Million Dollar Exits Ep.10 with John Rush: On Bootstrapping, Acquisitions, and Market Trends

Since our vision is to grow, you know, by the community, with the community, with the other founders. So when I say community, I basically mean other founders.

Hello everybody, I'm KP and welcome to Million Dollar Exits, a special segment on my main show, the Building Public Podcast. In this interview series, I sit down with founders who went from an idea to building a business and then eventually selling it for over $1 million. Hence the name Million Dollar Exits. This is something I'm very passionate about and curious about at this point in my career. I wanna learn and share all the insights, mindset shifts, lessons, and non-obvious tips that are part of this unique journey with the next wave of founders who wanna take this path as well. So buckle up and get ready to be inspired and informed. Here's a special shout out to our episode sponsor, paralect.com. Million dollar ideas come from every possible niche. If you're a busy domain expert, in marketing, sales, finances, or healthcare and don't want to spend 6 months just to build an MVP, you'll find Paralect super valuable. Paralect is a venture studio built to design, build, and launch a product for you that is ready to sell in under 2 months. Start with no-code or go full-stack right away. Simply focus on growing your early adopter community and build in public. And they'll take care of the rest. Build your million-dollar startup with paralect.com, P-A-R-A-L-E-C-T.com. In today's episode, you'll meet John Rush. John is the founder of MarsX, which is a dev tool that's making $1.5 million in annual recurring revenue with over 20,000 developers in the ecosystem. John has been a software expert and entrepreneur for over 15 years and has made multiple acquisitions, including Unicorn Platform. By the way, you might remember that we had the founder of Unicorn Platform, Alex Azoura, on the podcast earlier this year, and that's how I got to know about John. On this episode, we went deep into the trenches talking about AI, John's vision for MarsX, how and why he acquired Unicorn Platform, and his advice on how to validate new ideas for bootstrap founders. I had a blast recording this episode and hope you'll find value in it too. With that said, I bring you John. Hi, everyone. Enjoy. Hello everybody, welcome, and I'm your host KP, and today we are here to, um, do another interview, um, on the Million Dollar Exits podcast. I'm super thrilled to invite John Rush into the show. John, welcome to the show.

Hey, thanks for having me.

John is joining us from Istanbul, which is his, uh, home now for, for 4, 4 years almost. And we were just riffing about the tech scene in Istanbul. Like, so, John, let's start there. How would you describe the sort of the blossoming tech slash indie hacking scene in Istanbul? What are some observations you have there for any of the listeners who might be curious?

Yeah, 4 years ago when I arrived here, like, I think there was no tech scene as we know it now. So there were a couple of bigger big startups acquired by other big startups. But now I think the indie tech scene is really growing fast here because the new generation of Turkish people is looking at the West and they want to get this financial independence. And they use Twitter a lot. Twitter is one of the most popular networks here. And every few days I find somebody on Twitter who who is actually from Istanbul, while I was like not searching particularly for that, right? So that's kind of good. And I'm meeting people who are doing other projects and I like it here. So I think in the next maybe 10 years, Istanbul probably will be one of the best growing tech scene for indie projects, like not the VC-backed projects, but this bootstrap self-funded stuff in Europe. I have high bet on that.

Right. I love that vision because a lot of our listeners are self-funded or bootstrapped. I think I'm personally, I also had 2 bootstrap startups this year, you know, both SaaS, and we can get into it in a second. But like, I feel like you can go a long way in 2023 with just bootstrapping, right? With the rise of no-code and AI tools and sort of the You know, it was very expensive to run servers and like to have all these advanced functionalities for a lot of the SaaS ideas you may have in the past, but now the costs are going down, which means you can take more shots, you know. What is— I mean, you've been in the game for 15 years in the indie industry, right?

Exactly, since 2008. Yeah, exactly 15 years.

Yeah, so how would you characterize, like, if you have to divide this into 3 or 4 eras, You know, and you're from your own career, how would you characterize the trends of like 2006 to, you know, maybe 5 years later and then 5 years later?

Yeah. So it started very different than it is now. So I remember when I was in university, like we all want to build something. But, you know, we were not like the world back then was not the web, it was applications or games. So people would either think about the game, A lot of people would build game back then. And then others would think about some, you know, this desktop application that's a file at the end, and then you would kind of distribute it somehow, right? And that was the beginning. And then suddenly the web came in, Facebook movie came in, and we all saw the movie, The Social Network. And then we all started building some web stuff. And yeah, so it was like 5 years of building web stuff. It was easy, so you wouldn't really need funding then. We didn't even think you can get funding. So until 2012, getting funding was nearly impossible for most developers. Like, in my network, and I had large network, like hundreds of developers, there was nobody who got funding. We knew one guy who got funding, right? And like he would tell the stories in university about that case, how they got funding, but it was very rare. So the VC scene was quite small. But then in 2012, things changed. So the VC scene just exploded, especially in Oslo. Like, Oslo was having no scene at all, and then suddenly the government put money into startups, and then VCs come in. And from 2012 to 2018, like this is large kind of chunk, like 6 years, it was very easy to get money. Almost everybody I knew was running a startup and almost everybody I knew managed to raise VC round, at least seed round. So things were going quite well for that sake. Right. But one observation that was quite interesting. So the people who started early, who went through this struggle of not having VC funding and, you know, doing it on their own. And then when we got the money, it really helped us.

