The Next Wave of AI Companies in Latin America | Tako Founder on The a16z Show

0:00 If you look at the Americas, you find that the US is the largest economy in the world. It has 300 what 30 million people. If you look at Latin America, we're talking about 650 million people, third largest economic block. We have engineering talent, GPUs, we have abundant energy, commodities, biodiversity. It's so hard to do business there. The Brazil cost. Some people say it's 10 20%. You're talking about some of the most complex regulatory frameworks that exist in the planet. I believe that if you build a system that allows companies to deal with all of this complexity, companies are going to be able to do things that were unheard of before.

0:38 We're here today with Sebas, uh, serial entrepreneur from Latin America. In 2015, Sebas started a company called Rappi, who became the first, uh, delivery app in Latin America and the first super app, and later became the first Colombian unicorn. Um, after nearly a decade of scaling that business into nine countries, 250 cities, Sebas stepped away to start Tappsi. And we're going to talk today about Sebas's background growing up in Colombia and traveling the world before becoming an entrepreneur, building Rappi and Tappsi, and how now is the moment that we're going to start to see more global AI winners coming out of Latin America. Sebas, welcome to the a16z show. Happy to be here. Thank you. So, Sebas, uh, you grew up in Cali, Colombia. Both of your parents were entrepreneurs.

1:28 You want to tell us a little bit more about your journey having grown up in in Cali and how that shaped you in into the business person that you are now? Mother, very very entrepreneurial. Um, what I saw in her was the intersection of of culture and business. My mother ended up founding a um, you can think of us as Cirque du Soleil blended with salsa. Um, Cali was always a very culturally rich city, especially around music.

2:00 So, I think that's that's that's very very pretty much part of who we are. Um, on the other hand, my dad was, um, an also very entrepreneurial, uh, character. More of the type of starting one thing, moving on to the next with maybe maybe little little little structure. So, I was always exposed to that. Um, I have lots of memories of of um, just spending time with them as they were building their companies or or or pursuing any any business. My parents got divorced when I was five, so I grew up in a very independent way. So, just imagine you go to one's house, you go to the other, very flexible, not a not a a very rigid uh, structure.

2:45 And and I think, um, what I saw was a deeply creative culture. Um, we kind of have to because we also grew with a lot of, uh, um, tough circumstances around us. Um, it's not necessarily the most affluent place. So, so I think the the way to get ahead is just by the application of resilience, creativity, uh, and being able to do things that that the common person is maybe not pursuing. So, so that was pretty much my my my upbringing. And then, um, I ended up doing something that changed my life a lot and and and and it's it's it's the beginning of of a path to becoming an immigrant. Um, and and I I went to to Barcelona when I was 17 18 years old.

3:37 Um, that was a um, a big change and and at the same time was it was it was And what led you to Barcelona? My father was living in Europe at the time and I was traveling often even when I was in in in high school. So, um, my dad was the type of person that would say in the middle of school, like, "Hey, we're going for a trip." I was like, "Okay, where are we going?" "I don't know. We're going to remember a trip to Rio in in Brazil."

4:07 I was like, "Dad, but I have classes." And he's like, "Yeah, but you're going to learn more with me than in a classroom." So, so that was quite common and and there were trips to Europe and we ended up pursuing something that wasn't that common at the time. Like, I would say most of my my my friends ended up staying or others, the few that left came to the US. So, going to Europe wasn't necessarily like a like a thing.

4:32 Um, but for me it was very interesting also because it was a it was a decent school, but more than that, I think what Europe gives you is access to to the world because um, you're broke, you have little little little money, but you always have 50 euros to get on a plane and just go to Berlin for a weekend or Paris or Istanbul or whatever. And that's just a great way, I think, to to to learn.

4:58 So, so that was that was very very important. I think my life as a as a as a person of the world started started then. Awesome. And, um, you know, after Spain, you decided to come to arguably the greatest city in the world, New York City, where we are now. But, um, you want to talk a little bit about that experience coming to New York and, you know, after New York, you decided to go back to Latin America and start a company, which is not usually like the common path. And now you have started two companies in Latin America.

5:35 So, you know, we have this thesis that like having exposure to the rest of the of the world obviously makes you a great entrepreneur. You want to expand a little bit on that as well and how that shaped you into building the companies that you're building now in Latin America. There's more to the story, so so I'm going to have to to to maybe expand on that on the first chapters because as as as I arrived to Barcelona, um, two years in, I lost my father. And and that was a um, quite the the crucible moment for me because not only you're dealing with the the personal, um, drama and family drama that that entails, but also you have to stand up for yourself. And and it was a moment where we just didn't have the money to just continue to pursue what I was doing. And I ended up going to Madrid because uh, the only thing that I could afford was public school and I didn't want to do public school in Barcelona because I didn't want to do classes in Catalan.

6:32 So, I grabbed my bags, I went to Madrid, landed in Madrid, and this was in the middle as I was dealing with my father's passing. Um, so it was a pretty pretty tough moment. And as soon as I I landed in Madrid, I started what I would say is my first company. I was like 20. Um, and I I I I call it the the the the Spotify without the internet because uh, in essence, what I did is it was the You guys remember the 1,000 songs in your pocket? You know, classic, one of the best marketing slogans ever.

