He Quit Uber, Beat ChatGPT At Harvard, And Went Solo Building AI | Rahul Sonwalkar

0:00I have this idea in my mind and I just kind of want to go build it. And I don't want to let having no co-founder be a blocker for that. I've met countless people that are really smart, really competent, and they're blocked on starting a company because they don't have a co-founder. Eight out of 10 people that have co-founders, when I talk to them in private, they're always kind of trying to work through some issue. Startups are a game of outliers. If you're following all the conventional wisdom and conventional advice and all the best practices, you're going to end up being very average. A big misconception with solo founding is if I'm solo, I don't need to listen to anyone.

0:31People just will listen to me. That's not true. Ultimately, smart people, they're not going to listen to you just because you're going to do something. You have to convince people.

0:38Rahul Sonwalkar:What advice would you give them in that situation? The easiest way to convince people is to start moving in a direction and then show progress. The right way to do it is to be very generous with the early team in terms of equity. If a candidate is willing to take less cash or more equity, it's a big green flag. it's incredibly low friction to get started. If you want to build a company, you can get started tomorrow. Well, listen, I'm really excited to have you here. As you know, like the goal of this is to talk about solo founding and you've done a lot of podcasts and they're really interesting.

1:09Rahul Sonwalkar:And you talk about the history behind Julius and you as a founder. But the thing that you don't really touch on too much in any of those other conversations is the journey as a solo founder. And I think it will be really interesting for people who are either solo founding or are considering being solo founders to hear about your experiences. So maybe we can just jump into it and we can talk a little bit about sort of how you came to the decision to start a company solo and what that was like. I actually worked on a company back in college that didn't work out. And it kind of stemmed from an idea I had from going to hackathons.

1:42This was just me seeing a problem and just chasing the solution to that problem and then ultimately building something and trying to get into the hands of the users and it didn't work out. but what it ended up being was like I was a solo founder you know I didn't realize that but I just found a problem and first thing I did when I saw the problem was to try to build a solution for it and then fast forward to like three years ago I had a few friends from college that I kind of convinced to come work with me on this new idea I had and then within a few months they told me I don't think startups are really for me and so just just through through fate I ended up being a solo founder, you know, a few months into my second startup journey.

2:26And the decision was never to have co-founders or not. What I was really motivated by was, hey, I have this idea in my mind and I just kind of want to go build it. And I don't want to let having a co-founder or having no co-founder be a blocker for that.

2:47Rahul Sonwalkar:That's really interesting because I think that happens often, where people have these co-founders that they're exploring with. They maybe make a little bit of progress together, but ultimately it doesn't work out. And sometimes it works out. It doesn't work out in the sense that people are still friends at everything which happened in your case. And other times, obviously, you have these big implosions. But when that happens and the person ends up effectively becoming a solo founder, sort of by circumstance of it not working out with the other co-founders, I think there's really two paths to go, which is going and trying to find more people or saying, oh, maybe it's not meant to be and shutting the company down or not pursuing it further.

3:24Rahul Sonwalkar:I guess the third one is, of course, the solo path. So I'm curious, when you were there, when you were making that decision, I think that a lot of people probably come under similar circumstances. What advice would you give for people when that happens, when you're really excited about something, your co-founder says, oh, gee, or co-founders say, oh, gee, it's not really what I'm excited about anymore, or I'm not thinking that being a startup founder is the right thing for me right now what what advice would you give them in that situation yeah well in if if anyone ends up being a solo founder they're probably also the ceo so um i've heard this advice from ben horowitz and brian cheskin some of the other greatest founders is you know being a ceo is like the most lonely job ever even when you have co-founders and especially when you don't have co-founders you're a solo founder and you're ceo that's just the loneliest thing ever and so my advice to those people would be look if you have a if you have a vision or a idea in your mind that you think needs to exist and you have enough conviction start making progress you know the easiest way to convince people is to start moving in a direction and then show progress show that you can actually prove the ideas that you have in your mind by doing things and making things a reality and so um what i see oftentimes what happened is if if people have co-founders they let those disagreements kind of be the end and all be all of the company and just sort of shut down the company or just be in this state of like no progress i think the the the most addictive thing about startups is momentum and forward progress and so i have to use the analogy of football you know in football you don't really need to score a touchdown on every play you just need to get the first down you need to get a few yards and advance the ball to the end zone and then score a touchdown um so you know if you have a co-founder if you're talking to people about co-founding really the the most important thing you can do is look you have a vision you have an idea start making progress start finding proof points or start invalidating things that you think don't exist.