Right.

But then the people who came into the game and got money straight from the beginning, like the failure rate there was huge. Like in my batch where I was in like 2010, like the first batch, it was basically the first 10 startups of Norway, I would say. Like we had 4 unicorns. Wow. Yeah, it's Kahoot, probably you know the company.

Yeah, Kahoot, yes, wow.

Yeah, that's our batch and we have same investors and then Xaneta and, you know, 2 more. So it was really good batch and very successful. you know, 4 unicorns and the rest did well as well, right? But then the next one was worse and then it was worse and worse, right? So that's what I basically had in this regular startup world until 2000, um, like 2018, I guess. And then in 2018, I entered this indie world, like bootstrap, self-funded, solo, etc. Uh, and I did that because I made some money before. I had, you know, this portfolio of investments I made. I've been doing angel investments and then I had startups I've co-founded. So I had this revenue stream coming and I didn't really want to have an exit because if you have an exit and you have a lot of money, you have to think how to store it and that's not easy, I would say. And then I thought it's better I have revenue and I use it for my next ventures. And then I turned into this world And there I am, like, 5th year I'm in this. I'm not solo. I have co-founder, I have large team, I have a lot of people around me. But first 2 years I was alone completely doing everything on my own. And it was cool. Like, after having, you know, large teams and expectations and investors and everything, running the stuff yourself on your own pace, own speed, it's kind of fun, but not for too long. Yeah. Right? Like once people start using your product, it's not fun. It's funny when nobody is using your product.

Once you have people that expect, you know, great things and you're like, oh, now you gotta, you gotta ship at that velocity. Yeah. You know, deal with the support tickets. You get a boss. Right. Yeah, you get a boss. Micro bosses, right?

Exactly.

So, you know, that's fascinating. So let's talk about, you know, your biggest focus area now, right? Which is MarsX. And I wanna understand how would you sort of describe what Mars is for the listeners?

Yeah, so the main goal of Mars is to accelerate software development. And we do it through reusing ready-made microapps. So all these years when I was building stuff, every project we would have to build kind of from scratch. Like, you know, the user profile, the search, the gallery, like all the features. We would manage to reuse some stuff, but you know, as time goes on, the stuff you've written 2 years ago, like nobody wants to use that code anymore, right? It's just like not a cool code. And I just realized that there's no good way to reuse your own code and the code of the others. And then I thought like, it's kind of really, really bad for the overall progress. It's like in science, if you do something, you can use all this stuff other people have done, all the papers, everything, right? Right. But in software, you pretty much start from scratch every single time. Like, you know, 1,000 teams in the world right now sit down and write Uber code, like another, like, you know, 1,000 versions of the same thing, right? Yeah. Airbnb, Uber, and all those things. And I felt like it has to be changed. And then, you know, that was like why we started Mars. So it is the microapp reusing platform, I would say, but it's a lot more than that, right? It has its own IDE. And then the second big thing we have is that we realized that people actually struggle on picking no-code, low-code, and professional code. You have to pick one at the beginning and then you can't really change. And then we thought like, what if we do something that you can actually switch within the same project at any time between the interfaces? And then you don't have to choose. Wow. Like, you know, some parts of your application are really, you know, going well with no-code and some need Pro code experience, and then now we added AI there. So it's kind of like you know three interfaces on top of one project, and below the project you have hundreds of micro apps that you can take and use. And then you basically save like 95% of the time, and then 5% is left to actually build what's not there and to connect things together.

Right. It's brilliant. I mean, so I'm just reading out some of the stats around Mars. It's 1.5 million. ARR, correct?

Yeah.

It was the last based on your Twitter bio. 20,000 devs and 300+ micro apps in the showcase. That's crazy. That's amazing. Like, I'm curious when you say 300+ micro apps in the showcase, is it that these 20,000 developers or whoever, like basically these are apps that were built by other people using Mars? That someone brand new can come in and then also reuse that? Yeah. Or— yeah, okay.

Yeah, so the idea here is that comparing to other platforms which want to, you know, build everything on their own and then kind of provide it to the users, like Supabase or like, uh, now all of them, uh, we realized that it's really difficult to actually do that because Time moves on and the code gets obsolete really fast. Yes. You know, and then

if you centralize the approach as opposed to decentralized approach.

Exactly, exactly, exactly. So the only way to do well is to centralize this so that you can move really fast, or you have to burn insane amount of money to have a lot of people creating stuff. And we had that option. Like we had actually, you know, in 2020, we had enough investors interested in Mars who wanted, who kind of pitched in the idea that let's raise in a lot of money, like, you know, $50 or $150 million to actually have, you know, one team per microapp. And then we have, you know, like 400 developers building that stuff. And we didn't go for that. And we didn't go for that, why? Because, you know, it's not like you have to just build those things once and then you have them, you have to keep building them. Right. It's like little products, right? Right? You need owners of those little products. And then I realized that there are a lot of people outside Mars who wanna build products and own the products and make money on those products. And here are the products you can kind of build and own and make money on.