7:04 And when I saw that, I always loved music, but when I saw that, I said, "Well, like, the consumption of music just fundamentally changed. It's not going to be physical devices any longer. It's going to be databases." So, my great idea was I'm going to get my hands on as many of these iPods as possible and I'm just going to fill them with tailored databases depending on the business. So, you can be a bar, you can be a restaurant, you can be a business, you can be an individual. And I would basically mark up the price of the of [laughter] the of the iPod. So, I have like a like a a team of five people.

7:38 Imagine a room with five or seven computers, all of them connected to the internet, Napster equivalents, and just like download the download the music and and and send it to to to people. So, um, that was the way that I was able to stay in Madrid uh, with that small business. Um, when I when I finished school, the world was uh, a bit upside down because it was the great financial crisis and I didn't want to stay in in Spain. I wanted to be in a place that was more dynamic. So, where else to come than New York? Uh, but as all great entrepreneurs do, you come with a complete misunderstanding of the challenges and the risks and and how hard it's going to be, which is one of the the blessings of being a optimistic founder. And I landed in New York in in winter 2009 um, with very little savings. Um, but it it then started five six amazing years in in New York where I continued to build companies. I met my my my wife, found my first investors, uh, built a very global company. Um, and that global company, um, what we were doing is just enabling supermarkets to sell groceries.

8:59 One of my clients was Reliance Industries in India. And I ended up spending one and a half, like, roughly five to six months in in in Mumbai. And and that for me was very inspirational because I got to see how technology could be applied in large emerging markets. And I I heard things I had never heard before, which was, "Listen, we're going to connect a billion people to the internet through smartphones. We're going to bypass the cable. We're going to just we're we're going to bypass the traditional internet stack.

9:28 We're going to go directly to mobile." Um, and that was so inspirational for me because I saw how it was being built. I saw how they were bringing 757 planes full of smartphones from Korea and China because there were no smartphones in in India. So, I think that that allowed me to see the possibilities of Rappi. And I think that's when I started to to think, "Well, what could happen if we apply technology to the emerging world?" So, then we start going more into the into the into the Rappi chapter. Yeah. And so, anybody in Latin America had like 100% name recognition has heard of Rappi, but in case we have any listeners not aware of Rappi, can you describe what is the company and maybe a sense of the scale that it's at now?

10:15 It's the operating system for local commerce across Latin America. Uh operating across nine cities. Uh sorry, nine countries, 100 cities. Um a product that allows you to have groceries, restaurants, pharmacies, liquors delivered to you in minutes, in some cases under 10 minutes. Um because we do picking from restaurants, we do picking from groceries, but we also do our own delivery. Um it's one of the most uh special stories of resilience and entrepreneurship that has come from from the region. Um it's been the place where more than 100 companies, 150 companies have been founded by fellow Rappi alumni. Um it's a company that has impacted the life of the lives of millions of people uh that are thanks to Rappi earning multiples of their minimum wage in a flexible way.

11:09 Um it's a company that um also has uh a very mature operating system to be able to to to run a business that is incredibly hard and complex. Uh it's multi-vertical, it's low margin, it's competitive, so that basically made us like very battle-hardened operators. Um it's a company that has a very very special group of people and talents uh working in in the company. Some of them coming from the best companies in the world. Um A company that is is still founder-led today by by my two co-founders, Felipe Felipe and Simon, who who are leading the company with incredible love for the mission and and the impact and and it's amazing to see them operate and I'm still super close. I'm just helping in a in a in a different in a different way. I left the company around a year ago. Of course. Um and I think it would be it'd be interesting actually to talk a little bit about the founding story of Rappi, especially cuz it really it relates a little bit to how you're navigating what you're doing now. But Rappi, I think was iteration three. You had you had two companies before, and one of the things we often talk about is how sometimes the big ends up being an infrastructure problem of an idea before. And in this case, the problem was logistics. Like you've now created like you've very complex logistics business. So, maybe talk a little bit about the the discovery and how you guys landed on the idea. Simon and I uh had built a previous company that had brought me to to India. And and that company was uh in essence um a mobile experience integrated with uh supermarkets or a retailers uh order management system. So, we we started solving the problem from a more customer-facing experience, but we didn't want to do the fulfillment piece.

13:04 We didn't want to touch the the moving of the atoms. Maybe 2 years in into building the company, we started realizing that if we wanted to really change the customer experience and offer something that was orders of magnitude better, we needed to own the whole uh operation. We needed to have the mobile experience, the front end, the payments, the fraud, and specially the logistics. And and and and the connection with the local content or or or inventory.

13:38 And and I think Rappi Rappi was born of um the possibilities of creating a novel experience, but it was also born out of frustration because we're seeing these large incumbents that just couldn't think away from bringing people to their brick-and-mortar stores, and and it was just very hard to change that. So, so Rappi's first delivery was around October of 2015. We initially launched uh one market, which was Bogota in Colombia.

14:10 Maybe 2 3 months in, we went to Mexico. And that was the the classic investor telling you you shouldn't go international too early. Yep. >> We didn't listen at all to that. And and in partly we didn't listen to because I think we wanted to build an important company, and and we understood that being in the right markets mattered, and being in the markets that had scale mattered. So, so we did something that was probably not the most common thing to do.

