Why PostHog's 6th Employee Went Solo | Yakko Majuri, Railcode

0:00Yakko Majuri:Are you holding your team back? There are companies out there, we're talking about being AI native, and they are holding employees back without realizing that they're doing so. They're like, hey, we've given everybody Cloud Code, we've given everybody Codex, but what's the next step? How are they being held back? They had to wait a week or more to get something that they've built live because it needed to be approved by DevOps. Most of the software going forward is going to be internal software. The week that we closed, I actually spent 10 hours this week supporting this, and now the CEO can build everything himself.

0:27Yakko Majuri:I just fired off a cloud code and I just said add dark mode support, wait like five seconds, and then the deploy is live. What's the advice for solo founders? Get in touch with yourself and things you want to build. It's fun for me to build real code and fun is useful. Welcome to Solo in Prod. This is a new experiment that we're running where we highlight startups in the solo founders program just as they launch. So you can get a sense of why the founder is working on the company in the first place. the insight they have, and the inspiration that caused them to start it. This first episode is featuring Yakko, who is building RailCode.

1:05Yakko has particularly interesting insights into internal tooling at companies, and he actually built the first version of Session Replay with a team at PostHog, where he was one of the first employees. Yakko believes that most of the software that's built in the future will actually be internal-facing versus external, and Railcode is his solution to help these teams build much more effectively. I'm really excited for this episode. I would love to get your feedback on Solo and Prod in general. And if you're interested in Yakko and his company, you can check out the show notes to get all of the links.

1:43Also, if you're building Solo or you're thinking about it, you should apply for the Solo Founders program and join other people like Yakko who are building Solo together. You can check that out at solofounders.com. Jaco, this is really exciting. What is RailCode?

1:59Yakko Majuri:Yeah, so RailCode is a platform for building internal apps and agents in a secure environment. So what we're seeing is that people get access to cloud code these days, and they want to build software that works well for them. So we're providing a secure place for them to both experiment and actually build serious apps that go in production internally, and that help other people, and that can be shared around the company. And I think that like historically people have used tools like Retool to essentially build like different interfaces and things like that for people to use internally at companies, internal tooling, right?

2:35And historically also that meant that like all of the code was kind of like on Retool and like all of it was sort of like in their walled garden type situation. in this case though you essentially had an insight that there was something that was fundamentally at odds with like the way people are building these apps these internal apps these mini products can you share a little bit more about sort of how you came to that insight like what you were working on the types of conversations you had that led to that so the best harness is my harness

3:09Yakko Majuri:right so everyone's building their own harness and their product you go into every product and there's like an AI box there that you can talk to. But I use Cloud Code every day. Some people use Cloud Desktop. People use these different things. I use Codex as well. All my skills are there. All my workflows are there. It can browser test for me. It has all kind of my testing set up. So what I'm building is basically enabling people to stay within that workflow and just have a secure place to put this software that they're building. This came from speaking to a lot of people where they were either moving off of retool or they had retool but like people were spinning things up in other other places like they were putting things like on versell but then someone puts it on versell someone puts it on cloudflare it has no idea what's going on so i'm going the difficult route which is hey build on your own harness i don't have control over that so it's really hard to build but that's where i think the future is going i don't want to lock people in the code is their code put it on github run cicd but still have a place to kind of deploy this and share it with your team really fast and govern everything that's going around there so like i'm curious you know if you were to have a hypothetical question that you'd ask someone to really help them understand like and help you understand whether or not rail code actually solves a real problem for them what would that question be it would be are you unable to keep up with your team's development with AI today?

4:42Yakko Majuri:Basically, people are moving fast. People are experimenting. They've figured out that they can do a lot that helps them in their day-to-day job and the company's unable to support that. So that would be my question. Are you holding your team back? How are they being held back right now? So the customers that I have, most of them, they had to wait a week or more to get something that they've built live because it needed to be approved by DevOps. It needed to go through kind of a specific process to make sure that it follows the standards, that it has the right authentication, that it can connect to a database, that it can use these API keys.

5:19So somebody that builds something

5:23Yakko Majuri:that is really useful and they want to share with the team, they get discouraged. I don't want to go through this whole process. Some of the companies I've spoken to, some people have just given up. They're just like, okay, I just run it locally. It's fine. like people don't i teach people how to prompt their own thing into existence and it's not multiplayer so there are companies out there where we're talking about being ai native and they are holding employees back without realizing that they're doing so they're like hey we've given everybody cloud code we've given everybody codex like you know we're you know ai forward but you know what's the next step it's really interesting because you know when you talk about just like, hey, it's a week delay from somebody vibe coding some internal tool to it going live, that's actually the least painful part, right?

