Recorded 2024-03-26. Published 2024-05-15.
Wouter and Joachim interview Arseny Seroka, CEO of Serokell. Arseny got into Haskell because of a bet over Pizza, fell for it because it means fewer steps between his soul and his work, and founded Serokell because he could not get a Haskell job. He speaks about the business side of a Haskell company, about the need for more sales and marketing for Haskell itself, and about the Haskell Developer Certification.
Wouter Swierstra (0:00:15): Welcome to this episode of the Haskell Interlude. I am Wouter Swierstra, and I’m joined today by Joachim Breitner.
Joachim Breitner (0:00:21): Hello.
WS (0:00:22): And our guest today is Arseniy Seroka. Arseniy will tell us a bit about how he learned Haskell on a bet, how Haskell means fewer steps doing your soul and your code. Arseniy co-founded the Haskell Company, and he currently spends a lot of time working on establishing a Haskell developer certification process and pushing for wider enterprise adoption of the language.
Okay. Welcome, Arseniy. So, our usual question to any of our guests is, how did you get into Haskell and functional programming?
Arseniy Seroka (0:00:55): Well, it’s quite an interesting and maybe long story. So I was studying as a Computer Science Engineer for six years, bachelor’s and master’s. But when I was in the second year of my education, I was thinking that C++ is amazing, but I want to learn something else. I was a super black-and-white student who was like, “Okay, C++ is the best. Everything else is not good. Okay, Java is not a solution. Python is not good. Ruby, maybe.” So I was trying different languages. And once I was talking with my friend, and he told me like, I bet that he will never, ever be able to learn Haskell. It’s like a super complicated, very complex, unusual, unpopular language. And I was like, “Okay, let’s make a bet for a pizza.” And you know what? Like maybe six years after that, I found this conversation in Facebook, and I’ve sent him a screenshot like, “Man, you owe me a pizza,” because that was already a moment when I wasn’t only like a Haskell developer, but a founder of a Haskell company.
So yes, Haskell in my life appeared just with a bet. But actually saying, when I started to learn it, when I’ve started to get more into this language, I’ve realized that it’s the way for me, how should I speak, how should I talk with my computer, let’s call it this way. Because when I was – I used to be a professional C++ developer, and for me, that was like my idea, my thoughts. Then there was a box, like a converter to some code, and then like code and execution. And when I was using Haskell, for me, it was just like my idea and code. There was nothing in between. For me, it was a straightforward way to communicate.
WS (0:02:41): And when was this? When did you start learning Haskell at first?
AS (0:02:45): Oh, okay. It was like 2012, I guess.
WS (0:02:50): So 12 years ago. And then you finished your degree, or –
AS (0:02:56): Yeah. So I finished my bachelor’s in 2015. Then I finished my master’s in 2017. But actually saying, already in 2015 in summer, me and my co-founder Jonn, we realized that we should start a company that is going to use Haskell because I personally was trying to find a job to be a Haskell developer. For me, it was a crucial point because I was able to use any language, but I was trying to do in my life things that I really love with passion. And when I was thinking about programming, for me, it was just like Haskell and nothing else. I was enjoying all my, let’s go like games with this language, all my explorations, all the possibilities that it’s giving to me. And I was a huge fan of math logics, and I was like, the way how I can construct things with functional programming was the way for me. I kind of was playing with math logic statements and stuff. And yeah, I’ve realized that there is no joke with Haskell in 2014, ’15. And then somehow, we decided like, okay, let’s work together, find some customers, and split money in half, take like a freelance with Haskell. But it’s like a meme, like 10 jokes that went so far. And now we have a lot of employees working with a lot of companies, having several different legal entities. So, yeah.
JB (0:04:26): How did you go about it initially? Like, okay, you annoyed that there are no Haskell jobs, so you found your own company, but then what is the next step? Do you just ring some doorbell and say, “Hey, do you need a Haskell programmer?”
AS (0:04:37): That’s actually a very personal question because there were things happening in my life that told – so at the beginning, I wanted to be a researcher. I wanted to go into math logics, into category theory, like group theory—all of these abstract mathematical things. But then unfortunately, there was a tragedy in my family, and I realized that there is no way for me to support my family being a researcher in Russia in 2014. At the beginning, I was working as a system administrator in my university, kind of DevOps position. And it’s Russia. It’s not like I get thousands of thousands. I was getting like a few hundreds, maybe. I was also working as a service guy for vacuum cleaners in shopping malls, actually saying.
