Martin Rue (LangSesh): From Engineering Manager to Purpose-Driven SaaS Nomad

Rob: Hi Martin. Welcome to the fourth episode of the, Trial,

to Paid podcast. Welcome.

Martin: Hey Rob, thanks for the invite. Go to, uh, good to chat finally.

Rob: Yeah. Great. Thanks for coming on. So let's, let's just start by, I mean, we talked about this a little bit before, if you've got a guiding principle in mind behind how you, uh, sort of design and build your products, something that just keeps

you moving in the right direction.

Martin: Um, okay, so it's a good question and I feel like I've, I've got three or four answers, but the simple one for me, I think is, um,

is just a sense of the thing I'm working on. Is it helping someone?

Rob: Yep.

Martin: it making someone's life better? It reminds me of, um, I think it was Richard Branson, and I do not make a habit of quoting him actually, but he did say something that was quite cool.

He said that a business or a business is an idea that simply makes someone's life better and. Uh, it kind of stuck with me as being like a general, generally, right? I think obviously it also needs to make money and, and so on and so on. But, um, but it resonated with me because I've always felt the same way.

Like, I don't wanna work on anything that I don't personally believe in. And I think that is quite a challenge for a lot of people like us. People I meet that, you know, get offered very high paying jobs, doing things they don't care about. And, uh, I did a lot of that too. Uh, and working for myself now, uh, is a great opportunity in the sense that I don't need to do that anymore.

And I can start with the question, is this something that makes someone's life better? And be honest about that. Um, and, and sort of a second answer to that, I guess is just this, a personal set of virtues behind that, of like, if you're an honest person, you have to be honest with yourself first and foremost.

And, uh, answering that question honestly is hard. Sometimes the answer is no. Sometimes the answer is, this thing is not making someone's life better. Nobody cares about it except me. Um, I've been through that a few times and, uh, that's, that's why quitting is a good thing sometimes, if you can be honest. Um, so yeah, I would say, yeah, my answer is pretty much just.

Constantly assessing whether the thing I'm working on is helping someone else. And, and that comes with a bunch of other things, of course, of trying to figure out how to measure that properly. Not quit too soon, also not stay anything too long. Um, how to really answer that question. But yeah, I would say, I would say I wanna work on things that I know help people and, and, and, and also that help me.

That's another big part of that answer, I would say, is working on things for me

Rob: That's brilliant. I guess also just being, you know, having that contact with the users as part of that, isn't it? And it sort of drives, you know, the kind of desire to build something yourself, build your own product, um, as opposed to kind of being, being, I don't know, an engineer in a role where you're kind of like somewhat sort

of distanced from all that.

Martin: Yeah, for sure. I think, I think starting with like, what do I wanna work on? What would help me, what do I like, build for yourself first. Because, you know, as, as the, as everyone's said a thousand times, if you build something for you, chances are you're not the most, you're not a unique person. You're not the only person in the world that likes that thing.

It is a different question of like, how many other people really are there? And is it a market? Is it viable? Um, but there are definitely people out there that probably like whatever it is that you like. So I think it's a good way to start. But it's also a very, I think maybe people throw words around like lifestyle, business and too small and not scalable and it's not gonna be a unicorn.

Um, I've long since given up on those kinds of ideas anyway. I don't wanna build a unicorn. I don't wanna build something that's widely interesting to, you know, a hundred million people. I'm much more interested in what's appealing to a thousand, 2000 people. 'cause there's a lot of viability in that.

There's a lot of freedom

Rob: Well there is so much more nowadays, isn't there? 'cause you can reach those people. Yep. Love that. Um, okay, so we're gonna be talking about your, your, your current product, um, Lang Esh shortly, um, which I've had a little look at. Um, and I guess maybe I'd sort of summarize it as a sort of more, more sort of calmer, more sort of thoughtful, maybe a little bit more grown up take on the kind of, uh, Duo Lingo sort of concept.

That's how it looks like to me. But, um, I'd love to hear it from you, uh, shortly. So we'll get, we'll get into that. Um, but yeah, just, um, first of all, I suppose it would just be good to, um, just dive into your background a little bit. So, um, I suppose a good place to start, um, is it sounds like you've been coding for a long time, building for a long time.

Um, just how you got into that, was it as a

kid? Um, or a little bit later on?

Martin: Yeah, it was actually, I was, um, I was actually 11, so it was, none of this was really my choice. It was more just the situation. Um, and, and I do reflect on that a lot and think was quite lucky to get into something that was, and I didn't know this at 11 years old, but something that was then gonna kind of be very important in the world.

Um, being able to write code, build things and all of the things that come with that in, in terms of how it makes you think. Um, were, were very accidentally beneficial when I was quite young. Um, but the, the, the very short version of that story is I had an older brother and he was, I think 14 at the time, and he was just writing basic code on a Commodore 64.

And for anyone who knows what that is, I have to say this nowadays 'cause I'm like, actually, damn, some people have never even seen one. Um, it's a big blue screen with white text and he was writing code. So it was basically just imagine a blue screen full of white text and none of it really makes any sense.

It's not English. Um, it doesn't look like English. Every single line starts with a two digit number or a three digit number. And I just remember reading a little bit of it or trying to, and it's funny actually, I remember I read the word Go Sub, which is actually written as one word in ba basically it's G-O-S-U-B.

And I remember trying to read it as English and reading gossip. I thought it, I, I said it out loud as gossip. And I said to him, what, what does that mean? What is gossip? He tried to explain it to me, didn't understand a single word of the explanation, but it sort of, um, I would say it grew this fascination with what is this thing?

He's writing something that I don't understand. And, and for all intents and purposes, we're the same person, but he can do this and I can't. And what is this whole like, mystic thing? So yeah, I just got into it for that point of view. Um, I played around with basic myself. Um, started writing little programs in basic.

I have some silly stories about how I ended up destroying every single entire game that I played as a kid because I wrote password programs to the beginning of each of them that then couldn't decipher and forgot the password to. Um, but it was, yeah, it was all the kind of stuff you would imagine like kid programmer does.

Um, and then yeah, from there I ended up getting involved in more serious programming languages like C++, even some, uh, reverse engineering and assembler as well. Um, back when I was pretty young, when I was like 16, 17, I was very interested in compiler design, all this low level stuff that really doesn't have much of a place these days.

Sadly. Um, but it is, it is a definitely an interesting set of

skills.

Rob: lot of people talking about c plus plus and rust and things like that, you know, on. LinkedIn and whatnot nowadays. So presumably

there's, there's a fair bit of work going, going on there.

Martin: Yeah, for sure. I think it's, I think in the games industry there's definitely more of a potential for that kind of stuff. Um, but yeah, the lower level, even just a lower level of concepts these days. I mean, uh, you know, we, most people write software that fits into a certain sort of simple box. No one's really writing a new algorithm these days.

Uh, un un unless you're in a very, very specific job.

Rob: We're a lot of layers

up, aren't we? These days.

Martin: yeah. And which is a good thing, right? I mean, that's progress, I guess, as well, but I think it is important still to know those, those principles and I'm, I'm lucky that I know them just because those were the first things I actually learned.

Um, but yeah, that's how, that's how I got into it so long

Rob: Yeah. Yeah. And it's interesting talking about those, those early computers, Commodore 64s, and I got my first one, I think it was a VIC-20. 'cause I'm a couple of years, a couple years older than you. It was like a precursor to to that. Um, but it's like you turn the thing on and it doesn't, doesn't do anything unless you put some code in, you know, it's not going to, they can't download games or whatever.

And we used to get games on tapes or you'd even

get like a magazine and you'd have to type in the code.

Martin: Exactly. I

remember doing that as well. Yeah,

Rob: But, um, so yeah, it was more that you kinda start from that position of, well, it doesn't really do anything unless I solve some problems. So I think that was for me, I guess what, what drew me in.

But then also just instilled early on that, you know, computers were about,

you have to solve problems if you wanna make 'em do things.

Martin: that's a, that's a really interesting point. 'cause I think we're transitioning into a different kind of world right now where we, it's not what's the underlying problem, it's what's the request, what do I want out of it? And, and that's sort of like almost deferred. Um, and I'm talking about AI almost in

Rob: Yeah. Yeah,

Martin: um, there's, yeah, it does worry me a little bit the generation of newer engineers that will, will think that is the, that is actually the starting point.

There might be someone in 40 years now describing. That story as how they got into engineering. And it does worry me what's lost below all of that because there are still some important concepts there that shouldn't be forgotten.

Rob: for sure. Great. And I, I suppose, um, so just kind of leading a little bit into your, your, your product Lang Lang Esh, which we, we will get to, um, presumably at some point.

'cause you, you mentioned this, um, I think in your, in your LinkedIn profile, some of your posts this interest in, in languages as in not computing languages, uh, uh, sort of, um, languages we speak and write, um, that interest developed at

some point and became a passion.

Martin: Yeah, I'm glad you asked me about that

actually, because it's a funny story, um, in the sense of I didn't care about languages in school. Um, obviously we're both from the UK here, so our school, uh, education probably was a, a bit about the same. I think it's still roughly the same even now, where we sort of in the UK get forced to learn some foreign languages, but not that seriously.

Um, you get, you usually get forced for a year or two and you can choose to continue it, but learning language is a hard thing and, and no matter which way around you do it, you realize that, especially as a student. As a, as a, you know, 14-year-old making this choice that that's hard. And there's some subjects that are much easier and you're not, you're not in school to have a hard time, so you don't choose them usually.

And, um, I, I did Spanish for two years actually in high school, and I had no interest at all in it. Um, it was hard. I didn't care about it. I didn't see any interest in it at all. And I reflect on that with some sadness because I really wish my Spanish was a lot better now. And it might have been, um, had I been interested in it, but I wasn't that interest actually developed in uni.

Um, I was in uni, it was the second year of uni, so I was 19, uh, yeah, that's right. Yeah, 19. And I, so for, for one year it had a colleague, um, in, in uni. He was an older student actually, and he'd served in the British military and little did I know he'd served in the military and was deployed to Germany. Um, and I didn't know any of this for a whole year of knowing this guy.

And then randomly, in one particular moment in a, in a lab somewhere, there was some other students and they were speaking German and he just. Started speaking to them in native level. It wasn't native level, but he spoke the language very fluently. So he just switched into German and spoke. And it blew my mind because there was this version of this guy that I knew as my friend that I've known for one year and English was that there was that connection point and all of a sudden there was this different version of this man that I could not understand and follow and even felt culturally in some way slightly

different to me. And I was like, blown away by that experience. I was like, wow, that's, that's very much more than a language. I feel like this person's different. I, I feel like this person's got two persons. I've only got one person, he's got two. And that, that was the moment I think, when I was like, wow. German school. So I went off and in German for a while.

Um, that took me to working in Berlin and, and then other places in Europe working for language learning companies. Actually some, some well-known ones that I won't mention. Um, and uh, yeah, that's definitely where it re, well I say reignited. It was never really ignited at all. But that's, that's I think where it started.

Rob: Yeah. Do you, do you think it was, um, the way, the way it was taught at school and perhaps not having that kind of, uh, that cultural dimension to

it, that just meant that it just, you couldn't see the point in it?

Martin: Uh, yeah, my, my theory is that for Brits, um, living in the uk, we just, I mean, maybe this is changing a bit now, but, but certainly if you go back to when I was in school, which is like 20 years ago or something, uh, no more, oh God, more, um, there was no obvious cultural reason to, to speak Spanish or French or German.

Um, and I, and, and that is the, the, you know, the unfortunate truth about most of the languages in industry is something like half of the entire revenue of the entire industry is in, in education of English for other people. Not in, in in some of those, you know, English learning another foreign language. So I think that's a big reason.

And I think the other reason is, is as I was mentioning, you're a kid in school, you know, you don't want to take the hard path. So, and you learn pretty quickly, um, that it, it's not that easy to pass this thing. So I think those, those are two big reasons, but yeah, similar to the US and there's not that strong of a culture of multilingualism like you find in some European, like Switzerland for example, where it's very normal to speak four languages.