6:08Because the painful part is the thing that doesn't happen because people are, as you said, discouraged. Or, you know, I think another thing that's really interesting about this is like, if somebody doesn't even know that something is possible or permissioned, they just won't even think to necessarily do it. So it's like until people had access to all these great tools to you know build things and like a lot of people just weren't building things because they didn't know how to code right but now we have this new thing which is like you don't even know that you can actually build like the types of internal tools that are possible or or you wouldn't even just try because you don't think there is this way to actually get them into other people's hands or to even easily use them yourself so i think there's something really important about this, which is not just it's sort of saving people time or making it easier.

6:55I think by introducing something like what you're doing with rail code, it might actually help encourage companies and individuals at companies to start to build a lot more that will actually help them and help the team, you know, do the things that really matter for a company, which means, you know, serve customers, right? You know, get new customers and also like coordinate internally. These are important things. And once you give people the ability to do that with the tool by essentially enabling more people to build and deploy stuff internally, that seems like a really big opportunity.

7:30Yakko Majuri:It's two sides of the same coin. There's people that are discouraged and then there are the people that aren't encouraged. And it's really hard to measure the second one. It's like, what are the potential gains you can have when people see what everyone else is doing. This is something that I actually lived at my previous company where basically we didn't have a customer-facing product. We were developing algorithms for gesture detection for wearables. And so basically everything that we did was internal. And I came in to kind of basically build that whole internal platform, like both the data platform, but it ended up being a lot of internal tools.

8:05Yakko Majuri:And there wasn't a place for this stuff. And once I set up the infrastructure in a way that, hey this is a place where we can build these things so much stuff came out of it that made model training faster we collected more data we were able to get better models leadership had a sense of what was going on and could direct the company better it was like it made things exponentially better um and company was acquired by Oda now um and uh having that internal engine moving was something that was really important in that case and in that case you know when we talk about internal tools, sometimes we think about an external facing product and then internal tools as well as like a separate thing.

8:45Yakko Majuri:But for some companies, the internal tooling is the product. Like we didn't have a customer facing product. Well, not an open one that people can kind of sign up self-serve to. We were selling algorithms to companies that were manufacturing these devices. Most people think that starting a company without a co-founder is a bad idea. We believe the opposite, that solo founding will become the default way great companies get started. And that's why we built the Solo Founders program, where solo founders spend three months in San Francisco building solo together. Our alumni say it's like having co-founders but getting to make their own decisions.

9:22And four cohorts in, we're starting to see just how much solo founders can accomplish. One secured a formal partnership with a major U.S. government institution just five months after incorporation. Another hit 20 ,000 GitHub stars and 3 million downloads. And a third got 2 million ARR before making a single hire. The demand for the program continues to grow. Our first cohort got 1 ,000 applications for six spots. Our fourth cohort got 4 ,500 for 10. And today we are officially opening applications for cohort five. We'll select about a dozen founders. I'll work closely with each of them and we'll invest$100 ,000 in every company.

10:02We kick off September 10th. If you're building Solo or you're ready to start, apply at solofounders.com. It seems extremely interesting to realize, and I didn't really realize this until we met and we started talking about this. And also we should talk about sort of the history of kind of what you were working on before this and what led you to work on this specifically. But it does feel like in many ways a lot of things, there's sometimes we build things that are customer facing. And a lot of the abilities of models and code gen have really gone to us building new features and new products for customers.

10:44but in a way it feels like in some in some ways we've kind of like um not spent as much time in giving the uh the abilities for people to do things internally um and it feels like this is something that's lagging behind the customer facing it feels like a really big opportunity and one that seemingly the people who are using rail code seem to really be lighting up on can you share a little bit more about you know you don't have to speak about the specific customers by name but in the abstract you know there was this there's a customer who is working with you know ceo um cto um it's a relatively small company what was that like when they were using

11:22Yakko Majuri:the product yeah so just to tie back to what you were saying i mean i think most of the software going forward is going to be internal software so we want to enable that um as far as the company that you're speaking about the ceo is versed and well versed in like using cloud code but But the CTO was not confident with the CEO deploying things. You know, you can tell a coding agent to deploy something on Vercel and it will do it. The CTO was not confident that it was going to follow the right security practices, that it was going to have the right authentication. And wasn't confident like handing the CEO like, hey, here are some like database keys that you can build on top of.