5:34And all that is just more ammo, right? Gives you more conviction, convinces your founding team, your co-founders, or future people that want to come work with you to come work with you. One of the things that I see people, especially new founders, when they're starting out the company, one of their biggest blockers is, hey, I have this idea, or there's this space I'm really excited about, and I want to go build something in that space. but I'm looking for a co-founder and I've met people, you know, six months later and they're still looking for a co-founder and maybe they have changed the space that they're exploring.

6:13Maybe they did some co-founder work trials. Um, and I feel like that's a giant waste of time. Um, because if you really believe something should exist, um, and you know, with AI, even if you're non-technical, you can start building a prototype. You can start talking into customers. You can start like doing some sales, some marketing, some initial validation. There's so much progress you can make without a co-founder. And that doesn't mean that you should not look for a co-founder, but really just start making progress. And as you make progress, as you get more momentum, it becomes a lot easier to attract people to come work with you, whether that's a founding team, whether that's a co-founder, whether that's investors.

6:56But without any of that, when you're starting with a blank slate, the space of exploration is kind of infinite. And so what a lot of these co-founder trial or conversations kind of end up being is, hey, what space are you excited about? Let's first align on the space. And okay, cool, we align on a space to explore. Okay, cool, let's align on a problem. Okay, cool, we have a problem. Okay, let's align on a solution. And in the moment one of those things, there's a misalignment, things just kind of blow up. And so I feel like that should not be a blocker when you're starting a company.

7:27Rahul Sonwalkar:Yeah, it's funny because it feels like it's one of these beliefs that people have about startups. Maybe it's a sort of a truism or something that people think that think is just is required, right? I think that people have these maybe false beliefs about how you should do things, or maybe it's generalized advice. and maybe there's some merits to generalized advice, but I think that it ultimately just generally puts you sort of towards the average, right? I'm curiously how you think about generalized advice and maybe how that frame applies to solo founding. Absolutely. Well, you know, in general, like startups are a game of outliers.

8:10And so every great company is an outlier in some sense. you know Figma for example they took forever to launch and took forever to make revenue and on the other hand you have cursors or all these other other products that are just making revenue within the first few weeks of launching and and have instant PMF and so it's a game of outliers and so if you're following all the conventional wisdom and conventional advice and all the all the best practices you're going to end up being very average and so each great startup ends up kind of violating a few of the conventional wisdom of startup rules, right?

8:49A Rippling, for example, we're going to build a compound startup. We're not going to start with just payroll. We're going to start with payroll insurance and like laptop management and HR software, all these things bundled into one. And that's very different from like start with the MVP and minimal viable product that barely works, right? And so I think all great companies end up violating one or two core principles of building companies. And so oftentimes you realize that as you're building a company and when you're new, you get a ton of advice. Advice is free. And oftentimes when people are giving you advice, they're giving you an advice from their perspective.

9:26Like, hey, this is what I wish I did right. Your scenario might be completely different. Your problems could be completely different than the person who's giving you advice. And so if you end up following all this advice, you end up distracting yourselves with with things that don't matter and so kind of like in the journey of every founder when you're like a few months in you realize okay i gotta start getting really good at saying no or pushing back on advice and like hey let me cook right as long as you're doing the ethical things and working hard and um being genuine like people will people will not take offense to that you saying no to advice and so ultimately you know you have to find what are the few pieces of advice you're going to live by and then what is what are the things you're going to ignore and continue down your path and i think in general if you have a great co-founder that you have a perfect alignment with man that's a superpower but i feel like eight out of ten people that have co-founders when i talk to them in private they they're always kind of trying to work through some issue and in their head they're pretty clear about the direction they should take the company in but then their biggest blocker is i have to align with my co-founder especially in the early days when things are kind of like up in the air you don't know what product you're going to build or whether you're going to pivot next month or whether you're going to reposition the company and so having a great co-founder that you just like perfectly aligned with is is a superpower but most people I talk to don't have that.

