Почему ИИ заменяет front-end, но не back. Прособесим, чтобы не остаться без работы

Okay, folks, I'm live. Good evening, everyone. Wishing everyone a good mood, a productive Saturday. For some, Saturday means parties, revelry, and rest. For others, it's bachata parties, and for Dimka, it's work, as usual, folks, keep up the good work. I'll tell you all about it now. I've been grinding away for probably three months now, roughly. Without days off, approximately. In what mode, you ask? How much do I manage? Well, about 60-70 hours, yes, close to 80 working hours a week, I guess. So, I wake up, sit at the computer, and go to bed from the computer, well, I only get up for the bathroom and to eat. Three months of just this kind of seclusion, maximum effort. So, in that sense, I'm a bit happy, of course, because, you see, I'm not tired, so to speak, although perhaps I am tired, but energized. Therefore, I wish you all productive work on what you enjoy. Choose what you like. And we'll chat about this topic now, because I have some big topics lined up. They've started writing something in the comments now, some not-so-good things, this and that, you know. Well, you know that, in principle, I'm like, white and fluffy, like I'm on the side of good, on the side of honesty and justice, blah, blah, blah, blah, blah, blah. In short, you can't fault me. Therefore, nevertheless, I always welcome all haters. Write constructively what you don't like, what specifically, for example, seems, um, controversial, unclear, fake, said by me. I gladly accept any constructive attacks. Because, for the most part, everything seems to be on the surface for me. What I think, that's what I broadcast. And in this sense, there are very few, it seems, hidden cunning schemes in my head. They exist, but they are not aimed at some grand conspiracy, deception, at some, I don't know, personal enrichment at any cost. This applies to all my activities and just working for hire. I'm currently working at Galere, and we have one task left. And I haven't worked on it all week. Yes. I always, I, in short, I, in short, track every hour I work. What I don't work on, I don't track. Recently, someone asked me: "Are you really not tracking time beyond what you, like, haven't worked?" So, are you honestly tracking time? I say: "Well, if I were, perhaps, working for hire, I would, of course, allow myself to browse some article during working hours, read something, even if it's not on topic, but to rest during work, or I would think: 'Okay, I'm on lunch for an hour and instead of eating, I'm chewing, but thinking about work. Am I eating or am I thinking about work for this hour?'" Accordingly, in this sense, I understand that no, if I dedicated my mental activity not to jokes and laughter, but to work, then, well, well, in short, that's how I got used to it, yes, like some kind of slave mentality, not free, that you need to try to be, like, a little honest and all that. And this, as it were, is reflected in the activities of the guys that are happening in our incubator. Therefore, it's very strange for me to hear some things directed at me. Well, if people are just passing through, of course, from the outside, anyone who sells any kind of training, any courses, is, of course, an infobroker. Well, in general, in this sense, perhaps even respect to the guys. They run to everyone, as there are probably many malicious scammers, so they run around and call everyone a scammer, who is being fooled. But if someone follows the activities of my incubator, then, well, it's a bit strange for me to hear such nonsense directed at me. But oh well, it's not the end of the world. It's not the end of the world. People are different. I'm already used to the fact that people are different, adequate, inadequate, cool, not cool. And, in short, everything will be, everything will be if it moves forward at its own pace. Immediately, what I want to show you. Well, let me show you something cool right away. We, uh, did I joke about it on Telegram or not? Now, yes, let's find out. We hired actors. We hired actors and are launching the start of a new season of the project Samurai Life. Remember our old project where we hired actors who pretended to be studying with us, studying, studying. For a year and a half, they pretended to be studying, played this role, played it well, and then played the role of being employed, employed. Now there are thoughts, they'll ask for small fees to invite them for a meeting, to find out how they are, as if they are real developers now with some experience, who are already working. Well, that's it. >> Friends, a super important announcement. Next week, we are launching a new season of our show Samurai Life. Every week, our samurai will record a short video report about how they live, how their studies are going, how they learn, how they work, how they rest, what problems they encounter, and what new things they learn in the incubator. These are our samurai heroes: Ilya, Vladimir, Vika, Yaroslav, Vadim, and Alexander. Guys, please choose your hero and follow them. Subscribe to our channel, follow our samurai. You will see how learning in the incubator really happens. Support the guys with warm words and kind comments. Like. The path of a samurai is not an easy path, and it's long, but I hope that we will all see how all our samurai receive their coveted offer. >> Follow our super show Samurai Life. >> Victor wanted to say the coveted Oscar, but it turned out to be the coveted offer. But Victor, probably, is also not aware that these actors are hired. In short, honestly, I give a big shout-out to Igor for doing such a thing once. We had a star teacher on the team. Today it's simple. Just updating the text. Mom saved it. Dad saved it, erased it. >> Okay, stop, stop, stop, stop, stop. That's it, for now, I'm turning it off. So, and now the guys have gathered as a team and say: "Let's restore this thing," and they come to me with a proposal: "Let's restore it." And I say: "Well, if you want, restore it, I'm happy to, but I just don't have the strength to do everything." And a big shout-out to my team, who are now working on this. And we will now show you, guys, that this long, this difficult, not easy samurai path, yes, but it, which is guaranteed, nevertheless, an important point, it is guaranteed to lead to you going into service to the Shogun, the Shogun pays you a salary, and, in short, you have everything, your samurai career has taken place. That is, it is guaranteed, guys, still in the market, as in all things, the road is paved by the one who walks it. If you like it, you go, you go, you go in the right direction. It's desirable to wander minimally, to formulate the topic more clearly, like a goal for yourself, why am I going? For employment. Where? Well, as a confident, where? Well, into the industry, as a confident developer who more or less understands where he is going, with a foundation, so that, for what, to stand firmly on your feet with a foundation and build a career on this foundation. And here, guaranteed, guys, everything will be fine with all of you. With all of you, it doesn't matter if you are studying with us in the incubator independently, somewhere else. If you diligently go, go, go, go along the right tracks for a very long time, you are guaranteed to reach 100% of your, again, I wanted to say Oscar, your Oscar. And now I'll see what you've written here. And gradually, yes, I'd like to have a dialogue. Guys, I will, of course, tell you the announced theses now, why I wrote this, where it came from, yes, about warming up the hamsters, as they wrote, why I even thought about it. And I'll even open some Miro board now, maybe, and draw something for you a little. Newend. I actually wanted to tell you about our new knd that we're launching. Honestly, I really wanted to. And I probably will. I'll just open my cheat sheet. What Vanya and I came up with. And what we'll do, what we'll see. I probably will tell you. I won't tell you 100%. I will tell you. I will tell you what I'm passionate about, guys. I'll tell you what I'm passionate about. I'll tell you. I'll just open WebStorm now, because we have notes in WebStorm. WebStorm, bim, bimbim. Programmers don't work in notebooks, everything is in code, everything is inside WebStorm, inside Git, in IDEs. M. This Uber project, I think. And here we have notes for the full program. I'll peek here when I'm visualizing something for you now, drawing. Okay, okay, okay, okay, okay. I like what I'm doing so far. So, guys, first of all, I'll greet you, I'll look at what you've written there, with pleasure. So. Well, let's go quickly. Boom. Okay, I just want the