Subject Line Blank E4 - AI killed SaaS? Mailtrap CEO disagrees

Everybody's talking about the AI revolution and how that revolution has killed the SaaS business model. We invited Sergiy Korolov, Co-founder and CEO of Mailtrap and the true veteran in the SaaS industry, to discuss whether AI is really threatening SaaS, what's changing in the industry, and how SaaS businesses should evolve. 

Transcript

Copy Transcript Copied

SPEAKER_00 0:00

SAS is dead. Long live SAS. Service as a software might be dead, and uh it's probably AI that killed it. And I really wanted to know if this business model is really out of the window. But for that, I needed to bring an expert to talk about it. And I invited Sergei Karlov, who is the co-founder of Product Studio Railsware, which happened also to be the parent company of our beloved Mailtrap. We had a great discussion trying to find out the answer to the one question. This SAS really did. Welcome to Subject Line Blank. My name is Marcus. Let's do this. Hello, Sergey, and welcome uh to Subject Line Blank. Man, it's a pleasure to have you here. Hi Marcus, thank you for having me here. And it's a and it's a very interesting topic because I mean I I spend a lot of time on LinkedIn and it's not only there, but everybody is talking about SaaS is dead. And I know LinkedIn likes to kill everything, but especially SaaS is something that is very relevant for me because that's that's what I do for a living, that's what you do for a living. So I want to really to find out is SaaS dead or why people are talking about SaaS being dead.

SPEAKER_01 1:23

Yeah, I mean, right now there is a lot of hype, um, and there are ups and downs, and this is something that we observe uh for the last couple of years, right? Once uh AI topic became uh really hot. Um, so one day uh the whole internet is screaming that AI will kind of change everything tomorrow, and then tomorrow everyone says, no, no, no, it's not gonna change everything, it's just a bubble. It's a bubble, uh, and it's gonna be like uh wait for it, it is gonna collapse, uh, and then you observe for a month or two months kind of this downtrend when everybody is very skeptical, and it's it's not like everybody is skeptical, it's like it became popular, it becomes popular to be skeptical, and you know, those uh folks who collect likes and views on uh social medias, they talk about this because it's popular, it's trendy, talk about the collapse. But then next day, you gotta squeeze that trend. Yeah, yeah, you need like to get maximum subscribers out of this trend, and then the next day, some new version of uh model appears, uh, like delivered by Claude or like OpenAI, uh, and then everyone starts to say, Wow, no, that's the game changer, it like kill everything. Like SAS is dead, everything is dead. And um yeah, I was uh I I I would say that you know I don't try to jump uh back and forth, like go follow those ups and downs, although I cannot say that I didn't have some emotional ups and downs, but like sleepless nights. Uh it's it's there are like different like the changes on the market are huge, so that there is no doubt. Uh positive or negative, it's just a question, but they are huge. Uh, but yeah, when I start to hear this like SaaS is dead, it like it, you know, it starts to trigger me just deeply inside, not just because, like, as you say, kind of we we do SaaS products and uh this is something that potentially can harm us, uh, so it's not like a self-defensive position, but uh rather um pragmatic. So when you run SAS product, there is a lot of things, there are a lot of things that you should care about. It's not just coding. So, why this at all appeared like SaaS is dead, right? So people start to wipe code something like smaller apps here and there. Uh, they start like users who would never be able to build anything themselves start to gain certain results, and those first results make them believe that uh if I am not kind of a professional software engineer and I can do that such things, then as a software engineer I can you know uh build like things, huge products myself without a problem. So teams will not need SaaS products, they will need just uh a few engineers who will wipe code uh the whole infrastructure and the whole automation of the company, and it's gonna be like uh paradise for the on the earth.

SPEAKER_00 4:50

And and that's where I wanted to go too, because I'm uh uh when we hear about SaaS, right? Like uh uh software, it's all about the software. But I mean, you've been leading um Railsware for for uh many years, and you know that there's more than just software when we talk about SaaS, right? What what do you think is something that AI just cannot cover? Or or in other words, how much software it is in a SaaS and how much everything else, because it's not just software, right? There's a whole thing around it.