So when— it's actually quite fascinating because I don't know if I've ever came across— come across anything like this, John, you know, especially fully coded. That just is very new to me. I mean, the nearest analogy that I have from my no-code world is, I feel like, you know, Webflow for a while, I don't know if they still have it, but they had this concept of clonables, right? Where they would, they would like highlight, you know, community coded Webflow templates that anybody who just joined can go in and then clone, you know, and use it, right? Mm-hmm. I don't know, it might have morphed into something now, but that was a brilliant move, you know, by Because they're not relying on just the core team of Webflow developers to do this on the team. They're relying on the community, you know, to come together and like come up with some really creative Webflow websites. I think, I don't know if I've seen anything with the code side of it, right? Like that's just branding.

Well, you know, when we started that and when I was pitching this to investors, and I pitched this to 200 investors, like—

Wow.

Yeah, because I was from that world and I wanted to go the VC way. Yeah, yeah, yeah. And I kind of went that way and I just kind of, you know, canceled in the middle because I just realized that it will harm the project. I realized that if we have VC-funded company, we can't execute on the vision of sharing, you know, the whole project with everyone. Because now it's not just that others can build stuff, but others can build and kind of co-own the whole platform and the system with us, right? And that's different comparing to other kind of plugins, et cetera, because your plugin can be just kicked out of the system and you're out of the business, right? But in our case, we kind of have different model. But now, this year, so my prediction back in 2018 was that in 4 or 5 years, All, you know, all ID slash no-code slash platforms will realize this. They will realize 2 things. First, they have to, you know, involve the community to build the stuff, right?

Right.

That's one thing, right? And we see that, like, you know, Replit has bounties, and then Supabase now is paying people for creating plugins for Supabase, and Vercel is, is just putting high PR to everyone who builds stuff for Vercel. So it's like, it's really big now and will just grow. So we will see a lot of, you know, the pro code environments start doing that. And that's happening already, right? But we've been doing that from the beginning. So we're kind of far ahead of the game, but it's not difficult to explain this stuff to people today. It was extremely difficult. It sounds obvious now, but— Right. 4 years ago, when I talked to people, they said like, Like, why nobody's doing that? And, you know, if it was good, people would be doing that, et cetera.

That's brilliant. This is a good segue for us to talk about Unicorn Platform, right? And I brought on Alex on the podcast, you know, which is something that you said you watched. So he gave his version of how Mars acquired, I mean, you acquired Unicorn Platform. It was a great episode. We had so much fun. And he, in the middle of the episode, like 4 or 5 times kept saying, Kippy, you have to bring John on for this. You have to talk to John about AI, you know, all that. So here's us now giving a shout out to Alex. So tell us the story of the Unicorn Platform acquisition, you know, from your lens, John.

Yeah, so since our vision is to grow, you know, by the community, with the community, with the other founders, So when I say community, I basically mean other founders. So for Mars to succeed, we have to gather like 300 founders under one umbrella where every founder is actually independent and, you know, gets all the good part of being a founder. And we kind of take away all the bad parts of being a founder. It's kind of like the dream place to be. And And then we had, you know, we had a lot of micro apps to be built. Some were built, some were not built yet. And then I had the profiles of the people I needed. Like I had profiles of the people with the type of the product they're building. And one of the profiles was a no-code landing page builder. So I just went on internet and I just tried all no-code landing page builders. And I really loved Unicorn. It was like You know, it was not as advanced as the others, or, you know, like, it was not having so many features, and it was quite strange. It was like, you know, just one guy, and it didn't look like a real company. But, you know, from the other hand, I kind of felt like there's somebody who really, really puts thoughts into every little click and a feature and text on this product. And those kind of people are really, really good at building products. And I thought like one thing that I want to acquire the no-code tool that already has users so that we could just swap that with our micro app and basically our micro app would just get faster to the market. Because the alternative is to just, you know, promote them one by one. But then I thought like, if we, you know, built one and then we buy one and then kind of merge all of those things together, then we suddenly, you know, far ahead. And then the other thing is that I don't just need the project, I also need the founders to join our, you know, we call it Marsverse. And then I followed Alex for one year. Great idea. So I follow him for even, year and a half. He didn't even know that. So I read every tweet he writes. I read everything, basically. I was like a spy on him, right? I don't think there was anyone in the world who was watching him more than me in that year. It was not only him, it was like 10 people I kind of shortlisted. Right. And he was like, in my opinion, so much better exactly at that landing page And also I saw that he was so good at some of the things I'm not good at. And then I felt like this really good fit where I need the project that's kind of popular so that we can merge it with our microapp because we had the same microapp within Mars. And then I want that stuff to be run by someone, not employee, but someone who actually cares. Right. And then I needed somebody for Mars who could help me with certain things. And I saw that Alex is the guy who can help me with those things. So it was like perfect. Like there was no one in the world better than Alex for that kind of avatar I've built in my head. And then I started talking to him. So, and it was quite quick. So I had one chat.

Wait, did you, in your first chat or in your first conversation, did you bring up the concept of acquisition or was it just to kind of get a feeler? Of the founders?

You know, it was funny, like, it felt like we all, we both knew what we're gonna talk about. I don't know why. And in one chat, we just talked about everything. So we just started talking. And by the end of the conversation, we agreed about the deal. And that was really fast. I think it's like— No,

so I'm actually, the other thing I'm curious, did he put any feelers out there about Wanting to exit?

