Yes, we can go do that. But what we decided to do is actually say, let's redesign the entire process into it. And then what we redesigned is the entire onboarding process. And when you redesign the entire onboarding process, what you realize is there's a very important thing that happens in the beginning of the funnel, which is the old qualification. Like is this customer even remotely qualified to be a Brex customer? But when you have KYC for free, you can, you can KYC lead versus a customer. So you start to have risk orientation up in your funnel and that changes who you even target, because you know who's gonna qualify and the same thing should for credit to some degree. So now the bounds of the problem have changed. And you can go in and say, and I think a lot of, including a lot of our competitors had this approach of saying, oh, I have this entire old process. Let me go and like latch on a ion top of it or latch on a ion top of our product. And I think the biggest is continuity in a positive way that we've had, where when we said, hey, let's keep this old way here, keep putting it in a corner and like, how would we design it if we sort of the company to get it from scratch? And then just doing that. It takes a little bit of founder energy to do that, but I think it's the only thing we've seen working to really sort of inflect. - I think that reminds me a lot about this is sort of a way back. I don't know if you ever tried to compile arc distributions of Linux. The culture within power uses of arc Linux versus Ubuntu is very different. - Very different. - I think that Ubuntu people kind of feel more like people that try chat GPT. Stuff kind of just works out of the box or some stuff that you can get up and running. There's still not a lot of people that use Linux, by the way, which I think it feels where AI is. But with arc, you're like super hardcore. And I think that's what open claw and hermies feel like, because you have to really customize it to your own use case, maintain your skills, have all the mark downs. And if you get it working, you can build something awesome. One of the most impressive things I've seen people build with arc is actually I don't know if you know a valve, the steam engine, the operating system that runs, that makes it feel like an Nintendo Switch, is actually built on top of arc. - Oh, interesting. - They customize all the drivers over the air updates. It works with all consoles, it works with all sorts of hardware out of the box, but they super duper customize it. And I think this is kind of what's happening. If you get your open claw to work really well for you, you could kind of build your own custom Nintendo Switch for whatever you need to do. - I always have this thing that I tell people, which is funny, which is, think about your time two years ago. Like, I feel like you're working a lot more now than two years ago, right? And probably same for everybody here. So then the argument is, what about the productivity? Where's the productivity? Right? And I was talking to this, say I feel a very large public company this week. And she was telling me that we see all the silicon consumption and we're trying to measure like, like, product velocity. And we're seeing like more lines of code pushed. So yes, maybe that's the way to measure the ROI, but is it really there because people are spending so much on tokens? And I think this analysis, yes, of course, I think having a sense on ROI on tokens is important. But I think it misses the point that you're standing in the timeline of history and it's six months after electricity was invented. Like thinking about, imagine someone saying in like, I don't know, 18, the 1800s, like, oh my electricity bill is so high now, like gosh, let's use a little less. Let's keep this push this team engine to come like maybe 20 years later because the cost savings, like, yes, of course, like don't bankrupt your company and tokens. It's actually a perfect analogy because I don't know if you know this, but when electricity was first invented, it didn't work very well. And the ROI was actually bad. And so if shortly after the invention of electricity, some of the accountants had done this analysis, they would have been like, this electricity thing is like, is it never going to be a thing? The ROI sucks. Why do people stick to it? And it wasn't the cost savings. It was just because people were curious about it. And I think the point of like, why, you know, like I was yesterday until 2am playing with slash workflows and Open4.8 and all that, is because I think of a unique exact same thing if I wasn't making any money because you just see the possibilities and you see what it can do to technology. And that just drives people to behave differently. And I think that to me is the ultimate lip mistest. And it's a good separator. And sure, if tokens are so expensive, they're going to be, I think, over to four and it's a time probably free. If it project is, I don't know, 100 years online, almost compared to what electricity now, we don't think of electricity costs in our day to days, but unless we're in a data center. But I think there's something similar for sure. We talked to a lot of founders of later stage companies who wish that their companies could be like, as AI-pilled as possible. And you run this big company now with all of these employees. And that's only the brexit. There's also the, like, how do I want to