SPEAKER_01 5:21

True, true. And uh this is actually was like a just uh uh pre-story that I started with that people who are not familiar how SaaS is kind of functioning as the machine, uh, they start they mostly consider SAS as a software, but they do not consider like the whole infrastructure behind this software, and infrastructure I mean not only like servers, but also a lot of processes uh that are hidden uh behind the scene. So let's talk about meltrap, let's say, right? You need to build the whole infrastructure and a lot of algorithms and a lot of knowledge uh for the spam shield uh project. So we had in Mailtrap, right? So when you create an SMTP uh relay, then everybody in the internet, I mean those uh folks who are spammers, scammers, they see that oh, there is an option to send email, I will try to squeeze this opportunity and send my spam uh through through through this hole, and they start to do it. And you need as the leader of the product to build the system, and it's not system of software only, it's a combination of processes of people who are sitting kind of and managing those processes, a bunch of automations, a bunch of algorithms. Uh, so it's a you know really, really complicated uh mechanism uh of people and software and automation and uh relationships uh with uh email service providers, which you need to build over the years, balance this system so you do not uh block some good users uh and you do not let bad users to send too much. So it's it's it's a very complicated thing, and it's just impossible to uh build anything like this on the knee within just uh you know a few weeks of work, and um say that okay, I have built like yet another mail trap um uh within two weeks. Uh I mean you can write such posts, but in fact, uh you won't be able to reach such results. Also, scalability, uh, kind of when you will need some billions of emails to be sent, it's just uh not that easy. It's kind of standard architectures do not work. You need some complicated uh solutions that need to be implemented, and then those solutions need to be maintained and improved. And this is something that AI may advise to someone who is knowledgeable, but just you know have like AI as a pair, as a teammate. So maybe like it's I I try to explain this situation like um in more you know in a in a longer way, but um I want to give you an example like uh which I personally liked. Uh so everyone can cook something on the kitchen, right? Kind of you can boil eggs and like uh fry some sausages, and uh you can put some pizza in microwave, but it doesn't make you a chef, right? It uh does not kill it doesn't kill restaurant business at all. You have a car or you have a bike, you can go to you know to chipotle or kind of Starbucks by this bicycle and buy a coffee or a pizza or like taco, you can buy it uh in the restaurant, but somehow you order it through the service, uh like Uber It's or Doordash or kind of Glava or whatever. So you use service, which basically you can do yourself, like why because it's convenient, so you remove a lot of additional activity from yourself, kind of to buy some ingredients, to prepare a dish, then to clean uh the you know all those plates and forks and so on, like and cups. Um you just go to restaurant, uh, you say that I'd like uh let's say pat thai or whatever, and you have pat thai uh in your plate, and you don't care, like you pay for this, that's it. Uh you get it, yeah. Yeah, but like you can open Google Maps in your city and search for restaurants, and you will see thousands of restaurants, right? They are not dead just because you're able to boil eggs uh on the kitchen. Uh, so that's the same situation with SaaS products. Uh the regular company uh uses more than a hundred uh plus uh different SaaS uh uh inside to run the operations. Kind of every department has 10th plus of SaaS applications that are used. Imagine that you need to build them yourself, but then you also need to support them, make sure that someone will wake up in the night uh and fix the problem if the server goes down, and also someone needs to follow all those changes in the technological stack and update all those libraries and improve all those algorithms. I mean, this is crazy.

SPEAKER_00 10:55

I don't think you can do that as a prompt that quickly.

SPEAKER_01 10:58

Exactly. That that's the problem. Kind of um even if AI does a lot of good work, you still need somebody to operate it to kind of integrate those systems together. So, yeah, I that's that's why kind of I've I've decided to write even a post on LinkedIn which became uh pretty popular. That uh SAS is a king, it's still a king, even when you uh pay to Claude, it is a SAS, kind of right. So right, I mean you're still asking someone to do it for you. Yeah, you pay your monthly bills uh for ability to use this software, and then you pay additional uh for tokens uh that you burn. So that's a SAS.