Rob: Yes, yes. Okay. And so, so at this point, you, you, you

know, various languages fluently

Martin: I know bits of some, I know more bits of others and some fluently, yes. Um, but the, the sad and ironic thing is that I spend much more time building things in this space than I do studying languages. And it's such a hard task to become fluent in a language that's proximate to your own. Uh, actually stepping back, there's this more general question of like.

Which is the hardest language to learn, right? Which is one of those questions that's a bit like, which is the hardest programming language? Well, if you're asking someone that currently writes Haskell, then maybe they think another purely functional programming language is the easiest one to learn. So it's very much dependent on where you're at now and what you already know.

There is no general answer to that question, but there is actually, um, some bits of documentation research you can find that do measure from the perspective of a, a monolingual English speaker, what's the hardest language to learn. Um, so yeah, like some of those proximate ones are a bit easier to pick up some of the other ones.

So at this point, yeah, I think, uh, my Vietnamese is okay. I, I lived in Vietnam for the last year, so I, a year ago for one year, and I chose very actively to learn Vietnamese there because English, um, is of course very, very, um, the level of English is very low in Vietnam. And if you can speak a little bit of Vietnamese and you are living in that place, then it makes a big difference.

Um, and then some other ones, like my German is okay, my Spanish was better, but I haven't stood at it for a year and a half now. Um, and I speak another, another language that maybe some listeners have heard of, but probably not, um, called Esperanto, which is a constructed language, which we can get into later if you

Rob: Well, I was just gonna ask about that because I, you've mentioned it in several places on your profile, and so I'm

curious, tell me about this

Martin: Okay. So this ties back into the other things quite well actually. Um, and when I speak to software engineers about Esperanto, there is this additional dimension of interest that I myself experienced. Um, so Esperanto is a constructed language. Um, you could substitute that word for other synonyms, like made up, invented, created, like artificial, all of those words.

They all work at the end of the day. It's not spoken by a specific country. It was invented 140 years ago, so actually not that new. 140 years is a long time. Um, it's been spoken pretty much for that entire duration of time. Uh, and it's still spoken. Now there's about. People might, might not like that. I'm gonna say this number, but it's only about, I would say, 60,000 fluent because of the language across the world.

That's not a very big number at all. In fact, that would absolutely be an endangered language if it wasn't for the fact that it's just esperanto. It's a whole different thing. It's not, it's not endangered at all. It's definitely not going to die. Um, people throw around bigger numbers, like 200,000 a million based on Duolingo stats of how many people started the course, which of course is not the same as people learning the language.

Um, especially in duo's case. Um, anyway, I discovered this language because I read about it on Wikipedia. You know, that whole Wikipedia rabbit hole thing that you do sometimes and you end up learning some random stuff. Uh, I found this language and um, I just briefly read that it was a constructed language and it was designed to be very, very simple.

And at that time I was in sort of in the depths of German grammar. Uh, for anyone with any familiarity with German will know that compared to English grammar, German grammar is particularly more difficult to think about a lot more things. The sentence structure can change quite, quite a lot more than English.

Um. Through genders, noun, decension, all these, all these fun things if you're into grammar. But if you're not terrifying things, I was in the depths of that when I discovered Esperanto and suddenly it was such a contrast because Esperanto was this language that was redesigned to not have all of those complexities.

So my first question was, okay, well this is not real. It's probably, it's a bit of a toy thing. It's like hearing about Klingon or Navi or these like languages from TV shows. I kind of thought it was about the same sort of thing, just a hobby project somewhere. Um, and then, um, I ended up going to an event in Wales actually, where there was a hundred people and I imagine most people would speak in English and then you'd hit bits of people trying ESP Brando, not really speaking it, but maybe trying it.

And, and to my surprise, it was completely opposite. Everyone spoke Esperanto fluently and nobody spoke any English at all. And I could only say my name and how are you and things like that. Um, but that did definitely ignite this whole idea of like, wow, people really speak this language. People really can reach a level of fluency.

It at that particular event, people were talking about Bitcoin even, and other subjects the entire day. Um, so yeah, I, I went back from that event with a sort of slightly different perspective of a, this is a bold idea. You know, it's taboo to change things like monetary systems and radically change like healthcare and radically change language.

Language often never gets changed. There's only a very few number of cases in the world where languages have been dramatically changed. Turkish Hebrew was revived, but mostly language has been the same for a long time. It evolved very naturally over a very long period of time, but it doesn't get actively changed.

ESP Brando was this thing from the ground up that was created from nothing. Um, of course some things are taken from other languages, some ideas, some some root words and things, but for the most part it's a very unique language to itself. So it's a bold idea and I really just love bold ideas. Like if it's, if it's taboo, especially, I like it 'cause I'm.

If people are like, oh no, this is sounds like a bad idea. My first thought is, okay, so it's new then. It's very new if people think it's, it is a bad idea. So I wanna know if there's something in that. So yeah, I went away, played around with it, started going to international events, meeting people. I then, then COVID moved to Barcelona with my ESP Brando for speaking friends for around six months.

I was like, yeah, for about six, six and a bit months. And, uh, that's all I spoke. Um, again, another reason why my Spanish is not as good as it should be. 'cause actually in Barcelona where I could have at least improved my Catalan or, or Spanish, I actually spoke in an invented language with a bunch of people.

Uh, and yeah, I now speak that language fluently and sometimes dreaming it. I have very large number of friends in that language and it's pretty crazy. Even here in Asia, I just sometimes meet up with people that are over here and we just hang out and that's our, some of the, most of those people I've never spoken English to.

So it's, it's quite strange to hear them speak English.

Rob: Interesting. So, okay, so your initial interest in, in language sort of sparked, um,

I think you said, was, that while you were, you said that was

while you were at uni.

Martin: No, it was, no, it was quite a bit, quite a bit. Oh, sorry. The German, sorry.

Yes.

Rob: German Yeah. So that was sort where the language, the, the light bulb went on. I, I, I kind of, I actually languages, so there's an interest there.

Um, but then the Esperanto thing, which sounds like it was that sort of part of the moment where you thought, um, maybe there's something I, I could build here. Something I could go. Kind of deep on, and, and did that happen? Did you come across Esperanto? We haven't really talked about any of your kind of roles after uni, but was that something that came along the road when you were out

there in your, you know, so building your career?

Martin: Yeah. So I think that happened, um, it didn't happen that long ago. Actually. I've only been speaking Esperanto for about seven years,

um, seven years. So eight years ago, I didn't, couldn't speak a word of it. Hadn't heard of it. Um, the German thing of course was, was just after uni. So in between then was this, this gradual interest in language.

Um, and eventually Esperanto arrives on the scene as a, as a, as a cool idea. Um, but in the meantime, yeah, I started to learn Russian, Hebrew, Turkish, some other random languages. Um, and it sort of grew out of that first experience for sure. Uh, and yeah, I mean, it's like any hobby, right? Like when you really get into something, because I've had this with other things too, like the gym, and I also a big fan of the gym.

And, and you spend time in these environments and you understand people, you see how people behave and what's important to them and how they do things. And as much as that applies to language, it applies to the gym, it applies to any situation that you actually like, really put yourself in and, and you can't help but see.

Um, so yeah, for sure, that's when I started to think about, uh, well, I would say add to that, the fact that I was also at that, around about that same time, um. I was working for another language learning company in Berlin as well. So I was also seeing from the inside, like how a product is built, how they market that, the kinds of problems that they deal with in trying to grow that business.

And it did go from a startup to a very successful, uh, language learning business that pretty much any language learner is probably used at this point. So yeah, I think both of those combined and eventually, you know, I'm sure you'll re it'll resonate with you very, very strongly is as an engineer, if you spend a long time in a field working for somebody else, you, you eventually, once you've seen the whole business though, you see how the, how the sausage is made.

You're like, I can make this sausage on my own and more than that I can make a very Yeah, exactly. I

can make a much better sausage or a very specific sausage for this specific group of people that

Rob: Yeah, so you see the, you see the things they're not doing the different angles that they're just sort of ignoring 'cause they're a big company and they've just, you know, gotten complacent. Um, so, okay. Um, so I'm kinda jumping around a little bit, but I guess o over the kind of, uh, over the kind of years, obviously you had this kind of language interest developing you, you'd gone to uni, you went and sort of built a successful career in software engineering.

Um, were you, um, were you sort of building side projects along the way? Did you always have that kinda entrepreneurial

bug?

Martin: Yeah, so, so I'm gonna, I'm gonna strike from that sentence, the word entrepreneurial, because again, it is virtual. Being honest, I don't think I was entrepreneurial for a very long time. Actually. I was just creative. I was very interested in building things and helping people, but I wasn't interested in the money side of things.

And, and maybe that was a bit of a privilege because I was working in an industry where I could always find jobs if I needed them. Um, and I've sort of done that a lot. I've been in and out of like, okay, I'm working full time right now 'cause I need to, and as soon as I don't need to, then I'm much more interested in being able to take this solo route.

Um, but yeah, until very recently, I would say recently until maybe five, six years ago, I wouldn't have put the entrepreneurial word onto that. Um. At times, I would call them startups, but they, they honestly weren't because they weren't, um, not only these ideas started as how do I make a profitable business?

That wasn't the question. The question again was, was how do I make something that people like? And there's, there's lots of art in the world and there's lots of hobby projects and side projects and communities and grassroots events and open source that people like, and it helps them, but it's not a business.

So, yeah. Um, I did a lot of that and I've done that since the very beginning. Um, yeah, even going by ride to when I was like 12, 13 years old. I have projects there that I do reflect on some of those and think that's, that's kind of like an early version of GitHub. That's an early

version of this. Now, you know, we've all built things that were way ahead of its time in truth would never have ever become a successful company anyway.

But yeah, all of that kind of stuff. I built like a, a 3D, well, it wasn't 3D it was, it was two dimensional, but isometric, um, game essentially to teach my friends. See where it was like a classroom and a little house. You could walk around and chat to people, and then you could go into the little auditorium, and then I would walk upon a stage.

Literally, you would see it as a little isometric. 2D characters. And then I had a little whiteboard and you had little commands to write on the whiteboard. And I would teach my friend C, which was basically what I did. I'd learned C from these guys in this, in this chat when I was younger. So yeah, I've done so many projects and worked in so many different fields as more just a hobby for me to learn about the technology, for me to play around with it and, and build something that I think is useful.

And there's so many examples of that stuff across loads of different, um, di um, loads of different domains

Rob: yeah. Um, I

Martin: way to learn.

Rob: yeah, and it, yeah. Absolutely. And is, is, is Lang, is LangSesh your first kind of business

that spun out of one of these side projects?

Martin: There's been a few

things. Um, if I go way back, um, I did something in mobile that was capturing, um, statistics about the way people use their mobile phones. This was a long time ago, or maybe 10 years ago. I sold that, uh, it wasn't for a huge amount of money, but that was the first official like commercial thing that I did on my own.

Um. And I think maybe that was where I thought, okay, yeah, people will buy software. And actually, here's a side point I want to make as well. Actually, I think I see this with a lot of people in say, my and and your, and our generation of engineers. We grew up on an internet where you couldn't really pay for things initially.

And then we grew up on an internet where it wasn't really very common or or popular to sell things on the internet. And now we're in a point where everyone buys things and it's hard to sort of sometimes remind yourself that, that is the internet we work in now, the and back. And I sometimes, I think I've carried that forward and thought people are not gonna buy this.

And then you see what people really do buy. And there, there really is a lot of commerce on the internet. People do buy things, people do buy apps. But it took me a long time to really convince myself that I could build something that someone would pay for. Um, but that was probably the first one. And, and maybe that's going back 10 years now.

And then since then, um, yeah, I got involved in some startups with some friends and we did pretty well, but we did the whole startup journey where it was like build hype and not value. And we got a lot of interest. We, we, there was a good story, some crazy stories, fake investors that never existed. Big investigations that led to people that, uh, it turns out that it was someone screwing us over and all these crazy stories.

Um, but none of those things were really successful from a, in the way that they were, they were meant to be. Um, and then language comes onto the scene probably about, yeah, about seven, eight years ago when just after leaving the company that I was working at, I thought, okay, I wanna try and create my own value in this domain, in the domain that really interested me as a, as a hobby, as a personal, as a subject.