11:02And that means should you let finding a great co-founder be a blocker? Not necessarily. Get started. Start making progress and see if you meet awesome people along the way. Bring them onto your founding team. And yeah. It seems like choosing to be solo

11:21Rahul Sonwalkar:or being solo by default ultimately has a bunch of implications for how you run the company or how you think about things like hiring. I'm curious how that's played out for you with Julius. Yeah, you have to hire much sooner than other companies that have co-founders. And that's number one. When you're starting off, you have to find smart people to come work with you. And I think that in general, the recipe for building a great startup is find a problem that's really big. And a lot of people have, and they're willing to pay for it and then build a great product that solves the problem and then hire the smartest people you know.

12:01When you have a co-founder, you can kind of delay a couple of those things. But ultimately, the recipe is the same, whether you're solo or you have co-founders. When you're solo, you still have to convince people. A big misconception with solo founding is, hey, if I'm solo, I don't need to listen to anyone. People just will listen to me. And I'll tell people what to do and they'll just do what I say. And because I control most of the company and that's not, that's not true. You know, ultimately smart people, they're not going to listen to you just because you told them to do something. You have to convince people and to get the best work out of people, they have to be excited about the thing that they're working on.

12:45It shouldn't just be, I'm doing this because the founder told me to do it. It's like, I'm doing this because hell yeah, this should exist. Like this makes so much sense. and doing that over and over for like every feature and every part of the product and every part of the, you know, whatever the customer's facing, getting people excited about that is work, right? Why should smart people come work for you if you're just an asshole, you know?

13:12Rahul Sonwalkar:You have to be ultra good at convincing. You have to essentially offset that. Yeah, essentially offset that, exactly. And you have to essentially offset that and be ultra convincing convincing because you're going to convince somebody and then like shift your focus onto something else and you have to put the faith in that person that okay now that i've convinced this person they're going to go execute and convince other people on the team to go execute on that so you have to find people that are multipliers so the way i think about hiring is find people that you can put a lot of faith into because you're going to focus on one part of the company for some time and then shift your focus to another part of the company and then you have to put the faith into the people um to go execute in the direction right so when you come back like next week or like a few days later a few hours later like there's progress and you have to find people who will sort of multiply you and not just you delegate a task and they go execute the task and then they wait for the next instruction they just sort of like take a direction and just kind to go execute in that direction so people you can put a lot of faith into and then people who are ideally better at you than the function that they're responsible for so pretty much every engineer at julius is better than than me i used to be an engineer before this and one of my things was i i need to be the worst coder on the team um um everyone that does like growth stuff they they have to be better at me than at growth than me.

14:43And so to find people that are better than you and convince them to come work with you and put a lot of faith into their work because when you have a co-founder, you can kind of like split the company functions into two, right? As a CEO, you can do sales and marketing and maybe product and engineering, your CTO. and if you have three co-founders you can even split even more but as a solo founder you kind of have to put a ton of faith into your founding team now I think a lot of times when solo founders think about building teams they think about hey I have doubled equity because I don't have to split the equity with a co-founder I think the right way to do it is to be very generous with the early team in terms of equity because they're taking on meaningful risk when they're coming and building the company with you.

15:44And what you have is you don't have the burden of having to align co-founders on major decisions every step of the way, but you still have to find smart people to come work with you. The right thing to do there is to be generous with equity.

16:00Rahul Sonwalkar:I'm really curious if you have any specific advice around that or if you've talked to other solo founders and gotten their feedback about this. Because I think that we can talk about this more later, but I think that there's this really interesting thing. When you actually look at the numbers at a high level, it's very surprising how solo-founded companies don't seem to do much different in terms of equity grants. But then when I actually talk to solo founders like you or founders like Paul or Adeel, you actually see that at these more forward-thinking solo-founded companies, they're using this as a huge opportunity, right?