say? I'm curious, would you've done to bring the rest of the company along with you on this journey and if you have advice for other people? Well, I think there's a lot to do. Like, it's not an engineering team thing. It's not like a pride team thing. It's like, you have to understand the bounds of a technology better than anyone. I would argue that unless you really experience the limits of technology every day, I think it's really hard to even understand what it can possibly do. Oh, you know why? It's because nobody can say no to the CEO, except the board and the board won't be in the weeds per se. That is 100% true. When you go think about, like, you know, the whole example of KYC that we were saying. Like, the KYC team would never think, of using the KYC technology to score a lead. The only people that can think about the organization of the system itself is if you have the context of the whole. And to me, like the single most important question that any CO needs to answer is forget about the competitive landscape. Imagine you could get the state of the technology today and transports to the moment you started your company. The opportunity was still the same, but just the possibilities of the way to build a company are totally different. How would you do it? And then dip this versus what you have and then for suffer in silence for a little bit because you will. I mean, I do every day. But then the second thing is, okay, what do you do about it? And how would you do it if you were starting from scratch? You would be the one figuring out, okay, how do we design our onboarding process or how we design our growth engine and our customer acquisition and the way we talk to users and the way you synthesize that data. All of that would be redesigned from scratch. So I think it's almost like you have to sort of re-found the very concept of what the company self identity is and the way the functions and people's sense of success get structured. AI is a number all that I think has like three things. The way we talk about it internally, there's product AI, the product actually ships customers. There's operational AI, which is things that directly affect our ability to serve customers at scale, like think of customer success, risk onboarding operations, etc. And then there's corporate AI, which is how people work internally. The three agendas matter and they matter in different ways depending on the timing of the company. And I think people sometimes sort of pigeonhole themselves in one of the three. But in reality, I think you have to take a step back and be like, you know, the same thing we were talking about earlier, like, why can't you solve everything away? At a limit, that's the question. And then sort of start from there and sort of problem solve around that question. It's a turnaround almost. I think you have to assume that if you're a big large company, design AI native, you're doing a turnaround to some degree. I guess we've been making fun of Foxconn factories for some time. But on the other hand, like if you look at them, they're like this paragon of like very extreme efficiency. Yeah. But they also are designed to be that to like create like one thing perfectly back to back to back. And so you have to build a factory like that. And most companies are designed that way, right? I think like processes are designed not to change. Yeah. There is a certain amount of broken glass required. The question is how, like, I think it's 10X easier for the CO2 brake last than an executive. Oh, and 10X easier for an executive than an employee. So you know, a lot of times like someone comes to me and says, I'm trying to do this if I have a somebody saying no because we haven't tested this in this use case or in that thing. And I'm like, okay, what are trying to do? Like, do you understand the risks or understand the guard rails? Yes. Okay. It takes me literally 10 seconds to solve that problem. And it would take someone 10 hours. And then I would say, I think it's going to be a lot easier for me to go into the meetings and escalate and understand, okay, we do it as if I, or maybe never, and I think the conclusion is probably never because most people would say, you know what, I'm just going to like build this soft, this product in the old way because like why wouldn't we just work so we know it's that guy's going to hate me. And then I have to look at that person in the lunch line every day. And it's like, yeah, I want people to be happy and like me. So I'm just not going to do that. And what I tell people is I think I think the escalation paths need to be like the monetized in the system because the company puts antibodies against any sort of, you know, disturbance to the social cohesion of the company typically gets like rejected by the antibodies. And I think making escalations faster and being like, hey, we're going to go try this thing. You know, I understand the risks. Let's take this risk because the biggest risk is not taking that. It is, it's just literally missing the opportunity to rethink a problem from what would you do if you started the company today? From the corporate AI sort of leg of that stored specifically like do you buy into the like the Jack Dorsey view of every company is essentially trying to like build its own little company, a G I do, but but maybe in a slightly different way. I do think the main specificity matters. So I don't believe in the like, well, I'm