Uh, and I have some, some ideas about now. Uh, and then I worked on an app, a product called Yak actually. Uh, and anyone that knows me that maybe listens to this will think, oh yeah, you've been working on that for 10 years. Um, in name I have, but it's actually been three different variations of an idea. It's just the same brand because there's, it doesn't matter in my mind, it's the same, it's the same goal, which is to make someone's life better.

Who. Is aiming to learn a new language. And there was three variations of that app that were very different to each other. One was an iOS app that was based on just tiny conversational practice. The other was some kind of marketplace where people could sell tuition skills to people who wanted to learn language, and there'd be some clever match.

And then the more recent incarnation was, um, like digital interactive online books. And I've been working on yak actually until about six months ago. Uh, six months ago was the point where I sort of pivoted away from yak and went into Lang Sesh. They, they're not that different. Um, but then they are, it depends to what level we discuss it, but they're both language learning products.

Um, yak is out there. It's been out there for a while. Um, but the, the mistake I made with Yak, and again, very recent mistake is just not thinking about the business model really too much at all. Thinking if I create something that's valuable and that people like, then the money can come afterwards. I mean, I think there are cases where that happens, but there's this little bit in the middle no one talks about, which is someone at some point said, okay, it's useful, but to, for it to be useful and profitable, it has to be in this shape.

And sometimes there's no shape. You know, I remember hearing a great talk from the guy that ran, um, four Chan. He did a talk years ago, and he stood up on the stage and said, look, I had 10 million active users and I couldn't get a single one to pay me money. I was like, yeah, you know, you have, you've got users, you can convert it to Profit is like a nice idea.

And in some cases is, is definitely true, but it's not true. By by definition. It's only true if you can do that. And there are cases where you definitely can't. Uh, another risk of like freemium, and I did that as well, a lot of freemium of like, just put it out there, just get users, just validate the value and, and later you can get the money.

And so I, I, yeah, yak was a another case of that, of like not thinking about the business model and then arriving at a point where it's a cool product and I got a lot of great feedback about it, learned a lot from it in terms of what people like, um, and, and even what people will pay for. But I, I built a business that was sort of a bit of a scalp.

It's like the cost of acquisition of a user plus the cost of production of the thing that they wanna buy because I'm selling content. So, in other words, the content production cost plus the cost of acquisition of user was so close to the amount of money you could make off this thing. That the, the only way that business works is, is at scale.

And I didn't wanna build a business that can only work at scale.

Rob: Okay. Yeah. And so,

but Yakk's still alive, isn't It You're still running it?

Martin: It is, it's still alive and there's still this potential for it somewhere down the line. But I think honestly, I'm just going through that process of, of seeing whether this pivot makes sense. And if it does, I'll probably direct all my attention to that. In, in all honesty, I've not worked on yak for about yeah, about eight, nine months now.

Um, and I, I don't intend to go back to it unless there's some, some big reason to do that. And like I say, I, I don't think that's gonna happen because, um, the conclusion for me is that, you know, when you think about business model, if you've put yourself in a situation where the only way to scale it is to pull a lever of, of capital and say, if you inject capital into this, then we could go bigger.

But it's a bigger risk too, because if we can't make those sales at that scale. This capital just gets burned. I'm not sure I really wanna play that game. Um, I, I, I, I want a bigger answer to the question is this, is this valuable to people right now? Um, and, and Yak was also a pre AI product, so there was this big transition there as well of like AI and language learning, of course, have a very, very close connection there.

And, and it feels like that you can, like in many business domains now, small teams of people with AI can do a lot more than they could ever do before. So yeah, that plays into it to us.

Rob: Yeah. Okay. So the, the language company that you worked at, that wasn't your,

your most recent role, was it before you decided? 'cause, 'cause obviously now you, you, you, you're traveling and doing this, the digital nomad thing. Um, so. Even though you developed the interest in the language company, that wasn't the point where you decided, right, I'm gonna go for it.

Then it just, the timing wasn't quite right or the, or the idea wasn't quite,

hadn't quite sort of come together in your mind as to what you wanted to build.

Martin: So right after leaving that company, I did start to work on, on, on yak, let's say yak, um, iteration or incarnation one, and then later two and three. And in between doing that, I was bootstrapping that, um, I didn't, I didn't go for funding on that one. I was, at that point I'd realized I wanna try and build something on my own, on my own terms.

Uh, so I was bootstrapping, I was in and out of work in order to fund the lifestyle of working on that thing, but I was making a lot of mistakes and taking a long time. And one of the biggest lessons from that whole experience was, um, that I was moving far too slowly and didn't think it was a problem.

And it was a problem because I just wasn't learning quickly enough. Uh, it took me many months, maybe over a year to get something out there to begin with. So I'm learning too quickly, too slowly at that point. But, you know, this is one of the eternal problems with engineers termed startle people, entrepreneurs, let's say, is that there's so much joy in the building that you can easily distract yourself from.

The reality of you're not here to build, build is a part of it for sure, but it's not the only part of it. Um, but you get so much joy and so much creativity and so much freedom in. Out of that part as an engineer that you can really just mislead yourself. And I, I did that a lot, uh, in, in that time. So, yeah, to answer the question, I did that almost straight afterwards, about a year afterwards.

But while working for, uh, various different other companies, uh, at the time I got into a lot of, um, educational stuff. So, um, yeah, a lot of mentorship helping newer engineers. And then eventually my, my, my latest role was engineering manager at a, a big company in London for about 18 engineers. A again, not a role I particularly enjoyed, because I think it's a hands off role.

You switch engineering problems for people problems and realize, oh, engineering problems are not so bad.

Rob: No. Okay. So, at what point was it then that you decided to, to make the change that you have? Because, uh, you, you are, you're out in, um,

in in, in Vietnam at the moment, did you say?

Martin: in Thailand right now.

Rob: Thailand. Yeah. Right. Okay. Um, so, but you've moved around a little bit, haven't you, since you

Martin: that's right. Yeah.

Rob: Yeah. So what was the, what was the impetus, uh, that, that sort of led to that, that quite more big decision

Martin: Yeah.

Rob: did, and how did you make

it work sort of financially and everything else?

Martin: Okay. So, um, if we go back about three years, I, uh, my girlfriend and I both had, had the same sort of idea. Um, she's pursuing the same path as I'm actually, we're doing it in parallel, in entirely different businesses and sectors, both doing it together. The plan was let's, let's make enough runway for ourselves.

So we sometimes jokingly refer to it as raising our own seed round because the cost of this kind of stuff is so low. And, uh, when I say that, sorry, let me clarify specifically what I mean. Building is not expensive. It used to be, you know, I remember days where I'd have to like hire a service at $200 a month and, you know, that kind of stuff was expensive.

Just on the basis of that alone, let alone all the other things you would need, then you would like hire a designer maybe, or a UX engineer or these days, a lot of those skills have sort of become simpler through frameworks, patterns. And, and, and actually if you just, a lot of full stack engineers can do a lot of that stuff on their own.

Um, so we're at that point now where the building part of it isn't actually the expensive part. It may be the time consuming part, depending on what you're doing. If you're building a game, for example, it takes a long time anyway. Um, but the building part is, is very, it is very there. It's accessible. The, the expensive part of taking that route of, I'm gonna build my own thing, is that life is expensive.

Um, right. Like, you know, if you live anywhere in the UK at this point, you know, rent's way more than it used to be 10 years ago. And, and, and utility costs are way more, and food and everything. Life's just expensive in the western world. And we kind of have to mostly go work on things we don't wanna work on because to pay for that bit, it's not really to enable the building anymore because we can definitely do that bit.

It's to free up the time, uh, and, and, and then, and then pay for the, for life. So we both basically decided, okay, if we can't raise money from someone else, which is a bad idea, let's not do that. If we just save money and then pull the other lever instead pull the cost lever way down, sort of like the personal equivalent of like.

Let, letting go of 50% of your staff to, to improve your p and l, right? It's like the personal version of that. It's like, if we don't have to pay for an expensive apartment, we don't go and have an expensive wedding, and instead we go and live in Vietnam where the cost of living is a lot cheaper, then we buy that time back and then we have a real chance to do something.

Because, you know, as much as people talk about, you know, five six KMRR and you've had some great stories on your podcast, actually, uh, those are the exceptions. There's many more people trying to reach that level and not doing it in 52 days. Um, as much as those stories are incredibly motivating and inspiring, um, I think it's better to go into it with a bit of a, a more realistic, um, uh, realistic viewpoint and think, well, maybe it's gonna take me a year.

You know? And I already know my own history is a slow one. So, um, so yeah, that's, that's what we did. We decided, okay, let's save. So we worked for around a year in the UK and just saved every single penny we could. Um, and then we decided to move to, to Asia, and we came out to Vietnam originally. Um, the simple reason was just because, uh, Vietnams got just the right amount of infrastructure for it not to be.

It's too much of a developing country compared to some neighboring countries you might go to instead, uh, but also is not so developed that it's, it's basically like switching one problem and getting the, the same one back on the other side. Uh, living in Bangkok for example, is, I mean, actually that is still much cheaper, but it's, it's not as cheap as, as Vietnam.

So yeah, we moved to Vietnam and then we've been here for basically almost two years now. Um, living the lifestyle over here and there's been a huge surprising set of results from just doing that alone actually. Um, but the, the main reason was to buy ourselves the time, and that's what we've been doing for the last year and a half, is just pursuing our own ideas and trying to learn and find that founder and product market fit at the same time.

Rob: Absolutely. So, so are you, um, I mean, at this point, although we need to get, so get into the length session and some of the details there, but I mean, are you, are you still living off savings? Have you got kind of like a, is there a runway to what you're doing where you need to kind of hit a certain MRR

to be able to sustain what you're doing there?

Martin: Yeah, we both have. Um,

so we combine our efforts and we're, I would say we're about 40% of where we need to be for this to be sustainable. So we're not at sustainability yet, but getting to 40% I get into anything that's above zero is, is a huge, a huge thing. Um, because it's just so hard to not get there.

And I've done that so much before. So, um, yeah, we we're getting closer for sure. Um, but we are mostly still living off what we've saved to do this. So, um, yeah, not a glamorous reality. And, and I think that's a really important message is you can do this, but it's a sacrifice. It's not, um, it's not easy. It's not a privilege.

It's you choose not to do the other things. You can do this thing and it may not work. Um, but yeah, we both pick up jobs here and there when we need to sort of top. Up the money that we've got. So we could make it sustainable that way by doing the thing we don't want to do. Um, so, so that's a good situation to be in.

Um, but yeah, we're both seeing some early revenue now, which is, which is great. Um, but it feels like it's still gonna take some time because, you know, you just don't get it right the first, the second, the third time. Um, the more iterations the better. And the reality is we're still sort of like single digit iterations here.

So I'm not expecting to find that long tail hockey stick ramp up just yet. And that's good. And that's it. It's nice to be honest about that. It's nice to go into it naturally and say, okay, I'm gonna put the time in though, and I'm gonna put the effort in to make sure that iteration cycle improves, speeds up.

The learning is there and, and we're in it for the long haul. So that's

Rob: yeah, absolutely. And, um, I guess briefly, you're, you're

enjoying life out there, you'd recommend it.

Martin: Yeah, I mean, I'd recommend it for people who already think that they might like it. Um, I mean it's definitely, it comes with downsides too, right? Like. Um, I write about this on LinkedIn a bit actually. 'cause for me it's quite a powerful subject, but it's lonely, you know, especially when you take, you say, I'm not gonna raise money and therefore we're not going down this, you know, this route of make the company look bigger every year by hiring more people.

It's definitely the, almost like the polar opposite. It's like we're keeping costs low and that mostly means I'm working on my own. I do have some communities of friends and, you know, that does help and you need that soundboard sometimes, but it's not the default. And, uh, having worked in lots of incredible teams with people that are far smarter than me and incredibly motivating, um, it is definitely a downgrade from that.

Uh, just in terms of how you feel in terms of, uh, the kinds of, sort of creativity and ideas, problem solving that comes out of having those people around you to bounce ideas off and stuff. So that's the downside. And, and then there's a bit of a downside in terms of personal responsibility as well, because you're out here on your own too.

So you've given up a lot of the network and a lot of the infrastructure in terms social infrastructure, let's call it at home. The assault of happened by accident because you grew up there and you've got friends from school and from ex-colleagues and just, you know, the proximity and all that stuff.