16:38Rahul Sonwalkar:Because you may have raised less than other companies at that point, but generally, if you're evaluating two opportunities, you're both excited about both companies, they've raised the same amount of money, and once a solo founded company, they can give you a lot more equity, potentially, than a company that has three co-founders already, where they might, for very obvious reasons, feel like there's less equity to go around so i'm curious if you've heard anything or had more to share about that just for people who are solo founding because maybe that maybe this is a good precedent to set you know for for people to understand this absolutely well in general people having more equity in your company is good because that means to have more skin in the game right you're um oftentimes you know founders conflate you know equity being like the most valuable resource with like hey i'm never i don't want anyone to have any equity and so you have to find the right balance right if a candidate is willing to take less cash for more equity good sign that's a great sign exactly it's a it's a big green flag it's like okay this person is willing to take less cash and they want more equity instead it's like okay this person wants more skin in the game they want more upside and in general so when we make offers sometimes we'll like present two options which is hey here's a higher cash option and here's a higher equity option and sometimes we'll have candidates they'll they'll say hey can we go even lower on the cash and get more can I have a little bit more equity and then the higher equity option and in general that's good I've I'm not sure what the data says yet but anecdotally I have the solo founders I've talked to like Paul and others um what i've noticed is they are more comfortable you know giving equity to to early team especially if their contributions in the early days end up being more than what they were originally hired for right so the first engineer you probably hire might end up basically being the de facto cto right they they'll probably end up going and building things that you didn't even ask for or like there's a major outage and they'll like drop everything and go fix that outage and so those people you want to make sure you end up rewarding um over time um when they are filling in shoes that they weren't necessarily hired for but they're just like treating this as their own baby right and so you want to bring on people that treat your company as as their baby you know they care deeply about it and when people care deeply about something you know one thing that's going to happen is they're going to have opinions about how something should be done and when you're a solo founder that's something you're signing up for is hey people are going to have opinions about this company and I have to be really good at hearing their opinions and communicating what direction we're going to go in and just because I'm a solo founder doesn't mean I can just tell them what to do because the smart people will leave so people thinking all the time Thinking about your company, thinking about how to make it better, how to make the product better is good.

19:58You should reward those people with more equity because equity is good. It gives them more skin in the game. You have plenty to give out as a solo founder.

20:07Rahul Sonwalkar:I think there's something that's quite different between co-founders and solo founders when it comes to culture. Let's say that there's two co-founders or there's three co-founders at a startup. That second person and the relationship with the first or those second two people in the relationship with the first, they kind of ultimately create a culture with just that small group, just the founders. When you're a solo founder, though, I feel like when you add that first person, it's much different than adding the first person when you have co-founders already, right? Because culture, I think, is often best thought of as the way that people relate to each other and in their environment.

20:48Rahul Sonwalkar:And if you're solo, that you don't really have a culture yet. You have sort of the inputs that would inform a culture once another person comes into the fold. So I'm curious how you've thought about culture and what you've learned about culture when you've brought on people and sort of built out an organization. Yeah, culture is really important. And you're right, when you're a solo founder, and when you're like two people, the other person is 30 % of the company. When you're three people, you're a solo founder, the team is like 66 % of the company. And so they're going to have a lot of influence on how the culture gets formed in the early days.

21:27And it's true, like whatever your culture is when you're three or four people, that's going to end up being the culture when you're 20 people or bigger. I like to think about what Mark Zuckerberg said about building his early leadership team at Facebook. And one of the advice he got from Peter Thiel was, you know mark you should make people that you like spending time with you know your core executive team um and that's kind of how i think about recruiting and culture in the early days too is and i'm going to spend 14 hour days sometimes um just in the office i might as well work with people that i look up to and like hanging out with because if i'm going to spend that long in in the office five days six days a week with somebody i might as well work with people i like hanging out with um and i learned from and they like hanging out with me so ultimately like what you want to end up doing is finding those people that you like hanging out with um you you learn from um it depends on your company culture how you want the culture to be but if you kind of end up finding those people then the culture that gets set in the early days oftentimes will be a an extension of who you are as a person um as opposed to if you find people that are very different than you um you know and you don't enjoy spending as much time with but it's more like a practical marriage as opposed to like a marriage out of love you you you will have this culture that's

23:05Rahul Sonwalkar:very different than who you are as a founder how do you think people would describe julius ai's culture we're very like hacker oriented so um the early culture that you know i used to go hackathons back in college and i like to go from idea to like something live within a day or two at max and so it's a lot of people in the company you know they have idea like go build it feature in julius um go make it live in the product even when it's like not fully tested that are working and see how people use it and then use that to inform whether we should continue building it or iterating on it. And so a lot of the culture is just like tinkering and hacking on different features and different ideas as opposed to here's a central roadmap in the company and this is the exact thing that we're going to work on.