14:39 And and then um it was a um um effort of understanding what is the operating system of a company like this. How do you At its core is a is a optimization problem where you need to have perfect balance in of customers, couriers, merchants in every micro-zone, and you can almost kind of operate these micro-zones in a very independent way because otherwise the the the the the flywheel breaks.

15:09 If you don't have enough couriers or you have a lot of customers, the experience is not great. If you have a lot of couriers and not not enough customers, it's not great for the for the courier. So, so so we we went in that direction, and and and we made sure that we were um pretty much uh obsessed with with working with builders, working with founder profiles, um being very very close to the audiences. Uh we have a lot of a lot of stories of of of just how hard it is to build this this company in every city, in every country. What do you think like this would be I think a very complicated business to build anywhere. Was there anything about being in Latin America specifically or, you know, Brazil or Mexico or Colombia that like specifically made it more challenging or different?

16:00 Uh where do we start? I mean, there there's there's just no existing stack that you can build upon. There's no as a service. >> No, there's as a service of nothing. So, so [laughter] You are the service. >> Yeah, so so I mean, we could argue that that Amazon utilized a lot of FedExes and UPS infrastructure, right? We we didn't have that. So, so it applies to logistics. Um Payments was another one. Um there's a fairly It's depends on the market, and some markets more than others, but payments are still cash-driven.

16:35 So, we had to build our our our capa- our capabilities to actually accept cash, right? Or or fraud, right? Like you can work with any fraud company, all of the sudden you see these weird behaviors, and and and and they were like just not being able to spot it. So, we had to build our own fraud uh uh systems. And on and on and on and on, you know? So, so on on in perspective, I think doing the hard things and and and owning as much as possible of of of of your operation and going vertical, it's hard, but it's also very defensible.

17:11 You know? So, so it's it's it's hard to hard to compete against that. It allows you to provide much better customer experiences, but it also it also makes it really really hard for someone else to do. So, so I think durability and defensibility are two things that I think constantly about. We can we can expand on that, but but but it has a it has a silver silver lining, and if you look today, Rappi is the only independent, founder-led company in this category in the region. And and we're talking about the third largest economic block in the world. And and that's uh I think that's a a special special place to be. To save us, um you left Rappi about a year ago.

17:53 Um and then you started a second company uh in the region. And, you know, you want to talk us a little bit more about this company. Um what does the company does, and why did you pick pick this this problem specifically? Tako um what we're building is um an intelligence layer for companies to run their workforce operations. What that means is that we build um agents that understand all of the labor regulation.

18:27 And we build applications to be able to support all of the operations that are related to just running running and managing people. Um I'll give you I'll give you examples of what that means. So, we have uh we build this uh product where I think the only company in the world that has built a complete data set that has all of the corpus of labor of Brazil in the world. May- maybe give us a a sense for the non-Brazilians of this like complexity of the labor >> it in to put it in context, uh some people say that the the the Brazil cost is a fraction of their GDP. What that means is that some people say it's 10 20%. I I I don't know I don't know exactly of the total GDP of the country, but that means is that it's so hard to do business there um that it partially explains why um you have a country with incredible talent and incredible stack, a super forward-thinking regulatory body, uh very talented people that have built companies, but at the same time you see an economy that's still sluggish and doesn't grow. And if you bring that closer to the space where Tako operates, that's we're talking about taxes and labor, um you easily go into the 20 30 billion dollars of of of cost.

19:48 And and and that's because you're talking about some of the most complex regulatory frameworks that exist in the planet. Brazil is definitely top top three. Some people say it's the most complex country in the world. And and if you start understanding what this means on labor, you have tens of regulatory frameworks from the constitution to what's decided by a judge at a at a regional level that contradict each other, that change often and that makes it incredibly hard for companies to to operate there today.

20:22 It's a very hostile environment for businesses. So I believe that if you build a system that allows companies to deal with all of this complexity you are completely changing what companies can do. Not by a 10 20 30% I'm going to make you better. No, companies are going to be able to do things that were unheard of before. And I think now AI allows you to just take over a lot of the work that is in essence low value services that are provided by a lot of a lot of people that are leaving from this let's say not of complexity.

20:56 So so we took on on on something that was quite complex. We took on something that I find it fascinating because I think complexity and doing things that are hard are great places to find durability and defensibility. And what that means is that all of these all of these data sets now exist within our system. We we built the the the yeah, it's it's the richest the only data set that exists. >> the regulation, turned it into code.

21:28 Updating it all the time, making sure that you have the data pipelines, making sure that you build a proper taxonomy to understand how the laws interact with each other, making sure that this is easily readable by by by agents. So we have one product that is called Auto and Auto in essence is for those that understand and think of it as open evidence for but for labor, right? Auto not only understands any law with benchmarks that are much better than your general LLM.