comments to be displayed on your screen. Boom. Oh, in 2022, there were headlines like "AI won't replace frontend." Let's wait a couple more years. M, there were headlines and there still are headlines. Who will replace whom, who will not replace whom. Yes, the question, guys, is very, in short, we are living in such a revolutionary era that it's just, well, I'm in shock. I'm in shock. Although, naturally, everything is developing exponentially, but roughly speaking, the internet didn't appear yesterday, and Google also seemed like something at some point. How so? Any information, damn it, and you find it, in the sense, no more libraries, no more books, no more cards. Ah, and now another one: "How so? In what sense?" And this is like one-two. And it talks to you like a person, knows everything, can do everything. How did it read 50 files in 5 seconds, figure it out in 15 seconds, and fix a bug in 5 seconds? How so? How? Yesterday, I was working with very legacy .NET code on an old Windows. I installed this cloud-code with a debugger in some old terminal. I fiddled and fiddled and fiddled. Well, because very rare fixes are made, and in general, the code is so scary, it's not just legacy in its essence, that it was written a long time ago and the technologies are old, there's also a lot of disgusting, horrible code. So, I was, in short, fixing a bug, fixing, fixing, couldn't fix it at all, cursed. And to check everything, you need to deploy. In short, naturally, there's no normal test environment for these large, complex microservice interactions. And I was like: "Damn it, I was stunned," and then I was like: "Wait, let me try cloud-code now, wait, I say: 'Here, here, here, here.'" It analyzed for a long time, but then it says: "There's an error here." I say: "Well, yes, but where did this code come from? How did it work?" Then I remembered that I hadn't deployed or updated in a long time. In short, I thought, well, okay, I'll give you a try. It's likely true. Well, damn it, it works. I say: "Fo-foto, how did you do it, damn it." I say: "H-hey." And yesterday I continued working positively until 5 AM. I closed this task at 12 PM. I spent another 5 hours setting up Linux servers, migrating projects from Heroku to my own server, well, to a dedicated separate Linux server. Setting up Caddy so that it supports multi-domains on one project, in short. And in 5 hours, I did everything, sorted it out, tested it, but without chatbots, I would have spent a week, 2 weeks on this. It's just fantastic. The acceleration is so fantastic that I'm just in shock. The video titles amuse me. Alexander, why are you amused? What bothers you? What's your opinion on this? I'll be glad because I'm interested to hear your opinion, guys. Ah, a great thing, it turns out. All frontend courses were only to then sell backend courses, and then, guys, backend courses are only needed to then sell architecture courses, and then to sell DevOps courses, and then to sell company tech lead courses. Do you understand how it works? Of course, it's an evolutionary step by step. This is the essence of each of us. Therefore, of course, of course, of course. The second thesis: frontend courses only, but to learn backend, you can go for a full-stack vacancy. Oh, to my surprise, yes, full-stack developers have always been valued, always been in demand, but complex projects. Complex projects. Well, for the most part, not much has changed. And large projects cannot be done by full-stack developers, because there are too many nuances in UX/UI, too many nuances in backend. Two independent worlds, so to speak, where you need to focus on different things, completely different things, and keep it all in your head. Only if needed, well, add input and a new field to the table. You need an additional registration form and send an email. That is, support, some small refinements, bug fixes, a full-stack developer will do it perfectly, because, well, a full-stack developer has AI, because before, even for support, you always had to have a frontend developer, a backend developer, and they could smoke 95% of the time, 5% of the time, patching something up, cool. But now, why do we need such specialists? You can take an adequate full-stack developer, and they will work on three projects. Here they will close tasks with a small percentage using acceleration, guys, it has accelerated cosmically. Everything, I'm still amazed. Sergey Ivanov. Sergey Ivanov. M. Waiting, yes. Waiting, waiting. It's even interesting how they scam hamsters nowadays. Interesting, yes, how hamsters are scammed. Tell me, if you know. So you can write how hamsters are scammed. I can't, unfortunately, scam hamsters. And I don't consider any of you hamsters, friends. What, what do I consider myself? So. Ah, online recording. Online recording. Hi, Seryoga. You've thrown in some funny emojis. I'll click through them now. Will AI replace frontend? That's the question. Will frontend be replaced or not? Look at this idea. There was a recent stream on the IT Incubator channel. M, Vanya and Viktor were talking about backend. I chimed in that AI makes coding even easier. In short, coding with AI makes backend coding even easier than frontend, because there are no unclear things on the backend, like how to decompose, how to build architecture, interaction, transactions, and so on, to break down into contexts, in short, well, there aren't, as it were, it's quite clear in most cases on the frontend, starting from any UI, what manager to use, what version of React or Next or TanStack Router or and only this zoo of everything makes you do this. Okay, so, so or so. Because, uh, accordingly, on the backend, as they once came up with writing on the backend with standard, like, clean architecture, hexagonal architecture, use cases, commands, query separation, in short, everything is broken down, transactions have been around for 30-40 years since SQL was invented. Okay, and we throw events somewhere into Kafka and prepare some read model. Redis, in short, everything is very, like, well. Let's go. And backend coding, in my opinion, is even better and easier. Well, I, uh, a lot of the project, which I'll show you now about interviews, yes, the link is in the description where you can practice interviews with each other. Students are already interviewing actively, and soon HR will be picking from the platform. Be sure to follow the link and start interviewing right now, because if you don't start by the end of January, you don't conduct at least one interview and don't conduct one, it's all free, then you'll lose this free opportunity. So, let me show you, I'll show you right now. Go here. Go here to itincubator./interview via the link in the description or through our main website, you can probably also get in somehow. I don't know if we've put the link up yet or not. I think it's interview practice. And here you choose frontend engineer or backend engineer and go through 1,000 questions in the interview room. There are already 186 participants and you go through nine grades. Junior 1, Junior 2, 3, Senior 1, Senior 2, Senior 3. And I'm currently at the trainee level. I have 100 questions open to prepare for junior. And the system tells me what I need to train, what the interviewer will ask me in the next interview. You see, I have 19 out of 20 interviews left to get to the next level. Guys are interviewing actively. Look, Artyom has conducted 20. Ilya Pikhach 15, Alex 14, Gadel 10, Alexander, in short, they are all working hard. You're all working hard, it's absolutely free, but hurry, well, hurry up and do it now before the end of January, so that this freeness remains for you too. Later, we might close the freeness and there will be a subscription. Or maybe not. In short, but in any case, don't drag your feet, because while you're dragging your feet, someone else is improving their skills, their interviewer skills, their candidate skills, and getting jobs. Therefore, this is a key skill, we'll talk about this today too, we'll talk about this, because it's very important. So, this project, both frontend and backend, I wrote 90% of the code. Well, there, due to the fact that we already had a codebase and many different microservice integrations, specifically this wipe-code service was about 60%. Only because, only because, as I said, there was already a codebase. The frontend was completely created by AI by 95%, meaning there are microscopically few lines of code written by hand. In short, it helped me create this project. My team, to some extent, I did most of it plus cloud-code. And on the example of this project, which needed to be created dynamically and quickly, backend