No, no, no, never. So there was no even a little sign of him wanting to sell it, right? And like—

So your pitch was really persuasive then on that call. You must have really convinced him.

Yeah, I think it's next after that, do you want to sell sugared water or help me change the world? I think it's the next one after that.

Right. It's Steve Jobs. Yeah, I remember that.

Yeah.

That's awesome, man. Because, you know, it's very hard to get a founder who's committed and who's so like him. He's so obsessed and committed and driven about the platform to, you know, to say, all right, just sell this to me now. I'm sure.

Yeah, but he wouldn't sell it on a traditional acquisition. Like, he wouldn't sell it for cash and that's it, right?

Yeah.

So, and, you know, I watched him for such a long time that I had very good feeling on what he really wants. I saw him struggling with stuff he didn't want to struggle with. Right. And I felt like he would really benefit from, from us taking that struggle away and giving him more space in the stuff he really loves doing. Yeah. Like product and marketing.

Yeah. So, you know, one thing that came to mind when you were talking about sort of bringing some of the best founders into Marsverse and letting them be themselves, It's a very refreshing idea, John, because I think, I feel like it has a lot of potential even beyond the scale it is right now. I'm trying to think about it for the next 20 years, 25 years.

Yeah.

It reminds me of what Disney has done for— if you say Marvel series is independent on its own, right? Marvel's got its own set of movies, but it's still Disney, right? And like, there's so many like that, right? And the creative freedom for the director or for the people who are working on that is so theirs. It's not, you know, it's not so influenced by the acquirer, right? Or whoever's like above them. So I think that's a very refreshing thing, especially in tech. Because I don't know if there's any other examples that I'm trying to You know, I'm drawing a blank.

You know that there's a really good example that you will really like. Uh, think of labels, like the music labels.

Yes. Okay.

So every single singer in the world has, you know, he is a part or she is a part of the music label, but we don't care about the label. Like we see the person to be, you know, themselves and they look really independent. Right. And they are independent. But then there is a label who takes care of all this stuff that, you know, no singer wants to take care of, like production of the, in old times, production of the CD. These, right? And now like distribution, etc. So I think basically the label logic applied to tech world, and not just to the whole tech world, but actually to the indie world,

right,

might be one of the biggest things that's going to happen to the indie tech world in the next 5, 10 years. Why? Because, you know, in that case, basically the whole indie world becomes like one large label that can just beat everyone who is like all the big corporations. Because otherwise, every single indie hacker is facing a large corporation or VC-funded startup, and that's not a fair battle, right? Right. And here, there is an army. And I think this is like, I'm sure there will be a lot more than we doing this. It'll be a really big thing soon. I'm pretty sure. And I'm seeing this happening now. in Twitter, unofficially, I see these little labels, kind of like coalitions popping up among, you know, 2, 3, 5, 10 builders and makers.

Right. And it makes sense to collaborate, right? Because then you have shared resources and shared distribution.

Exactly.

I've been saying this even like in the earlier sort of episodes of this podcast, I was just saying, I feel like one of the things that you'll see is, we may see is, creators who have an abundance of distribution collaborating with an indie hacker who loves product and loves sort of the creative side of it, but may not care about the distribution as much. That is a deadly combo. That's right. Because then you have someone like Alex who loves the details, the, you know, the microcopy and like, you know, all the fine craftsmanship. And you have someone like me on the other side, you know, someone, not just me, but like someone like me who loves distribution, like the, the You know, flywheels and putting it out there in the world. So I think we'll see a lot more collabs like that happening. It'll hit the indie world. I don't think it fully hit the indie hacker world yet, but I think it'll hit soon.

I think indie world is a bit concerned about this, and their main concern is that, you know, there's a pie and we're having, you know, each our slice. But what indie world has to understand is that we can— it's not zero-sum game. So we can actually all win because we're gonna take the pie of the corporations.

And then we can expand the pie too, right? We can expand the pie. Pie is not a fixed thing, right?

Exactly.

I love it. So tell me about the other acquisition that you were part of, like a few years ago, you were talking about there was an interesting story there. Tell us the full story.

Yeah, so I have 2 stories, one quick and one not that quick. So I will start with a quick one. So my first acquisition was in 2014, it was an ICON project. So I just wanted to buy a project. So I made some money on startups, and then I didn't know what to do with the money, and I didn't want to buy a house. Or actually, I bought a house, and then I had some left, and I thought like, I want to buy a tech company. And back then, the prices were really low, you know, like for a million dollars, you could buy like a really good business back then. Now you can buy like, you know, barely profitable Very profitable startup. So, and we bought this. So we, we looked at all these marketplaces where they sell stuff, and we found one startup that had a million monthly visitors. It was the icon project. It was called Icon Search. It was like the biggest icon search project back then. Wow. And we thought like, since it has good traffic, and people who look for icons are either developers or designers, So it's a good asset. We can, you know, build other stuff on that Odin's later. So we bought that, and then, you know, our plans changed. So we spent 6 months doing improvements there, and then we realized that we don't want to really work with this stuff anymore because, you know, it was a lot more interesting stuff happening, and it was just icons. Mm-hmm. It was quite boring. So we sold it after 6 months. So we bought it for I don't really remember the price, but we sold it for double of that price, which was good.