12:01Yakko Majuri:And like maybe you're going to put them out there. And like we never know if this thing is actually secure. so what would happen is the ceo would go to the cto and they would say hey i need this built um and they're a small company you know you're taking time from the cto that should be spending time on product and sometimes the cto was spending you know it's like oh it's it's going to be a two hour thing and then it becomes a four hour thing and the cto the the week that we closed he was like i actually spent like 10 hours this week supporting this and now the ceo can build everything himself um and the cto is fine because he knows that real code is like a secure environment for this um there's all the permissions are there all the connectors he's kind of audited all of that and he's happy with the ceo being able to build things on top of there um and actually he's been looking at what the ceo is building and said well actually i can build things here too and it's going to be faster even than even me as an engineer it's going to be faster than me trying to set it all up the right way to deploy it elsewhere um so they've been building a lot they're kind of making rail code the place where work gets done in a lot of ways particularly now that we have agents and they can interact with the data that they've brought in and the apps that they've built on slack for example so they're doing some really interesting stuff with rail code it's really cool because in a lot of ways when you think about permissions and things like that you think oh this has got to be mostly used by larger companies that really need to like tamp down on things right make sure that everything is secure.

13:26But at the same time, small teams, small companies, as soon as you have customer data, you need to be really, really mindful of like who's getting access to things, how they're getting access to it. And ultimately it's really great to see that in this example, you're helping the CEO get unblocked and you're also helping the CTO not have to spend as much time on building things that are important, but aren't core product. I think that that was another thing you've told me is that, you know, all this stuff is important, but it's not core product. So sometimes the people who, you know, who might have to, you know, like deal with the requests and the work associated with getting these non-core products, these internal, important internal tools out there are, you know, essentially getting pulled away from core product work and customer work, which, you know, is not great.

14:15I'd be really curious to hear and for you to share a little bit about kind of, you know, what you were doing when you and I met and you joined Solo Founders Program and a little bit of the backstory before the company that you were working on most recently that you joined and then was acquired by Aura. Can you share a little bit more about the backstory and then sort of like the context in which you joined Solo Founders Program?

14:40Yakko Majuri:Yeah, so I'm an engineer, always been an engineer, full stack, but always had a focus on backend and infrastructure. I found that dealing with complex infrastructure is really motivating and particularly because if you do a great job, people don't see what's done, but you're simplifying something really complex into something that just works. Was one of the early employees at PostHog, owned a lot of the infrastructure there and also built internal tooling there. joined this company, finished company called Double Point, where I was owning infrastructure. And again, it was building this whole kind of internal product.

15:19Yakko Majuri:And when we met, I was early in my founder journey. I had pivoted off of another product and I was validating an idea that I invalidated for a few kind of different reasons. and I was doing a little bit of soul searching and I was looking back at all the notes that I had from people that I had spoken to about the previous ideas that I had and products that I'd worked on and correlating that with my experience and things that I've done in the past and this was like a clear fit it's um it's enabling people to do things that they weren't able to do before and making it really simple for them to do so.

16:07Yakko Majuri:There are really complex infrastructure problems that I have to solve in the background that hopefully you don't even see them or you never feel them. They're just solved for you and that's really motivating. And it ties back to things that I did at PostHog and things that I did at DoublePoint. So it was quite a natural fit. I looked back at my notes. I had people mentioning this problem in relation to other things that I was working on. So then decided to jump on this. And coincidentally, actually, or maybe not, some of the stuff that I was working on in the past actually builds up to what I'm building now.

16:42Yakko Majuri:So I think in some ways it's actually an evolution. The previous product was also a security infrastructure-minded product. And I've actually reused some of the code in RailCode. It's really cool. And I think that there's something really important to note there, which is it's that through line, right? You talk to a lot of people who build companies and oftentimes maybe they're skating around various ideas, but usually the thing that they end up working on is some sort of continuation, some sort of thing where we call it like pulling the thread. You continue to pull the thread. You continue to sort of explore an area and see where it takes you.

