#21 - Rahul Sonwalkar, Founder of Julius AI

0:00The thing about big companies is one no can kill your idea. So if you get you need to get yes from everyone you need to be buying from everyone and one no can kill your idea Think about startups is you can get 50 no's doesn't matter as long as you get that one Yes, that one person that believes in you could be an investor could be a first user first customer first employee first hire You just need that one yes This is Raul Sun Walker Let's start off with What was the first project that you can remember building when you were a kid? I signed up for Gmail with Rahul Sunwalker at gmail.com. And then I forgot my password when I was like 11.

0:46And Gmail wouldn't give me my account back. So I tried to create an email service with a different domain so that I could get Rahul at the domain.com. and then I got like 12 of my friends to sign up for it too so they could all get their first names I guess a couple of decades later or a decade and a half later we can see hey.com tried something similar but it was like hey.com before hey.com yeah do you remember like was that before the social network movie came out it was literally when the movie came out so yeah I watched the movie when I was like 11 or 12 and I think that movie is what got me into like tech I had no exposure to tech growing up you know my family wasn't really into tech um but the movie is what got me into building things with software like you can literally build something in your computer and then millions of people use it that's kind of fascinating so it was like right around the time when I uh when the movie came out yeah do you remember like so was this the first project that you worked on directly after watching that movie or did you have like any, I don't know, things that you tried when you were even younger?

2:02There were a lot of things I tried when I was younger. So for this one time, somebody got me like a small aquarium when I was a kid as a gift. And then I tried to flip that for like more money. I would do stuff like that all the time. But after the movie, I felt the agency to like go solve my problems with like code. before the movie i just it didn't really like sit with i didn't click for me that you could actually like go write code and like can make money with that and like you can actually solve your problems like oh gmail took your email like you can just like go make a new email what was the domain that you got i think it was like a five letter like something with a g um with a just a shitty domain name that i could find for like eight dollars um because i was broke um And it didn't matter if it was like a memorable domain.

2:56What mattered was that you could get your first name or a shorter email. It could just be yours. It's just yours, yeah. Yeah. What was the next thing that you worked on? Naturally, after that, I tried to build a social network because I'd seen the movie. And I had this thick book on PHP and MySQL. and it had simple things like how to make a profile on a website and how to have login authentication. So I took all those pieces and barely glued it together and no one used it. It was like a lesson that you can't copy an idea and... Like hope it will succeed. Hope it will succeed, exactly. I mean, later on, you learn all these fundamental things like network effects and you need to have a hook to bring people into your product.

3:49Back then, I didn't know any of that. But I still learned how to write login and authentication code. And that turns out to be much useful later. Yeah. So how many times have you basically tried to launch something or create something and had basically just no response from the world? Oh. Fortunate. Because I remember you talking about the podcast comments, and you just relaunched that, and then nothing happened. Yes, yes. So this is, you know, fast forward to COVID. I was pretty lonely. You know, everyone, all my friends were logged, you know, logged in in the lockdowns indoors. And I had just moved to San Francisco out of college in Texas.

4:32So all my friends were still back in school. And I was really this new person in this new town. And I would spend a lot of time on podcasts, just kind of like filling in the time in my life with like people talking. were there um i did some joe rogan uh back when he was on youtube i did um acquired podcast a lot of those like startups founders podcast david sandra is the goat um so just the classics you know and one thing i would always find interesting when i would watch joe rogan on youtube is i would go to the comments and read the comments and the comments were like an additional add-on to the content you know there was a discussion there were some jokes and i would miss that when joe rogan and somebody's podcast moved to spotify because they don't have a comment section so i tried to build podcast comments.com because you could literally just you know take any podcast link and just have a unified comment section it made so much sense in my head the idea sometimes you just need to get the ideas out there and you realize like it only makes sense to you and nobody else um and how much time did you like spend on that before you launch it and realize like damn oh i this was the first mobile app i built so um it took me like a month to get get get this thing bare bones version of it out and the the sad thing was um it's like precursor pre ai so like you just read the documentation and try to learn how to write mobile code um i wish i did some validation before that like i wish i shipped a bare-bones version or shipped some marketing before i actually wrote the first line of code to like get validation that someone else someone would actually want it yes yeah yes and so you like to answer your original question which is how many times have i launched something and failed countless times it's kind of every time i launched something new i was at a point where failure was default yeah and in some ways that's a very scary position like why would you even try if you're gonna if you know you're gonna fail but at the same time it's kind of liberating it frees you to like try things because you feel like you won't necessarily look stupid because no one will see it yeah this guy's just gonna he's gonna launch this new thing and it's not gonna work yeah so it's super liberating because then you stop kind of like being in your head too much and doing analysis paralysis and you just go try things um so i think that that early failure is i'm very glad that that happened to me um and it was very liberating yeah do you have any other really good stories because i know that you like was it before you joined uber that you were basically trying to track trucks or was it after after after yeah But I had a startup before that in college.