When you give all that up, um, you have to sort of rebuild it, or of course have a very, very lonely life, um, which I would not recommend. So, but one of the surprising things about that is that it's actually very easy to do that here. Um, I didn't really know this until I came here, but, and it actually totally makes sense when you think about it, is that it's a self-selected bubble of people that are all doing the same thing.

So it is, it's like when, when you talk about Haskell, you also probably find quite a, a lot of mathematicians, or when you talk about S Brown, you find people who are really interested in languages because it's a selection criteria already. It filters the, the, the, the, the set of possible people. It could be down to a much more niche set of people.

And when you come over here, you go to like nomad meetups or you go to, um, you know, there's a lot of indie hackers, sort of SAS meetups here as well. And you meet up a bunch of people doing the same thing. They've taken exactly the same route. They've left their home country. They're in a new one. They're probably learning a local language.

They've been here for a few years. They're making a little bit of revenue. They found a way to make it work, and they really want to connect. They really wanna make friends. So I found making friends over it, actually really surprisingly easy. Um, so there were definitely, yeah, there were downsides. Upsides.

Um, I would say just, um, there's a lot of quotes about this, like, uh, having a second language gives you a second soul and those sorts of things. I think travel, real travel does that as well. Uh, real travel, being actually living in a place, not just being there, like integrating into it and, you know, living in the places.

Like some of the places we've lived in, in Vietnam especially, certainly eyeopening. They're not the same life that you're used to in the West. Um, but yeah. Yeah, it's, it is, it is fun for sure.

Rob: Yeah. I mean, the photos I

keep seeing pop up in your feed are amazing.

Martin: Yeah. Well, it's like, but, but bear in mind

it's, it's just like any Instagram, uh, you

know, those are the nice, those are the moments, the one moment in the day

where you're like, damn, that's a really incredible sunset. But it was raining half an hour ago and I didn't take a picture of that. So.

Rob: Yeah. Uh, great. So what's, and kind of your, uh, I guess your, your, your kind of medium to long term plans there, do you think it it is probably not gonna be forever. It, it's a great experience for now?

And, or, or what, what are your thoughts?

Martin: it's a good question. I,

I think it's really important to stay open-minded about these things. Like, uh, I meet a lot of people out here that absolutely have already decided that they will not go back to the uk. Um, because the, it just, the economics don't work for them. And I, I get that and I, I sort of feel a little bit of that myself too.

I will actually be back in the UK at the end of the year, um, for a few months and I have no plan for that. So, uh, yeah, it's nice to have no plan. That's one thing. Another thing you have to embrace, I think, in this lifestyle is not really knowing what's gonna be next and being okay with it. I mean, I say this lifestyle, you could take that to mean the world of, um, business or the world of travel.

'cause they're very similar in that sense. You really don't know what's coming next

or what you're gonna decide. Um, but I think, I think it's probably likely, and it wouldn't surprise anyone I don't think, that knows me. To find that I'll probably be back, back in, in Thailand or in Bangkok at some point. I think Thailand is a perfect country for the balance of cost, culture, um, and the sort of mix the, the sort of the international mix of people that you get in a place like that.

It's, it's an incredible city. So I would like to be there. Um, but yeah, it depends. It depends on so many things.

Rob: Absolutely. Okay, great. Alright, let's, um, let's dig into LangSesh a little bit more then. So if I could just start by asking you, um, I guess if I don't have the, the first clue what it is, what is it, and sort what does it do?

Martin: Okay. So, okay. Elevator pitch time 10 seconds ago. Um, so it's, it's a tool. I describe it as a tool. Right now it's a tool to help people who want to learn a language. Um, and it's aimed at people who are, I would say ideally slightly beyond complete beginner, although I'm trying to address that. 'cause I think it would be helpful if it, if that wasn't true.

Um, but right now in the, in the form that anyone listening to this, this second, for example, would see, um, slightly post beginner, um, but absolutely not, um, post intermediate. So it's that, that gap in the middle of, um, let's, let's just for simpl simplicity sake, say beginner to, if anyone knows the official names I'm talking about basically to about B one level, which is an inter, a weak, intermediate level.

Beyond that, I think, um, there apps have very diminishing returns because once you reach that level of any language, the best thing you can be doing is probably watching tv, meeting people, talking to them. Even any form of real immersion is gonna be better than some form of synthetic immersion via an app or via a book or anything like that that you consume.

So it's, it's a tool that helps people learn language. And the way that it does that right now is it generates different kinds of exercises. Right now, only reading the plan is to go into audio very, very soon as well, or listening and, and reading are both very important skills, in fact listening, uh, particularly so, and, um, it generates small exercises where people are encouraged to read them and then.

Complete a series of challenges based on comprehension of that reading exercise. So for absolute beginners, this is a great way to learn a language, is, is see, the real language used other, excuse me. Other apps have the problem of, um, here's a random sentence. Dual lingo famously did this for, for many, many years.

Here's a random sentence, what does it mean? Um, there's some stupid, usually something to do with having a bear or a tiger eating your grandma or something like that. Really, you know, not very useful phrases. Um, the Lang sesh goes one step further in the sense that it's saying, here's a, an actual story, a very short story, um, that you're meant to read as a, as a basically a beginner, and maybe you won't understand the whole thing, but through the exercises that follow, you'll de decompose the sentence into its pieces, uh, sorry, deconstruct it into its pieces and actually start to see connections between the relevancy between certain words, certain grammar, and things like that.

Um, it's not in its final form yet. Lang session is an MVP and um, and, and of course an MVP as an experiment is you think it's viable. Uh, you even think it's minimal. It's probably not even as minimal as you think. That's another lesson from recently as well. Um, so, so I put out Lang session. I would say it was like MVP iteration one, and what I've learned is that it can be more minimal and version two is more minimal.

And in terms of viability in some senses, it hit a little bit of that. I got a lot of great feedback from people about how it works and the experience that it gives them. Um, but in other ways it's not viable because people are not using it reg regularly. People are not coming back and using it again and again.

Um, there is a drop off and that that's not ideal. And there's big question of why. Um, so yeah. Uh. That, that's what Lang session to answer the question. That's what line session is. It's a tool that generates short reading exercises and challenges off the back of them. And then it has a little bit of gamification in there to show you what, where you've come from, like how many words you've learned, how much time you spent studying, and one or two more gamified elements of like streaks.

Um, streaks actually in Lang session are a lot, uh, carefully designed not to be sort of, um, the painful streak that you get elsewhere and you know, you will learn for 30 days, then you lose it and then you think, oh, I'm not gonna do anymore now because I've just lost 30 days. It's not, it's ironic because it sort of motivates you until it doesn't and then it demotivates you exactly to the degree that you are motivated, but in reverse because you're like, I've just lost this thing.

I feel terrible. Of course, Duolingo commercialized that and started selling it back to you, but I'm not gonna do that because going back to the virtue of the beginning of the conversation, this is not an honest thing to do. Um, but what it does, what Lanchester does with streaks is it actually shows you your best streak.

So it's like a PB in the gym. If one day you went in the gym and you deadlifted 200 kilograms, that will always be true. Always, every time you go in the gym, you can celebrate that you were able to deadlift 200 kilograms at some point. Now if you ever beat it, great, but if you don't, that will still your best.

So streaks work a little bit like that in Lang ses. Um, but yeah, that's, that's what the current product is, that's live now. Um, but I, I've been in the depths of, of iteration two for the last two, three weeks.

Rob: So, 'cause I had had a little play with it. I mean, learning sort of foreign languages isn't really my, my world right now, but had a little play and I've, you know, I've obviously used Duolingo before, so I was sort of very aware of the difference, I suppose, between the, the two. Like I say, your, your product just felt a lot calm or it didn't have loads of gimmicks going on and kind of crazy animations and sounds and stuff.

Um, and it seems like, so it's trying to draw you in by you, you sit down and you choose some settings in terms of the language you wanna learn, and then it's, and then it's session based. Is that the idea? You sort of say you're gonna do a few minutes, sort of commit to doing

a few minutes of, of practice.

Martin: Exactly. Yeah. The name Lang session is sort of like a bit of a, a nod to that. It's like, it's just a quick session, um, to practice the language. That's it. And the, the value, uh, proposition right now is just turn three minutes into some language learning progress. Um, but I'm, I'm sort of shifting away from that.

Uh, as you mentioned at the moment is it does feel quite minimal and I'm quite happy with the ux, but there are bits of it that I'm not happy with. And, and, and in spite of trying to make it as minimal as possible, for example, with, you don't need to sign in, um, you don't need to do anything. You can actually click one button and just immediately use the val.

The value is right there. It's behind one button. It has to be one button because you might not, I can't assume which language you wanna learn. So it, I've put a lot of effort into making this work, um, without you having to do anything. Nevertheless, there is, um, there's this whole section where you're like, well, you can change your language.

You can see the exercise types that are coming but not implemented right now. You can choose the difficulty level and the subject that you talk about, or rather, that you read about is chosen randomly based on some emojis. So it's sort of like a, a one of those sort of like fruit machines that spins three emojis and it picks three random things.

And then that's what the subject that you read about is based on. Now this is random, and this is kind of one of the complaints, um, thrown around with dual lingo and, and it's fair to throw around with EIS right now, which is. I don't wanna talk about random things. I probably have a more specific goal.

Like if I'm learning Spanish to go on holiday, I probably wanna order a beer. I wanna learn and how to, or say, can you gimme one? Or can you gimme three or four or five? Um, you know, certain numbers, certain expressions, those are gonna be important to a certain context and that is not being captured right now.

And, and that's why I described it originally as a tool. It's a tool that was sort of an experiment of like leveraging AI carefully in the background to produce these, these reading exercises based on these, these parameters. Does it work? Can it produce good language? Can it produce language that's constrained in specific ways based on the difficulty levels?

I think I've learned from that, that yes, there's a lot of potential there, but the product is, is not productized that very well right now. And, and that's shifting to a core structure. Now. The problem with flying session at the moment is that people don't know how to continue to use it effectively because it's just a tool.

It's like if you give someone git and you say, go ahead and use gi well, it's like, well, I need to learn the patterns first. I need to know how to use this thing. Lang session is a little bit too much like that and a little less like. Okay, you're a beginner. We know who you are. We know how to help you right now.

This is how we're gonna help you, and this is where you are not right now. This is the direction you need to head and that's where you're gonna go. So there's not enough structure right now. That's the big, the big

lesson from speaking to everyone that's

Rob: Makes sense. I think, you know, you're learning a new language, you want that kind of roadmap, don't you? You've gotta get from A to B and this is how it's gonna go

to get you there. Um, which I, I feel like Duolingo does to an extent when I, when I've used it in the past. But as you say, it's a bit chaotic.

That's what it feels like. You don't really know where you are. You're kind of going through this kind of fake world of whatever it is, getting the, uh. I can't remember getting the badges or whatever, and it's just, you know, um, so, and, and, and I, um, and also like what you're doing with, um, you know, you're sort of saying, okay, I'm gonna commit to doing however many minutes practice.

And that, that's what I'm saying. It's not this kind of like, well, we're gonna try and just keep you on the app all day, you know, um, and get you into this kind of really addictive loop. It's like you want a little bit of that. Maybe you want people to come back the next day, but getting 'em into this strange

state where they're like,

it's like,

Martin: no, the gamification thing is like, it's, it's, overdone in Duo Lingo for sure. I'll give Duo Lingo credit for great UX for people who are very, very casual. They maybe don't even wanna learn language, they just wanna do something that's different from doom scrolling Instagram, uh, in which case Jolan goes way better at spending your time than doom scrolling Instagram is.

Um, but it is, it is still, I think in that, on that spec side of the spectrum of like, it's much more of an entertainment approach. Uh, and I'm trying to be definitely aiming it more serious actual acquisition because again, going back to my original point, I'm building this for me. So right now I'm learning, um.

Indonesian. Um, and I want to use Lang Sesh to help me get better at saying basic things in Indonesian, but that definitely requires structure because I come straight back to the question of, okay, what do I know now? What do I need to know from here to progress properly? I can't just jump to certain constructs and certain parts of the language because I'm not ready for that yet.