SPEAKER_00 11:47

And I mean you you do have experience. You you made a lot of um customers that you had over the years, I mean, not only with Mailtrap, but with Railsware as well. Um there is this idea, especially for the vive coders, right? That you don't need the support behind, you just need the solution. But uh in your experience, is that really what companies need? Because we we do have big names working with us, right? Like is it's just the solution? Do companies and customers think only about just solve my problem, or they really put value behind the support, behind the uh the teams, uh, behind the people that is building this software, or they just want it done, in your experience.

SPEAKER_01 12:30

Yeah, of course, uh as I've said, that's uh sales is much more complicated mechanism than just a software, and um kind of you as a company you need to have a trust uh in someone, right? So let's say when you sit into the so you buy a ticket for the flight uh to let's say San Francisco, uh you want to have a trust into the company uh that you buy a ticket from, right? You uh uh understand that there are some processes behind and regulations that make sure that the pilot is not a random person uh who is sitting there and who will not fly your uh airplane into the mountain because uh he doesn't know how to manage the situation. So uh that and to build the whole system it's uh it's a complicated thing, but for you it's you just sit into the airplane and you fly it. So it's that simple, right? Um kind of the services with with help of SaaS, they brought our activities, our chorns uh to a completely different level. So we're able to manage many things simpler in our devices, uh with help of our phones and even watches. Uh and this is because uh kind of thousands and thousands of people uh do their daily routines uh kind of building those applications, support them, improve them, uh closing uh security holes uh when they are when they appear. So all those things are important which kind of it's uh impossible to reproduce um kind of when you decide to build a hundred plus uh SaaS applications uh within your kind of processes in the company. So to cover your processes in the company. So that's that that's crazy. And yeah, think about legislation. So imagine you have uh QuickBooks or you have Xero or some some Xero or some other tool which uh requires constant updates uh when the legislation uh is gonna change. So who will follow those leg or not even like any application like GDPR appears or kind of some Californian uh Californian laws uh regarding the privacy of data? Who those things changing all the time, right? Yeah, yeah, they are changing constantly. Uh so like every year something happens. Uh so in uh in France they recently uh introduced the new law uh of that cover that helps consumers to hide their IP addresses, kind of and make sure that the IP address is not being exposed to to to the uh until they confirm it. So many things are changing and somebody should support it, somebody should follow those uh uh specific regulations for the specific country for the specific type of service. Like it's it's crazy, it's just uh impossible. And um I'd like to tell like my personal story about my experiments uh with vibe coding. So yeah, and this is uh kind of you you do have a background, right?

SPEAKER_00 16:00

Like you have the knowledge because right now everybody can be a vibe coding. I mean, it's just yeah, yeah. But yeah, yeah, please go ahead, man. Go ahead. Like, how is it going with that?