7:29Let's talk about that. So I ended up going to college at the state school in Texas because, you know, my family was like, hey, if you want to go to college, you can go. But you just have to find a way to pay for it. And so I found a school that would give me a full ride. So I ended up at the state school in Texas. And while, like, the full ride was awesome, the people around me were just not at the same level of ambition. you know they all they wanted to do was just finish the classes go home go to the career fair get a job at raytheon or all state or an insurance company um that social network energy like starves for ambition i was starved yeah it's just you know that social network energy of like hey you can just build something in your dorm room and get a bunch of users and build something big was just not there so i started going to a lot of hackathons you know go to these hackathons back in 2016 and meet people that were just trying to build things on the weekend.

8:28I made a lot of friends there that ended up eventually referring me to companies in Silicon Valley. But back then, what I noticed was a lot of people, a lot of these hackathons would spend time getting their project just set up. The fundamentals like login, database, backend service, Like all that spinning that up would just take up half of your hackathon time. And then the other half would be actually working on the project. So I tried to build something there. It was called Waterview. And it would let you basically spin up all that out of the box. So you can actually focus on the actual project and not worry about these details.

9:11It got some traction. It got some downloads. I would go hackathons and get people to use it and they would like it. an important lesson i learned there was the people that went to these hackathons would never actually end up working on the projects outside of the hackathons so the retention was basically zero um so that's where i learned oh retention matters like you know you need to build something that people are going to want to use daily or weekly um or as a part of their job um another user how did you like just get people to sign up did you just walk up to them and say you know download my thing yeah so this these hackathons are basically like you know like 500 sweaty college students computer science students in a big you know basketball converted basketball stadium or like a big hall and you spend the like first half of the hackathon just getting set up and i would go to these hackathon students at the start of the hackathon and just tell them like guys, you could literally just download Waterview and be faster than everyone else in this hackathon and win the prize.

10:19And it would take them five minutes to set up, get them off the ground. They would go build a project on top of it. You can think of it as like Supabase. It's like a startup analogy. Yeah. So it was Supabase before Supabase. I didn't have the technical chops to really build a Supabase quality product. And also I did not have the right target customer base and I was giving this out to hackathon goers that were building things outside of hackathons. Yeah. Has there been any specific person, for me, it was Warren Buffett that I really look up to. Has there any been any person you've looked up to a lot that you're trying to mirror their trajectory in any way?

11:01I think for me, it was Bill Gates. So if you talk about tech as Bill Gates, you're talking about outside of tech as Napoleon. My mom - What do you like about Napoleon? Well, my mom just told me this story this one time when I was a little kid that Napoleon, his ambition was so massive that he wanted to conquer the world. And when you're like a seven-year-old that you don't know what that means, but it feels like massive. It feels grandiose. and so and then she would tell me how he would work so hard that he would he would nap while he was riding his horse so that he could like continue working when he gets off his horse and you know those kinds of stories were pretty cool and then I got started reading about Bill Gates and Microsoft early history of Microsoft which I think was super fascinating because Microsoft I think is a is a really special company they've built a massive massive business you know how many tech companies do you remember from the 70s that that last today just to Microsoft and Apple and just his competitive drive you know his contrarian bet on software for the first five years the personal personal computer industry you know nobody wanted to bet on software because they thought software is just it's just code it's free you know today we think think of software as the greatest business model ever.