17:17It's really cool to see where this has taken you. And I remember sort of the time when you were sort of going from the last idea to this one and sort of that journey of like reflection and trying to look through your notes and think about things that you enjoyed. And I think that's another thing that's really important. It's like I remember we were talking. You really enjoy the stuff you're working on now. And I think that it's very easy when you're sort of in the idea search space. You can kind of sometimes get sucked into something that maybe sounds like an attractive business idea. but if you're being honest it's maybe not the thing that actually causes you to light up and I think that some people really need to not just have an attractive business idea even if it starts to get momentum they really want to have something that they have some like passion in or some experience in and it really feels like rail code is that for you oh yeah I mean that's why

18:04Yakko Majuri:like everyone's talking about my website and my website has like a little bit of a playground in it and it has like a roller coaster and everything and and what that website shows is like it's fun And the thing is, it's fun for me to build real code. It's actually, I'm having a lot of fun building with real code. And actually, as I was iterating on the website, I actually, I was having a conversation with a customer and I asked him, is real code fun? We were talking about like how useful it is and all these different things. But I stopped and I say, is real code fun? And he was like, yeah, it's fun because building things is fun.

18:38Yakko Majuri:And you're allowing me to build things fast and without having to deal with the stuff that I don't like taking care of, which is like the permissions and the security and the infrastructure. And you're not taking me out of my workflow either because I'm using, well, in their case, open code. So you're not kind of forcing me into something else. So it's fun. And I think the website shows that because I'm having fun building it and I'm having fun using it. That's awesome. And I feel like if I were to think of like a sort of a big company that's like high growth company right now that I feel like really exemplifies fun, at least in the way it presents itself to the world, is PostHog.

19:12so can you share a little bit more about like how I think that a lot of people are like wow you're super early at post hog like how did that happen like how how did you find the opportunity how did they find you uh because I think that would be informative for people who are maybe looking to join something that becomes you know the next post hog um which which is rare but it's like it's possible so I'd love to hear that and then I'd also love to hear sort of what you're thinking is around fun and like what the team over there does that makes it fun um sort of what you took from inspiration from them?

19:43Yakko Majuri:Yeah. So I got to them because of the handbook. I think a lot of people did. Um, and this was what, five, six months into the company and they had a whole handbook written out about how the company is going to operate. I think it wasn't even about just how the company operates, but about how it's going to operate. Cause there weren't that many employees to be operating in the first place. Um, so the handbook was what caught my eye. Um, and back then we were doing work trials like three-day work trials and you could see from the work trial that the culture was what was told in the handbook it wasn't like wishy-washy stuff um so that's what really got me interested and i think it matched really well with my personality um and i feel like i came in because i related to the culture but consequently i added to the culture as well um and i think what they do right i mean they do a lot of things right um but a lot of autonomy to people that They hire great people, people that align with their values.

20:39Yakko Majuri:The values are for real. And then people that work hard and do great work. And they give them a lot of autonomy. No teams there are larger than six people. People have the freedom to move around, build new products, start initiatives. I think most of the products, some of the core products have come out of hackathons. Session replay is one of the massive things they do today. I was part of the hackathon team that built that kind of v0 of that um I won't get most of the credit it was Tim and Carl but like I was I was there and it was built out of hackathon and so were a lot of products so yeah I think there's a lot of stuff that they do right but the autonomy is kind of like one of the top ones and and the transparency which is reflected in the handbook two quick questions on that before we go to the demo which is like I think something would be really fun for people to see even just the website we could scroll through um okay so what is uh what is something that you feel like you've applied from either that post hog or or the company after um and and then like also mascots like what's like how important is a mascot um where where's the rail code mascot yeah i don't have a mascot yet i need to figure one out yeah do you think it's important though like is there aren't that many mascots if you think about it is it important i don't know if it's important but it is it is i'm sure it's useful certainly fun it's fun and and fun is useful um yeah i don't know if i can say it's it's necessarily important but it's something for me to potentially think about um and um what was it what was it i mean in terms is there something that you've already applied in terms of how you're building rail code that you either took away from post hoc or other work experiences i think a lot of how I do engineering is very related to how they do it there.

22:28Yakko Majuri:Move fast, iterate really fast, respond to customers, learn from customers actively. I think there's a lot of stuff about how I want to build a company that has a lot of synergy with how they build a company as far as when employees do come in and things of that nature. But right now, I think just basically a lot of how I am as an engineer aligns with how they do engineering. And that goes both ways. It's not like, I mean, the culture influenced me, but also I joined because the culture was a certain way. So as a solo founder right now, that's kind of how I apply it. It's a really fast iteration and learning from customers really actively.

23:05That's awesome. Why don't you share a little bit about kind of what Railcode looks like today? And we'll wrap up, but I think this has been amazing. Thank you for doing this. Let's hop into it.