The funny thing is that I was working in Google for one week in St. Petersburg back when it was in Russia, but unfortunately, I had to do some things every day during the work time, and it was not possible to combine these things. And I was also finishing my bachelor’s, so huge problems on my way to just work in the office. And I’ve also finished a book called Atlas Shrugged, actually saying. And somehow all of these things, like Haskell possibility, idea to work just for myself and to work remotely, and the idea that if there is no job, then you can create this job yourself, came up to my mind. And I was talking with different people, like, hey, maybe something – I had no idea to make a company. It was just like an idea just to get more money for food, because there were problems like this sometimes. And I’ve told this to my friend Jonn, and we were not very serious when we started. We’re just like, okay, let’s have some fun. Let’s just do some Haskell projects because we love Haskell and we also love Nix, and NixOs a lot. That’s actually how we met. And we were making combinations of these things.
And somehow, we’ve started our first customers. They were, let’s call it, some criminals from Russia. By criminals, I mean that we were doing projects for them. And it was okay for the first month, but then they were not paying us for several months, and we owed a lot of money to our first employees. I was saying like, “Okay, I’m taking this project, I’m going to open source it.” And in two minutes, I had a call like, “Arseniy, we know where you live, and it’s not a good idea to open source it.” And I was like, “Okay.” So I gave them all the things I had. And this was the moment I realized that, unfortunately, even though I’m from Russia, I will never, ever have business with Russian businessmen because it’s a different culture and I’m not this type of person to make deals like this. And we owed money to our first employees. We owed several thousands, and we had to sell 21 Bitcoin back in the days for 300 each just to give some money back.
JB (0:07:40): Oh, no.
AS (0:07:41): It’s okay because it’s very bad if you have Bitcoins, but you owe money to people. So I never ever owed anything to anyone. And yeah, there were quite interesting times that were making me very – how do I say? Stiff in terms of how I make business, how I have some of my personal preferences. I understood that. I want to work with people from Europe, from States, from countries like Korea, Hong Kong, and Japan, because those people, they respect the business itself. They respect contracts, they respect the deals. And unfortunately, in my experience with Russian customers, it was not like this.
So that’s why we founded our company initially in Estonia. Now we have it already in France. We were pretty fast in terms of employees. For the first year, we’ve increased from me and my co-founder to, I guess, 15 people. I don’t remember perfectly, but I think that in three years, we’ve grown from – we grew to like 50, 55 people. Now we have around 80, 85. And we’re not doing like product, our own personal product. We do our service as a product. We do turnkey solutions. We don’t do outsource or outstaff. So we usually do not only like consultations. Basically, the most important things that we do, we do like proper research and development on the problems and questions that our customers have. And usually, they come to us like, “Hey, we have this question. There is no answer in the world to this question yet. Please help us with it.”
WS (0:09:22): But how do you find customers, right? I mean, assuming you don’t really have a name for yourself, it must be tricky. Or is it just that easy to find Haskell consulting jobs?
AS (0:09:36): So we exist for nine years, going to be 10 years in August. At the beginning, it was just like pure luck, friends of friends, some people that we know, or these criminals from Russia. But actually saying, then somehow, we got into the blockchain project called Cardano. Actually, we were there earlier than the beginning. We were talking with the founders, so when there were like two people in their team, and there were like three people, four people in our team. And actually, me, my co-founder, we were the guys who actually proposed Haskell as a main language for this ecosystem. And at the beginning, we were working mostly with the blockchain projects, but those that are related to research. So we’re working with several universities, we’re working with complex crypto or blockchain solutions. And then we understood that, okay, we have some customers and a pretty good name in this field, so we should grow. And it’s very important to notice that me personally, back in the days, I never, ever studied any business-related courses or how to make a company, how to – I was studying mathematics, I was studying computer science, like algebra. And for me, it was like, “Okay, let me think how to do this.” Because there were no advertisers or anyone. This is just me, my co-founder, and our brains and a lot of luck. I would say that any business could be calculated to some point because, at one point, you need to have a lot of luck. You need to meet some person on the street, you need to have some friends, or just something happens to you because we all know how to make a company. We all know how to make a business, but unfortunately, it’s just not possible in overall, I think, society to go just by steps like an algorithm. You also need this kind of human position, human energy, or something. I don’t believe in these things, but I mean, like abstractly.