Uh, and those, that course structure is kind of what yak was it that? 'cause those were books. Those are very curated, expensive to produce resources. This is a similar thing, but it, but the, the cost of production of that content is a lot less because it can be generatively produced following various different guidelines of like what a, a core structure should look like for Spanish A one or German B one and, and that kind of stuff.

So, and then another, uh, one other thing I'd just like to mention as well is, um, pricing. So another big lesson from this first iteration, this first MVP experiment, um, was that the pricing model was completely wrong. So, um, if you looked at Lang session recently, you'll see now that it's monthly, it has a monthly offer, but it didn't have that.

Two weeks ago. Um, so the version that most people used and most the feedback came from was a version that used credit system instead. And this is a common thing with AI products because as a, as a, as the business owner, of course, you are metered on your usage. If you fire a bunch of API requests to some model, uh, let's say it it 10 x is overnight, then you're gonna pay 10 x more.

So it kind of makes sense to build a cost structure for the consumer that at least mirrors that. And then, you know, one power user doesn't take away the profit of three moderate users, let's say. The problem with that though, is again, if you think about the user, they don't care about that. They don't care that it's me like that that's your problem.

So I made the mistake of going down this credit route and yeah, I was offering people like, okay, you can buy this many credits, this many credits, this many credits, and there's just too many buying decisions. At this point, people are like, what do I need credits for? And once I figured that out, how many do I need and should I buy more now so I don't have to keep buying them every month?

And best case scenario, the people that did, some people did buy credits and it was cool to see that. But best case scenario, those people would be presented with another buying question. Like as soon as they were out, they have to buy again and buy again and buy again. Nobody wants to do that. Um, it was also too expensive.

Uh, the, the business model itself was just pricing this too high. So moving to a, to a monthly model, I think is the right move, but not until the product is also right. So yeah, structure, structured learning and, uh, and, and, and pricing are the two big takeaways right now.

Rob: We, we touched on exactly the same issue actually in the last, last episode. And, and, and the founder came to exactly the same realization because everyone's pushing this idea of usage based pricing. It's great, great for you. Right? But I mean, from, from the user's point of view, it's a real turnoff.

And actually, I was just thinking now about like, chat GPT as an example. Um, so no, I'm on their pro for that. I dunno, it's like $20 a month or something. But if it was, uh, I had to buy credits and it was x whatever cents per prompt, I wouldn't, I wouldn't sign up, you know, I'd just stay on the free.

You just wouldn't, 'cause I don't know, like, I don't know. I mean, I'm asking you stuff all day, but I don't

know how much that's gonna cost me. It

Martin: don't even wanna think about it, right? No,

you, it needs to be there and just be like, useful. You don't think about it. And I think for that reason, the pricing model has to be on the monthly side, has to be right? Because if it's too expensive, um, then people are gonna be like, well, do I use it enough to justify this price?

But LangSesh says she's now priced at $7 a month. So I think it's at that point now where it's cheaper than most other language learning things you'll use. And if I can show you more value than I can currently show you, which I'm confident about in the next iteration, then hopefully that's a much as a overall picture, that's a much, um, a much bigger offer for people.

It's an offer that hopefully is a lot less easy to refuse. Um, but we'll say, this is the great thing about this whole world, right? Is that we build an experiment and we try and we see what happens.

Rob: Yep. Um, okay. So in terms of how you. Came up with the idea, obviously we can kind of get a rough sense from that on the basis of your interest in, in languages, the, the, the, the yak sort of story. But I guess what was the kind of, uh, moment of realization that something like Lang sesh was the

way, way to go with all this?

Martin: That came out of, um, pursuing this, uh, this product that I was working on just prior to this yak. I worked very for a very long time on that product. The, the, the last iteration of that was about a two year project. Um, it's way too big. It's full of all of the mistake. If anyone wants to talk to me about all the mistakes, to not to try and avoid that is a perfect, like source of all of them.

Uh, it has way too many features. It has way too much supposed value in there. And it was such a complex project, uh, product that when it was launched, most people didn't gimme much feedback about it because it was just too much in there. And there's another lesson for me is like, launch. The simple thing, not because, not necessarily only because you can do that quicker, but when people see a simple thing, when they see a page that's got two things on it, they're much more likely to tell you what about this, or you could do this different.

Because they can, they can already understand the whole thing. So they're empowered to, with the full understanding of what the thing is, give you their ideas about what it could do instead. But the second you present something that's like, so refined, I worked for a company actually once that did this.

They acquired another company in the exact same domain as them. I'm trying to anonymize this sufficiently. And the, but the, the business they acquired was like the, it was the Asian version of the thing. And it was actually from a UX point of view, terrible. Absolutely terrible. Um, and the western version, the western company that acquired this Asian company and brought it under the umbrella of the big company, of course, thought, well, we need to unify this experience.

We can't have two things that look completely different under the same brand. So the, they improved the UX so much that they lost the entire customer base. I. And it's one of those interesting things of like, maybe people actually liked the really bad UX because it signaled to them that it was cheaper.

And that's why they went there because they thought they were gonna get the cheaper prices from the, the crappy ux. Uh, and now all you've done is you've signaled that this isn't a really expensive Western alternative now, and they're gonna go look for the cheap one again. That doesn't exist anymore. So, um, yeah, that with Yak, I, I learned that lesson that I almost did that same thing.

It was like, here's the big refined product with like 18 features and all these things, and very few people had much to say about it, probably 'cause they didn't have the time to really sit and think and understand it. But, um, to answer your question, when La Yak launched and I started to get feedback and, and, and, and think, okay, what's the next step here to turn this into a profitable business?

Again, asking that question far too late, um, the answer was, find a way to produce this level of content at this quality curated language learning content. This as good as like taking any, you know, successful language learning book off a shelf in a bookstore and saying, how do we recreate this more cheaply?

Maybe there isn't a way to answer that. Maybe there's no answer. You know, maybe that is actually the lowest. It's already raced to the bottom. That book is already at $10. You're not making it any cheaper than that. Um, yak had that problem. And, and so that happened around the exact same time that, um, AI was, was really just present everywhere.

It was, it was, it was, it was suddenly capable of doing so many things. You know, the transform architecture had boomed into these huge big L LMS with in intensely massive amounts of parameters that suddenly could accidentally speak every language. And, and, and it became clear that there was an opportunity here potentially for some kind of entry level learning app to have content generated, not, not produced by a human, but generated.

Now you couldn't, some people said to me, why not use AI to just produce the books for yak then? Uh, that's not possible. For the same reason that that go atch GPT to write the next iteration of Harry Potter. And it's probably gonna be really, really bad. Um. Um, that's exactly the same reason you're not, that's not gonna work in yak too, because these are books.

These are books of like 50, 80 pages with a very, very structured approach that really someone who's gone through the processes understands. But an LLM is not going to be able to do that consistently. So Lang session was really just, okay, if you basically took yak and simplified it back down to its essence.

But the exercises that people could do again and again and again, were generat generated, would that work? In other words, the hypothesis is LLMs are, are one of the most incredible tools for language. They are language tools and l LM is a language model. It's a model for language. It's the exact perfect resource or capability that never existed before that aligns so perfectly with much of early language learning.

If you're learning Spanish right now and you go and sit and chat GPT and you say, okay, tell me I say hello. Okay. Actually I'm in Argentina. Tell me. I'd say in Argentina. Okay. I've said hello, I, I wanna say, how are you? And then you say, okay, what, what does that exactly mean? I break that down into individual words so I can memorize it better.

Like, what am I actually asking here? Am I asking, how's it going? Am I asking how are you feeling? In an Asian language, you'll realize that actually probably means what did you eat? And you can, you can sit and check GPT and get all of this. You can say, okay, here's the sentence I think I can say, is it right?

And it'll accurately tell you in almost all cases, no. Oh yes. And that'll break down what's wrong. This capability is incredible. We've never had this before because modeling language is impossible until now. And now we did it by accident, which is insane to think about. Um, but, and it breaks down at the higher level like it does with coding, like it does with many things At the very top level.

The context is just too nuanced and there's, and it does not fit the training set. So we get this disconnect where the result from the LLM doesn't match quite what you wanted because you now want something that's much more nuanced and specific that happens with language too. But there's definitely room for products in the way before that point for these beginner learners, intermediate learners that can absolutely produce good content that people could consume when it's not important that you read in the news or, or watching a Netflix show because actually you can only say five things in the language.

And what you're trying to do is learn 10, it can produce that kind of content. So yeah, it came out of that. It was like Yak is in this horrible position where it's very expensive to probably like churn this machine and try and make profit. There's this opportunity over here that employs AI and it sort of changes the product so much that it is definitely a pivot.

It's not something that you can sort of shoehorn into the same approach and, and that's exactly what I'm pursuing now with Lang Sesh.

Rob: Yeah. And did you go through a. A process of validating the idea in some way before you started charging for it, before you even had the kind of fully refined product to even just check that you're kind of on, on, on the right track. Whether

this idea in your mind even makes sense

to anyone.

Martin: Yeah. Um. I don't have a great answer to this really. Um, I wish I could say yes, I validated it perfectly, but I just don't think that's binary. I think, um, you can have a conviction about something and I think when you build for yourself, you know, you are part of that as well. So I would say initially it was just simply as a language learner, this would be really cool if it could do all these things.

And I had this idea in my head of, if it could do these things, it would, it would definitely help me. It would help me right now with Indonesian, uh, and it's basically what I was already doing with these lms, it's just prioritizing it and putting it into like a much nicer ux. You could do this thing directly with a model if you want to.

It just wouldn't feel as nice. So I was, I was convinced that the capability is there. What I, what I wasn't convinced is whether people will buy a productized version of it at the cost that I, I will offer it. Um, and I've always actually you are a great example of, I'm seeing your post on LinkedIn a lot and I like that you just say what I'm thinking a lot of the time and it's almost the controversial, taboo too simple thing, but sometimes it is just, it resonates with me as being That's true.

And, and this idea of. You know, building the landing page and saying, would you pay for this thing? It's kind of useless. Like, what do they even imagine that you're selling them? Potentially, because you don't even know. So how are they gonna know? And, and, and what you write on that landing page is probably gonna change.

Um, so I, I just think sometimes the best thing to do is actually answer a different question, which is, how can I just give them something quicker? Um, and, and that's what I try to do. So, so Lang session was about four months of work. Again, quite slow, um, but way less slow than, than the two years prior with Yak.

And I just put out there, and I've learned a lot from that. And I think I've learned more from that than, than any other kind of attempt at validation. Um, but that is with the context of, I did speak to a lot of language learning friends. I'm, I'm very active in a lot of communities thanks to Esperanto.

And, and everyone basically said to me, this is a cool idea, but I take that with a pinch of salt. 'cause I'm like, yeah, I think it's a cool idea too, but what's it gonna look and feel like? And how is it gonna specifically, uh, make you, uh, how is it gonna make you better at this language? Like those, those are the details.

Those are the things that you can only answer by doing them.

Rob: Yeah, I feel like, um, a lot of people sort of get, get a little bit scared off by this word validation and, and maybe we need to do something, um, about that because I think it's more about just, you know, trying to involve people like you who have the same problem as you as early on as possible. And just seeing if, uh, you know, just to check if it's just not, not just you, you know, and, and also, you know, to get their early insights that are gonna help you sort of shape, shape the product and get you to, you know, an MVP that actually you can, you can, you can go somewhere with, um, so I mean, it sounds like you were, you were talking to, to pe to other people who were interested in languages and wanted to

learn when you were, when you were sort of building

Martin: yeah, yeah. I, I, I hesitate to say validation, just 'cause I don't, you know, I don't want to, I don't wanna be like, yes, I did a lean startup canvas and then I did this and that, that I didn't do any of that stuff. I, I just spoke. I already have a good sense of what I wanted. And again, I, I, it needs to, it's important to me that I build for me here.

Um, but yeah, I did speak to a lot of people, uh, in my, in my friend group that are all language. I mean, I have some friends that are absolutely incredible, 16 languages fluently and this kind of thing. And you know, when I speak to them about it and they say, yeah, I think that would help me. I kind of do that already in, in chat GPT.