Nice. That's good.

Yeah, that was like the short story. So I kind of tasted this world and I liked it. So I realized that at any given time, there are quite a lot of good startups that you can buy below the market price because people just might say yes to an offer. Like, you know, even those who don't think they want to say yes, you know, when they get the offer, they may say yes. Right. And then the other project was, I was building this platform to list your apartment for sale. So in Oslo, it's really expensive, like you have to pay a lot of money to sell your apartment. And then we wanted to make this system where you just kind of enter your address, it just finds the apartment, and then you can just complete the whole process, sign the documents, etc. So it— and we started that project in November. So it took us 3 months to build it. So it was released on the 1st of February, I think. And 15 days after, 2 potential buyers came in. Nobody expected that. We didn't even have users. But we were just very, very lucky. So basically what happened, That year, it was 2016 or '17. At that year, there was big hype among insurance companies that they have to have some product, some digital product, because they were kind of offline companies. They were selling insurance, kind of boring stuff. And then I think there was an article from Deloitte or McKinsey some big consultancy firms. And then all of these large insurance companies and banks read that. And then they realized they have to build some product. And then they had just a few options. Like one was in the homebuyer space, one was in the health space, and one was in just consumer space. And because homebuyer has mortgage, right? Yeah. So basically, they realized that if they own the platform where people buy and sell housing, they can sell their mortgage there just by default.

It's actually quite clever, right? It's a quite clever move.

Yeah. And nobody thought of it until that year. And that year, like, you know, somebody mentioned that and then everyone from the banks and insurance companies like, that's a brilliant idea. And then for them, you know, go and build this stuff. They have never built any digital product, right? So some went And assembled the team and started building stuff. And some just started looking for acquisitions. And there were a few in the market who were doing this. So when they came in and they made an offer, and then the other company made a contraoffer, and then this made a contraoffer, and then the whole thing was sold, right? It was not a lot of money. I think it was back then like over $5 million. So it's—

That's a lot even back then, right? Yeah.

Well, it's not like a billion-dollar exit, right? So it is But for the time it took to build the project and also for the potential, like, you know, fast forward, it didn't work.

Yeah.

None of those projects worked out. Why? Because when people sell the house, it's such an expensive and important thing that they trust a person, right? They don't wanna really save, you know, money when the apartment price is like half a million dollars and saving $1,000 is like, doesn't make sense, right? So, And we had that feeling. That's why everyone agreed to sell because sometimes you do stuff you don't believe in 100%, basically. But that

was a remarkable stroke of luck, though, like you said.

Yeah, exactly. It was lucky. And you just sometimes have to build stuff and you might be just lucky and sell it. So it's not always things make sense.

Yeah. I also want to, I also want to note here that this is still a Million Dollar Exits podcast, John. You're raising the ceiling.

Oh, it's a million. All right.

I want to do a new series called Billion Dollar Exits, and I'll bring you in for that one. Just kidding. So tell us a little bit about sort of where, what are your thoughts on sort of the current acquisition market? You know, because you're looking at a lot of marketplaces like acquire.com, which is MicroAcquire, right? I brought I brought the founder, Gazdecki, Andrew Gazdecki, on the show. There's, you know, he was talking about sort of the rise now. There's a rise of activity, right? From both buyers and sellers about acquisitions. I'm not so deep into Flippa and other things, but I feel like from what I'm looking at Twitter, between Microns, there's another one which I think you're looking at. Yeah. So Microns, Acquire, and a few others. I feel like this is the time to build something to sell, right? It's just a great period. I feel like it's blowing up. You know, what are your thoughts on where we are right now?

Yeah, I think the acquisition market will, you know, grow a lot more than a lot of people expect. And the main reason is this. So in old times, there were very few people building startups. So the distribution was easy, actually. Like the stuff I built before 2015, we never really had a struggle of distribution. Like, now all the jokes are about like, people won't come to use your product, go and market. Like, we didn't have that problem back then. Like, actually people would come and use your product. Like, we launched this first movie app in the App Store and it was like the first of its kind just because, you know, now you have everything in the world, but back then like some things were missing. So, and we just got users, you know, We launch it and in a month we have 70,000 users. Like, now nobody gonna find your app, right? Yeah. So because of that, now it's really hard to find people who are good at both distribution and building. That's why people who are good at building, like we partly solve that problem, right? Like by putting them together, et cetera. But in the outside world, people who are good at building, they tend to, You know, not fall in love with distribution. So they can try, they get some little MRR, etc. And then they feel like, I want to get back to building. But, but it's not the right way to do, right? You have to actually market it. And then for them, selling it to someone who is the opposite is great for everyone, right? Somebody buys that stuff, and then, you know, falls in love with marketing sales. It's great. This guy goes back and builds new stuff like You see now people who sell something on acquire.com, etc., they sell a lot of stuff. Like, I barely see a person selling just one project, like, you know, 4th, 5th, 7th, 8th. Like,

it's kind of like, it's kind of addicting in a way. Just once you get the high of, oh, I see the journey now, I'm like, I want to go back and do this again.