WS (0:11:47): Serendipity, right? Where you happen to be the right person at the right time and talk to the right people or know the right people somehow. Yeah.
AS (0:11:53): It’s so funny, I hear this word like from the last month. You are maybe the sixth person who is telling me this word.
WS (0:12:01): Okay.
AS (0:12:03): So yeah, customers. At the beginning, we were pretty successful in blockchain. And then we were again starting with just companies that we know, with just some people like, okay, let’s try to make a project in something else. And we’ve tried to build projects in biotech. It worked. Now we have pretty good customer traction in the biotech field. We are working not only on research solutions, but we even write firmwares for biotech hardware. We use Rust for this. So we’ve extended our technology stack, and we pretty like this combination of Rust and Haskell. It’s like using Rust for hardware-related things and Haskell for, let’s call like backend related things. And we use TypeScript for frontend. We were trying to use Elm, we were trying to use PureScript or GHCJS, but unfortunately for us, real-world application, when you need to do a little bit more complex than just an interface, it was hard. By hard, I mean, it was expensive. And it was just a good solution for us to move to TypeScript these days.
WS (0:13:13): And a little bit more about this quick growth. I mean, how do you find this many Haskell developers? Was that a problem for you at all or –
AS (0:13:21): When I’ve started Serokell the same year, in September, I’ve started to give courses about functional programming at my university called ITMO in St. Petersburg. At the beginning, I was just giving lectures because my teacher, my master, I called him this way. His name is Jan Malakhovski, his nickname is Oxij. He’s very – he’s quite famous in NixOs and in Haskell communities. He quit his position of like functional programming teacher. And I probably was the only one in our department who was that much interested in Haskell and was also very good at it. That’s why I was giving those lectures. So I was 22, and my students, they were like 20, 21. So I was basically giving lectures to the same people of my age. I was giving them for about seven years, but now my colleagues are working on this course, and I’m still managing it. And so we try to get the best students from after the finish of the course, and they are becoming our junior developers, but they’re very high-level junior developers, usually, because it’s not only people who’ve passed half a year course of our way of functional programming that we need in a company. And we think that is a quite good standard for the industry. But also, I would honestly say it’s the best department in Russia and maybe one of the best departments in Europe for computer science. For the war, it was very popular that representatives of Google, Facebook, Snapchat are coming to this department like, “Hey, we just want to have someone.” Maybe like Olympic competitions like ACM ICPC. My university was like the top winner of all of them. I was not this type of student, but I was starting with these guys. Some of them are working in Serokell right now.
And yeah. So this is one of the traction where we used to get mostly developers. But now we don’t need that much new developers. We’ve grown a lot of of them. But the thing is that there are not that much Haskell companies in the market, let’s be honest. And there are not that much like pure Haskell companies. At least like you can be just Haskell developer or just Rust developer. I think we have a pretty good reputation among Haskellers. Sometimes I meet people on conference. They’re like, “Oh, you are this very fancy hipster Haskell company. I want to work with you.” Because I think that not only the quality of your work, but also the style of your appearance, the style of your representation is very important because we are all humans. I just don’t want to be as robots. And if you have a company, even if it’s a remote company, where you feel that you are in the way, you feel like a part of a very cool group of people. And it’s not only the thing that you have inside, but also your representation, like external one. I think it’s very important because, okay, you can motivate people with huge salaries, but if you motivate people with interesting projects with some kind of lifestyle in a way, that’s much more valuable than just giving a huge cash out. This was very good that before this meeting, you mentioned our team page. That was the beginning of my personal idea, okay, let’s make something cool, some appearance. And a lot of people notice this kind of style and positioning for our team. And yeah, that’s why we have just a lot of incoming requests from people like, “Hey, I want to work with you.”
WS (0:17:23): And are you fully remote or do you have offices, or what’s the situation exactly?
AS (0:17:30): So let’s call it this way. Officially, we’re fully remote since 2015. So when the COVID hit, we’re like, okay.
WS (0:17:39): Yeah.