SPEAKER_01 16:09

Yeah, this is like a separate story from those uh from the story I've mentioned before. It's like a different perspective, uh, what I want to highlight. Um, so I'm engineer myself. I I'm 43 right now, but I was uh in software development since 14. Uh so for quite some for and I played uh so many different roles. So I was engineering, then I was product managing and uh setting up companies and products, and kind of this is something that I do on the daily basis. Uh, for last uh couple of years, not couple, but just decent amount of years. I do not program myself like a production-ready solutions, I do experiments and stuff like that, but uh something that we deliver in Ralseware, it is developed by professional uh software engineers, product managers, uh, so those guys are doing a great job. Uh, but I definitely follow the trends and I experiment myself on the background uh so I can have better knowledge and understanding uh of the direction where I should drive the company. And um wipe coding is a thing that appeared, and of course, you cannot just miss it, right? So you start to pay attention uh on it more and more, more hype about it, and so it was 2025 uh uh summer, I guess, when I tried to wipe code something myself, and uh it was like a pretty simple web application, and uh it sucked, like it was impossible to build anything reasonable. Uh while uh YouTube promised you that you're already able to build calendly within 15 minutes. Back then it's yeah, it's such a weird declaration, but anyway, so I tried, it didn't work out. All right, so I let go and monitor situation. Like I was constantly communicating with engineers how they use AI, and it they used AI heavily, but it was not vibe coding, so it was like a prompt coding. Let's say they they were uh in the process of development, but then instead of writing tests themselves, they were asking already AI to write those tests, and then they were uh following the logic of those tests, kind of changing things if there are needed, uh if changes are needed, and so on. So AI was already heavily in the development process, but on the uh vibe coding level, it sucked. So people who claimed uh like that they've built an application with Lovable, usually it was like a one-pager website, which you were able to have with uh with WordPress or Wix or any other tool uh within a few seconds, uh and uh those applications would kind of applications actually websites would be much more stable on on those, but it wasn't cool uh because it was old school, like new school, you wipe code everything yourself. Yeah, but uh that story uh was 2025 mid of year, but then um uh I guess in the autumn 2025, I tried to build first uh CLI for mailtrap. Uh it's a console uh interface uh for for mailtrap, so you can query uh and ask Mailtrap to to do some some work uh through the console through the terminal. And I was pretty surprised that uh so from the engineering standpoint a pretty simple work. So you have documentation, you have API endpoints, and then you have uh you need to write a wrapper around those endpoints, uh kind of that's uh make uh this application work and execute uh certain expectation like execute certain commands. So it was pretty easy, but in fact, it was able to build this CLI which worked and it was written in Go, and I didn't I didn't have any experience with Go. Uh it was easy for me to read through the details because like I came from C and C background, so Go was not a problem to read and understand that the code is clean and nice, and uh there were some bugs, and we kind of uh together with AI were able to fix it, but I haven't touched any code myself, so I was fixing those problems, explaining uh AI that there are there are some problems. So, anyway, it required me to be an engineer to explain those things. Um but but that's fine, that that's something that I've built that way, and then I tried to build something with UX and I wasn't able to achieve result. Then there was some pause, and uh when uh Opus 4.6 and 4.7 appeared on the market, uh my partner Yaroslav came to me and uh showed me uh the app he did, like a test uh test. Sorry, uh it was a task management tool. He was able to hack uh within a weekend, and I was pretty surprised because I was not able to achieve before like uh my experiments uh such uh precise application. So I did uh experiments myself, and I was kind of like that actually the period when I had sleepless nights because uh it was really shocking for me that I was able to build an application. So I started with the simple application again with the CLI, but kind of built a new version of it. Then I decided to build Like a desktop application with uh from Ltrap, just an experiment. It that product uh uh never appeared on the market. I mean, this this application never appeared on the market because uh didn't make a lot of sense, but it was my experiment. So, what can I achieve? Will I be able to build an application which works on both like Mac, Windows, Linux? Uh, and I actually was able to do that without writing a line of code. So it was like an application, you put your talk in, kind of you see your emails that have been sent, you see your statistics, and so on. So it was pretty impressive.

SPEAKER_00 22:37

But sorry, um sorry to interrupt the story for a second, but uh that's um that's because you know what you're doing, right? Which it goes back to your little to when you were talking about like you can cook, but you don't you're not a chef. I mean, there is a knowledge that you have in how things work that you were able to apply to AI, right?