12:24You have 70 % gross margins. This is the best business model ever. You have multiple trillion dollar companies that sell just software. Software used to be hardware on floppy disks and stuff. Yes. And so he had this super early bet that software is going to be the differentiator. And then the way he went and captured the market by getting every PC provider to offer Windows and make it Windows compatible so that all the application software that gets built on top of that is windows compatible and then you just can't compete you know um so i think that level of prescience that level of um you know drive is really really special um and i i think bill gates is great i mean today we think of bill gates as this guy who's you know doing a lot of charitable work in like vaccines um and um but you know up until like the 2000 early 2000s like he was like this conqueror yeah and the drg had to come in and neutralize him it's exactly it's crazy right it's bill gates versus the people of the united states and the world yeah yeah it's a they had to step in and neutralize but um i think he was a very dominant figure do you have any good stories of doing things like I remember on Turner's podcast, you said something along the lines of getting in people's trucks or something.

13:51Trucks. I believe, and tracking trucks. How did you, what was that business? Yeah. And how did you start it and why did you start it? So when I was at Uber, I was on the pricing team. And you do pricing, you figure out how much to charge for a ride. and all that is so simple because uber has real-time data on where your driver is and where you are as a rider so they can track wait times pickup drop-off etas and all that data that real time location data they have that which simplifies so much overhead turns out logistics doesn't doesn't have that they're so old school that they pick up their phone and call the truck driver um if you have your own fleet you know if you're amazon you have like devices on the trucks that will help you track your trucks but amazon is a small percentage of all logistics most logistics like 80 is um 3pls like third-party logistics providers and those truck drivers they you just literally call them and you ask them hey are you gonna be at drop-off at 4 p.m and if they're not at drop-off at 4 p.m they're like 15 minutes late then your next shipment is also late it's also delayed because you have one loading dock and that you can't just park a semi truck on a street you have to go park in a parking lot somewhere else or how many trucks are there like on the road is it like tens of thousands or hundreds of thousands oh hundreds i think there's like millions yeah Yeah.

15:28And so, you know, the idea was, well, the U.S. government in 2018 created this like new regulation where all trucks have to have this tracking device called ELD. So they could track the number of hours that each driver drove and they could regulate that, you know, not more than 11 hours because it's not safe for the roads and all that. So the idea there was, hey, if you could get that data from all these trucks, you can make it really easy. You don't have to make new hardware. You can just make it really easy for these logistics companies to get real-time location data. And the logistics companies, the shippers, brokers, all of them are super pumped because it solves so much overhead for them.

16:19the downside was it was basically impossible to convince truck drivers to share their location data you know they just weren't they just didn't really want this didn't want it yeah yeah you know it's it's hard to convince somebody that's done something a certain way for a few decades how did you how did you like get in touch with so how many like truck drivers did you actually end up talking with and then how did you get in touch with them and what was their response yeah well uh the The response was just, hell no. Yeah, so it was, the response was hell no. I talked to like hundreds of truck drivers.

16:53There's spots around San Francisco that are like trucking hubs. So Lodi, California, in the East Bay, in Stockton, and all those places have like big hubs for trucks. I would just go talk to those people. You know, as an engineer. You sound like an early Sam Walton where he's just like walking, just staying in the distribution centers talking to the truck drivers to see what's going on at each of the stores yeah it's funny you mentioned sam walton because people think of walmart as this you know old company that is not technology forward walmart was super original the og yeah yeah they were they were they were they were they were the first company to like really adopt like computers and set up like their own private networks of the different distribution centers could talk to different Walmart locations and all the POS data gets sent back to their inventory management system so they can like ship the goods fast they were just the pioneers of that and I really wonder if you know how the Amazon story would have played out if Sam hadn't existed if Sam was around when I in the internet age so Sam died I believe in 1991 someone and Amazon was started like three years after that?