23:51It's very fluid. Luckily enough, it lends itself perfectly to like the AI stuff that's happening. The models are constantly evolving. There's new capabilities that are getting smarter. And so things that weren't possible a few months ago couldn't have been conceived a few months ago are not possible with the latest models and so like this constant tinkering and like building and and like hacking on different features and ideas is kind of like what the culture is another part of the culture is um in general people just like are always just making jokes in the office there's like some pranks going on um or um people make fun of me all the time in the office you know they'll and and so i'm totally okay with people like making jokes about me and you know so there's this culture of like hey it's okay to you know say whatever about rahul like he's not gonna take it personally and so it it makes everything very like light-hearted um so there's like you know we don't have a memes channel in slack but there's like any there

24:47Rahul Sonwalkar:could be memes in any channel yeah there doesn't need to be a memes channel there's no hr approved you know memes channel there's just memes you know there's an infrared debate going on and and means will just pop up about Vercel or about Cloudflare and so on. And so that's just who I am as a person. If you and I were having a serial discretion, there would be sometimes where I'll just go off on a tangent and just start making jokes about something, and that's just who I am as a person. And so that's kind of what the culture ended up being in the company. The sort of tinkering and sort of DIY aspect of all this seems extremely beneficial for kind of what you're working on.

25:24Rahul Sonwalkar:But I also know that you went through six ideas, something like that, before you got to what you're building with Julius today. Maybe you can share a little bit more about that, because I'm guessing you were doing that mostly solo through those ideas, if not the entire time. And I think that it would be really interesting for people to hear just, you know, what that actually was like, how you knew when to move on to another idea. and just I think that very often we hear the story of oh there's one pivot and then and then we find this amazing thing people people know about pivoting obviously and they and they they sort of glamorize pivoting maybe but I think that six is is actually not a crazy number I've heard a lot of people over the years who've gone through similar journeys where they're pulling threads and they're hopping to new ideas so I'm just curious what that was like for you and particularly if you were doing it solo yeah the first few were definitely solo um you know before i started even started the company um i spent like half of covid building different mobile apps you know i got really into podcasting and listening to podcasts and during covid because it was just so lonely you know you could barely hear any humans talk uh you're just in your apartment all day and so i listened to podcasts all the time and i was building different ideas in that space and my my plan was hey once something starts working i'm gonna go all in on the idea and after about a year of doing that i realized man nothing is really working and i'm kind of just like tired of waiting for something to work so i'm just gonna quit my job and just go all in anyway so this is all this is all sort of pre quitting your job you spend about a year sort of working on this stuff part-time yeah um and but then and then the six pivots happened after that got it so that was started the company, I was seeing GitHub Copilot kind of like starting to work and it was helping me as a programmer.

27:22And I wanted to see if it could help people write queries and analyze data. I didn't know what that was going to look like, but this was like the loose direction I wanted to explore in. It took six pivots from that like loose direction and a lot of improvement in model capabilities to arrive at Julius. And along the way, you know, the hardest pivot once you have started a company is the first pivot because you've told everybody you're going to work on this idea and it's going to work and you've told your friends your family your former colleagues um your your you know girlfriend or your boyfriend and then a few months in you realize oh it's kind of not working um and shit i gotta i gotta i gotta pivot now but i've told everyone that i'm gonna work on this thing and man i don't want to go to this thanksgiving dinner and like tell people I'm working on a different idea now it's going to be kind of embarrassing so the first pivot is the hardest but once you get past the first pivot when you're a formal company then you're kind of free to fail again and what you want to do is keep exploring, iterating until you feel a pull and then the moment you feel a pull you double down now I've talked to solo founder friends that are still pivoting and because they're completely solo, what has kind of happened is they don't have much of accountability.