23:15Yakko Majuri:Cool. So yeah, this is my website. I'm pretty excited about it. I mean, of course, I built it. And so it's a reflection of you. And I think that's honestly like what the most important thing is when you're getting started is you want somebody to represent who you are. Yeah. And I want people to see that. You know, this is something that I care about. So this is the website, but this is the app, right? This is the more interesting part. So this is like a demo organization. And like we talked about, Railcode is this secure environment where you can put these tools. Tools meaning apps and agents.

23:44Yakko Majuri:And they weave pretty magically inside of Railcode. like apps are immediately accessible to agents without any extra work. So you build something custom and an agent can already operate on top of that data. So, because I think that's the future where we're going. Right now, it's a lot of UIs that have agents. And I think we're going towards a place where it's maybe agents that have UIs. So we're kind of going towards that. But I was preparing this demo here and I noticed that this Kanban app is actually on like light mode and it doesn't seem to have dark mode. This is like a demo organization.

24:14Yakko Majuri:so I spent I just I just fired off a cloud code and I just said add dark mode support um and this is like you know a small thing it's not like an actual but I just wanted to show like what the development work looks like with real code where I'm building in cloud code and then I'm going to deploy I could tell the agent to do this but just for the sake of of the demo I'm going to go ahead and say real code deploy um and then we're going to just wait like five seconds and then the deploy is live and now it has dark mode that's cool and essentially what's happening here is you're using cloud code which is the thing that you're used to if you have any skills set up i mean maybe you're using impeccable which uh which paul was on the podcast uh for your design stuff like there are different skills and things you have set up the way you do it you're not having to use some other interface um and then also you're able to pull in your own data sources that have been permissioned across rail code exactly so that's exactly right so basically you're building locally and we're governing all the data connectors um with granular permission so you connect your postgres and you say hey julian can access this on our postgres instance and yakko can access this um or hey we connect google calendar and these people can access it or these people can access it so we do have another demo app here which actually has some data um which is this oh we just went to kanban i just clicked it again um which is the crm so this crm here kind of showcased a little bit of like all the real code can do building this didn't require explaining any technical details around hey get a database hey get file storage it just works so you just say hey i need a crm i want to store files on my customers i want to store data on my customers and it just has the primitives to build on top of so that's kind of the magic part that we do building with cloud code but we do have this magic sdk that connects to these provided services which are like database file storage llm apis for example so the crm i can manage it from this chat right i can i can ask things and i can create deals right from this ai feature and a non-technical person could have built this because they just need to say what they wanted and it just has all the primitives to build it But at the same time, an engineer that wants to own the code and have it on version control, they have it too.

26:32Yakko Majuri:So it serves both purposes. And we were talking about connectors. This CRM specifically connects to Granola, for example. So I can connect my Granola account and import data from there. And all of these connectors are governed within this real code space. Super cool. I mean, like, you can just imagine, and this is not, this is, you know, this is real. Like, where people go and they just vibe code stuff. and then they're like put a little password line in there you don't need like a password input it's not like actual secure stuff and then they deploy it somewhere on like versell or something and you have this you have all these issues with like people you know in these companies like building their own little apps and then trying to deploy and like this is the this is the thing you're making it so that people can do it login is already handled for them right all of the all the access to the right things is handled so you make it super easy for people it's it's really cool.

27:24This is awesome. Like you just launched like two days ago. Or if you're if you're watching this, maybe it's maybe it's about a week ago. But this is like the official demo. Anything else you want to say is like words of wisdom, or comments, like where should people check this out? What's the URL? It's going to be in the show notes, of course. And then also like any advice for solo founders.

27:45Yakko Majuri:Yeah, so basically, there are a lot of teams out there that are doing what you're saying, like where they have like these tools where it's just like a password input, and there's a shared password that they share around on slack and it's like single player like everyone acts as the same user like real code by default it's like multiplayer like it already knows like who the users are so you can find it at real code.dev um so that's yeah real code pretty self-explanatory how to spell that and advice for solo founders i think it would have to be to just get in touch with yourself and the things you want to build because that's the way you're going to build something that means something to you that means you're going to work harder that means you're going to stay at it longer that means you're going to enjoy it more and consequently you're going to serve your customers better it's clearly been like that for me early in my journey very very clearly right but it's working on something that I'm excited about and that's like an extension of me it's it makes things a lot easier.