18:0894. 94, yeah. So I really wonder if Sam is prime in the internet. Competing against Bezos? Yeah. Yeah, it's funny because Bezos, I believe, walked around handing out the Made in America biography to all of his executives and all the Amazon team members in the early days. I love it. So yeah, dude was a goat. I love it. Okay, so you basically hear a whole bunch of no's from a whole bunch of truck drivers. when did you decide like this is just not going to work and there's no there's nothing that i can do yeah to make it work so um it was we had a lot of yeses from not truck drivers the people that bought these truck drivers to move their truckloads yeah they all really wanted this thing they were they're like take our money you know we want to use this thing um so we closed some sales before we even thought of talking to the truck drivers.

19:01Really? Yeah. And then... So you got some false hope traction. False hope, exactly. And I think the need was there. And then after months of just banging on doors and trying, it really came down to whether this is the most important thing I can spend my time on. As an engineer, I was... this is back in 2022 pre-chat gpt pre um you know like the boom but i was spending a lot of my time playing with gpt3 i was spending you know i was just fascinated with like github copilot and how you know when i write code for tracking and like location data it just it would just know what python module to import in my ide and i didn't have to like look it up on stack overflow and to me that was very cool and I was just fascinating fascinated about that and spending a lot of time playing with that so I decided to go do that instead yeah when did you decide to start Julius because I know that I think you were in like the summer batch of YC yeah 22 yeah so what did you decide to like start Julius and then this was like pre-chat GPT so you're just working with like GPT-3 I think yeah right yeah it wasn't called Julius back then yeah we had a bunch of different names.

20:17The initial idea was like, you have these language models are getting really good at writing code. And I had seen at companies like Uber and Facebook, they're very data science driven. So Facebook, everything from feed recommendation to like, you know, content recommendation to your friends, all that is super data science driven and ML driven. Similarly at Uber, pricing, routing, all that is super data-driven.

20:50And I wanted to see if you could leverage language models getting really good at writing code to get that power of data science into the hands of everyone. Let's say you're a finance person. Let's say you're a CFO. and your company collects a fuck ton of data across all the different parts of the company from sales to product to recruiting. All the data is right there. You just feel helpless because you can't go analyze it on your own. You can't ask questions about the data. So could you use these models getting really good to help you write the analysis code? That was the initial idea. We didn't know what that was going to look like.

21:30We ended up launching four different iterations of that. Yeah. We called it text to SQL and tried to get big databases of the internet that we could demonstrate. You can actually ask questions about the data. How long between these different launches? A year. A year? Oh, so you started a year before? Yeah. Oh, okay. And we got a seasoned assist from Microsoft. Really? Yeah. Because we launched this thing called Excel Copilot. So I still own ExcelCopilot.com. Beautiful. Which is like a Microsoft product. Yeah. And we're initially like using that name and they sent us a, they're like, you got to stop right now.

22:15Apparently the best thing you can do is just kind of ignore it. Ignore and just continue building. Yeah, because the lawyers, they'll always try to stop you from, you know, creating value in the world, right? It turns out when you start building things, there are always going to be people that will tell you why something won't work or shouldn't work. Either it's because it hurts their interests, it's because sometimes it proves that they lack agency. Yeah. It's the whole Paul Popper syndrome. What is the syndrome? It's basically like, I forgot exactly where this comes from, but where like if the grass, you know, if you're a blade of grass and you grow up taller than the rest of the grass, someone, you know, cuts your head off.

23:09Yeah. And that's the tall poppy syndrome. So if you like strive for more and it's more than what people think is like reasonable or normal, like actually I'm going to try and shut you down. Yes. Just to stop you. Yes. Because I feel like I don't feel good about my own place a lot in life. Yes, exactly. Exactly. So spot on. So like there's, you know, incumbents will try to tell you why you shouldn't do something or people try to tell you why you shouldn't do something because it shows their lack of agency. Did you experience that at like Dallas, the university where you were like wanting to start a company or something and people like were kind of against it or?