going to have like a single company model that has like every piece of data like in a single like with no judgment or lens into anything. So and the way I think about it more is like is more the sort of the virtual employee analogy sort of speak, which is like, how do I build an agent or virtual employee that is exceptional at understanding everything that matters about this customer? That is a well defined problem with clear boundaries of like, like, APIs, who people who need, who depends on the data, who interacted the data that is self contained. Then there's another, another agent that can be, okay, given all the customers that we have and the problems they have, how do I manage my product world map? That can be a separate agent, but that that builds on top of this customer world. Like a virtual exact team basically. Exactly. Functional and domain knowledge still matter, right? These things are not going to go away. And I think the way knowledge is structured, I think, still true, right? That doesn't necessarily change that much. And you should separate the agent that and the systems that are actually emitting code from the system that is talking to customers and the system that is reasoning about the conversations of customers and translating into product world map. These three three separate things. We're kind of like a Tesla for AI. We're like, I don't believe in anything that doesn't have real usage. So it's like, yeah, I build this great model and I'm like, okay, how many of you are using it? Is it actually displacing the need to hire a person inside a company? Is it actually displacing the need to, you know, spend literally hours, like how many hours is this thing saving? And I think a lot of times people say, well, you know, it's a cool model. And I'm like, yeah, but like, that's not, that's not going to cut it, right? Once you have that orientation, I think customer world model, okay, like your, or, for example, our client sales team now runs on our customer world model. So I know it works. I'm actually having a lunch of a customer tomorrow. And I don't know the state of that account as well as I probably should customer world model answer the question for me. And I now have a report with including things that the team didn't know about that came through support tickets and, you know, an executive that was traveling had an issue at an airport with their cart. So all these information awareness, toward information awareness, right? That is a well defined problem. That is working. So I think the rest is building block as part of my company model as a whole. And you can have evals on it. Like we know like, you know, I think a very, we should talk about evals. There's a bunch of learnings on this and how to build evals into the fabric of the company. But, but anyway, I think it's more of like, you have to decompose a problem a little bit. Yeah, my favorite thing about evals is just running cross-modal evals against each other. So one of the things that we're doing that I, I, I, it is related, but I think it's really fun, which is how do you have every single human interaction in the company becoming an eval when you have any agents? So for example, we have the onboarding, onboarding agents doing something. And then you have a team that actually goes in and looks at KYC exceptions that the model can't figure out how to make that a breaking change. And, and okay, like this, this manual interaction will become any eval case, you know, we have an expensation and, and, and Brex. Whenever someone has a conversation with the agents that is the flags an issue or, or a bug or, or, or, or, or something that feels like the conversation didn't go as smoothly, that creates a bug, that bug triggers the nation that it's going to go and modify the code base and the prompts and everything to make that eval pass. And if that doesn't break, then engineers going to go in and figure out how to make that pass. Because the goal at the end, I think, is to make the whole thing a self, being a self learning system, right? And I think the, a lot of what I see with, with companies is they spend a lot of time getting an Asian working, but never thinking how to make the agent improve every day. And I think that's like always the biggest unlock. You need a dream cycle. Exactly. Exactly. Exactly. Everything, every night. Exactly. And it's like, oh, what's going on there? I need to put this over here. Exactly. What actually happened? Is there a pattern? How do I cause this? So how do, how do, how do I bake the dream cycle into the products and into the agent into the things that ship? Right now is I'm building like three or four agents for my friends. Oh, interesting. And some of it is like, this is a user research for me for G-Brain, because it's like, I have one. It's working really well. I have three or 50,000 markdown pages in there now. What a crazy, like, I thought it was this wild, you know, pie in the sky thing. And it's like, it's going to happen in our lifetimes. You know, I remember when, when Neuralink came out, and I used to think about it, I was like, I don't get it. I was like, yeah, of course I get it conceptually, but why is it a thing? And then now you say, I, and you're like, yeah, it makes sense. Yeah. I'm the bottleneck. Yeah, yeah. Typing is so slow. I don't know if you use a lot of addictation. I use a lot. My most used developer, developer UI right now is like voice memos