So again, I don't know what they're imagining and I still have to present something that hopefully maps to what they think I'm talking about in that conversation. So it's not really validation, it doesn't really validate it, but it to, to your point though, I did get some ideas from that of like, okay, the emoji thing came outta that conversation.

Incidentally, I'm taking that away now. But yeah, some ideas come out of that and, and, and it gives you a little bit of an approach that gives you a conviction that, okay, I'm gonna try this.

Rob: Uh, I think you put it really well earlier, that idea of just putting the simplest thing in front of people as soon as possible. So that you're not sort of hemming things in and kind of making all these decisions upfront, you are offering to 'em and saying, well, look, you can kind of be part of this and you can, you know, I haven't got it all figured out yet.

I kind of need your input. Um, and that's just gonna be more, more, more inviting to, to sort of, to early users in

terms of giving you that feedback you, you need, you know?

Martin: true. true. And if you're building, if you're doing that with the people that are potential customers, that's even more ideal because if you really are gonna produce value that doesn't exist elsewhere, then those are the same people that might buy it. And, and I'm in, actually in that position. I'm glad I'm in that position because quite a few of the people that have bought, um, already, um, it, it LangSesh is not making a crazy revenue.

Rob: Okay, let's, let's move on to talking about the, progress of the business itself. When did you actually, I suppose when did you launch? But I mean, you, you've probably effectively launched a few, few times in different sort of iterations, but what was the point at which you started charging for LangSesh?

Martin: So I launched about. I don't know a specific, I think it's about seven weeks ago. Um, I'd been building up until that point, but it was public about seven weeks ago. I'd done, uh, quite a bit of testing actually internally until that point. So some people had seen it, but there was no way to pay for it and I didn't wanna launch something that didn't have a way to pay for it.

I was, you know, trying to avoid this, um, eternal mistake that I've made in the past. Um, so yes, I implemented, uh, payments. I also added, uh, referrals and voucher codes. I was insistent on that being a useful marketing strategy afterwards. Turns out no one cared about that, by the way. Uh, a, a perfect example of just wait for it, you know, or do it manually until, until you see that there's a reason not to.

And even someone that gives people that advice, uh, fell into the trap of, of doing, of doing the opposite, building the automated thing that in the end, nobody cared about. And that delayed the launch by about, uh, three weeks at least. So yeah, I launched it after testing internally with a bunch of language learning friends, and then it was possible to pay for it from day one.

That was about eight weeks ago. Um. And then, yeah, from, from that point there was a bunch of people that, you know, went there from where I, it was, it is all driven by me. So talking about it in basically every place I can. It's funny listening to some of the other episodes actually where, uh, people talked about the same, you know, eternal problem of everything, self-promotion nowadays, like the second you even, I think people are almost at like the, the pre self-promotion detection stages.

Like, I think this guy's gonna talk about something he's working on Ban him now. Um, it's, it's hard these days to, to, to talk, talk about stuff. Um, uh. Um, which is fair. I mean, I guess there's a lot of spam and, and, and rubbish on the internet, so I understand why communities try and protect that. Um, but it, it is a hard thing to get people to pay attention to anything these days.

Um, so it's all, it was all me, but that was good. Uh, there was a lot of people that went to it from all the various language communities that are already in some of the, some, some things spread out that network as well. Uh, and then it went over the course of eight weeks, it went from just me and a couple of people to around about 300 users.

So it was quite sign ops and, and, and sign ops. Actually, in Lang, Phish means something because you don't have to sign up. It's an active choice that someone makes specifically because they're asked to, so they don't lose the progress. Uh, but you can continue to use the product without doing that. Um, so yeah, about 300.

And from that, uh, there was a couple of sales that, it was about six sales of credits, uh, and the cheapest credit. But some, someone actually bought the most expensive credit by the, and then a couple of other people bought smaller ones. Um, and there was about 50% over overlap in that group of people who know me as well.

So that's another useful thing to, to, it doesn't really mean anything. It's, it's very nice support, but it's not necessarily a signal of value, uh, in the, in the, in the purest sense of I want this, and I, yeah,

Rob: so you, you've actually only had the option to pay for about, about eight weeks?

Martin: Yeah. Since it was publicly available, about it. Yeah. I think it might, it might be, it might be eight weeks, it might be 10. I can't remember, but yeah.

Rob: Yeah Yeah Okay. And, um, and are you still on, on the credit based pricing or did you change that Um, quite quickly.

Martin: So since then, I actually got into this incredible, uh, it, it's the perfect situation, right? Like you launch something, a bunch of people use it, and then a bunch of people give you feedback. And I was very lucky to get feedback from like, not everyone, but I, I suppose maybe 30% of the people that used it sent me an email, responded to things.

I I also reached out to some of those people as well. Uh, and people were very, very willing to gimme feedback. Some people wrote me pages of email feedback, um, incredibly useful about what they think works, what they think doesn't work, what I could do to promote it better, and all these other things. So I had a bunch of feedback and some of it was very specific.

Some of it was like, this particular exercise is not helping me. Um, which, which is very small feedback, but it's useful to know that. So I basically composed a huge list of like, okay, iteration two work. These things all need to be right for me to present. In theory, the best thing. And, and of course I have to agree with these things, you know.

Uh, just because someone says that they should have this thing or not have this thing, doesn't not, you have to assess that somewhere yourself of like, what's the vision of this thing? Do I agree with this feedback? But for the most part, the feedback was, was either confirming something I was doubtful about or sure about anyway.

Um, so I, I made the big list of work, went through it, and I was releasing as, as quickly. Some days I'd released four times. Uh, as soon as the the work's there, ship it, you know, a big fan of. Um, these principles of, of lean, um, you know, all the way back from the Toyota production days of like, inventory is not something we want, it doesn't make sense to build things and not ship them.

Sometimes you can't ship them though, uh, because, you know, it's half of something that's not usable. Um, but a lot of this early work wasn't like that. It was, it was like three or four hours of, okay, now this exercise works properly. Ship it. Um, you know, funny, the, the, the bottleneck then becomes how do you tell people about this stuff?

You can't be telling people four times a day that there's new stuff to check out. Like there's, there's, there's a whole different problem of like spam. So some of that stuff people learned about much later and, and that I gradually talk about it a little bit more often. But yeah. And then more recently the, that work, um, was the sort of low hanging fruit followed by the harder things and the harder things were these new sort of experiments which are still very much untested of does this pricing model work?

Is this new course-based structure better, uh, more effective? And, um, the pricing model was changed and that was released immediately. So you've got this weird middle ground of like, the product has a different pricing model, but it's not for the product that it's meant to be priced at yet. Uh, but it was work that was done.

And, and if people think that they wanna pay $7 a month for the current version, I can almost, in my head, I'm saying to myself, I guarantee they'll wanna pay $7 for the version that's coming because it's much better. Um, but yeah, I shipped the pricing change as soon as it was, I was able to do that, but it was a big change.

Uh, I'd gone from like, you know, single, single purchase. Um, workflow to, to now recurring where you have various different states you have to deal with. Like people's, uh, people can cancel the membership on like the Stripe side. They can update the payment details. Money isn't taken, how do you withdraw service at a certain point?

All that kind of stuff. Um, so yeah, ship that. And, uh, I think I'm probably about a week away from being able to ship this, this new update to the, to the entire workflow, which is gonna be a, um, not only a course-based structure, but it's gonna focus much more on conversational learning as well.

Rob: So, so I guess only a couple of, couple of months in the market and quite a few things in flux in terms of pricing and, and also functionality. Um, so probably still fairly on, early on in terms of MRR.

Martin: Yeah. Oh yeah, definitely it's, it is the very early days. Like I don't think right now it's possible to say this is not gonna go anywhere. Um, and it's definitely also not possible to say this is definitely gonna work. It's not clear yet. There's, there's some early signals, but there needs to be much more data of people.

And in the early days it's hard because you, it is so hard to drive attention to these things that, you know, partly the reason it's gonna look like hockey stick if it's successful is because, you know, you just have to keep building. You just have to keep saying, like, for every a hundred users that I get there, how many of those people are serious language learners Anyway, even if I'm getting 'em from the right places, I'm reaching out to these communities.

You know, how many of them have, are only checking out because they're curious what I'm working on. So it's hard. You have to have enough data to really make a, a good decision about that. But, um, yeah, it's, it's very much, it's very much early days.

Rob: There is, but there is, is is there some revenue coming in at this point? The sort of early

Martin: yeah, it was $60, about $60 a month. So it's, that's, that's big for me because, um, you know, the attempt prior to that were, were nothing at all. 'cause I wasn't thinking about how to monetize the thing. I wasn't thinking about this, you know, this, um, everyone talks about it, don't we? But it is a kind of a useful phrase of like an offer.

People can't refuse the Alex Hall Mozy approach of. And, and I'm like, was that enough that people can't refuse? Um, no. You know, the credit, the credit system and the tool and, you know, they use it once and they're like, okay, this looks nice, but how do I get, what am I supposed to do with this thing? And, and as much as I think as the builder that people might be able to sit down and ask themselves that question, they won't, the it, there's, it is almost like a sub thing.

It's like they're not actually asking the question. They've got a feeling. And the feeling is, I don't really know what to do with this. And they just close it and they're gone. They never think about it again. So yeah, without that structure, I think, uh, it was not, it was not enough for people. Uh, couldn't refuse.

Uh, and hopefully it'll be closer to that with this, this, this one vote. I have to wait for the data. It's still very much early days. Yeah.

Rob: Yep. Okay. And how are you getting new users to the product, to the site at the moment? What's your sort of main channel, as it were?

Martin: So I thi this takes a lot of energy actually, but I am, I active in a lot of different communities that are all in some way language learning communities or, or adjacent to them. Um, there's so many more things I can do here. Um. Podcast episodes is something I like to do more of, uh, writing to some degree.

I tried, I write, I did a language learning newsletter for about 12 weeks. It was slow growth, uh, and I've not got the capacity to do that properly. It was, I grew to about 130 people. Uh, I also had like two other language learning products before, um, with several thousand users. So that's a future thing.

I'll reach out to those people at the right time, but not now. Um, so, you know, there's some, there's some. Future ways I'm gonna do that at, since, since the launch. It's basically been me in those communities. And that's a mixture of Masteron Twitter, uh, sorry, X Masteron X uh, even Facebook has some big language learning groups actually, so that's also been quite useful.

Um, personal Instagram, not so much. And Reddit not so much. Uh, sorry. Uh, LinkedIn, not so much. Uh, Reddit's hit and miss, I think. Uh, some stuff just gets immediately deleted and. And once you, once that happens a few times, I think you're basically like screwed your, your account's like marked as like low calmer and no one's gonna see anything you post anyway.

Uh, and it's also very disproportionately, um, useful because you have to engage a lot on those, on those things. And so, yeah. Um, I'm trying everything, I'm very open to, like, if there's a bunch of people I think I can help, then I'll, I'll, I'll try and help them. Um, and, and incidentally, quite a unique one is, is virus bronto because.

It won't surprise you to find that many ESP zos because they're interested in language in general and they're learning another one. Um, so it, and there's a lot of them. And so the product supports Esperanto actually, and I've had to work specifically hard to make that possible because LLMs, unlike every other language that they do support, don't support ESP Brando perfectly.

There are some edges that are not so clear because the trading data is gonna be much less for ESP content on the internet. Um. Uh, but, uh, Langhe does have support for ESP Brando, and it works relatively well. In fact, I've even learned some random ESP Brando through using the pro my own product, which is a weird thing for me to, at this stage run into.

But, uh, yeah, uh, through ESP Brando, um, friends that, that's, that's been helpful as well. There's a few people that, uh, I know entirely through Espan and they're like, oh yeah, I'm learning Spanish or German, and there's another avenue here as well of more languages. AI can support more languages, but I'm being very conservative and not just adding like 18 languages on the site because I want to make sure that.

It's mostly working properly. There are some difficulties with certain languages, ones that are written right to left. For example, ones that don't have clear word boundaries like Thai. Um, I can't just, like, some of the, some of the code is not gonna work properly with those things. There's assumptions in there that, that is not true.