Yeah, exactly. And on the buyer side, it's also great. Look, like, if you want to test, you know, this, this builder world, like you want to build some stuff and you are not developer yourself, That's the worst place to be. Like now we are having this no-code stuff and you can learn, but not everyone, you know, wants to do that. And if you don't learn some way to build stuff and you just outsource it or hire people, that's like the best way to fail, right? Like, I almost never seen that work. Like almost never.

I have a personal experience and I had a tragic failure story in 2017. Where I spent $17,000 or $18,000, John, on an outsourced agency to build an MVP, which took 8 months, which I think should have been the— That's a lot. But so I wasted 8 months and then almost $20K, which now in 2023, I could just go and buy something from, you know, from Acquire or any of these places and then just market them, right?

Exactly. So, you know, it's a good synergy. And the iteration speed goes up. And, you know, if you're a tiny influencer or large influencer and you have, you know, $20,000, you can actually buy a couple of projects and test them whether your audience, you know, buys it. And then if they buy, then you put more money into improving that stuff.

Yeah, I couldn't agree more. I feel like over time, as sort of with an abundance of projects, over time, what I think will tend to happen is it's kind of like how before the internet, like in real life, right, in the sort of in real businesses, um, what do you call them, um, brick-and-mortar businesses.

Yeah.

A lot of that activity is about buying and selling, right? Like nobody ever goes and starts a brand new, um, brick-and-mortar store from scratch. They usually, like, usually, like, there's— it's very common for someone to just buy a gas station, right? They know they don't go and like install the gas station themselves. I mean, rarely they do, but generally they just go and buy a gas station, you know, or 7-Eleven or some convenience store or something, and then you renovate from there or do something else. So, but it like, that kind of activity is very common— buying, selling, buying, selling, flipping, or whatever, right? Um, same thing with houses, right? Yeah, people like buy and sell all the time, right? Multifamily homes.

Secondary market is often, you know, greater than Building stuff from scratch.

Right. So which I think has never really hit the tech world because it was only limited to like, you know, very, very high-end acquisitions, like, you know, $20, $30, $40 million acquisitions. I feel like now as it's coming down to affordable acquisitions, you know, I think we're gonna see a lot more people try it.

But you know why it never happened before? Like, I think the main reason why these small acquisitions never happened before is that we didn't have non-developers in this world. And developers don't really buy other code because they kind of, you know, they have opinions about code.

Yeah, they feel like I can do this on a weekend.

Exactly. Exactly. Like, if you don't think so, you're not a developer, right?

The number of people on Hacker News website who think that they could build the next Dropbox on a weekend is proof. That they don't think about distribution. Yeah.

Exactly. And now we have people less technical. And for them, like, they just don't like the coding stuff. They just want to get, you know, work done. And then zero work is better than any work, right?

Any work. Yeah. That is so true, though. Developers are the most, one of the most opinionated, you know, sets of people. I love them. I'm sure you're one of them, right? But It's just objectively, they're very like, it's kind of like chefs, you know, like when the chefs don't like anybody else's cooking, you know, they trash them. They're like, they're very opinionated. So that's funny, man. So actually, let's go to AI. I really have to do justice to Alex's request of me getting to ask you the AI questions. He put this in my ear that you've been in some fashion working on some You know, AI for definitely more than the last 4 or 5 years. So I kind of want to understand, where did you catch that bug? Where did you get, you know, started about, you know, being curious about it and kind of walk us through the journey, you know, through your lens and where is it going, you know, in your view?

Yeah. So AI was my, you know, my first thing to ever do in the tech world, basically. So— Really?

When was that?

Yeah, it was like 20— 2006.

Wow.

It was university. So my project for the bachelor was OCR. So we were— yeah, I was making this neural network to recognize handwritten text. And I did that because I was so lazy to go to lectures. And my friends went, and I wanted to scan what they write there because I couldn't read. You know, it's like when people write, like the handwritten text is impossible to read if it's not yours. Right. And then I really wanted to have the application that can do that. And there was just 2 in the world back then. It was ABBYY and Adobe. Like they had this and it cost like fortune. Like back then everything was really expensive, like the application stuff. And you had to buy a license and it was like huge app, et cetera. And I thought like, you know, I can try to make one for myself and, you know, everybody would love it because everybody would benefit from it. And I started building and, you know, I was the type of developer who thought that he can build that in a weekend. Yeah.

Oh, that's so on point. Yeah.

Yeah. But actually, you know, I built it quite fast because, you know, Like at a certain age, at a certain, you know, time, like you're extremely productive. Like back then I was like crazy productive and I just sat on my chair like for 40 hours straight from Friday to Monday morning, I remember. And I had first version. It was like, you know, 3 layers of the neural network and it was doing fine. It had like 80% accuracy. Which is fine, but it's bad. Like, you can't really read the text still if it's 80% because it misses the important words, not the not important. But then I realized that I have to have a project for my bachelor eventually. So I just take this and I just, you know, keep going with it. And I did that. So I just kept going with that since 2005, actually, until 2009 when I finished university. So for 4 years. I've been building this, so, and I had to learn everything about AI because it was, you know, I ended up actually building an algorithm where I had a couple of neural networks doing independent work. I had other algorithms and I would exercise all of them and then I would get this, you know, guesses from all of them about this word, for example. And then if 2 of them agree I would say then that's the word, right? And that was quite a cool approach actually. So I came almost to like 98% accuracy where FineReader and Adobe were 99— no, 98.4. So it was like 1% difference. When I presented that at my, you know, at the day where you present your stuff to the committee, people didn't believe. They thought I made up the number. So I had to actually run the whole thing to show that it actually works. I even asked them to write something, write as bad as you can. So it was really good stuff that worked. But, you know, my— I kind of regret that I was too young because I had no idea, like, believe me or not, I had no idea that I could actually build a business out of it.