AS (0:17:39): But we used to have an office in St. Petersburg. But it was mostly just for some brainstorms and, let’s be honest, parties. After the war, everyone who was in Russia, they moved out. We’re moving a lot of our employees right now to France, even though there is no need. But we’re thinking to open an office here in Paris. But again, I think it’s going to be more just like, if you want to come, just come. Let’s work there. If you don’t want, stay wherever you are, because we have all our processes, all our management systems, and flows 100% dedicated to remote style of work. And it’s not only about the tools, but it’s also about some mindsets. Because even now, and even like six, seven years ago when we were having an offline meeting or a call or something, we were summarizing what we were talking about and publishing it to related channel or issue so others can read about this. Now with AI tools that are making summaries of your calls, it’s much, much easier. And it’s amazing, I would say. But we were doing this like six years ago manually.
JB (0:18:54): So you’re saying you’re actually recording all your internal calls and have them summarized by some AI for the rest of the employees?
AS (0:19:00): Most of them, yeah. So if it’s necessary – so of course, there are calls like just for chatting, I don’t know.
JB (0:19:07): Sure.
AS (0:19:08): But those that are important. And even though there could potentially be all the necessary people on the call, we have pretty open policy inside the company, and you can just go and check the call, what was it about. We have a pretty open policy also that we have regular reports from the business side of the company to the tech side of the company. We have all the company meetings where we just give the news, what we’re doing for several months, what we’re going to do. And each representative of each department, like marketing, HR, business, technical, is just giving those summaries. We try to keep people informed, even though it’s hard. I would say, even when everyone is remote, you have all the Notions, Slacks issues open, still, a lot of people are out of context. And we understood that even though we make a newspaper about our internal things every month, this is not enough. And every four months, we are trying to make like just a regular call when we are speaking, when we are using our voices to give the perceptions and the ideas of what we’re doing.
JB (0:20:24): And the other thing I’d like to come back to, which is maybe strange for this podcast, because you usually don’t do much politics, but you did mention that when the war started, all your employees moved out of Russia. And I was curious, is this something that your employees all wanted or that you as a company asked of them? And how did that work? Was it hard?
AS (0:20:44): So, first, I’ll talk about myself. I left Russia the first day that was started. It was coincidence, but it was my personal decision to not come back without a reason. And with our employees, we were not forcing anyone. It was like, if you want to go, we help you with everything. We help you with – I don’t know, tickets, place to live with visas, all of this kind of stuff. People moved out, unfortunately, I would say. Everything that happened is a nightmare for me, honestly saying.
WS (0:21:18): Yeah. I think it’s something we see with quite a lot of Russian tech companies. JetBrains is another good example where they’ve really left Russia almost entirely.
AS (0:21:27): Unfortunately, I was never ever thinking that I’m going to be an immigrant, but now I am. That happens. I mean, it’s life. Unfortunately, there were a lot of ethical problems. There are a lot of business problems. There were a lot of just personal problems about this. And it’s just, for me, not a question. Will I go now back to Russia or not? I will not. It’s my country. It’s my favorite city in the world, St. Petersburg. And I always say that not that much Russia, but I’m much more St. Petersburg person. I hope that this situation will change and that peace will cover us all with its huge fluffy hands. But let’s pray for this, honestly saying.
JB (0:22:14): Yeah. Right. Since we’re jumping topics wildly, just to to interrupt with that, I noticed on your team page on Serokell, you say that in your spare time you do extreme sports, but you don’t say which kind of extreme sports. And now I’m curious how extreme the sports you take doing this.
AS (0:22:34): Oh, actually saying, I’m an adrenaline junkie myself. For me, I have a lot of work using my brain. I have a lot of conversations, calls, messages, emails, conferences. I’m this type of person who flies so much that I actually hate flying right now. Before 22, I’ve never ever been to any city except for St. Petersburg, and it’s like countryside. But then something happened, and I can have, I don’t know, 10 airplanes a month or even more. For me, the moment when I do – even I go to the gym, it’s the moment where I can turn off my head a little bit. But if I am jumping from the helicopter with my skis on top of some mountain, it’s the best moment when you really can turn off your head and you just focus on the moment where you are specifically. I just do a lot of skiing. I do a lot of surfing. I used to do a lot of downhill mountain biking. I will probably come back to this sport again too. I do kite surfing. In my spare time, I try to play tennis. I used to be a professional tennis player when I was a junior. So yeah, my family, my father, my grandfather, they were always forcing me like, “You must do sports to stay healthy.” And you know what, my grandfather, he is 93 years old and he’s still playing tennis.