SPEAKER_01 22:57

True, true. And later on, when I've built already some experimental things uh more complicated, uh so just to make long story short, uh I've built uh pet project which is uh just had absolutely nothing uh to rail activity or meltrap. Uh I was just purchasing a car and I wanted uh to have like a better comparison of the cars, and so kind of just you know, wipe coding helped me to build a pretty decent tool for choosing a car from like mobile D uh website, and so I was able to build it without writing a line of code, but you are completely right that uh I was I faced a bunch of problems during the development of this tool, and I've built it through you know a few weekends. Um I faced a bunch of kind of questions that I needed to decide, like about technology stack that I need to choose. Then uh there was a question about optimization of requests uh to a to uh open AI API because there were huge delays, and it was just you were sitting and waiting for minutes uh when AI will process uh the data from the website. So I've built a bunch of optimizations, I will I parallelized uh the processes and kind of optimized the many things because I knew as a software engineer how it works uh under the hood, and so and later it helped me, of course, to set up the server, kind of set up the domain, connect everything, launch it as an application, and so uh, but happiness from my side was that I was able to build a pretty decent tool without writing a line of code, actual code, yeah, actual code. Um, and it works. I mean, it works not only for me, it kind of you can if you want to to buy a car on mobile DE, you can go there and uh kind of create those collections and so on, so you can play around with it. Um, it's not something you know that I consider it some for commercialization, but it was an interesting task for me. If I as you build something that works, yeah. I mean, do I need a software engineer right now to accomplish this task? And at that level of uh AI and at that level of my understanding of software development, it's a this combination is enough to build this type of tool, kind of, but it's still relatively simple. When I try to build a feature inside Mailtrap, I failed to wipe code it. So, because that was my next step experiment. All right, am I able to build a feature like a decent feature, not just a fixing button or something, but build introduce some decent feature into Mailtrap uh with wipe coding without help of engineers? And quick answer is like it was a complete fail. Uh so I was able to good for the engineers, though. Yeah, yeah, for sure. I mean, so because you need to understand where are the limits and what can you achieve. And my kind of summary of what I've after multiple of those experiments for the last uh I would say five months, uh yeah, I did all those different experiments myself, and of course, communicating with team members, like their experiments. So I've came to the conclusion that uh for prototyping for building MVP, uh, it's a fantastic tool that you are able to, you know, have a really small team, you don't even need some engineers at a certain period of time, like being a software product manager or product designer, you are able to wipe code something and actually bring this something to the market to validate the idea. But when you need to grow it further, when you need to build kind of more complicated solution, when you need to move like to push it, then it's absolutely, at least from my perspective, impossible just to stick to the concept of wipe coding. You're basically getting back to the processes that we have in software development before AI era, where you create a code, there are pull requests, uh, there are team members who review those pull requests, accept them, and kind of and by this, they continue to keep quality of the of the software. They write tests, but AI tool in this perspective, it became like it's already prompt coding, it's not wipe coding, but prompt coding. So AI helps a lot engineers uh to remove those routines, but they are not operating on the high level, like create me this feature. No, they are consulting with it, uh, explaining what type of feature they want to build, they consult with kind of with AI to decide architectural things, and when they implement, they slice those uh big plans into the smaller tasks, they deliver those tasks with AI, but then they review it with their own eyes and understand what actually is done, and if in if this piece of software can be committed and pushed to the mark uh to the main branch, to the major brand, the main branch, and um become a production ready solution.

SPEAKER_00 28:46

So there's one thing that you mentioned though. Um throughout your story, you you tested it in and like sort of half 2025, later in the 2025, you've been testing more in 2026. Have you seen a significant advance? Uh, like for example, you mentioned um Opus 4.7. Have you seen an advance that is saying, okay, for now, vive coding can get you here, but very soon it will get you here very fast. Like, how how do you see the speed of development for AI at the moment?