So I'm not

Rob: be easy to just switch, switch all those on, but they've gotta actually work well, having the experience that's gonna be right for

Martin: And it's really important to me that, I mean, this product does not exist if it's not a good experience. Uh, that's probably true with most products, but this one, especially because the capability is. Being provided by AI in, in a significant way. The thing on top of it is, is the product, the experience, the structure, uh, the, the, the choices of the exercises.

In theory, you could sit down with a, with with GPT or a different model, whatever you want, and actually just say, well, I'm going to generate me read an exercise now. Generate me another exercise that takes out three words and, and, and tell me whether I put 'em back in the right order. And then generate another exercise that breaks some words out into grammar groups.

And I'll tell you which ones I think they are. I'll tell you, that's a verb and that's an adjective. Um, the product is, is, is producing that structure and, and those assumptions for you and, and productizing it into a nice ux. Um, so it's really important that that works well for me.

Rob: So let's talk a little bit about the, just, just, just briefly about. The MVP product itself. Um, so obviously I appreciate it's been through sort of several iterations. Um, and you're working on one on one now, but I suppose in terms of the kind, the, the, the product that you are offering up for sale, uh, for people to sign up to at the moment, um, how long did it take you to build that start to finish?

Martin: So on the calendar about six months. Um, and, and of that, well, that, that's the best answer because there's no way to really measure it otherwise. But yeah, on the calendar from conception of idea to like, I mean, literally a conversation with a few people of, oh, AI could probably produce good read exercises and people buy short story books, so maybe we can just convert this into a product, blah, blah, blah.

From that point, uh, about about six months. Yeah.

Rob: Yeah. Um, great. And what, uh, you able to share a little bit about what, what's under the hood in terms of sort of the, the stack frameworks, tools you're using?

Martin: Yeah, sure. Um, it is such, it's a good question actually, because I sometimes mention this and I don't think non-engineers really get it, but it's really important when you go down this route of saying, I want to build my own value in the world. I want to produce something that I am proud of and want to work on that this is really important because if you just caught all the corners, and I do think this whole vibe, coding nonsense is, is gonna produce a lot of this, that people are gonna get a really nice fast start and then they're gonna lose all hope because it's gonna be an absolute pile of crap.

And who wants to work on that? Nobody wants to work on that. And as engineers, we've all done that plenty of times. We've taken a job at a company and realized it's a in a big mess. And, you know, engineers like Greenfield projects for a reason, because everything's nice and clean and all the possibilities are wide open, and we can make all these cool choices that are definitely gonna be right and we definitely won't regret 'em and all this kind of stuff.

But, um, so for me, uh, the tech stack is really important and it's very simple. So another guide in principle I could say on, on this question is just keep it very, very simple because, um, I, I always remind myself of the Lindy principle, um, and if people don't know what I'm talking about, the Lindy principle is just this idea that.

If something's been around for 10 years, it's much more likely to be around for 10 years. Uh, and really what that means is if something's only been around for one, then you should probably take it with a bit more of a risk thing. I was like, maybe it'll be around 10, but it's not clear. React. In year one, everyone hated it.

People were like, react is a terrible thing. It mixes data and structure and behavior altogether into one thing. We've, we don't do that. We can separate them. This is a terrible idea. And now that's completely different because you're like, react is definitely gonna be around for another 10 years. Might not be the most popular choice in 10 years, but it'll still be here.

So, um, with a tech stack for Langs, it was important that I was choosing boring technologies as people like to call it from the infamous, uh, post by, by Dan such and such. Uh, so, so Lang sesh is, it's react, but it's actually Preact, so anyone that knows about that world, the Preact is just a very minimal version of React without some of the other features that I don't care about.

I don't care about server components, and I don't care about half the other things in there. I only want to use it as a transformation from state. To visual parts. So I use pre-ACT because that way it's much smaller. I can bundle it together and the entire workflow is very simple. I don't need to do complex things like chunking it because I've got this huge framework bundled in there.

Um, LangSesh is a single page application though, and I think that's important because I'm, I'm focused very much on, um, the UX here. I want things to be snappy and fast. Um, so yeah, it's a, it is that the architecture generally is SPA and it's very simple. Uh, it talks to a backend that's written in Go, and it does so over HTTP and it doesn't even do it over any kind of rest thing.

It actually uses a custom library that I worked on a few years ago that some to simplify and, and, and not entirely accurately summarize it. Um, it's a simple version of graphic ql. It's basically like an RPC thing. So it's like, just call this thing on the backend. Do a thing, do a task, give me a result.

That's all it does. It doesn't even yet batch them. That was the idea that it would be some minimal version of GraphQL in that it could at least batch those things. But it doesn't even do that because it doesn't need to right now. So it's very simple. Go backend, that's on a virtual host. Uh, it doesn't need to be expensive.

It can probably serve, I would say it could serve so much more than it does now, maybe a hundred times the traffic that it currently serves without even upgrading the server. And that's $10 a month. So, um, it does not need to be complex. The deployment process. Uh, and the provision is also simple. The, it is a series of bash scripts.

This literally sounds like something from history, but Bash is a pretty useful thing, uh, and it works well. And a friend and I, years ago made a bunch of bash libraries that you can import into new Bash scripts that produces a ruby like kind of, um, language, which is basically just a bunch of conventions of how to take something from one machine, cross compile it, and then via SCP and SSH.

Uh, provisions and deploy those things onto an on provisioning machine and then deploy the application onto them. So it is agentless. So it's really simple and it's using technology that's been around. Again, if you're talking about Lindy, those things have been around for a long time, right? Since like the seventies So, uh, I can deploy the thing from anywhere. It could be. It's not if you can deploy from a GitHub pipeline 'cause it doesn't need to be, it's deployed from my machine because it's SSH and perfectly secure to do it that way. Um, so yeah, really simple. That's, that's literally the whole thing. It's like we've got the front end, the backend backends go front end is TypeScript, by the way.

I didn't mention that, but. Yeah, I, uh, I've been writing JavaScript since day one. Well, maybe not day 1, 95, and now I was nine, so I wasn't writing since I was nine. But I was writing it since very, very early days, back when JavaScript was actually a huge mess, more than it's now. And, uh, I do like JavaScript, but the truth is, uh, any large project, it benefits hugely from some additional structure on top that you do not get in I, I worked on a co base, it was over a hundred thousand lines of JavaScript, actually, about 6, 7, 8 years ago. And it became, it became horrendously difficult to manage. Uh, most things with that even

Rob: funny to think, I, I remember my first experience of JavaScript just knocking out a basic HTML site and it was just to do, do, uh, image rollovers. Remember just getting the thing where you hover, you obviously do it in CSS now, just where you hover over a button, it changes, you know, the couple hover over an image and it swaps it for another image.

And that was. That was pretty much the extent to which we used JavaScript in the beginning and now look where we are.

Martin: I remember we used to call it something different. I mean, we called it D-H-T-M-L back then it was like dynamic HTML. It wasn't an app at all. It was just, it was HTML. That kind of moves or responds in some way. Um, yeah, what a, we still, and it's crazy. We're still basically building on all that technology now.

We just have a bunch of frameworks and pretend that we know what we're doing now.

Rob: That's pretty much it\ Great. We've covered a lot of ground, I think. Let's, let's move into, um, just summarizing if you can, um, a few wins and learnings just picking out a few things from your, from your story. So if we could start with, I know we probably talked around this a lot, but just, just to distill it a little bit.

Excuse me. Yeah. Just to distill it a little bit. Um, could you pick up one thing that you feel, one thing in particular that worked really well so far in this, in this journey with, with Lang Sesh?

Martin: For me, I would say what's worked really well is, is me taking responsibility for the fact that I don't know the answer, and the best way to get it is by asking people. Um, in the past, I've just not done. I mean, it's safe to say I've basically not done it at all because I've done it so infrequently of just reaching out to people and saying, what do you think of this?

How can it be better? In particular, I even encourage people to gimme bad adv advice, bad feedback. I'm like, tell me why you hate it. Tell me what it is. Just absolutely sucks at, and the reason you definitely won't pay for it because it completely disarms the whole thing I'm not asking. So, and it also is like, don't, I'm not looking for you to say nice things.

I, I just want you to be honest and the chances are you probably don't like it as opposed to do like it, you know. If I let you don't like it, um, if I enable that in the conversation. So what's gone well is I've just, I've, I've said to myself, listen, you know, no one cares about this more than you do, and no one really cares about it at all outside of you So there's no reason not to just ask people. They're gonna tell you. And, uh, and they're not gonna feel too bad being honest. And you are gonna learn a lot from it. And as much as I'm a dogfooding this thing, because I'm also a language learner as well and love language, that doesn't mean I have all the right answers.

For sure. I have con, I have certain ideas, I have certain theories. Um, and at the end of the day, I can build all of this stuff like I've done in the past many times. Spend a lot of time doing that, a lot of energy doing that, put it out there and maybe learn, maybe learn the answer based on whether people use it, but it's just so much more effective to be like, put a simple thing out there and then ask people about it.

Um, so I'd say what's gone well is I've done a lot more of that and it has definitely, uh, revealed again and again all little things that I either hadn't thought about or agreed with, but now it produced an absolute conviction that I was right. Um, and then I should act on it. So, yeah. Uh, in summary and to simplify talking to users a lot more

Rob: Yeah, that's, I mean, that, that's exactly it. You, you know, ask, ask people for their opinion. But yeah, if you say it's it's so powerful. If you can, if you can disarm and get 'em to flip into that mode where they're actually gonna feel. Um, sort of comfortable with giving you their honest opinion, you know, so like, don't sugarcoat it, you know, really don't, you'd be wasting, you're wasting my time if you tell me, you know, that it's great and I go off and build this thing.

And actually you were just, you know, you were just saying what I wanted to hear. There's, there's no point.

Martin: Yeah, I, I

Rob: the other thing

Martin: that's, that's, sorry to, to, to interject there. I I have a friend who's particularly good at that and he reminds me again and again that I need to, I need to really embrace and just copy his style of like, he. We, we have a good, um, relationship. We've worked together for a, a long time, and he will just tell me straight and he'll be like, this isn't working.

This is a bad idea. And that's okay. That doesn't mean he's right, it doesn't mean I'm right, but it's new information. And if he really believes that, then that's useful to know. So yeah, this, and he asks me sometimes for my advice, he'll say the same thing he'll say. Tell me your, your honest opinion as though you are aiming to make me cry.

And that sounds silly. It sounds like a little bit exaggerated, but it does com because, because it's so easy to go the other way and say, I'm gonna try and be nice here, or I don't wanna be a bad person or a nasty person because that's our default most of the time. For good people, most people are good people.

It sometimes useful to swing so far the other way and say, try and make me cry. And then maybe you land in the middle somewhere and you give really nice, honest feedback that just says what you're actually

Rob: Yes. Yeah. But at the same time, putting it that way, you can also kind of have a sense of humor about it, can't you like sort of roast my idea. You know what I mean And if you can find a way to like legitimately roast it, then well, you know, I. Fair enough. There's, there's probably a problem with it, but, but also, um, I think it's important to, um, it should be obvious really, but try and try and try and ask the opinions of, of, of, you know, people who might actually be your users at the end of the day, rather than you just ask a friend or family or whatever.

And you know, even then you ask 'em to say, well, you know, don't sugarcoat it. And they might not, but what do they really know? And they might tell you it's rubbish and actually they're wrong. 'cause they, because they're not, you know, they're not your users. So.

Yeah.

Something within all that. Um, great. Uh, so now one thing, um, uh, there are probably one or two things you've already mentioned.

1, 1, 1 thing that that didn't go well, maybe we we'd call it a failure. And, and, and what you, what you learned from that.

Martin: I think probably what's not gone well, um. It's very early days, right? So I I, I don't think there's any like, major thing of like, oh damn, I really screwed up here. Um, and I've tried to address some of the things I've made mistakes with in the past, and those are really good answers to that question. Again, with hindsight though, you can find those nice answers, but I would say that one thing that hasn't gone well, and I think it's pretty typical actually at the stages, you know, you, you, you care about the products a lot and you think this is gonna be cool and I've built this thing and there's a lot of energy and, you know, sleepless nights and stressful moments.

I got into making this thing real and it now works pretty well. Um. But people don't care about it. And that's fine. That's completely reasonable. And, and you should never have any resentment about that. And it's a pretty typical story of like, you know, I, I think I'm trying to hold onto this belief and, and remind myself of the why, and at the same time acknowledge that it's not perfect yet.