I was gonna say that you could have turned that into a company.

Yeah, obviously. I just, I didn't have a slight little idea, like zero at all. Like it never came to my head. I thought I'm studying, right? This is for study.

Right.

And that was kind of sad. And, you know, I don't even have the program anymore. After study, I just deleted everything because it was finally I finished this university. And then I did some more stuff with AI. So my next project was called AIMS AI. So it's It's kind of big stuff now. It's the biggest monitoring software in the world, I think, now. So that was my first big startup. So I was doing all the tech there, and the whole project was about monitoring complicated systems such as banking systems, ATMs, like the systems where you have thousands or tens of thousands of different kind of components which are offline. Like, which are physical connected to the digital, right? And in that world, basically, you don't know if somebody— if something is broken until somebody calls you and says, this shit is broken, right?

Right.

Because now you have cameras, et cetera. But back then, it was just standing like a vending machine, for example. You have no idea if it works or not. Right. And then what we built, we built a software that was Monitoring everything that's happening with that machine. Now the sales, the time it takes, like all the parameters, everything, all the numbers. So now they call big data. Big data didn't exist back then, that word, but we kind of did that. So we were collecting a lot of data, everything, the raw data. We were applying these regression models and neural networks on that, on top of that, and we were predicting. The behavior of all those, you know, points. And then whenever we see deviation from that prediction, we would say something is wrong there. And then we'd correlate that deviation with the previous historical data. And then we would kind of confidently say that this machine is broken because at this time of the day, you know, at that hour, It's strange that it's selling so little or something like that. And it was like huge. It was really huge. We got clients who were like those, this vending companies and ATM and on all the stuff that has, you know, the large corporations basically. Right. And for them, it was saving a lot of money because they wouldn't have the machine that stays for weeks and they don't know it's broken, right? Because the problem is that some machine can still work. Let's say you can buy soda, but you can't buy the chocolate bar. And you can't know that, like, well, people buy some stuff there, right? Nobody's gonna like look at every machine and compare like what they sold yesterday and today, right? But our software manages

it. Except the data. Yeah, except the data will tell you the truth. Yeah,

yeah, exactly. So that was the second project. And then the third project was the movie recommendation system. That was—

Is this the Film Grain?

FilmGrail, yeah.

Yeah. Oh, FilmGrail, sorry. Yeah, I think I saw the website of that. That was pretty cool.

Yeah, it was—

is it still going?

Well, now it's in the B2B space. So basically we built software and it's white labeled and then it's running as that brand. So it's big. It has 15 million users using all that kind of, you know, white label software combined. So it's Very big. And, but back then we were doing just an app that helps you find a movie to watch, right? That was kind of our first app that we made that got 70,000 views and we got the best app of the year back then. And it was so good because we have built this AI model that could basically recommend you a movie that's really good, right? And it's not that easy. Like now it's really easy. Now everybody knows how to recommend stuff. Like you go to Netflix and they're quite good at it. But back then Netflix was just a catalog. Yeah. Like everything was just a catalog and we just scraped all the catalogs, Netflix, iTunes, you know, HBO, everything. And then we just put it all together and then we, you know, get all the data about you, about your friends, about you. Like back then Facebook could give you everything. So we would just take everything about you from Facebook and then we would just run the models and then we say like this movie, is really good. But then if you're watching that with your girlfriend, for example, we would just take both of you and make the model for both of you and then say like, this movie is good for both of you. Like, you're gonna like them both. And so yeah, those 3 projects I worked and it was quite heavy use of AI before like anyone that was really using it.

Yeah. So when you look at the latest sort of application of AI now with ChatGPT, And, you know, DALL-E and all this. What do you feel now? Like, you feel, where do you think this is going? Yeah.

Yeah. In 2018, we started using, or 2019, we started using GPT-2. It was, yeah. Wow.

And that was under, it was not open for anyone, I think, right?