SPEAKER_01 29:20

That's uh I guess the most tri like the trickiest question because you never know what will appear in the next uh in the next day. So that's that's what I've uh that's why I've compared what happened in 2025 and then what happened in the end of 2025. So you start, observe the uh abilities of AI, it does something, you understand that there are a lot of problems, and then next day you see the new version that solves a lot of problems, and then next day you see the version that solves uh more problems, and there was like a huge spike, as I said, between those beginning of 2025 and end of 2025. It was just significant boost. So, as I've mentioned, like not able to build anything reasonable to ability to build a reasonable software without any line of code. Um, yeah, but um what happened next? I personally don't know, and I don't want to you know predict that it's I may say something right now, and when you will render this video, it it will become invalid, it will become invalid. Uh so I guess only guys in um in those big companies like Anthropic and uh OpenAI and Google they may know potentially where it leads. But I had a conversation with the guy who works in Google, so they're in absolutely similar situation, like uh the other engineers. Uh so they use inside the same models uh as they share to the outside. So there is no there are no secret models uh that they keep there. So the latest thing goes uh uh outside, so goes public. And um they they're searching for those limits. Uh and uh I I like the rule they had that if you cannot resolve something with AI, try to resolve it with the eye. So meaning that uh without trying to okay, AI doesn't work, I will write it myself. No, they actually try to find the limits of AI themselves, so that's why they try to change the mindset of engineers. So they uh for them, it's also like a journey, experimental journey right now. And what will happen? So we all heard about this mythos uh right model by uh Anthropic, and um that it shows some significant result, but I haven't had a chance to test it. Kind of then I don't know uh what will be the results. Maybe it will fix the problem of uh the pro because wipe coded code right now far from being great, let's be clear. So sometime and this is something what happened during my experiments. I've built a certain app, uh, and I came to engineers, they open it and we looked through together. There were good places, but there were a lot of crap. Uh, and then when you try to fix this code by guiding AI without changing code yourself, it sometimes introduces even more crap. It's like it already went the wrong way, and then instead of getting back and start to build new routes, it just tries to get worse. Yeah, it's just getting worse and worse and worse and worse, and then maybe if they will resolve this problem, uh then we will have new chapter when you will have uh just even more possibility to build something yourself, but it will not eliminate kind of this it's like Tesla uh self-driving, right? So you are it's it's fantastic. Like uh I've been recently to Los Angeles, and uh I was driving uh with uh with Yaroslav, my friend. Yeah, yeah. So self-driving, like we were drive, we were driving the full day on self-driving, and I was totally impressed. Like it's it's uh it drives you like a ideal driver who understands that all right, there is a car on the left which is speeding, and then you need to uh pose, but not pause, like you shouldn't stop fully, but you need to just re reduce the speed, and then when it goes by you, you just uh get into the lane and speed up as well. So it does really predictable moves, and uh, I loved it a lot how it drives. But what I wanted to say that legally, if this car will kill the person, you will go to jail. Uh so legally, the the driver is responsible for everything that self-driving car will do. And this legislation is works this way. In software, it's more or less a similar situation. Like uh in the end, if there is no person who understands what code has been shipped to production uh and what are the consequences, uh then it's it's a problem. So, person and responsible who put his or her, like you know, stamp that all right, I guess I put thumb up, this is software that I reviewed, I understand what it does, it is it is ready to go to production. So that's that's a key element still uh in the chain of the software development.

SPEAKER_00 35:06

And based on your experience, like you've been playing around with um AI a lot. Um, how much influence do you think AI is gonna have in the future for SaaS? Um, how many more areas uh AI is gonna get in there? Because I I totally agree with you, and I think we mentioned it before. I mean, AI is great as a companion, it's great, great as a tool, but do you think eventually we'll get more than that? We'll leave, we'll create like unassisted. Like, how do you see uh AI specifically taking over certain things in SaaS?

SPEAKER_01 35:44

Well, this is uh what we observe, right? It's not about software only. Um, so speaking about the software, you have this uh quick possibility to validate, which is create. Um, so different prototyping uh designers should not spend too much time in Figma, but they can just do upgrades uh in on the real product, uh but in a separate branch, and then they just can kill this branch uh afterwards after they re after they present the feature uh to the team. So that's that's very useful. If we talk about marketing parts, uh there are a lot of automations that you can build, and this is something we do uh across the board in different products uh which help like uh competitors analysis, uh introducing kind of different approaches how to create content. We still uh do not kind of follow trend that sometimes uh you know hackers, uh gross hackers show on the LinkedIn on LinkedIn that you can generate everything with AI. I've built an automation that kind of creates everything themselves, like everything by AI, nothing by a person.

SPEAKER_00 37:02

Yeah, but but then they try to sell you the course, that's it.

SPEAKER_01 37:05

Yeah, that they try to sell you the course. So we uh we of course we use uh AI for creating content, but behind this content, there is a person who understands the value, the story, who collect this data, and then with the help of AI, it like he or she shapes uh the article or the the the page itself. So this is something that uh but definitely that's uh that's a powerful tool. Will AI be able to analyze everything later on and uh create the website for you fully? In theory, yes. Um, kind of it already can do this, but if everyone uses the same approach, like who will win? Um, like there always should be like should be some secret source, right?

SPEAKER_00 38:00

And that's exactly what I what I wanted to ask you on that. Um where we can all access software creation, right? Everybody will be able to create the software that they have in their mind. But in your experience, dealing with real customers, real people, real life examples, what do you think will make the difference? What do you think is the thing that's gonna keep SAS alive? The secret sauce. What do you think is that secret sauce?