21:57 But it also helps you as a company understand if you're staying compliant. Right? And and and staying compliant is is is is incredibly important because if you don't, you're talking about people's money, you're talking about additional payments, you're talking about fines. So so so that product is is one of them. We also we also build a system of agents that basically allow companies to recruit, to onboard, to pay to send all of the events to government, to integrate with your accounting system to let go of an employee and all of this is basically agents that are taking over the work that is today being done in a in a very low low low quality way. So so our customers are medium sized companies but also some of the largest companies in in in Brazil.

22:52 Talk to us a little bit. I think you know, one of the things that's very differentiated about you is that it's almost as if you think about talent first and you know, everything else second. You focus about being surrounded with the best people and we seen this in Brazil how you recruit people in Brazil but now recently, you know, Taco has opened a hacker house in in Silicon Valley and I think that that's pretty innovative for the region. You don't see a lot of companies that like think that way and make the effort to kind of build a brand to recruit the best people in Silicon Valley to to work for it. So maybe you want to expand a little bit about how you think about talent and I know that you have great anecdotes on on on recruiting people. So it'll be great for the founders in the audience to to understand what it takes. You recruited a lot of people from DoorDash for for Rappi and I think that a lot of those learnings show in Taco now but just curious yeah, if you can share a little bit more about how you think about talent.

23:52 I think it comes from my my personal experience also as a as a as an immigrant and having built in in Europe or having lived and built in New York. I think when when when you see those things, you start um understanding that great comes from different places. But you stop thinking that what is greatest is in essence what you know because you come from a specific place, right? You're not you're not necessarily thinking about well, this is locally great. No, you immediately go to globally great because that's what it take what it takes.

24:31 So I've always been very very interested about identifying who these people are and very interested in identifying talent that is overlooked that is probably not the the the the standard looking person. I interview in a in a in a particular way. Maybe some people do it but mine is is really trying to understand people's childhoods, origins mommy and daddy stories, resilience stories, hardship stories and I go very early. I go to childhood.

25:08 I go to teenage years. I think there's just a lot of signal by understanding a person. And I also think that when you're hiring really smart people, they're they're very good at selling themselves and and very good at saying all the right things that you want to hear. So I discount that a little bit more and I I try to focus on on on the person. I have this little rubric that every time that I interview in or or looking for candidates, it looks like a like a like a diamond and I always have it in my head.

25:35 So it's a diamond with four corners. So in in in in one corner, you have IQ, obvious. We want we want smart people. Another corner, we have EQ which is the emotional emotional intelligence. Sure, very important. The other ones are a little bit different. So so one is AQ which is the adversity quotient or resilience. So I spend a lot of time evaluating for that. Then you go to the last corner which is energy. So some people call it agency but but but at the end you want people that are enthusiastic with solving problems that are just able to take things all the way to the end that like to overcome challenges and problems. I think that requires a lot of energy.

26:20 And then at the center there's a heart and I like to work with people that have high integrity, high values and and and I think yeah, high trust. So so I always I always do this and that you take that global part and you take this this this assessment of of talent I think if if if if if you look at building a global company um which is is always my my aspiration you need to have people that are globally great.

26:48 So in in the case of of of Taco for example, Taco has incredible talent. It's a highly talent dense company very very very profound on engineering and what we started realizing is that the team was incredibly talented but we also wanted to connect them with what's the cutting edge of how to apply artificial intelligence for real. So we we have this connection to the US, more concretely San Francisco and and we basically have it's a house where our engineers spend a lot of time. We've recruited people in San Francisco as well.

27:28 And and I think it's been fascinating to see how they they learn and I initially started seeing this at at Rappi where um at one moment I was probably investing 70 80% of my time on on talent and I remember very clearly at Rappi maybe around 2020 21 like we were riding the the the COVID wave. It was wild like all of this demand that came forward. But it was just so hard to operate a company and I remember sitting in a table and looking around our team and I said these guys are incredible.

28:03 Incredibly talented, hard working but the complexity is just insane. So I want to think about where the best people in the world doing these companies. And and I I mean I think DoorDash is a company that I that I deeply admire and I think it's one of the best run companies in the US and in the world. So I started meeting people there and I was incredibly impressed. Um Little did I know that the consequence of that was like probably around 25 people from from DoorDash that [laughter] started working in Impressed and relentless if I remember something like >> Yeah yeah, they they ended up working at at Rappi and and they were incredibly impactful in the company with our team, right? I think our team was able to were able to take this a lot of these lessons and also execute at par with this with this with these individuals and that required a lot of a lot of very intense recruiting. I I There's not a lot of precedents of companies in this part of the world that are able to bring some of the best talent from some of the best companies in the world. Wow, which is maybe one more question on that cuz I know you're still very focused on recruiting now for Taco. Both at Rappi and at Taco, there's a lot of future founders. Is it just you know, future founders or people who want to start their own companies mapped to your rubric or is that another sort of lens that you're you're looking for especially attractive place?

29:29 If a candidate comes to me and says I want to start a company, I love it. And and I I took that deal all day long. I It was basically something like this. Look, I want to be an entrepreneur. I want to start a company. I'll say amazing. Give me three four years of your life. We're going to give you a lot of learnings lessons. You're going to make mistakes with other people's money. It's going to be a great way for you to learn how to be an entrepreneur and I and I think if if I look at Rappi for example, because the business was so hard and I thought about this. I think I think PayPal is a is a great example. It's a hard business, right? And I think that's why you get so many of these like incredibly talented entrepreneurs, right? PayPal being one of the probably the best the best group of entrepreneurs that ever has ever existed.