coding is easier than frontend. It's easier to code than frontend. It's really easier with AI. But the important point, the important point, the cost of error. The cost of error. On the frontend, whatever AI does, I say: "Doesn't fit." Redo it, redo it. I can do 10,000 iterations of redoing, because the cost of error is zero. Well, none, zero. The main thing is that it works. That is, you need to test that the functions work. The payment button doesn't work. The cost of error, yes, maybe 100 payments not made, redirection didn't happen, but the cost of error is very low. You fix it, it all goes to production, and everything works. What is the cost of error for the backend? While you are developing a project from scratch to version one, to release it into production, the cost of error is zero. You work, you work, you work. Even if your migrations don't work, the database constantly drops, the schema regenerates, so you don't care at all. But as soon as you go into production and data appears, you can't say with your eyes closed: "Hey, make me a new feature." And you say: "Well, probably it did it, probably everything is fine. On the frontend, you can do that. Probably everything is fine. I'll go poke around. I went, poked around, everything is fine. I'm not even talking about end-to-end tests, unit tests, storybooks. Just go poke around, everything is fine. On the backend, you can't say: "Make me." And then I'll poke around and everything is fine. When you already have production, the cost of error is huge. You say: "Make me something wrong, do something." And you didn't fully formulate the requirements, what AI will do. It will drop a column, a table. Reset counters, denormalize incorrectly, write something off from somewhere, and then you can't say: "Well, what kind of bug is this, figure it out, figure it out." It can figure it out in the code, but you can't fix corrupted data. And the cost of error is, yes, why backend developers have gray beards, because coding on the backend is no harder than on the frontend, but the cost of error is much, much higher. And, accordingly, trusting AI. Well, that's debatable. Now I've been grinding at Galere for 3 months. Well, grinding, lightly grinding. Okay. And now I have a final task that I thought I'd finish. Damn it, damn it, damn it. Again with AI on the frontend part. Damn it, damn it, damn it. But it turns out no. I have to go back to the backend configurations. The server is written in something, something. I have to somehow go into the base default Docker container, patch it somehow. Well, in short, make my own, or go through Kubernetes, there, provide configmaps, other configuration files so that they are picked up. And I, not being a DevOps engineer, understand that it's simply impossible to tell AI: "Set up my config maps, Helm, this one, and deploy all of this." Even if I don't take it to stage, I won't take it. I'm scared, because the cost of error, if I mess something up now, close something, then I'll have to go to a DevOps engineer and restore all of this, an expensive DevOps engineer, in short, and the cost of error is high. And, accordingly, why, yes, it's harder to replace AI and backend developers, it's harder to replace DevOps engineers, because the cost of error is like, excuse me, remember that meme, right, exactly, when ChatGPT performed surgery, it says: "Oh, yes, you know, you're right, I should have removed your right kidney, not the left." Now, now, now, now I'll fix it, sorry. And this is somewhat similar to working with AI on the backend and on DevOps. The cost of error is very high. On the frontend, you can do whatever you want. I haven't encountered a task yet where it's important for me to do something pixel-perfect according to the design on the frontend. Accordingly, AI and various MCs and Figma MCs, how do they do it? Recently, a friend of mine told me that she's not succeeding with AI yet. Write, guys, if you succeed, if you know how, I'll raise this question, by the way, on my Telegram channel. It's interesting what everyone will share. Is AI really succeeding when it comes to making a design exactly as per the Figma layout, according to the MC, according to some descriptions, I don't know if it's succeeding or not. And, accordingly, it turns out that, yes, frontend can indeed be delegated in theory. Well, first of all, you can generate 20 frontend developers for one person and not delve into the details at all, what's in the code, because most projects are tables, CRUD operations, screens, pop-ups. AI will do this excellently. And the cost of error is very small. Accordingly, so that AI doesn't replace you on the frontend, what do you need to do? You need to, 100%, start teaching frontend programming with AI as early as possible. How beginners now are not afraid, at all not afraid to, well, like young children are not afraid, you know, to mess something up. They mess around, they mess around with something on their phone, anywhere, and they think: "Well, what, it won't explode, everything will be fine." So on the frontend, you need to feel this: "Well, what, it won't explode, everything will be fine, and AI will fix it, and AI will find it." One AI will code, the second will do code review, the first will get the code review, say: "No," refactor it, everything to production. Like, you loop this cycle, and in principle, what? The cost of error is small. Therefore, frontend developers, so that you are not replaced, right? And at least you should be performing, just with the help of AI, just from scratch, without even being afraid that you don't understand something. Naturally, naturally. Gradually, you try to keep up. What did he do? What library did he install? Is it necessary or not? Okay, I'll install it now, I'll make a feature, and then I'll read about this library. Oh, it's taking a long time to install. I'll read about this library now. And you read about this library to be aware of it. What are its pros and cons. You read. Uh-huh. Damn, it's heavy in the cons. Uh-huh. In the cons, it doesn't support some browser. Damn, is that important to me or not? Damn, it's important, cool. Okay, give me an alternative. There's an alternative. Damn, cloud-code is going to do it inefficiently, probably inefficiently or efficiently. Well, cloud-code. I read about it and let's go. Let's go. Research. This is if, for now, just like this, specifically for information. About operators and, guys, yes, an important point. This, when, in principle, well, closer and further and further. Deeper and deeper to this. Ah, although, you know, how? Look, let's do this. Operator and backend DevOps operators and AI are not operators, they are engineers, yes, I would say so. Because to simply entrust AI with work and say: "Do it, update the cluster config, update the DB, give me a new feature." Without understanding, it's dangerous. Operator and frontend developer, operator and, as it were, not possessing engineering skills, but understanding well enough, conditionally speaking, in UX, for example, a conditional designer, a cool, sharp designer with analytical thinking. And here, yes, he will replace you, frontend developers will be replaced by designers who will use AI and generate real components and all this stuff. Make it work for me. Make it work for me. They understand business requirements well, how it should be to satisfy the business. And what should designers do? They need to learn to use it and to take your job, frontend developer. So, how soon do you think AI will replace frontend developers? Guys, look, the question is, what is the topic here, what is it?