23:46um my professors thought i was um so i i basically i would basically pick courses where most of the credit was the exams so if you just if you skip all the assignments and homework and projects but just crush the midterm and the final um and you get an a minus like that's good enough for me you know it's like the 80 20 rule right so i would i i never had like a straight like 4-0 but my whole thesis was like you know you could really like pull an all-nighter before the test and just study the semester worth of content in a night and like get a minus you're optimizing for like good enough good yeah because i could spend all my time yeah there's a semester just doing projects that i want to actually work on so i must just thought i was crazy for like missing those assignments and but you know I optimally chose courses where the assignments didn't have like read it I think I'd big even I said big companies like Facebook and uber you know I always looked up to these companies because if their founders you know Travis built a great company you know there's some great stories from uber in the early days and I want to go work there I want to go work there and learn how from like Travis yeah or like thumbs mark yeah yeah uh and i think facebook in many ways is still a killer company because it's still founder led uber is not so when i got to these companies i still i've i felt so much bureaucratic inertia on everything you want to do you know the thing about big companies is one no can kill your idea so if you get you need to get yes from everyone you need to buy in from everyone and one no can kill your idea think about startups is you can get 50 no's doesn't matter as long as you get that one yes that one person that believes in you could be an investor could be a first user first customer first employee first hire you just need that one yes you can get 50 no's and all that matters is that one yes and then you file your way from there yeah i i like the uh whole like you launched four or five different times uh because i remember this line from brian chesky where he says like if you launch and no one notices you can just keep launching yeah launch airbnb three times yeah um so when when did you actually get that like first believer at what point so we um we we weren't really we were on the right thread but we weren't really building the right thing you know we were you know we're trying sql sql is kind of limiting because you know you can't really do a complex analysis with sql and there's only so much sql code on the internet that the models have been trained on python there's just like a lot more python code it's much easier to do with iterative analysis in python and python you have this ecosystem of python modules that are really this is really powerful for analysis so that haven't clicked for us yet so we the first four things we tried were sequel and then the fifth thing we tried was python we called it julius um and why did you call it julius well i've just been obsessed with roman history so um when i incorporated the company i called it caesar labs because it you know julius caesar and i just read a lot of roman history just a classic meme you know i'm not unique in that sense i think i'm sure there's a lot of your listeners um a lot of founders like yeah a lot of founders like emperors yeah in the point so exactly um so that was me um and coincidentally we ended up launching on julius caesar's birthday july 12th we were supposed to launch july 10th and because of a last minute delay we had to push to launch by two days so we could fix a few bugs and just ended up being julia cedar's birthday we realized like a year later just on accident on accident that's crazy yeah that's well was that when uh were you in yc at that point a year out of ic a year out of yc yeah wait so did you get basically zero traction until 2023 well there was traction it wasn't durable traction okay yeah what was that experience of just like launching stuff having it not work and then like learning from it just keep on going Yeah, I think it's, so one of the, so it's part of it is intuition.

28:18Part of it is like, you know, data. But basically it's one of the hardest things is whether you keep going on the same thread or you start over from scratch. And you're going to be wrong many times. You make that call, you know. sometimes you would you know even in julius we decide whether to like build a feature or not or if you launch a feature does it take off right away or do we need more iteration on that or do we need to focus on something else it's just like you're a place of yes it's like you're managing resources right you have like time energy money i think the most important resource is morale people i think startups die because of like lack of morale like you know you you read about like napoleon's like waterloo campaign and it's like they were so outnumbered but they had so much morale and they just wanted to fight for napoleon like you get excited he's excited right you can do a lot more if you are super high morale and momentum but low on other resources but you can't really like use money and other things to build morale.

29:35Yeah. So you're playing this video game. You have like these resources and you have to make something work, right? And you decide where to allocate resources, where to pull resources. So in the super early days, it's like, do we keep building this idea or do we like change the idea or like pivot? Yeah, pivot. Yeah. Yeah. So luckily for us, we ended up building on analysis, like analysis with AI. We just didn't know what that was going to look like. We had some ideas, like maybe AI or SQL for you, we'll query a database for you. Maybe it will make charts for you. We weren't super sure what that was going to look like, but...