30:19 And and these folks were just touching real things. I mean, high stakes, like budgets, recruiting, firing, operations, commercial, product, engineering. >> payroll. And now now now now now now payroll and and and tough stuff. So so we we I love that and and and I think those those type of profiles I I I love working with them and and giving them as much of learnings that I can. I have to say also that I I I learn a lot from them too. I Something that we we've done with Gabe and and Gabe and I have been on quite quite a few crazy trips together.

31:00 And and I I take every introduction that Gabe send sends my way and some of them are just very young potential founders and maybe they learn a few things from them but I sure learn a lot from them. So I think over the last years I've spent especially if the last four years I spent a lot of time just being a student again. I think I find that The best trip so far has to be the North Pole.

31:28 Yeah, if you tell me that we were going to be doing Arctic skiing, stuck in a hurricane in the North Pole and then doing Nordic skiing, I wouldn't have I wouldn't have believed it but we we we survived it. Well, awesome. So let's maybe closing the chapter a little bit about Taco and then we can jump a little bit more into opportunity of AI in Brazil and like what what would it take for Latin America to produce not only great local companies but like global winners.

31:58 Um but maybe tell us a little bit what what what is coming up next with Taco. Like what what is happening in the next 12 to 24 months? Taco today has in in in essence four products with four ICPs. And it's the the labor um product where you basically you can understand um like no one can the the law and untangle this super complex system. That product is being used by end customers, by mid-size companies but also by enterprise companies.

32:36 Then we have a a product which is a bunch of agents that are doing end-to-end work for companies. So from recruiting to onboarding to paying processing payroll, sending the payments, doing all of the integrations with accounting, all of the integrations with the government for mid-size companies. We we built another product which is a voice recruiting agent which is again thinking about about defensibility and doing things in the right way, um this voice is also trained on Brazilian regulation. So it's not your just your your voice agent but we take all of the operating principles of the company, the typical interviews and that agent creates the script of the questions that are going to be asked.

33:24 And we we ensure that the the the agent stays compliant with AI laws and non-discrimination laws that are incredibly important in Brazil. Again you don't just do well you get sued, right? So so so it's it's incredibly incredibly important to to to to do it well. This product is typically used by large companies in the tens of thousands of of employees that have a lot of churn and they need to hire a lot of people. And then the fourth product is an enterprise motion where we're going to the largest organizations and we're solving some of their hardest problems.

33:57 Um and and we're starting with the people related problems because we believe that if you crack that and you understand a people and you build a people database that is properly constructed then you can have these people running whatever type of agent. And we build our own agent OS and and and and we can build the agent, select the tool, select the model, do the evals, all of the telemetry, who is using it, who's not using it, the the the whole the whole thing. So I think those those four four big pillars are are are what we're going to be focused on for for for the year and I think um the application of AI in enterprise requires this. It's it's I I I wish that we stop talking about features, stop talking about models, stop talking about agentic disco pilot that um co-friend AI this and I think we should be talking about in essence what the work that the that the agent is doing with a clear indication of of the outcome. And we speak a lot about this at at Taco like you should be able to explain it in layman terms to anyone without using AI.

35:06 Don't talk about AI, don't talk about models, talk about what the the agent is going to be doing for the company in terms of work. And we also talk a lot about Taco in in in in in terms of um building things in an inside-out way. What I mean by that is there's this museum in Paris which is the Centre Pompidou and it's it's a building that is completely hated by all Parisians. They hate it. I I don't hate it but I remember the first time that I saw it I was very impressed with this building because it's a building that is inside-out. It's it's it's the opposite of other buildings. So so so you see the pipes, you see the stairs, you see the air vents, you see the AC and each of them has a individual color, right? I remember red red is for elevators and stairs and transportation. I don't remember exactly what it is.

35:56 But I love that concept because I think if you're going to build an agent, you should be able to see it and to understand it. This is a problem. A lot of these things are black boxy, enterprises don't get it but if you're able to see it and you say, "Look, an agent is a system of software. At this moment is calling this tool, is using this model, is working on this data set. This is this is where it's at at the moment and you see the progress." I think that's a incredible important unlock to be able to apply this in enterprises. So I think Taco has a critical role in making sure that companies don't stay behind in terms of AI. Like they all get it, they all understand that they need to use AI but then you you try to bring it down and they are still very much behind.

36:38 That is a perfect segue to the I'll say AI opportunity real large in LatAm. And one of the things we talk about often, like it is probably the most exciting time to be building like definitely since we've been in technology. And the AI opportunity anywhere is very large because you're not only tackling the software market, you're now tackling the software and the labor market. And if you go to LatAm, one thing I think many people there don't realize is software for a variety of reasons never really took off in a lot of sectors and in a huge way it was just labor was cheaper. Like I remember like time like, "Hey, you should buy my fraud software." I was like, "Why would I do that? I've got 200 people that cost less over there."