Soon will it replace? How soon will frontend developers be replaced? Again, why don't you ask this question? Which, of course, numerically, by itself, will not replace frontend developers. Other frontend developers, you give them this notification, other frontend developers will replace with AI, they will replace you frontend developers without AI. That is, how is it possible, right? Look, generally, just remember this, guys. Surely, you who write like this, you are sitting somewhere on the sidelines and watching, somewhat fearfully. And a giant boulder, just an avalanche on you. And you're like: "So what will happen when this avalanche reaches me?" And on this avalanche, someone is on a snowboard: "Hey, cool," in short. And you also need to ask yourself the question: "Well, I won't be able to stand on a snowboard like that, but I can, in principle, take some children's sled, jump on this avalanche too, and it won't cover me, I won't wait until I die, but I'll ride with it somehow." And the avalanche is this AI, right, the avalanche. And what needs to be done? Like, ask yourself the question: "Wait, how can I, well, replace frontend developers?" That is, you, a frontend developer, must tell yourself: "How can I, with the help of AI, take away the jobs from those frontend developers who cannot do anything with the help of AI?" How can I always, guys, yes, like, not always. Okay, there were times when more developers were needed than there were developers on the market. Therefore, there was no competition between developers as such. A super genius sits at one table, and a super, well, just a simpleton sits, who has learned to paint buttons. And they sit in a company, and both are needed in the company, because, damn it, the work has to be done. Accordingly, there was no competition between developers, in principle. And now there is great competition between developers, because there is little work, little, little work. Therefore, you need to think about how to efficiently, effectively complete tasks. And this, guys, well, we don't need to think about how to do it more effectively, what to do. Master AI, AI, AI, naturally, along with fundamental education. Frontend fundamentally, backend fundamentally, and simultaneously with AI. Guys, I just met with students, we also discussed the Night Show during the day, which was the Night Day Show. Oh, guys, I'm saying this now, even though I'm working tirelessly without days off, I'm so charged, it's so great for me to meet with students, it's so great to code, and when you code a lot, it's great to teach, a lot of ideas on what to teach, what to educate, how to reform the program, how to help, in short, a lot of everything, everything, everything. I'm just like: "Wow, cool, I want more." Therefore, students, uh, take my energy, take my expertise, I will gladly share it with you. Just wait. Now, yes, the code phase has ended. And now there is a phase of many transformations in the incubator. Wait. Ah, Dimych, ah, I'm studying in the incubator. The only place where I wasn't deceived. What they promised, they do. They don't cheat. They reset the program. What's good, I can say. So, only good things. From the heart. No name, no name, it is what it is. We try to be human, not like a business, not client-oriented, just human. Well, we got involved in this thing with guys to teach them, to take responsibility, yes, we take responsibility. I personally, guys, take responsibility, even hyper-responsibility, direct hyper-hyper. They reproach me for this. I take hyper-responsibility for the team, for the students, for their fate, what to do next, how to teach them, all that. Ah, well, yes, this is also my glitch, yes, my psyche, something. Well. But I got involved, I promised, I must, I must, I must see it through. That's why, guys, those who came to the incubator, you must understand, you have one single goal - employment. And, well, I told the guys today, and we will do everything possible for this. By rearranging some blocks, some functions, turning off some functions, adding them. Naturally, as a living organization, as a living, yes, you understand, a living organization, a living organization, where people are alive, and in fact, even software, you understand, software is a living, it's a living organism, software is happening. I actually thought yesterday that perhaps great doctors are engineers. Well. And I didn't say it out loud, that I wanted to tell my son about it, but I was too lazy to talk. I kept it to myself, just thought, yes, like, a doctor is an engineer of a very complex machine, right, the human body, where there are inputs, outputs, all sorts of parameters, uh, volumes, forces, pressures, blood flows, fluids. Oh my God, what's in our electrons, uh, neurons, that is, we are doctors, engineers. That is, and those engineers who deal with hardware are such primitive engineers, who have a mechanism simplified by many, many, many times. That's what I wanted to say? What did I want to say? What was the thought? The thought was that I wasn't deceived, we try to do everything well. And like a living organism? So, the program, yes, it also changes like a living organism, it worked, it stopped working. Here's my hosting. Ah, I turned off the IP address, the IP address from which the bug came, which forced me to move. I turned off the IP address. Well, it writes a letter: "We are changing the IP address, prepare for it." Well, prepare. I went to DNS, changed to a new address, while two supports, well, they activated two simultaneously, I changed it, wrote a letter, everything is fine, everything is done. They: "Okay, cool." Bang, after some time the service stops working. Incoming, incoming messages to the service are going, but outgoing are not. I write them a letter, they, ah, check, the network parameters, blah, blah, blah. And indeed, the gateway is wrong, the routing from inside to outside is not working. Something, ah, while it wasn't working, I was like: "Damn, why isn't it working? Maybe the service crashed? Maybe it's in the code, I started deploying, updating." So look at this thing. Everything was working, everything was fine, right, but the server as an organism, as an infrastructural organism, in which something else is spinning. It malfunctioned, it's like immunity. It malfunctioned, got sick, got sick, the brain started working poorly. Runny nose, runny nose, breathing became difficult, difficult to breathe. And it's the same in programming, everything falls apart. It's like, that is, there's no such thing as it was made statically, it works. There are always factors affecting the operability of our system. This is cool, because this is precisely, perhaps, what, by the way, a thought comes to mind now, that comes with experience. It's hard to simulate, it's hard to teach. M, well, it's understandable, it's possible, in principle, to pull all the cases from your head that you've had in your life. Like these. And, uh, if you formulate them, present them and say: "Well, this can happen, this can happen, this can happen." And the person will feel all these situations and will guess them intuitively, right? But generally, this is professional intuition, when you're like: "Okay, maybe this, maybe that, maybe?" And again, paired with Jarvis, you go and invent like Tony Stark. What's there? Palladium or what did they invent, I don't remember. Palladium, I remember. There was some word. Ah, yes, nothing will replace it. On the contrary, now there are more startups that want to tidy everything up after AI. Ah, well, to tidy up after AI, also with AI, right? What's the idea, guys? I can't imagine a scenario where, like, let's take developers, forbid them to use AI, so that they tidy up what AI has done. In no case, right, because no. With the help of AI, you will tidy up the same order, right, they messed things up quickly, and then it turns out that the product needs to be developed, and it turns out that it's simply impossible to redo everything, to rewrite it, even, guys, on the frontend, because there can be many interdependencies in a complex frontend product. And when you say: "Well, rewrite it, everything is fine, we'll figure it all out again later." Well, you won't figure it all out anymore. That is, tests need to maintain backward compatibility, a new feature, and then the movement becomes complicated. But again, this doesn't mean that it will only be done by hand. In no case, only through, only through AI. I don't believe that there will be another type of coding, like not coding through prompts and pseudocodes, through instructions to the operator. Just recently, on a call, Sergey, our graduate, he is now the head of the frontend department at Kaspersky. And, well, initially, when I conducted interviews, ChatGPT appeared, I already told the guys: "Let's use ChatGPT, use it right during the interview in front of me." I want to see how you solve the problem. But still, there was, yes, like, a situation, an approach that, oh, if you use a chatbot, you're probably a bad programmer. It wasn't welcomed at interviews. And now, if you don't use AI, you, damn it, are not suitable for us, because because. Ah, listen, this is not hate, this is humor. Your courses are good compared to others. What happened to the guys who studied, Dasha and the guys, they disappeared. They disappeared. The results were not shown. Weren't they? Didn't we talk about their employment? I don't believe it. I'll write now. Look at the latest reels. All the guys should have been there. All the guys got jobs. About Dasha, honestly, I don't know, because Dasha left the show. Dasha left the show. And where will I write now? Well, let's write, I'll write, I'll write, I'll write here. Ah, and