30:19It was a general thread, just helping people understand their data and visualize things just in a more like simpler like simpler way yeah well it turns out like data is very ubiquitous people just have data you know you're a marketer you have marketing data your finance person your finance data your ops person of ops data people the the need for insights is there you know you always have questions about the data but what was missing was that how do I get the insights out of data and turns out like ai can help you with that um and so if you stay on a thread i think the learning's compound you launch don't get users you you try to debug why do we build the wrong thing you know then you launch again you get some users um and they retain some of them retain some of them don't the ones that retain you find figure out like what were they trying to do in the product like what did they need this for um and i think nikita beer had a similar thing where they launched like eight times before i think it was like seven different apps or something crazy and they were like just about to give up and then i think it was like six years or something crazy like that yeah i'm glad i didn't have to do it for six years it was just one i know Yeah.

31:42Yeah. But I think it's kind of the same thing where like if you're building stuff since you're a kid, you are kind of working for six years or 10 years, but you're just working on different threads. Yeah. When was the first time, I know that the GPT store was a big inflection point for Julius. When was the first time that you launched some iteration of Julius and it actually worked? People were excited. It was July 2023, and we did get a lot of our early users from there. And then OpenAI basically shut it down overnight. And so that's where we pulled up Brian Chesky, where we just relaunched. We realized at that point, we had users, we had paying users, and we had used those users to tweet the products that actually worked really well and serve the users really well.

32:41And we had found something that had like PMF. We just need to like get it out there. And so our main source of users is like disappearing overnight. What do we do next? How many users did you get from the GBT? It was like a few thousand. Okay, so it was reasonable. Yeah, yeah.

32:58And it was the GPT plugin store. Yeah. And, you know, we had already launched it in July, but we just called that the beta. And then we're like, all right, we just, we're out of beta now. The official one. Yeah, exactly. The official V1 launch. Exactly. Right. So, so we did that. Yeah. Yeah. Okay. So building off of like that initial momentum, how did you get the next, I think you like pretty quickly went from, I don't know, that first few thousand to like 500 ,000. And that's what you had roughly until like a year ago. Yeah. So, you know, we definitely like sharing what we're building with people has helped.

33:44I think if you put yourself in the shoes of the user, I think the customers, they have this fatigue of trying a thing. And then you realize that it's not being actively iterated on and they aren't building new things in it. They aren't taking the model advancements and turning those into features that help you use the product better. So in Julia's, for example, you can click on like, when you're kind of data set, you can just click on like columns. And we will take the point and click data and have AI generate like a chart for you before you even write a prompt. That level of like attention to detail, it only predicts what you want it to do.

34:23Yeah. So let's say you click on revenue and like customer lifetime. lifetime. And we'll just show you, oh, you probably want to look at customer lifetime and how does that affect revenue. So we basically, because that attention to detail, it only comes when you have built something with a lot of love and attention to what users want. and the opposite of this is like users get served slop it's slop and then they pay for the product and they realize that six months later it's the same product and so we actively launch new things on social media we talk about what we're building we integrate with users and that compounds because you know then people start spreading the words today like word of mouth and seo end up being like the two big drivers for julius yeah do you know like how much time you spend a day probably like talking with users it really depends on the day you know there's so for example when you were launching workflows we did not know if we hadn't quite figured out those micro interactions and workflows and so my priority became just reaching out to use our users offering them to hop on a call like i'll literally make a workflow for you and um making a workflow for them understanding like what are the pain points there and then switching that where switching that to like letting them make the workflow and like observing how they use and what are the things that are confusing to them like you know it really depends on the day like there could be a day where if something is like the top priority is like we need to get user feedback on this thing like let's go do that we all we don't have any support people on the company the the engineers do support everyone in the company does support i've personally answered 5 100 support tickets matt our cto has answered like three and a half thousand support tickets is this like a kpi It's we should probably make it a KPI.

36:40And so I think that's... Yeah, well, that was one of the things that I was like very impressed by, but was you just seemed to like really care about making people happy.