WS (0:23:56): Okay.
AS (0:23:57): So for me, that’s a very, very good example of control of your mind and your body in a combination of physical activity.
WS (0:24:05): So we usually ask our guests a little bit about the technical side, but I think for you, I’d also be interested in the business side of a language like Haskell, and what could we do to make it easier for people to found their own company using Haskell, or to ensure that Haskell flourishes not only as a like academically interesting functional programming language, but also as a successful industrial language.
AS (0:24:34): We always have to convince some of our customers to use Haskell, or at least Rust, because, first of all, no one knows what is Haskell. Second thing is that unfortunately, it has this kind of reputation of very specific mathematic-related nerd language, you know? And people are like, “Hmm, I don’t know. I mean, if I can find Haskell developers, if I can have a reliable development workflow, if there are well-established practices, libraries, frameworks for Haskell, if it’s going to be fast and cheap for me to implement a solution using Haskell, et cetera, et cetera, et cetera.” And all of these questions, they arise, in my opinion, not from the Haskell itself. Haskell itself is a language, is an amazing language. You can make so many things, so many solutions just using it in a lot of different fields. You can make a super simple application that is going to be in your web browser. You can make very complex solutions to do a lot of algorithm optimizations using your bank system. You can work a lot with your scientific research where you just need to do a lot of data analytics. Or you can just have fun and even make, like, I know there are games using Haskell, or even though it’s not like a specific feel for this language, but also like a lot of common line tools written with Haskell. We just don’t know this. A lot of high-performance systems are also written in this language.
But unfortunately, the problem with Haskell right now is around its ecosystem, around its representation in the world. One of the reasons why I’ve decided to make Serokell was to make Haskell a famous language to make Haskell as popular as Java solutions. We’re trying to aim big. But the thing here is that a lot of people just use it without a real way of popularization of this language. Not among the people who already know what Haskell is, but among people who don’t know what Haskell is. Even now, if you go to haskell.org webpage, it’s quite complicated for a newcomer to understand, what can I try, what should I do? What are famous frameworks? What are famous projects? Because even if you compare it to a new language like Rust, and it’s also a functional programming language, it’s also a complicated one, but it’s very friendly. If you go to social media, you will see that Rust, in some way, it has some kind of human form around it.
For me, it’s important to try to make Haskell very nice in terms of appearance to people, very easy to go inside. Because for me, when I was having this bet with my friend, it was like, okay, it’s not possible to get into Haskell if you don’t have a degree. And even though, like three of us, we know that is wrong, a lot of people still think that Haskell is just for a PhD people. This is the part where we just think about the newcomers. How can we popularize this? But we should think of Haskell and any language as a tool and as a product, and we need to have proper marketing. We need to have proper sales. Even that’s not a commercial tool, but that’s still something that we need to spread around. That’s why we try to have these educational videos for super newcomers, for super newbies for this language. That’s why we try to convince our customers, like, “Hey, that’s a good solution. You shouldn’t be afraid.” There are enough Haskell developers in the world, and it’s good that most of them, they really have some good computer science knowledge. It’s a good correlation.
But unfortunately, then we go to another field of problematics, field of enterprise applicability of Haskell. Okay, you can use Haskell – no, I would say that it’s really hard to use Haskell for a startup because you really need to find what is the best framework. And if there is none, you need to create it yourself. What are the proper ways and processes? How can you write something? Because when you use JavaScript, you have so many well-established flows, how to make something. And in Haskell, unfortunately, if you change the CTO in your company, one of them is going to use free monads, another one is going to hit them. And you need to rewrite everything because this approach was not valuable for this person, for example. That’s maybe not a super real-life example, but it’s very close to what is happening sometimes in the companies.
JB (0:29:38): But why wouldn’t that happen with other languages? I mean, they have more than one framework of doing one thing. So wouldn’t the CTO in a JavaScript company also just bring their own favorite news framework?