about our previous guys, we told the whole world that they received offers and all that. Exactly, please. Let's, guys, if you suddenly hear this, please, let's have a call, let's do something, some kind of trick. Exactly, please. Let's. Aha. Quiet. We're looking for everything. Uh, I want to talk to the guys. I, in particular, guys, those who have now subscribed to the Saurraai Live show, count on my support. I will be with you quite often, I will conduct individual sessions with you, well, with your small group, I will identify your weak spots to help you strengthen them. Well. And we will really work with you. And I think, now you've started, although you are at different, yes, already these, next week, we'll try to have a call with you. We'll try to have a call. You'll tell me about your mood, your plans, and I'll also guide you a little and keep an eye on you. Well, you are our exemplary, exemplary gentlemen, we will water you like flowers, take care of you, and, well, we help everyone find jobs. We help everyone, guys. There is no one we don't help. Employment. We even had guys who were, like, uh, a bit toxic, a bit, a bit, a bit, and even then, to get rid of these guys faster, we helped them find jobs faster. Oh, we laughed with the team, yes, about this, that, like, Max, let's get rid of him already and let him work quickly and that's it, so that our eyes don't see him. It's difficult for toxic people to find jobs too. Do you understand what's going on? This toxicity, it manifests at all stages, in all communications. It's visible to experienced people, experienced, team leads, hiring managers. Toxicity is visible, it's read very easily. It's very easy to see how a person talks about others, how he thinks, how he, I don't know. Well, I, for example, read it quite easily, I think. Although, of course, I can always be deceived. Ah, thank you. I envy hus. We will definitely try to find the guys now and share their results. They are already senior, some are senior, not by name, my God. Someone was even hired directly by Sberbank for a senior position. Someone was hired for a senior position, then they weren't hired, and then they were hired again. Something like that was the story. We'll tell you, we'll tell you. This whole AI agent thing demotivates me. Ah, look, yes, I understand this pain. For beginners. I understand it. I myself, yes, was demotivated for a large chunk of time, demotivated, not understanding what would happen next and how our profession would transform. And how would they not pay me money for a set of letters now, and for my experience, because now, like, you have good experience, say, in 10 files you could look and guess what the problem is, but AI will look into 10 files even faster and suggest the problem even faster. Do you understand what's going on? And I was like: "Damn it, oh my God." But nevertheless, like, engineering hasn't gone anywhere from the profession. Creativity, yes, a certain creativity, when we create products that come to life like organisms, are born, and we're like: "Wow, we are creators, we created something that worked." And with the help of AI, this activity continues. That is, we, guys, once wrote commands at the processor level, then allocated memory, then high-level languages, then low-code platforms. That is, engineering remains, because we need to connect parts, know protocols, know weak bottlenecks, and connect all this so that it works. That is, this remains in our profession. It's just that we can do it faster now, right, not a page of functionality per week, not five endpoints per day, per week, not a use case per week, but ten, right? That is, much faster, faster, faster, faster. Ah, wow, so we're faster. That is, I can now. And now, guys, developers are being forced to go, well, they are forced to go into DevOps, because they closed the task quickly, and now they need to adjust all this. Go figure it out. And this is all engineering, this is all about our profession. That is, there is something to learn, there is something to do. The focus just shifts a little. And it's not yet completely clear where this foundation, this platform, is, so that I can stand on it. Everything. And this is further, and here, here it's not yet completely clear. But, Nadya, everything is quite positive, and as beginners, yes, I told my guys in the incubator, what to pay attention to. We do wipe coding on Sundays. Almost every Sunday in the incubator with students, we do wipe coding. We do some projects there, in particular, we've been doing interviews lately. Sometimes we analyze tests, sometimes I show something from myself, just share my wipe coding experience, so to speak, which I've been accumulating over the past six months of working with cloud code, maybe even a little more. Uh, so get inspired by AI, guys. It's like getting demotivated by Google and the internet. With each new generation, there are more and more opportunities for each generation to learn something new, see something new, create something faster. What dinosaurs did in Pascal Delphi for years, then started doing in months, because there was more information, more frameworks, and so on. Practices have already been developed, and now in days, right? Okay, cool. So, imagine how many more cool interfaces there are. VR, AR, uh, displays, spaceships, baby pots, uh, cups, vases, everything will have displays, with displays, with some interfaces. And all this will help us live in a digital world, including, which, like it or not, we all like. Uh, and this is, well, cool. I can't say that it's really cool. If it scares you, then you need to go to the village, right, to live. But since we are consumers of all this, it should be cool for all of us. No other way. Uh, uh, the internship will be updated, they ask. The internship is being updated, guys. From this cohort, we already have eight, instead of four classes, there will be eight classes. The first class is on wipe coding, the second on DevOps, the third on UI K, specifically on Storybook, the fourth starts with Next.js or Next.js, and then on UI Storybook. We are discussing this with the team. Uh, in short, instead of four, there are eight classes. That is, uh, we are constantly updating something, constantly doing something, because our task as a company is to offer a product that is, well, unique. Unique, yes. We are currently struggling with the cost of the product, because we are simply spending more than we earn. And, accordingly, with the economy, this is again, yes, why is it like this? Because from the moment the incubator was founded, I've been constantly thinking about how to ensure that a person is guaranteed to come to us and find employment. I, in fact, having created a business, never thought from the position: "So, how do I set up cash flows in such a way that profit comes in." Yes, the standard, correct approach from a business perspective. I still, after 8 years, don't think about it. Only because now we are in a certain cash gap. Our economy doesn't add up. Only because of this, like, damn, stop, we can't just think about the product anymore. We need to, like, make sure that we have the resources to survive and move forward. That's why I have to bother about, ah, and about the price, and about the flow of people, and all that. Otherwise, like, further, further, further into the product, product, product, products, integrations. We have a lot of interesting, tasty things there. So come to the incubator, and we will lead you to victory, to employment, like a friend. Okay, now I'll try to comprehend what Victoria said. Now I'm finishing my internship in the incubator, the title is: No words, so much effort has been put in. No words, so much effort. Punctuate, guys. Punctuate. Execute, do not pardon. Here too, it can be interpreted in different ways. That's the trick, that a programming language is good, it's interpreted clearly and unambiguously. But human speech, here it is. Oh, however you want. I can interpret it. But no, look, Victoria, what meaning? Uh, the effort was not in vain. It is precisely those who put in effort, you will later be those very AI operator engineers and create software products. Those who don't put in effort, well, they won't be needed in the market. That is, everything, yes, there is only one option now. Put in effort. By putting in effort, you then find the right, effective application, so that the business says: "Wow, cool, we use you well, we exploit you well." You, in turn, as a proletarian, should think about what? So, I am being exploited, but I want to remain in demand in the market, so that the exploiter wants me further and pays for it. I will think about