37:21 And so I think the opportunity now is you're not replacing sort of software and labor. Like there's all of these industries that are, you know, arguably 15, 20 years behind. Um but just to say that productively, if you look at the large AI companies that have been scaling, they haven't yet come out of LatAm. And so do you think is it just a matter of time? Does something else have to change? Obviously, you know, you're you're building one but like what is it that for the the region to really take advantage given that they skipped the last generation and so this opportunity is twice as big if not more.

37:59 What you say about about labor is so true. I I call it HaaS which is humans as a service. And and that was the the that was the that was that was always an issue and then the issue is even worse because a lot of these HaaS are incredibly poor and then it's a bunch of screens with humans with HaaS running running the products. It's just terrible. >> [gasps] >> It's it's also it's also true what you say that the markets are very large and and if you look at the economies of Latin America, it's they're very service oriented and generally they're not high value added services.

38:33 These are low quality services that are being provided. And and I think there is a lot that AI can do to go after those big markets. So previously we were talking about the size of the problems of uh Brazilian complexity for taxes and labor. That's a classic example. I mean, when when we think about those the size of those problems, those are opportunities for companies that are taking over. Um so your question on what it needs to happen for us to have real quality companies is A being able to serve those markets. So I think the application is going to be incredibly important.

39:13 Um the application of something so complex takes a little longer. Um I I I go back a lot to the building of high quality businesses. Um and I was I was studying this this this this book that I that I like a lot. It's called Engines That Build Markets by Alistair Nine if I'm mentioning his last name correctly. But the the book basically just covers the last 300 years of technology revolutions, right? And you go back to the railroads, and you go to mainframe computers, personal computers, everything in between, mobile, the internet, cloud, on prem, the whole thing.

39:52 And at the end of the book, he basically talks a lot about the the characteristics that he found, right? And it's it's it's great to see because you see some patterns, right? So, there were people that were hyping it. There were people that are were under hyping it. There were people that were saying that if you go on a train, your lungs are going to explode, and the internet was not going to come to anything, and and there's some people talking about it in different ways, but my favorite one is that the companies that really created value in each of those technology revolutions were companies that were the very durable, and in many cases they had some monopolistic, it's a tough word, but I'm going to use it, dynamics. And in essence, it it mean it meant that the companies were big, but they also had durability, and durability is a function of defensibility.

40:38 And it could take many shapes. It could be network effects, it could be scale, it could be switching cost. Um it could be that you're doing very hard IP, that you build a great brand. All of those things are I think very important. So, going back, if you're doing if you want to set serve this very heavy services sectors, like in our case, it's very complicated. And you need to do it well, and you're basically taking over people's liabilities, if you think about it.

41:10 I'm absorbing all of that liability because the company is saying that you're going to manage this for me, and if you I don't manage it properly, I'm going to be on the hook. And I love that. I can take it all day because I understand the risk, and I understand the risk because I built the product, and I have the systems and AI that understands the risk, right? So, so I think we're going to be seeing companies that do that very well, and that are going to have scale, and they're going to have defensibility.

41:34 I also think we we need to continue to invest in being as sovereign as possible in terms of the AI stack. Uh and what I mean by that is um I would love to see more AI clusters. I would would love to see more GPUs that are hosted there. I would love to see more even even integration of American companies in the stack. It's it's another topic, but I'm a little I'm a little worried about the the the US not being closer and have a lot of Chinese uh companies taking over the stack, whether it's cloud, whether it's telco, whether it's applications of AI, whether it's models. I think it's really important that this stays in the in the in the American hemisphere. So, I would love to see more of that. Um I think we need to be doing more in terms of research.

42:24 We're doing that in in in in Tago now, where we're doing fine-tuning of our own models. I think we need to have a point of view on that because um again, defensibility. I think you cannot just be relying on someone else's stack, especially when you're talking about very highly important information, right? You need to have your walled gardens, as you guys say, right? So, so so I think that also needs to happen, and then you need to have teams and founders and engineers that are really really ahead, uh studying all the time. A lot of the the knowledge in AI is is on papers, and it's out there in X, but a lot of it is very tribal.

42:59 And you need to be learning all the time, and you be talking to the companies that are applying AI in the right way. Um so, so I think if we if we see all of that, we're going to have uh great companies being built then. I I think it's I think it's it's going to start to happen, and I wouldn't rule out that these companies are going to go beyond Brazil. I mean, if Tago solves these problems in Brazil, I think the other markets are going to be They definitely seem easier. Yeah, hard, but but I I I think they can be they can be doable. Yeah.

43:31 Earlier you mentioned, you know, how Brazil is ahead in certain frameworks, or especially in the regulatory side. Um you know, the Central Bank is an institution that we have spent a lot of time studying, and it's kind of impressive. It's very very impressive what they what they've done in Brazil, and how that drove uh fintech innovation in in the country. Um you know, one of the questions that we have is that when people abroad think of Brazil, they still still think of Brazil as an emerging market. But um what do you what would you say to to the people listening on what is it that they're missing about Brazil?