SPEAKER_01 38:28

Uh like uh we already can we already can notice that people like start to like uh emails with uh mistakes, uh kind of post with mistakes, uh because they feel that there is a person behind who actually written that. And people already train their brains to see if this is even written by Claude or by ChatGPT, so those patterns, and I guess in the nearest future, kind of the game changer will be how human your service is, like how many human communication you receive uh uh through through this uh kind of through the use of this service. Like if it's gonna be only the chat uh bot which reply to you in uh in support, or it's a real person who will write to you even with mistakes. Uh so I guess uh the this this balance of using AI and human touch uh is gonna be the the secret source. Um yeah, that's that's a and brand, of course. And brand won't be I mean so when you deliver something right now on the market, it's very complicated to build a brand. Uh so still the old brands uh they have values, of course. Sometimes they fail, but if they follow the trends and they're able to adopt uh this powerful uh powerful inertia that uh big brands have, they're able to use it and still still be kind of on the top of the market. So it's great that you build the software, but what happens next? Kind of uh all right. You who will know about this software? How you will reach your audience, how will you sell them the idea, uh, how will you collect the feedback? Uh so all those questions start to appear, and uh it's it's it's it's not gonna be a problem to build some simple software, but it's gonna be a problem to deliver it to the community so community understand it and uh start to use it. So that's that's a challenge.

SPEAKER_00 40:59

And I think especially like what you mentioned before, too, is the is the trust, and the people get to trust humans, and when they feel that the humans are behind, uh it makes a big difference. Uh Sergei, I have uh one final question for you, uh, which is we started with the idea of SaaS is dead. Um do you think, hand in your heart, that AI is gonna kill the SaaS model eventually? No.

SPEAKER_01 41:28

Uh it's a complicated question. Uh it is because I'm scared too. No, I I I don't believe that uh it's gonna replace SAS uh by the reasons I've mentioned before. Yep. Um yeah, that's uh so far. I don't see this uh this this thing that is coming. So good new that's long story short. There are gonna be new brands, new applications that will grow from uh from from this uh movement. There potentially some kind of like 10 tools might be replaced with one tool. So it's not like SaaS as a concept won't be dead, but some SAS applications gonna be dead for sure. Like even uh if you think right now about those um tools which allowed you to create your own uh website, like Wix uh or like Square something square, I don't remember what like square something. Um so like right now you go to Lovable, let's say, and you create your web page uh pretty quickly, uh so just by saying to it uh something, and it's cheap and it's no problem. So I know that uh those folks they start to lose uh traction, of course, because there are alternatives, but that's that's everywhere, it that's always happened, like when there is a change uh in the technology, the new opportunity appear, kind of the new brands pop up on the market and they start to grow. But to us as a concept, uh I don't believe it's gonna die.

SPEAKER_00 43:05

That's good to hear. That's exactly what I wanted to hear. Probably a lot of people around. Um, Sergei, well, thank you so much for your time. It's been a uh a very cool conversation. Um, is there anything else you want to cover or anything else you want to mention?

SPEAKER_01 43:20

No, no. Um, I would just uh recommend everyone, doesn't matter what you do, uh, if you're a marketing person, content person, you know, you uh support team member, sales guy, uh try to go and wipe code something. It's just uh refreshing. I mean, like you don't really need to be a software engineer, at least to see some powerful elements and uh just don't block yourself, go and try to wipe code something. Uh it's it's it's gonna influence your understanding kind of how software is created.

SPEAKER_00 44:00

That's that's a good advice. I'm I'm doing it. I'm actually building my app. I'm gonna bring it to Railsware later one of these days. There you go. Thank you so much for everything. Uh, for you guys watching, thank you for joining us. If you have any questions or anything that we can cover after, make sure you drop it in the comments. And remember to come back, subscribe, put the notification uh button. And if you want to know more about email marketing tech news, come back to subject line blank. My name is Marcus. Thank you, Sergey, again, and we'll see you, Marcus, again. Take care, brother.

SPEAKER_01 44:33

Thank you.

SPEAKER_00 44:34

Bye.