how to increase my expertise further and how to be more effective again. That's all. So, Vika, everything is just wonderful for everyone. Continue. You invested in yourself. That won't go anywhere. Now you need to, the next, yes, moment. Come tomorrow. Tomorrow, rather, follow, I will write in this. Tomorrow, due to my rearrangements, most likely, we will meet for wipe coding on Monday. Oh, yes, now I'll put it in my calendar. One second, please, because if I don't put it in, I might miss it. And tomorrow we have a meeting with tech. I will conduct it in the evening, I will also postpone it a little later. Now I'll do it. Ah, so, guys, sorry, sorry, sorry. For now, please give some likes, while I will be very grateful to you. I will postpone the tech consultation by an hour, and the test wipe coding analysis will be postponed to Monday. To Monday, to, ah, let's not make it too late. Oh, damn, I don't like Google Calendar because it's inconvenient to move like this. Come on, come on. Oops. And I will move this. Oh, thanks for the likes, guys. I see them flying. I will move it to. Look at the calendar, guys. I might make a mistake with the time zone now. My calendars sometimes go crazy, but I think I've moved it to 6 PM. 6 PM Minsk time. So, for wipe coding. Vika, that includes you, come. We will now, guys, we already, yes, like, this wipe coding is an introduction. Look, especially our first classes in the recording, where I smoothly introduce you to this whole topic. And now Valera is preparing content from his perspective, how he sees working with AI, how he does it. And we will now further accustom you. This is a large part of the innovations that we will be making in the incubator, in the career booster. This is learning to integrate you, you, young engineers, with AI, so that you are productive, needed, productive in the market. Now finishing. Why was the girl's comment deleted just now? Mikhail writes. I don't know, guys. The message here is deleted by the sender. Who deleted the girl's comment? Which one? Write again. Uh, is there a moderator acting now? No, it says the message was deleted by the sender. If it said it was deleted by some moderator, if one of my guys is moderating now, but I haven't agreed with anyone. Ah, write, write, write everything that's on your mind. Write. I will talk about everything, including the inner workings of the incubator. Do you think everything is ideal there? Of course not. And this is generally normal. This is generally normal. It's like you're not ideal, I'm not ideal, nothing is ideal. The question is, what is our vector, right? What do we want, where are we moving? That's a good thesis. Fullstack is more of a fairy tale on large projects. More of a fairy tale on large projects. In short, uh, I don't know, yes, guys. M, I'm now, ah, before AI, I had doubts that there couldn't be many fullstacks. Projects are indeed large, complex, on the contrary, there can't be a fullstack, because there's complex frontend, complex backend, a lot of everything. And there you need to close with separate specialists. In small startups, a fullstack can exist, because you do this, you do that, you quickly create an API, quickly generate code, a Data Access Layer, quickly make a form on the frontend. Now, thanks to AI, it seems that before I saw that fullstacks were profitable, effective in startups, but not effective in large companies, in enterprise. Well, unless it's some very support mode, just support mode. Ah, and now, it seems, AI allows fullstacks. Well, for example, me, yes, in frontend, ah, well, I haven't coded in frontend for a long time. Very long time. My last 3 years were JavaScript backends, .NET backends, and just like frontend consulting. I consulted, consulted. And now I'm coding frontend directly. And I don't understand much. AI does everything. That is, I set the boundaries, what I need. That is, where I want the logic concentration to be, what decomposition, everything. And then I just look approximately, well, yes, approximately okay. Is it right or not, like, that is, the idea is that before, to code effectively, to type characters, I needed to be good at it, to do it quickly. I needed to know the API of libraries well, the API of some tool, even the API of React, and now good prompts and general engineering understanding are enough. And in principle, and in principle, it's okay. So, I, for now, I'm thinking along these lines and I understand that fullstackness is gaining a bit more meaning in the era of AI. Like, engineering itself, right, as in large companies, like Google, they never hired, like, frontend developers and backend developers, they hired software engineers. And this software engineering, it seems, today, thanks to AI, erases all these differences, erases them. Well, in theory, like, well, look, on this channel, on this channel, I literally already have the technical specification, I will start studying this technical specification tomorrow, on Monday. I plan to write a Swift application by myself, specifically with the goal of teaching my iOS developer guys wipe coding, to write an application by myself, right up to its publication in the App Store. I will record the first video, upload it here to YouTube, upload the last video. Most likely, all videos on this topic will, most likely, be available on Patreon by subscription. Well, on our platform. Not on Patreon, on our platform. Well, what am I getting at? I'm getting at the fact that I'm not a mobile developer. I've only seen Swift from the side, how guys write something. That's it. And my task is to write an application and get it to production. And I understand that I, being a software engineer, will solve this task quite quickly. Well, because the application should not be complicated, because I will clearly understand what problems I need to solve. And I will have a rough idea of what tasks to set for AI. Well, 95% of everything will be written by AI.

I don't know what I'll do with my hands there. Well, I don't know how layout is done in Swift. Something feels to me like it's done easier there than in the browser again with a million different approaches. Therefore, therefore, that's why there are many messages, guys. In short, let me tell you a little about the backend, tell you something about the backend, and I'll get back to your messages. In particular, look, those who won't go to our incubator, just look at this as "Wow, they came up with something cool in the incubator to teach competitors, you can look, take our concept and refine your programs." Learn from the best, as they say. Here. And what will we do? What's the point of our training at all? Let me explain now. I'll draw. We will have, I think, six sprints. Six sprints, that is, 6 months. The first sprint, oh, by the way, we are making an innovation. Our sprint will now be not 4 months, but, oh, not 4 weeks, but 5 weeks. A week for the fact that since we are strengthening the depth of the program more, more, more of everything on the frontend and backend, including, in particular, we are forcing students to go through interviews, yes, this whole interview thing. So, we literally force students from the very beginning to learn how to pass interviews, plus consolidate the materials needed for interviews, in short, and, accordingly, it will be harder to study. Accordingly, we are adding another week so that you can calmly breathe out, catch up on loose ends, and all that. So, 4 weeks, yes, in general. 