AS (0:29:49): I think that in other languages, there are quite some good, well-established ways, how can you solve typical problems. There are very popular frameworks to, I don’t know, manage network connections. And they’re popular in Haskell. And of course, most of the companies, they’re using pretty the same libraries. But actually saying, when we are making interviews with them, when we are making interviews with our customers who are already using Haskell, we see that there is no real consistency. When you check out Java projects, I don’t know, C++ projects, of course, they use different libraries, but there is still some consistency in the way how people are writing code, the way how people managing it. And unfortunately, in 10 years of existence, we haven’t seen this consistency in the world of Haskell.
One of the solutions that we are trying to do right now, it’s the first step in this field. We’re trying to make a Haskell certification program where we will try to establish this kind of basic flows, how people should use Haskell at the beginning when they try to solve common problems. We’re not talking about specific things, but like when you just need to make, I don’t know, a server with a chat application, I don’t know, like communication one, they’re pretty well-established solutions for this. And with our problem, we are trying to go even like in a smaller, more basic fields. So, we finished this a year ago. We’ve implemented certification for like middle Haskell developer who should know some basic topics about these theoretical things about monads, transformers, how should you use them, how should you not use them, what kind of popular libraries can you import in your projects. And we are trying to build this ecosystem around Haskell certification where we will have this together with other enterprise-level Haskell companies. So all of us together, we will try to create these kind of standards for this language in the business field.
So we’re not trying to touch any community standards. We’re not trying to change the world how everyone should write Haskell programs. But we try to establish these kind of standards among Haskell enterprise companies, that if you move from one company to another, you will probably have the same quality, the same set of libraries, the same processes and flows around it. So it’s, first of all, easier for employers to change jobs, and it’s very important. And it’s very important for businesses to employ new people, to employ new teams, to extend their teams. Because they don’t need to spend days to reeducate their developers, they don’t need to spend days to rewrite their code from one framework to another, to try different solutions.
And right now, we have conversations with the Haskell Foundation to make the certification as an official Haskell certification for the language and to make it the only one, because if there’s going to be several, it’s just useless. We would like to have a Haskell certification board that will consist of representatives of different companies. And all of us together, we will make decisions, what we want to see in this certification, and it will be open for any company. It’s not something, there’s going to be a small group of people making some small evil things. And we are going to have an examination board that will consist of different people who will help us to create those kind of theoretical and practical exams, those people who will help us to work properly on the materials, because we’re not going just to release the certification problem, and you just go and pass it or not pass it. We also want to have materials where you can study, where you can have courses, where you can do all-around education for yourself to prepare for this exam. And again, this will help you to get a job because these companies that will work with this Haskell certification, they will much easier hire more people because if they know that, okay, this person knows how to work with my X system, I will just spend much less money on the onboarding of this person. Here is where we try to meet business needs with the needs of the community, because we also think that this will gain more productivity for people, and they will know where to start, where to go.
WS (0:34:33): So usually, when I think about certification, I think about like Microsoft Certified C# developers, and this is usually that you have a clear, whether it’s a company or an entity or a foundation or somehow. And I think Haskell, the community, is more spread out still despite the work of the foundation, right? I think the hard thing would be to get end companies and universities and developers and the community in the broadest sense to agree about what does certification mean or what level and what topics should there be, this kind of thing. How do you find that process? I mean, that must be a huge challenge.
AS (0:35:12): Well, that’s, I guess, my best skill, just to find people and talk with them. But actually saying, yes, that’s the trickiest thing in this process. I’ve started from ZuriHac a year ago. I was talking to representatives of the Haskell Foundation. I was talking to Simon Peyton Jones. I was talking to some other valuable Haskellers in the community. Example, Andres Löh from Well-Typed. I was talking to representatives of Tweag, et cetera, et cetera. And people are like, “Yes, this is a good idea.” But unfortunately, when you discuss these kind of things, people usually are like, “Okay, that’s a cool idea, but we don’t have time, resources, or it’s really hard to implement a system like this.” So we’ve invested a lot of our personal time and money just to create this kind of system. And now we have it on our hands. We have something to show. We have something not just on paper, but you can really go and check this certification, even though it’s in beta stage right now. But it’s not just like a starting point. It’s just like MVP. It’s like the first initial version of the product. You can really use it. And we were making a real announcement only in ZuriHac 2023. We got around 250 applications, and we got around 10 people who really have passed the certification. I think from this kind of impact, it’s a good number. Slowly, step by step, with the help of such institute as Haskell Foundation, we can really move things a little bit forward. But generally speaking, yes, that’s a very serious, I would say, business development work for a product like this to engage people, to unite them, first of all. But I believe in Haskell community a lot. I know that everyone is very smart, is very intelligent. And we are not trying to build anything that can potentially harm Haskell itself. We really want to make it on the same level of popular languages because we know that it’s good. We know that it’s cool.