Yeah. It was like the beta we applied since GPT-1 and then we suddenly got in and we were using it and we built quite cool stuff on top of that. And we were quite amazed. But it was kind of, you know, it was failing too often to use it for, for things people use it now. So it would give you a good outcome once and then wouldn't give again. And then it gives you like, it was not stable, basically, you couldn't really ask it to produce JSON, for example, reliably back then. So you would have to build some stuff To parse what it, you know, outputs. But still, we kind of liked it. And then, you know, when I was starting with the vision of Mars, so we were kind of confident that we have to build a system that lets you reuse microapps. And then, like, the main concern for you reusing microapps is that what if the microapp doesn't fit my need? Like, it's kind of almost but not exactly. And then my idea was that we will actually manage to get microapps to be able to rebuild themselves using AI. So you just say, you know, you have microapp for e-commerce or for reviews, for example, and you want to change that somehow. And you just tell that what you want to the microapp because you don't want to kind of, you know, dig into the code, right? It's not your microapp. So you just ask it using AI and then it just, you know, reassembles itself and it's new. Wow. It's just for you. So we had that vision from the beginning. That's why we kind of made a high bet on Microps because we knew that eventually when AI gets very powerful, this combo will be really good. And then the time went on and then in last year when GPT-3 came in, we connected GPT-3 with Microps and it was amazing. So yeah, like the moment I run it, I remember that it was just Really good, and I and I had this lamp blink in my head like yeah, like AI world has come to where I want that world to come, right? But that was before the hype, and I felt like a madman. I remember like you can ask Alex. I was like telling everyone like like we have to do everything with AI, and I like people just thought I'm crazy because they all use Siri, and they thought like well. We saw AI. It's, it's not that good.

Right.

Yeah. But you know what I did? I built a lot of stuff with AI last year and I used it since then. And I think that was kind of one of the main reasons why everything I do grew up so well since then. So basically I built this little project I don't have a name for. So it's kind of agents where you program an agent with the prompts and AI and then the agent behaves like you programmed. Like I have a lawyer agent, I have accountant, I have developer, I have myself, et cetera. Wow. Yeah. And I used that since then. And, you know, it made me like 10 or 20 times more productive. And, you know, we have no people in the company who are not developers now. We have no lawyers, no accountants, like no, almost no support people. Everything is done by AI, everything. So we're extremely productive. And we were doing that for a long time, like since the beginning of the last year.

Where do you think is the biggest opportunity in the AI land now? For like, for the listeners who are like just thinking about, okay, I want to build something in the next 6 months or the next, I don't know, 3 months. Where do you think are some open gaps or open areas that you want to explore?

I think it's UX. I think people should try to come up with really clever ideas with UX because I wouldn't say there is any solution yet to the UX on top of AI. Like there are a lot of attempts, you know, Notion is doing well. Everybody copies Notion, but it is still like limited use. It works well for the documents. But if you look at almost everything else, like coding where we are doing that stuff, like that's, you know, not working that well. Like most developers don't use it, but the way we use it, you know, it's different from everyone else. But then there are more simple stuff, like for example, The regular stuff like accounting, lawyers, chiefs, cooking, little businesses that you can just go to the bakery business and see how you can apply AI with a good UX there. So it's a lot of opportunities and I would focus on not sexy one.

Yeah, I mean, that's, no, that's a smart sort of distinction, right? But if I were to ask you, Like from a layman point of view, one thing that I feel like I'm struggling right now is when I see the demos, I'm blown away. But when I try to use, when I go into my normal routines, I feel like it still hasn't percolated into my user habits. But I'm like mainstream sort of layman, not a very early adopter, right? Mm-hmm. I mean, I am early in a way, but Like compared to my wife, right? But yeah, compared to maybe you and like Alex who are living in the AI world, I feel like that is the, that is the challenge I'm having even with, I'm, I'm a paying customer for ChatGPT, but I rarely use it. You know, it feels like a very powerful toy, but it's not integrating into my workflows. And that's been a, that's been a challenge. Yeah.

Like, and that's UX, like you have to copy and paste stuff, right? Yeah. You have to wait. For stuff, you have to remember where exactly and how exactly you have to write that. So I think there's like enormous amount of, you know, possibilities to do good UX on top. And I would just not recommend anyone to do anything deep tech there, because that's just too complicated. And you need resources, but UX is easier. But when you do UX, try to do something that's not that obvious or not that popular, because if it's so popular, most likely the bigger brands will just include that as a feature. Do not build a product for the email, right? Gmail will come with that.

Gmail will compete.

Superhuman is there already with their— so just don't enter that. But there are other, like the second wave, Where like Google will not build anything for the bakeries, right? Or etc. So I would just go there and and I see people doing that. And if you look at them in the world and the race of Indies this year, it's crazy. Like I see people who were struggling to get any MRRs for years, and they built quite a little good app on top of AI. Some people say it's a wrapper. But who cares? And they're making like, you know, $5,000 or $10,000 a month.

Yeah. I mean, some popular examples, like, I mean, Danny Postma and like Levels.io, they've gone into, I think Photo AI is one of them. And the other one's like the, I forgot what the other one is, but I feel like there was an interior design AI tool, right?

Yeah.

It's a little niche compared to mainstream and it's a pretty good MRR, both of them. So I think, yeah, I think, I agree to your point about, you know, look at some of the boring niches or unexpected niches and then build AI tools for them. Yeah. Fascinating. All right, so I know we're almost at the end of the hour. I feel like I could go for another hour with you, John. It's been a blast. I wanna say thank you so much for being on the show and for such a delightful, reflective conversation. You know, I learned a lot. And, you know, I feel like we gave some playbooks for, for the listeners too. So thank you.

Yeah, thanks for the chat, KP. Invite me again after one year, maybe I have something new to say.

Yes, I would love to have you back again soon. All right. All the very best with Mars and all the other projects.

Yeah, same to you.

Thank you.