36:56It seems like that should be obvious, but it doesn't seem to be as obvious. You've been very good so far at least about basically building your own personal brand and how do you think about doing that i think um when people try to build a personal brand they um i'm curious what do you think my personal i'm curious like what as a third as a third person what do you how would you describe my personal brand a comedic undertone with like obvious seriousness like you clearly are making something real while also having a good time doing it and like it's enjoyable to watch the ride yeah love that okay so i guess my answer to the question is like how do you build that is is you just have to be yourself i think people try to copy somebody's style way too hard um for me like this is how i am outside of this interview like um if you work with me in the office like sometimes I'm making some jokes sometimes it's like serious I think it's I think it's important to like have fun keep it light-hearted but also you know work on things that are important they're serious like your life is not a joke you you want to work on things that are serious so I think that's just I'm just me right now yeah what a trace do you look for in like potential you know hires yeah traits yeah i look for people with high slope so um you know you could either recruit for a one very specific skill like i want i want somebody who has specific experience with this this one particular thing but the nature of startups is you're solving unique problems every day you know um for example like your biggest source of users decides to shut down overnight what do you do like this is there's no playbook for this you have to just kind of think from first principles but also like claw your way and like just throw hail marys and like try everything or um a really big trillion dollar company sends you a season desist letter like what you do uh you got any more of those uh can't say i can't say perfect yeah um but uh you know there's you there's you you want to work with people that can pick up new things and like learn new things and they're a multiplier so when they join a team you aren't linearly going to team you're multiplying you know um so you want to you want people that are multipliers they will take a finite amount of resources and make so much more out of those resources so high learning rate and like um yeah i think uh that kind of comes back also to like just bring people on that have lots of energy and have like the right vibes i used to like under price vibes but i i genuinely think that if some you know people enjoy showing up to work yeah that's like half the battle it's like the the thing that kills startups is literally just the founder waking up one day and saying i'm going to do anything other than the thing that is my company yeah and then the thing's dead it's over and morale exactly exactly i would so you are also really good at like getting in touch with powerful people and like making them your friends you've done with this with like even jensen and i think you know guillem or uh like punch i think that's how you say his name okay oversell yeah yeah guillermo guillermo yeah okay how do you think about like deciding who a who to look up to and then like making them your friends um i think you'd be surprised i think people are more approachable than you think people you know you so guillermo i met him so i was at this happy hour for like engineers back in 2022 and he was there talking about like his journey building rossel and after the chat he was just surrounded by this crowd of people so i take a photo with him they want to like talk to him about rossel and like I was the only one in that crowd that asked, told him about my company and asked if you want to invest.

41:29And all the people could have asked, and there was, there's no rule, like you can't ask. You know, turns out he's a human being. I just like sent him an email, hey, this is what I'm working on after. And he's like, I'm in. So it's, the worst that can happen is they say no, right? So you, I think it's people, people underestimate how much human these people are. Because, you know, five years ago, 10 years ago, a few decades ago, they were in your shoes. Yeah. And they remember. They remember, yeah. Yeah. I've noticed that like the best founders, not only, you know, they study the greats, but when they actually become great, they eventually also try to help, you know, other founders up.

42:17um have you heard a story of like steve jobs calling um hewlett packard hewlett packard yeah mr packard i think or yeah he's like he's like i think he wanted uh parts for what i forgot exactly which um it was something like measuring device yeah and he was like yes we'll give you the parts and also do you want to just have a job over the summer and i think steve jobs was like 14 yeah and he did and that's how he got his first job i love it yeah exactly so you can just do things you know just do things um okay uh final question what's the hardest thing you've overcome i don't know man i think um they all seem easy once you overcome them you know they seem super hard in the moment but then once you over overcome them it's just like you just forget you just forget you move forward i think the hardest thing that i have to overcome is yet to come yet i mean is probably around the corner, you know?

43:13But then once you overcome, it just seems very surmountable. The things that seem insurmountable. It sounds a lot like I remember Sam Altman saying that the first crisis that feels world ending, you're like, oh my god, is everything over? But then the seventh time it happens, you're like, oh, well, I survived the first six, so I'm probably fine. We'll make it through. Okay, that's a great place to end it. Thanks for having me, man. Appreciate it. Thank you.