And even going back to my personal story, I used to be a C++ developer. I also was working as an Android developer. I was working as a system administrator. I was working a lot with some DevOps practices. But for me, I was really happy when I was working with Haskell. I was really feeling that this is the way for me, how can I create my art of software development, because I still think that a real developer is an artist as well as a painter or movie director. Because real development is not just like two plus two equals four, but you need to create a system. You need to create this unique virtual world of your program. And Haskell, as well as any other tool, is just a way for you to make it better, to have less steps between your soul and your piece of art. I really want to see these happy faces when a developer is saying like, “Yeah, I feel very, very cool right now just because I’m using Haskell and nothing else.”
WS (0:38:30): So, I have another question where you mentioned that you use both Haskell and Rust as the main development languages. Do you notice a difference when you go to a client and they have a problem and they might be technologically agnostic and say like, “Oh, I don’t care. You tell me what should I do,” and then you suggest Haskell to them, or you suggest Rust to them? Do you notice that they respond differently to these suggestions?
AS (0:38:57): I think we’re not going in like, okay, we can use Haskell or we can use Rust. No. I think we know where it’s going to be valuable for us to use Rust, where it’s going to be valuable for us to use Haskell. I think we haven’t been into situations where we have this choice, like Rust or Haskell or something else. But I would say if we have this choice, we will suggest Haskell. It’s our personal preference. And it’s not like Haskell is better than Rust or vice versa, but we love Haskell. Yeah.
WS (0:39:31): Let me rephrase the question then. Do you notice that the languages are perceived differently by customers?
AS (0:39:37): Yes. And this is going back to my point about Haskell marketing, because Rust is kind of moving into direction of this, like hardware language, but cool to use, nice to have, fast to work with. And Haskell, unfortunately, is still not changing its perception. There were a little bit of attempts during the high era of blockchain projects, let’s call it this way. A lot of them were using this. And I’m really happy that me and my co-founder were the ones who were these like ignitions for this moment. But unfortunately, now even those projects are moving from Haskell to, for example, Rust, there are a lot of problems why they’re doing this. But actually saying, I think marketing is one of the biggest problems here right now.
WS (0:40:30): So I think Rust has a clear sales pitch, right? Saying it’s safe systems programming languages, it’s fastest C++, but it gives you better guarantees about memory safety. And it has a very well-defined niche where it says, “Oh, okay, if you want to do this, then maybe Rust is a good choice.” And despite being a newer language, it has a well-known name just by support of the Mozilla Foundation and things like this. And for Haskell, why would you use Haskell besides it being a fun language to program in? I think the sales pitch needs work, I think. Is that what you’re saying?
AS (0:41:08): Yes. Yes. No, yeah, there’s nothing I can add to this. Unfortunately, yes,
WS (0:41:15): No, fair enough. Yeah.
AS (0:41:16): Haskell right now, it’s like everyone who knows it, they’re like, yes, it’s amazing. But it is just very hard to get into this group of people.
WS (0:41:26): Yeah. I mean, if you look at Hacker News or whatever, right? I mean, it’s not the source of all truth, of course, but whenever Haskell comes up, it’s usually associated with, “Oh, it’s very hard to learn,” or “You have to know a lot of maths in order to use it.”
JB (0:41:41): There’s always the question when you have a consultancy company. I mean, there’s a stereotype with people at consulting companies, and they work at the consultancy. And then eventually, they have some customer, which is happy with the consultancy, and they start pulling the people off the consultancy and employ them directly. Does that happen to you? Like at Serokell, a path to being employed by then your customers once they have the Haskel stuff and they want to be able to work on it on their own?