44:09 I think um the most misunderstood innovator in Brazil is the Central Bank. It's it's it's it's incredible, not only on on PIX, which is what's most widely understood, but also on drugs, and how to utilize blockchain, and how to do um regulation the proper way, and and also a lot of people don't understand this, but it's it's the Central Bank of Brazil that is exporting a lot of their uh um a lot of their experience to other countries to have similar systems or payment systems. So, in addition in addition to them being highly technical, and I think we've all we've all met them, and and the former uh Central Central um Central Bank director of Brazil is incredible. I mean, you would talk about technology and cryptography, and it's amazing.

44:58 So, so so that's that's that's that's something that sets Brazil ahead, for sure. Um but I think also Brazil has other pieces of the Brazilian stack. Um I I I dealt with one of them, which is the the the regulatory body that governs competition. Incredibly sophisticated, incredible technocratic, incredible uh how they make decisions, how they think about markets. I was super impressed. Uh and I had a uh several interactions with them back in the day in in in Rappi. I'm seeing it today in in in Tago, for example, in everything that governs relationships with companies and workers. There's a central repository in Brazil called eSocial.

45:43 And I don't I don't know if there's any country that has it, but in eSocial, you have around 10 million companies with almost 60 million employees with all of their events of every type of event that has happened in a company with an employee from you get hired, you get paid, you take holidays, uh you take paid time off. Like all of that stuff is is is is registered, and that's that's that's that's incredibly important, and and for us, I mean, it's it's it's a huge asset because we understand those those those repositories. So, we can onboard companies in days, for example. We only need access to the to the to the system, and we're onboard a company in in in days. It's it's it's it's super powerful.

46:27 Um and and and you can go on and on in in in other pieces of the stack of the Brazilian stack. So, um money there moves so much faster uh from what we're like taxes move so like you pay taxes. I don't file taxes in Brazil, but I have friends that do, and partners that do, and my co-founder Fernando does it in Brazil, but it's just it's just um it's like you're basically you have everything ready, you just sign it on your phone, you tap, and you're ready to go.

46:54 >> Sending a shout. Yes. Yes. So, so I I think it's it's it's it's very very very modern, and I think it's it's important that we tell the right narrative about what what what this market is with all of its problems and all of its challenges, but but it it is true. Yeah, it's interesting that well, in all those ways and more, Brazil is so far ahead than many places in the world. Yet, I think one of the stats of the the skeptics often bring out is that there's been no per capita growth for a decade.

47:27 So, why? How would you explain that? Prominent a prominent investor, I'm not going to mention him, but he he asked me the same question, and it was framed for Brazil, but it was also framed for Mexico. Mhm. We're so close to Mexico, a largest economic block. What's going on? Why don't we why don't we grow? And and and I think uh doing business in these countries, right? It's a low cost it's a low trust environment. That's why we're solving this problem to give trust, so you have a a common language and a common infrastructure for people to to govern and hire and just govern workers' relationships.

48:05 Um so so those things I think solve part of solve the problem. It's hard to grow if you basically have all of these expenses on just doing the basic things that a company needs to do, like filing taxes and and and paying employees, right? It's it's obviously hard. And and and I think the promise is that if you have the efficient intelligence applied to these problems, you're going to remove part of that, right? That doesn't solve the whole the whole problem. I I I I think that these are just removing all of this complexity in which companies need to operate because I've seen how inefficient they are.

48:41 I think it's going to be an it's going to have eventually an impact on the GDP. I cannot calculate how much, but but it's definitely an issue. Um I I I also think that uh there we need to see bigger companies coming from new businesses, right? And um with all of the great companies that have come out from Latin America, and many of them are friends, mutual friends, uh we often speak about about this. We we never build these companies with the governments. We never had like uh such a so great to execute with this president. He helped us so much.

49:18 It was it was pretty much in spite of it, right? And and then you get incredible companies being built, right? And through all of these like recessionary cycles and and and macro cycles and political cycles. So, I think there is it's it's important to also uh maybe take that perspective these companies are going to still exist, and they're still going to thrive in spite of the macro, which is really important. Um and hopefully when as they get bigger, they're going to help modernize the economy because there's few companies that still have, I think, too much power, which explains part of the the the the lackluster growth.

49:52 Um and maybe the classic, right? Like education and we need to have people that are more capable doing works that are jobs that are more uh technologically focused uh apply to other parts of the economy at the end. Economies grow because of productivity and if we don't improve the productivity, we're going to still be seeing the same same dynamic. Sebas, you have [clears throat] this concept about or you often discuss the the airline Pan Am. And it's kind of tied to to your story and what you want to achieve uh in the in the years to come and you also describe yourself as a founder that you don't want investors to put you in a box. And uh it would be it would be great if you can you can share a little bit more about the two.

50:41 The box or both? If you if you if you look at at at this at the Americas, um you find that the US is the largest economy in the world. If it has 300 what 30 million people. If you look at Latin America, we're talking about 650 million people, third largest economic block. Um if you look at the assets that we have, we have them all. I think maybe some rare earth metal metal we don't have. I don't think we have like uranium, but I think aside from that, we have uh GPUs, we have engineering talent, we have abundant energy, we have commodities, we have biodiversity, uh we have two oceans that are fairly protective.

51:29 Um we have three languages broadly speaking. We have similar cultures, similar religions. We're in the same time zone, so for founders that are building teams in Eastern Europe or founders that are operating here in San Francisco with teams in Stockholm, like they're going to connect to what I'm saying. So, I I I want to see way more integration of this and and I want to see more talent flowing and I want to see more companies flowing from one side to the other and I want to see more capital.