to open clock. I said this before, like, it was maybe accidental, but I actually just really love the fact that like telegram works so well with, um, audio. Yeah. Because it's forced me to just put more stuff, like make the agent more intelligent so that you can do more stuff via voice memos. Because you have to sort of fight the natural instinct as like a traditional developer where you're like, oh, like, I can't quite do this. It doesn't do this. I need to go like build more client or more UI or like more functionality for it. Foxconn. Yeah. Exactly. We take a note. Yeah. Just let it do what it wants to do. Yeah. Give it some context and it'll just think about, you know, oh, like actually, what about this? What about the work to your point is the, how do you organize the context for the model? And you can use the model to help, but that is the bottleneck for most things. Once you have the context in there, it's actually you can do some pretty crazy stuff like my favorite new feature. I saw your brain LSD. Yeah. Latteral synactic drift. So you just bump the temperature on the surface? It's not just that. So you have the vectors, right? And so, you know, if you think about what conventional ideas are, like most people give you, like, oh, well, this idea with this idea and it's like kind of like in this cone, LSD mode actually says you can. cannot combine concepts that are within this cone. They actually must be orthogonal or just like, like seemingly random. And then it'll try like, you know, randomly hundreds of these combinations. And then it'll rank or to them into the ones that are actually the most coherent. And then if you do like a hundred of them, actually like the top five tend to be banger tweets. - You know what's crazy is like, I didn't tell Alfred X-Prisci to be dry. I went, I actually had like chat GPT generally like the sole file. It was like based on like everything you know about me, all the interactions here, like generate like a sole MD for like my open core agent. And it was so unerringly like accurate. (laughing) About like kind of what I would want from like an agent. And it was like, oh, damn, these models know a lot about us. - My open clock got really interesting once I, I just ingested my 60 gig Google takeout. I mean, you have to write a bunch of high coup code to like only get the emails that are actually real. But you know, there's like, extract like 4,000 emails out of 60 gigs, actually matter. But like those are like, oh, actually like a lot of your thinking and the consequential moments of your life. So Pedro, thank you so much for being with us. I mean, you're by far one of the most AI-pilled, farthest out on the edge, but also very practical CEOs who is, you know, playing with this stuff and actually building it yourself. What would you say to people watching who are founders, who want to be founders? You know, I think that you are sort of the model for the way people should start companies and run them with AI as your SLN buddy. - I really can't stop thinking about the electricity analogy, which is, you know, you're standing, there's a 200 year timeline of human history. There's a point in time where electricity was invented. It sucked in the beginning, your six months after that point. What do you do differently? Knowing everything that would be true about electricity. Knowing that data standards would consume electricity and anybody, right? Well, you do a lot of things differently, I think. So I think that's one of just marveling at the possibility of the exact moment in time we're now. I think the second is like, I have a post it on your computer, which is you wake up whatever problem you have in your life. Why can't you solve it if AI? And 80%, yeah, you can use a chatbot, but a 20% that you can't figure out why and go build something that makes you solve that problem. Less so because of the immediate usefulness that solving that thing at scale will have, because it gives you a texture and a feel for the possibilities of technology, which are really hard if you're not playing with it every day. And maybe the third thing is like, I think it's like just measure your total consumption and how much you're just pushing the limits of the company and starting with the premise of like, okay, why can't it just be one person? Like, why can't it just be me that builds the whole thing? And you're going to probably face a wall of the elements of what models can and cannot do. But at a limit, I think the question is, how do you spend your time on the things that only you can do as a founder? And these things to me are number one, which problems are worth solving? And two, and sort of the choice thing we talked about. And the second thing is, okay, given that the, given these choices, what are the limitations of an LM that they still cannot do? And I have to go in and do those things myself. But almost, to some degree, you're working for the LM, to some point. And if you're in a bigger company, you're gonna turn around and put the LM as almost a founder and a CEO, and you're almost architecting the entire company around that idea. But I think early on, so much of it is, choosing what matters, talking to customers, injecting the signal that the models don't have, and just rebuilding it the way you would do it in 2026 with electricity being six months old. Thanks, Pedro, this is awesome. - Yeah, thanks for having me. Appreciate it. (upbeat music)