AS (0:42:06): No, usually, we are working with our customers for quite a long time, usually for the lifetime of the project because we are in some ways substituting the technical team of customers. And again, we’re not doing outsource or outstuff solutions. We are really trying to solve parts of the project, or like the huge project in general, from the technical side of things. And it’s very important that, for the last six years, we also do a lot of combinations of technical decisions plus business decisions, how they’re going to influence each other, because in Serokell, we have a very strong business department. And sometimes we’re also consulting our customers, like, “Okay, if you make this technical decision, you should understand from the business side of things, it’s not going to be cool for your clients. You will decrease the sales, blah, blah, blah.” And we are acting here as a unique provider of these combinations of technical worlds, business worlds, marketing worlds. And it’s not just an experience of our company in code writing. It’s an experience of our company of making successful products for the customers, successful in terms of the quality and the speed of the technical solution and the quality and the scalability of the business solution of the project that are related to the technical sites.
WS (0:43:30): So looking back at the growth of Serokell, it started as an almost coincidental founding of a company. And then now you have 50-plus employees. Where are you heading or what are the plans for the company?
AS (0:43:45): So we have around eight employees right now. So we want to be this kind of unique point of experience where we are able to solve very complex problems and being experts, not only just like in Haskell or Rust. Even now, we are not experts in Haskell and Rust. We’re experts also in the fields of FinTech, in the fields of biotech, also in the field of blockchain. And for the last two years, we have our department of artificial intelligence. Before artificial intelligence, we used to have a department of data science. And now it’s grown to these kind of things. So we also do these kind of consulting services. It’s not as huge as other directions that we have, but it’s growing a lot. And we want to be this kind of company where you work on complex solutions for interesting problems. And the projects that we are working with this, they are impacting the fields of businesses and the world itself. And I think this kind of high-quality service is something that you will not get just anywhere. It’s something that is very, very valuable on the market. And we also think that maybe in five years something, a lot of AI Copilot tools will substitute a lot of developers, regular ones. And we already think about our position in this market. And we think that complex solutions, architectures, formal verified proofs, and projects, they will not be substituted for a long time. And this is who we are. This is where we are. And we hope that our brains are quite useful for the world.
WS (0:45:36): So it’s more like high-quality end solutions to complex problems in the kind of job, which isn’t so much build a website or something.
AS (0:45:46): Oh, no. We do website, but as a part of solutions, not like this.
WS (0:45:50): Exactly. Yeah.
AS (0:45:51): Yeah.
WS (0:45:52): But not a run-of-the-mill kind of software development team.
AS (0:45:56): Exactly. Well, we usually call us like research and development studio, where we are specializing mostly on backend solutions. Super simple way.
WS (0:46:05): Yeah. Very cool.
JB (0:46:06): Since you didn’t mention it, maybe I should, but it’s worth pointing out that Serokell also has a bunch to offer for people who are not paying customers, and that you have a very wealthy YouTube channel with lots of tutorials. You invite people from outside the company. From two years ago, I guess, there’s a blog that has lots of interesting content. And I just discovered, you also have a shop that has very nice t-shirts that cater to the nerd functional programmer. I took the liberty of putting in the advertisement block for the Serokell offerings.
AS (0:46:39): Thank you very much. I was there from the shop, but it was just my personal idea. I have a lot of friends who are like artists and designers. I was like, “Hmm, there is a cool idea to collaborate with them.” And yeah, we just started the shop. It was very successful. So we’ve sold for several thousands during the first months, and now we have a very constant frequency of t-shirts going out. And oh, we can make a promotion. Like we can make a promotion like Haskell podcast, I don’t know, 20% discount. Why not? So yeah, everyone who finished the podcast through this moment, they can use it.
And about the marketing, as I was telling, the idea behind Serokell was to populate Haskell. So we really invest a lot of money into our polarization of Haskell and functional programming in general. And we are very proud of our blog. It has very, very high numbers of viewers every month. We are working a lot on our YouTube channel. There are going to be a lot of new videos coming up soon. There are not released yet as well as this podcast, for example, right now, March. But yes, we really like to educate people about the technologies that we use. We really like to make them humanable. We really like to make them popular and easy to talk about.
WS (0:47:54): Great. Thanks so much.
JB (0:47:56): Yeah, thanks a lot.
AS (0:47:57): Thank you very much, guys, for this. Thank you for inviting me here.
WS (0:48:00): Yeah, no worries.
Narrator (0:48:03): The Haskell Interlude Podcast is a project of the Haskell Foundation, and it is made possible by the generous support of our sponsors, especially the Monad-level sponsors: GitHub, Input Output, Juspay, and Meta.