51:58 I think we do that, uh we're going to be talking about about just enormous abundance along this corridor. And and that's the optimistic side. There's also a side that is more alarming, which is I I think we're going to we're going to look back. I'm I'm here in the US, I'm a US citizen, obviously Colombian, I'm obviously super connected to Latin America and Brazil, so I'm I'm a bit a bit of a of a mix. But I think it's really important that we kind of own this stack, right? And and and we want to make sure that it's us doing it and and not another adversarial country.

52:34 And if I if I look at Let's Let's Let's take a look at the AI stack, for example. Um so we go start with GPUs. Well, maybe we maybe we start with energy, we continue with GPUs. Then we need cloud, then we need telco. I mean, at the end everything runs through telco, then we need applications, right? So, let's focus on those. So, energy, um we were not fully integrated and we had a lot of players that were just exporting energy and it still happens, right? Like China and Brazil are two of the largest trading partners. And it's Brazil's largest trading partner.

53:07 Um if you look at GPUs, um I still think we're not doing enough and I think we need to make sure that countries in Latin America have their own clusters, but they could be built with American technology. Um then if we look at telco, it's alarming because you have um a lot of people know about this, but almost 100% of the telco it's not 99, it's 100% of the telco infrastructure in Mexico is Huawei's. I don't love that, right? But there's not been much done to change that. Um then if you look at applications, um I want I want the models to be models that are culturally aligned. Uh I want open source models that are that are that are also built by like say us.

53:55 Um even even applications and I think I think um companies in the US don't really understand this. Uh many of the greatest companies are in San Francisco and I open a Brazilian office and it's just like and a Mexican office and it's like two people um super under invested. And I think that's a mistake in the long in the long term, so so I want to see way way more of that. So just super important for me.

54:22 Um and I'm kind of organizing my life around around around that. And then in terms of of of of of of the boxes, I I I think uh We're we're we're in a pretty hot box right now, don't worry. This might be the only time we put you in a box. >> [laughter and gasps] >> I just I just I just never um really believed that uh uh you should be bounded by the orthodoxy of doing things. Um I I guess it comes from my early beginnings of just being very free.

54:51 And and I think as as as as I see more things and I learn, um I also realize that I work the best when I'm the most authentic. And I work the best when I'm doing things with people that enjoy doing the things that I really enjoy and and working in a way that makes sense for me. Um and I I kind of organize all of my life to to be able to do that and I find myself that I I love to build. Um I find myself that I love to be very close to the teams. I love to be very close with my engineers. I love to be selling products and I still that do that all the time.

55:26 But also find myself that I learn a lot by investing. I learn a lot by maybe understanding other companies and maybe even building other companies. Um and I I I think that if you're able to build it in a in such a way where knowledge flows, talent flows, even capital flows and you keep it in a very concentrated way, I think that's a that's a very powerful powerful way of operating and took me a while to to to understand it and and to design it in a way that it fit me as best as as possible. And and and I'm trying to I'm trying to do that.

56:02 Amazing. Um I think maybe a a good place to tie it together is uh all of us clearly believe that um there's a large opportunity for global AI companies to be built out of LatAm. And one of the things that you, you know, have clearly enjoyed doing is just mentoring the next generation of founders. And so, maybe as you think about 20 25-year-old Sebas who's looking at the the sea of opportunities now, what do what do you have to the the next gen of founders that are looking at the opportunity set that want to build a global AI company out of LatAm?

56:37 I think um pursuing problems that are interesting to you is very important. Uh authenticity is very important. Um Make sure that whatever it is that you're building matters to you cuz building a company takes a long time and when you're doing it to start something that is not really true to you or authentic to you, it just becomes really really hard. Um another one is is to run toward what's difficult. I think difficulties, hard problems are really important and I want to think I'm thinking more and more about defensibility.

57:18 Um I'm sure you guys are as well, especially in AI. I mean, it's it's it's the the pace of change is is is relentless. And and and I think you need to have a even though it's hard when you're starting a company early on, you need to have a point of view on on what is going to make you more defensible and and even if you don't have it yet, like how are you going to grow into that? How are you going to lean into that?

57:40 Um as well as growing your business, but but that that I think is is is is is critical. Um three, just be always a student and a learner. Um and and not only learning about the next model, the next application, the next release, you know, the whatever some lab is is offering now, uh but also being a a student of businesses. Uh I think it's it's there's a lot of a lot of leverage by by by understanding uh what some of the greatest entrepreneurs uh did.

58:14 And lastly, just understand that the bar is really high and it's hard to build a great company in Latin America, but that's the only way to build something that matters. So so you got to you got to really understand that quality is incredibly important. And companies that don't embrace that quality and don't aspire to be also globally great are going to have a really really tough time. The good news is that I think we have the the the raw materials for that to happen.

58:47 So so I'll I'll stay optimistic. Amazing. Well, I guess to end, if you want to start a company but you're not quite ready, come join LatAm. >> Yes. Thank you, Sebas. It's a pleasure. [clears throat] >> Thank you. >> [music]