28:50And so at the slightest friction or sign of trouble, they're like, okay, it's not going to work. Let me go find something else. Turns out every great idea, there's a ton of friction to make it work. You're pushing a rock uphill and then when you hit PMF, it feels like it's all rolling downhill. But so I think having people around you as like a founding team or like a co-founder, I think builds up some accountability. Like, you know, we're not going to pivot at the first time in trouble because we have invested so much time and energy and mental focus into doing the first launch. And so I think, you know, when I brought on a couple of my friends to be part of the founding team, that actually ended up building a lot of accountability in me to like not give up on the or like go back to the whiteboard at the first sign of friction and to kind of like continue exploring um and ultimately like when you're pivoting in a direction the good thing is all the learning is compound right even if you build and launch something and doesn't get users or you get you get few users but they don't retain that's all learning that you can then applied to the next iteration of the idea and so kind of having a direction having like a couple people that can hold you accountable is is good what was the thing that kind of got you to want

30:13Rahul Sonwalkar:to focus on that direction because i think that that was one of the things that maybe when i was listening to prior conversations you've recorded with folks it sounded a lot more like hey you know i tried a bunch of different things but i don't think that you ever shared as much about the through line through all of those ideas? And what was the thing that kind of inspired that? And the reason I ask is because I think that oftentimes people, when they're trying to figure out what it is that they want to work on, sometimes they just start with anything. But it seems like you started with this premise that you wanted to explore.

30:50Rahul Sonwalkar:And then you kind of just almost, it's funny because I think that your logo is like a bust, right, of Julius Caesar. It's like revealing the the sculpture in the marble right where you're chipping away at the marble block you're kind of finding this uh this thing underneath but it doesn't necessarily look or present properly for a while until you actually kind of reveal the sculpture so i'm curious what that was like how you thought about kind of what you wanted to pursue and what advice you might have for people there in terms of finding that general area to sort of direction to go in i love that i've never actually thought about it as like you know taking a piece of marble and and you're kind of building, chiseling the sculpture.

31:29It kind of doesn't look good until it's the final product, but that's beautiful. I think I might use that again. But I think that's the perfect metaphor for building a company. You look at, one of my favorite things to do is to go to archive, internet archive, and look up websites of the products that we admire today.

31:51Rahul Sonwalkar:On the Wayback Machine. On the Wayback Machine, exactly. And like Figma, back in the day, or like Zapier back in the day or like, you know, OpenAI first website. Like they, you know, it looks so, it looks like anyone's hackathon project. And I think that's kind of awesome, you know, because all great things start as like something very basic or simple looking. You don't have to be perfect from the start. And so you need to have conviction in something. and the the way you know it evolves could change so for example um you know the cursor guys they were doing autocad and then there's michael has this like old hacker news post about you know cursor as like an email autocomplete tool and email is so different than coding um but you know that's so so they kind of like spend some time in even figuring out what direction they want to build in they knew ai was going to be transformational they didn't know like what their calling was and so everyone sort of has excited different paths to how they get to a place for me it was i was an engineer at uber and so i was obsessed with like you know like um data and like um how can data and data science help people and i was seeing program you know programming get really really good with tools like copilot this is pre-chat gpt and i was i wanted to see if they could help people analyze data right i thought data is like the you know the bare bones of every business decision out there and so if you can help people analyze data um in plain english that's awesome i started with like logistics data because i was in this like local minima at uber uh after doing stuff in like in that space for like three months i realized man it's nearly impossible to convince a logistics company to adopt ai and like use ai to analyze data but what we're building seems promising there's something here and maybe like we just need to build a general tool and so we built this thing called excelcopilot.com which was like a copilot that would help you write excel formulas um and we got this cease and desist from microsoft and they're like hey guys you're calling your thing excel and copilot that's double bad like both of our both are our trademarks and so you guys gotta shut down right now but the early usage on that thing gave us more conviction okay there's something with like people analyzing data with natural language and then so some of the next things we launched was okay what if we query every single data point about the u.s census and make it really easy to like analyze that data um with plain english and ai will just like write SQL and write queries and do the analysis for you cool okay let's go build censusgpd.com and so we build that launch as a demo open source the code um and a ton of people use that on the first day you know you can ask all these questions like hey what is the gender ratio for different neighborhoods in SF and people love asking that stuff um and so we build that we launch that and a lot of users first day, nobody uses second day.

35:03Because it's not a problem people have. People don't have questions about the US census area.

35:07Rahul Sonwalkar:It's a fun thing to play with. Yeah, it's a fun thing to play with. But that gave us some more conviction that, okay, the technology kind of works. And people are curious about data, but not census data. And so we kept iterating in that direction. And it took us another four months and a couple more launches where we just tried more public data sets until we realized, okay, the thing people actually care about is their own data, especially if you want people that use it daily or weekly, it's people that use work data. And so we build that, bring your own data, analyze it, launch it, and launch it, and then that became Julius.