5 weeks we have now. Well, probably from this intake it will be 5 weeks. I think we agreed on this with the team. I can be a bit vague on these organizational matters. So. Third sprint. Fourth sprint. I'll tell you now. In short, before this, our program was whole, but we did something, redid it, and all that, in short. So, conceptually, we dug deep, but the same thing, yes, we poured from empty to full, from full to full, from meaning to meaning, but functionally we did the same thing. And now we want to do it differently. Let me make a white background. I like it when there's a background. What do we want to do now? Where do we start? So, we will immediately focus on the key point that we will train our backend developers immediately. Do you see what is written there? No, because it's too small, and I write large. Microservices. We will immediately focus on getting guys used to microservice thinking. Microservice is a standard. You can see a lot of holy wars. Microservices are an excessive monolith. First, we start here and there. Well, even, you know, in large organizations, I can't imagine any other approach, not one that repeats the company's structure, not this one that isn't split into microservices, what can I say, even our small company, and we make small products, well, we make very small products, and even then I can't imagine how we could do all this not in a microservice architecture. How to connect all this together, how to deploy all this, update it, how to stabilize it, I can't even imagine. Well, I can imagine, but I understand that it would be a big disadvantage. Therefore, in short, microservices will be with us from the start. What will we do? We will make Samurai Taxi. That is, we will make a kind of analog, well, like Uber, so this is the kind of project we will have. Let me write it down, Samurai Taxi. Like this. I made it with a white background, but I didn't make everything with a white background. Well, let's do it this way. We will do this in the first month. So, the first 2 months, two sprints, they are still, well, like, we're warming up, warming up, backend developers came, we need to warm them up. There will be Express there, there will be this Express base, no matter what they say about it. Express is, well, you can say, a small, nasty, unpleasant monopolist in the market, and you can't get away from it. Introduction to Swagger. Swagger, well, basic basic, in short, concepts like that. I won't go into much detail now. Well, like, in Swagger, end-to-end tests, first of all, well, tests, I'll write unit tests. What else did we cook up? Well, errors, tests. Authorization, we have base, authorization base, basic authorization for the admin panel, Mongo. What else do we have? JSON API, this is a new development that we debated for a long time whether to do it or not. So. JSON API is a protocol, a specification, how to format responses, requests, and objects. In principle, Music Fun. A Music Fun thing that's on Samurai React without alternatives. It's also similar to JSON API. I will make it as close to JSON API as possible. A slightly overcomplicated format on the one hand, but on the other hand, it's definitely suitable for learning, definitely cool, because it structures, at least it will give some understanding that has already been packaged into a specification. Well, in short, I recommend looking at this thing. Next. Okay. So, yes, we didn't just make one microservice. So, here we are working, it's called, it's called the WRS trip microservice, I think. WRS ms. I'll write ms microservice abbreviated. Is it interesting, can I add padding here somehow? Padding. Padding, okay, probably not. Next, in the second sprint, we continue to work with this same microservice. For now, consider it a small analytics. Because further we are refining its concepts a little. Here we made something, did something. Then we add a little. A little bit we add. Now I will make sure not to deceive you. We study Secure Light Application Result. Yes, we introduce you to the concepts. We introduce you to the concepts. DDD CQRS. All lightly, yes? This is an introductory thing lightly. DI, hmm, DI, well, classes, yes, so let's say OOP, OOP, so we rethink this thing from, roughly speaking, a JavaScript format on the fly into a kind of enterprise Java C#ARP style, which, in my opinion, is the most correct way to learn backend programming today, because it will be much, much easier to jump into another paradigm of writing code from this. And there we also add, I'll delete this now. Besides this, we add Mongoose Mongoose ODM, yes, specifically from the perspective of how it can help us think in OOP, think in DDD. There is a machine there, a small one appears inside. Well, we make a special state. Finite state machine. I'll write the abbreviation. Well, these are like, so this is like a conceptual, let's say, rethinking for a month. And here, a little rewriting of what is there, a little in a new format. So, so, so. Yes, it seems so for now, just in case. And here is a more detailed thing. Mongoose. Boom-boom-boom-boom-boom. All. Yes. What happens next? Next, a new microservice is born. So, this is a white microservice. Already here, we are making a new microservice. So, we are making an authorization service. So, before that, guys wrote everything together, rewrote, did, rewrote. Now here we have Express. Let me write it like this. Here we have Express. I'll remove it from here. Well, here, of course, also Express. Well, TypeScript, of course. I don't even consider anything else. Here we already have NestJS. We had NestJS, I think, here before. Hmm, now it appears here, NestJS. And we are making an authorization service. Well, as it should be. SMTP, JWT refresh, refresh access tokens. And if I'm not mistaken, at this stage, we don't touch OAuth yet. I remember we discussed this point, like, why not do OAuth. And by the way, oh, by the way, I'll think now, let me write, because now we have homework, because it would be very easy for you to validate JWT for the most part. Listen, I would like to shove OAuth in here. I'll ask Vanya and write immediately. And immediately shove OAuth2 in here. And now I'll quickly write to Vanya. Bro, bro. Oh, OAuth 2. We don't want to do it as homework in the third sprint. I wrote this because I want to. Well, the idea is that we already have inter-process communication, IPC communication, between these nodes. So, let me remove this arrow, because it's the same microservice, but here they need to communicate with each other, because I have, yes, my bounded context manages trips, drivers, and authorization is somewhere else. How to do communication? Roughly speaking, yes, at first, we introduce guys to the fact that we have two independent processes running on two machines, in two independent namespaces, in short. And the first, yes, the first option is also synchronous communication, REST, well, through HTTP, HTTP, for example. So, we analyze why there are problems, instability, why it's not very convenient, blah-blah-blah-blah-blah-blah-blah. And we move on. In the next sprint, we will develop a large file microservice. So, this is already And attention, here Nest + Mongoose. Mongoose not plus Mongoose. Next, we will develop a file microservice. A file microservice that is Nest already plus, attention, SQL, raw SQL without anything, without TypeORM, raw SQL, so that you can feel it thoroughly. The idea is, what is the point of this whole approach, that at each step, working with this microservice, you will be forced to fix something here. You will already move to Nest with SQL, and you will need to connect these two microservices. Why connect them? Well, because authorization uploaded some document there, it needs to be displayed here, here, in short, the driver uploaded his driver's licenses, they are already authorization-related, you need to communicate between yourselves. Accordingly, you will constantly have to jump, jump, jump and constantly keep in mind, remember Express with custom-written, yes, some things. Here Nest with Mongoose, Nest with raw SQL. What do we have for files? Well, the main key thing is working with streams. Working with S3. Working with S3. Uh-huh, free already, yes, here Outbox pattern is introduced, Outbox pattern, cron pattern, speaking of cron, guys, immediate sale. Every month, the program gets more expensive by $100, so keep that in mind. But now, let's also be upfront, so that it doesn't sound like a cunning sale. Like this, yes. Cron. What else appears? File processing, we also want to shove it in. I wanted to say that we will add, well, this, no, well, I would actually like to add some FFMpeg for processing. So, like, it's not in the plans yet. I'll just cross it out for now, I'll discuss it with Vanya. So, not just uploading photos and documents, but also a bit more complicated. All. All. Yes. And here's the topic. Wow! We need to communicate somehow, communicate somehow, communicate somehow. And if I'm not mistaken, uh-huh, I see Kafka here already. So, at this stage, we are adding Kafka plus Kafka. So, yes, this is 1, 2, 3, 4 months of backend training. And you will come out after this, go to the market and say, "Damn it, look, patterns, S3, Kafka, SMTP, authorization, OAuth, DDD, CQRS, OOP, you're just going boom." Naturally, you've grabbed all this, grabbed it, grabbed it. Oh my god, I've already grabbed everything and forgotten everything, I don't remember anything. Nothing, like, we're going to consolidate, we're going to continue consolidating. Different color, yes? This is another microservice. I'll do it like this. And we go to the next microservice, which is called Payment. Payment microservice. Well, about payments. Boom-boom-boom-boom-boom. P. Oh, so. Command Payments Microservice NestJS. And instead of raw SQL, instead of raw, oh my god, SQL, it will be