Meanwhile, people are coming and going. They're literally coming into that, using it once or twice and dropping off. And even though I've got a nice theory right now of why I still don't know for sure. I think what's gone not well is just, you know, it's not been a runaway success. Um, but that's kind of why I expected, so I don't think that's necessarily going badly.

It's sort of going just exactly how you'd expect, which is you've not quite found market fit yet. If you had, you know, these people would be paying immediately and staying, uh, they're not. Um, but that's good though, because now you're learning something from it, you know, they're using it a little bit and then they're not staying.

So I'd say, yeah, for me, stickiness in the product is, is not right yet. That's, that's not gone well. Um, or maybe it's gone as well as I would've been expected at this stage, but there's a lot of improvement required there still. So that's where, really where my attention is on, if, if I reframe the question is, you know, what do you really still need to get?

Right? That is what I still need to get right. The, the, the overall value proposition, which I think is a combination of pricing and, uh, structured approach here as well.

Rob: Yep

Yep So actually, you know, the other thing I was gonna ask is one thing you're figuring out now, but I, I think realistically there is no, there is no failure yet. 'cause you're so, so early on and actually you're, you're figuring all these things out, the pricing and getting, getting the, uh, the product right?

Okay, well, tons of insights there and it's really interesting to kind of learn about the, kind of the, the broader story behind, behind Lang Sesh and sort of how you, how you got here. I. Um, I mean, I suppose in terms of the kind of future that you sort of have in mind, I know you sort of feel like it feel like it's unwritten, um, as it were.

Um, but is this kind of, do you, do you hope that this product might develop to be kind of the one for a good period of time? Or is this kind of maybe you thinking there's gonna be sort of portfolio related products? What's your

vision there

Martin: yeah, that's a good question actually. The current plan actually is to mo, um, is to, well, is to continue on this thing. I think this needs a year, um, of it being in a state where I'm fairly convinced that it provides value because it does to me personally, uh, and build, uh, more, well, not really build, just, just get better at telling people about it.

You know, this is skills that, as you know, at my core, I'm an engineer. Um, so on the marketing side of things. Uh, and knowing how to promote this. There's a lot of things there I have ideas about, but don't know if they're gonna work and any time to do that. So yeah, for the next year, my focus will absolutely be on this thing.

And I think one way or another, this can become a useful product. I'm hoping at a price point that makes it sustainable. And again, in my case, I, I had a great conversation with someone a few days, a few days ago, and they said to me this question of like, you know, if you like, can fill out a YC application.

One of the questions on there is like, what is. Your individual competitive advantage, and that's such a difficult question because the truth is most people will answer that with some attempt to convince the reader of the application that it's true when they probably don't believe it, because it's really hard to actually truly believe that you have some kind of unfair advantage.

Like what Unfair advantage do I actually have, honestly. Um. But one answer from a a that a friend gave me is a really interesting one, and it stays with me now, is, uh, he says, I'm small. Most people wanna be big. Most people want to take on a seed round, maybe the first big round. Most people wanna reach 10 KMRR.

The truth is one KMRR, and I'll be working on this thing permanently. So I can offer this thing at 10 times less risk than someone else offering it that isn't yet at their MRR. It's, it is a weird answer to the question. Most people wouldn't say that, but. I'm interested in this. I like the tech stack. I like the problem domain.

I like the current incarnation of the solution. If that becomes sustainable, it doesn't need to be a big company that hires 10 people and maybe at some point buys me a yacht. I don't care about any of those things. I just care about working in this domain. And that is an unfair advantage because most people are just distracted by the money side of things.

If I care about money, I go back to the corporate job, which will always make me a lot more money than, than this world will. So yeah, I think that's for that reason. I do, I can, I think I can stay in this, um. Game for long enough to know that, to see it work or to be fairly convinced that, fine, I just haven't found product market fit with this thing.

Um, both of those outcomes are definitely possible. Um, and then in the meantime I'm experimenting with another idea and I've not done very much on it right now because it was too much of a distraction at this stage. But I'm gonna, um, start opening up a bit of a private community for people in my environment.

Um. That are doing the same thing, you know, that are on their solo journey, feeling a bit lonely, feeling like they need conversations like this conversation and many conversations I have with with friends where really it's to bond to, to soundboard, to share ideas and to, to really just make new friends.

I think as adults, especially in like our age demographic, there's a lot of people doing things, but not a lot of guys like get together and share that experience and learn from each other and we can learn a lot from each other. So I'm going to, yeah, I'm experimenting with a community called Build Mode.

For people that consider themselves in some form of build mode. Um, but I'm, I'm nowhere near anything that's viable or interesting at this stage with that idea. So, uh, and I, I'm intentionally ignoring it because I need to get esh to a good point before I even consider it. Um, shiny object syndrome.

Rob: so that's not, not a community. Anyone can go and join just

Martin: No, no, not yet. But if, if you are listening to this and you're interested, please reach out and let me know if you think it's a good idea and that would be helpful.

Rob: Yeah. Great. Okay. Brilliant. Right, well, let's move towards, uh, move towards wrapping up. So, um, yes, we talked about this a little bit before the call. Do you have, um, a, um, a tip or a message for, I mean, you've already shared a ton of tips and insights, but kind of, uh, perhaps a new one for, uh, for a would be founder who's thinking about building something, starting something and just can't quite figure out maybe if they've got what it takes or how to, how to go about it

Martin: Hmm. That's a really good question. Honestly, I think there's so many ways to tackle that question. Um, for me, I have this kind of default of like any question that seems difficult. I sort of like devolve into philosophy. Start thinking about it from that angle of like, you know, what's the real meaning here?

At the end of the day, look we've not got long on this planet. Don't waste it. That's really probably my advice is like. Don't waste it, but that's not very helpful because you know, what does waste it even mean like, how do I know I'm wasting it? Um. I have a friend and, and on this island that I'm living on right now in Thailand, actually, there's many people like this, very spiritual people, um, very hippie people.

Um, whether that's good or bad, no comment. Uh, it's just different. Um, but people on this island are very present, uh, like very o maybe too present, you know, too intentionally present. But it's nice for the contrast because, you know, in the, in the corporate grind, you have the opposite and people just don't even realize the sacrifice in their entire twenties to work on something that basically, it statistically has very little chance of working.

So my advice would probably be that is like, just, just go back to the question of like, what do you really want out of life? Because go building a startup now, it, it's always been hard. I would say in some ways it's harder now than ever because we're in this point where. You know, the economy's harder.

It's harder to have that disposable income that enables you to do that with, you know, the, the state of our industry at this point. There's a lot of uncertainty about what AI will swallow, what AI companies themselves will swallow, where the opportunity lies. It, it is not to say people can't be successful, absolutely they can, but it's, it is not gonna be about by accident in most cases.

You're gonna have to really work hard. You know, the 10 year overnight success is still a thing, and if you're not prepared to be in it for 10 years, you're probably not gonna get the overnight success. It might come in year one, might come in year two, but it might come in year nine. I know people that literally worked in my company for nine years before they sold it.

Um, so, so my advice would be that is, um, it's just ask yourself, do you really wanna do this? Or you're being trapped by fashion. Are you basically doing the equivalent of thinking that you want a Ferrari or you want that really expensive bag or that really nice suit when actually what you want do is put some swim shorts on, go and move to a cheap island and actually make some really good friends and have a great time and enjoy your twenties and not waste it building something that probably won't work like.

That is intentionally meant to sound slightly cynical because sometimes people don't ask themselves that question, and I've spent a long time building stuff that no one even knows about Now. It's not currently available, it's long since deleted from the internet. I don't regret those moments, but it, but I probably should have asked myself a few more times, do I really wanna work?

Do I wanna work on this idea? Does it excite me? Another piece of adjacent advice is don't work on problems that you don't care about. Just don't do it. Even if you think that you might make a lot of money off it, you again, statistically you won't. But you won't even enjoy the experience and you probably won't learn anything that you find particularly useful.

Don't work in domains that you're not interested in.

Rob: Yeah. Yeah, that, that's an interesting one because you hear this advice a lot, you know, just like. Like pick a boring problem. And I always think, I don't know about that. And obviously what's boring is, is subjective, you know? And it's like, uh, you know, I think it's more about, okay, pick, pick maybe a kind of a less obvious problem.

A kind of, maybe kind of an underserved area, but it needs to be interesting to you. It might be boring to other people, but it is not like this idea of, let me just go and find the most boring thing 'cause that's what I'm supposed to do. 'cause no one else will do it. And you know, 'cause you'll just get, you know, you, you, you'll get bored and you won't see it through, you know,

so it's

Martin: exactly. Like you, you've gotta, when you say like, most people's advice about like, how do you stay motivated is you keep going back to this. Why? Well, if your why is I think I can make a bit of money off this. Well, you can objectively make a bit of money off other things much more surely. So you haven't even got a good why at that point.

So like, there's no point in, in, in doing that. Um, you, you think you're scalping, but you're not, you're playing, you're playing like. Intraday trading with algorithms that you're gonna lose. That's pretty much what you're doing if you just pick any old problem that you don't care about. But on the other hand, if you pick a problem you care about, it won't really matter to you much what your MRR is.

If you can keep doing it, you can keep doing it and you, you're gonna enjoy your life. You know, I know a guy that does a, that builds, um, a network for retired NFL players in the US to, to do that in Europe, and he makes great money from it, and he just loves what he does and no one cares except him and a small number of people.

And that's, to me, that's the goal. It's not like. You know, it, it's not to reach the point where I can piss off every resident of Venice because I have a big lash, uh, lusia wedding there. I'm referring to Bezos at the moment. Uh, I don't who wants that money? Who wants to be that famous and powerful? I mean, maybe some people do, but for me, I just think it's about making the most of your time.

Um, because time is the thing that you're gonna sacrifice. And also health, you know, there's definitely the whole grind of like, just if you're not working 60 hours, someone else is and they're gonna outcompete you. Well then let them, because the sacrifice is something that's a lot more valuable. Than whatever they're gonna get for it because they're gonna be like 40 years old.

They're gonna be like outta shape, probably some kind of like big health problem. It's not worth it. So that, that's my advice is like think about the other side and enough and, and then, and if you do and you say, no, look, I am actually really excited about this and I wanna work in this space. I wanna create my own path and freedom in the world.

Great. Now, now you, now you've got some useful things to think about. Maybe you can pull some different levers like I did and say, well I can kind of mostly do that actually if I just don't spend as much. Maybe I don't want the car then because that means I could work on this problem.

Rob: That's brilliant, right. Let's, let's bring things to a close. So, um, I'll, I'll get a few links into the, into the show notes. Obviously I'll get the, the LangSesh uh, URL in there. Um, and you mentioned, you know, if anyone wanted to kind of reach out to you if they've got any questions, what's the best way to do that these days?

Martin: So, um, the best way to contact me is my full name, which is martin ru RU e.com. Um, on the top of that page, there's a bunch of links to the various. Different social places. You can find me, of course, LinkedIn as well. Um, and my email address is also linked there, so you can reach me by whichever poison you've chosen to spend your free time.

Rob: Perfect. Right. I'll get, yeah, I'll get, get the link to your, um, because you've got a kind of small blog on there, haven't you, where you post some thoughts.

Martin: I, I used to write more on there, but these days I feel like nobody reads blogs anymore, so I, I tend to just post those things on LinkedIn. But yeah, there's a, there's a little blog there, there's some very old articles of funny coding stories if anyone's interested in that stuff.

Rob: Great. All right. I'll put that on the, in the links as well. And, um, yeah. Brilliant. Martin, it's been absolute pleasure. Really enjoyed it

Martin: It has. Thanks for inviting me, Rob, and thanks for the conversation. I had a, I had a strong feeling this was gonna be a good one because of our interactions on LinkedIn, so, and I've really enjoyed the previous episodes as well, by the way. So this is a, it's an exciting new podcast and it's definitely on my podcast list.

So, um, yeah, thanks for starting this and thanks for the invite

Rob: Thanks, Martin. All right. See you around. Cheers.

Martin: Cheers. Bye.

Rob: Bye bye.