TypeORM. RightORM. Well, naturally, I'm not indicating Kafka anymore, because, well, Kafka is already done. Here we got acquainted with it and then move on to asynchronous communication. So, let me write here. Here we have synchronous communication via REST, yes, don't be alarmed, like, requests are asynchronous, but the communication is synchronous. There's such a trick. And here is asynchronous communication. Via Kafka, and at the same time, it is allowed that most likely, well, and here you will refactor, yes, this communication via Kafka, because the more asynchronous communication, the better. The better from the point of view of gaining experience, the better, because there will be more hassle, more problems that need to be solved. And, well, at the same time, asynchronous communication solves a lot of problems, but from the point of view of coding and thinking, it also creates a significant load on the brain. But solving all these problems is cool, because, because, well, growth, growth, growth, engineering growth. Here. And then, then, then, then. Fifth sprint. Payment. Well, there with Stripe, etc., well, let's say plus Stripe, plus payment. Well, there also, the deepening into webhooks. Well, Webhook, yes, the concept appears, well, Webhook appears. Well, webhook, validation, so, and WebSocket, we connect it here. WebSocket appears, Webhook. Oh, and Redis. And Redis is still in our program. I'm ashamed, we have Redis, well, such simple things. Sometimes I have such a problem, I think it's more my problem. I'm like: "Well, these are already simple things, they don't need to be explained, guys will look at it themselves, do it." Although, like, we have some gaps like this. Now we will close them. And here we have Redis, and this is like we are dealing with payments, integration with an external Stripe system, Webhook. And, naturally, these payments are for what? For these guys. So, payment, a trip was paid for, then statistics, oh, no, we don't do statistics. Which microservice is this already? This is already the 1, 2, 3, fourth microservice. This is the fifth sprint. Well, and the sixth sprint. Sixth sprint. I'm just copying now. I'm just copying the topics that are here. In the sixth sprint, we will not introduce any new microservice, but we will simply have a set of topics that we will cover. So, now I'll find something. No, this is something small. So. Just paste the text. Well, it's like this. So, concurrency, yes, we start studying. Concurrency, optimistic locking, pessimistic locking, ACID transactions in SQL, we study. In short, in short, the idea is that these topics, indexes, more advanced statistics, advanced locking, in short, in short, statistics, in short, more in-depth work with databases, with transactions. And the thing is that, like, all these things, from the point of view of, from the point of view of, um, from the point of view of this, um, depth. So, we return to all these cases. Here, on raw SQL, we work with locking, with indexes, here on TypeORM, here on Mongoose, here on Mongoose. Well, there Mongoose, there and there, okay. And, in short, it turns out that we go back to all these things and look. Asynchronous communication, damn, concurrency, race, and here and there, here and there. And immediately, therefore, new learning repeats the class, that this is a cool question, well, like, a cool thought. New learning repeats the internship, as a result, there is no room for thinking. Guys, internship, yes, it will adapt to this. So, the idea is that in the internship, you will have to do this. Here we guide you by the hand, we put pressure on you with homework, we help you, and in the internship, your task will be to work through all this. So, the idea is that we will give you more instructions here on how to do it, so that you form a pattern of beautiful code, correct code, correct architecture, correct checks. And in the internship, you will go and be like: "Okay, I will now hone all this." Not only hone it, guys, you need to understand that you will have to hone it in the internship. With what? With, damn it, pair programming. And in the internship in general, if we talk about it, this is like, well, not that it will happen tomorrow, but actually quite soon, in the internship, we will do more integrations. In short, we will strengthen the internship project by doing integrations, internships with what? With AI APIs. We will build vector databases, connect vector databases. We will build N8N, these orchestrators of processes, set up the framework. We will work with building search systems. So, in the internship, we will modernize, let's say, assimilate the approaches that, like, AI has thrown into the industry, and that forces us all to understand this. And, accordingly, this is our plan, in short. Oh, cool. Honestly, when you do what inspires you, of course, it's awesome. You can throw in thoughts on this topic, when the new backend program is launched. The question is, in theory, everything is ready for, well, not everything is ready. We have the codebase written for all sprints to launch this program. The codebase, so quickly with the help of AI, we will prepare manuals. So, in principle, we can adapt. We have started writing tests, auto-checks, tests. Students know that our system checks a lot automatically. Guys do homework according to our Swagger specification, and automation checks. And we have started writing exams, homework, for this purpose. Possibly, so that it can be. In short, we are now thinking about how to seamlessly integrate this, guys, into the new program. So, you are already coming, you are studying with us now, we will start little by little. Most likely, we will introduce this new program as an addition, a little on the side. A little on the side. We need to reshuffle our videos for all this, understand what information is missing, what information is of disgusting, terrible quality. It needs to be rewritten, redone for this thing. In short, we have a big job ahead of us for a year. But, like, we will release all this, possibly even with this intake, with this intake. We will start not every 4 weeks. Those who are coming to us, in short, sign up, because the price will increase. If you were thinking about backend, sign up. If you are suddenly unsure if it's for you at all and programming is your thing, then don't rush, yes, like, I would never. The most disgusting thing is when you go into programming on some vibe, like, well, you can make big money there easily and simply, like, you can make good money quite easily and simply compared to other specialties, but you need a certain mindset, perseverance, interest in the profession, all that. Tests will be, guys, yes, tests will be. Oh, this, by the way, about tests. Well, for backend, we teach tests from the very beginning. On frontend, we introduce tests a little. And what I remembered. Sergey personally gave the task to work on this lesson. We have tests in the internship. For frontend, what I said, there were four lessons, now there are eight. One of them will be on testing. This is also an important topic, but we have always skirted around it for various reasons. So, guys, on TypeScript NestJS. Will there be backend in other languages, platforms? I promised, I intended to. And in fact, we are moving towards it very slowly, but we are moving. We are moving in this direction, and there will be directions in other languages. In the incubator too, when, I can't say yet, but there will be. We are moving towards this, including updating the program, yes, we have gained an understanding of what the program should be like, so that, well, so that it's just wow, just several heads above what is generally expected from a candidate. Several heads above. So, it's about training a junior, a beginner, who gets their first job, and it's like, damn, this is a whole backend university. And the task is like this, to be several steps ahead. In principle, we are always in the incubator. Because of this, we manage to employ guys, because we have always been several heads above the requirements. Now, naturally, the requirements are insane. I can't say that we are several heads above. We are at the level of market requirements, so everything works out. So, everything works out. This question is just wonderful. Should I study backend immediately after frontend, or is it better to work as a frontend developer first? Before, I always said 100%, and now I advocate for the priority that you first work as a frontend developer, understand that, okay, I've consolidated everything, I understand where and how it's applied. Then I'm like, what else can I grow as an expert? Aha, I'll grow into backend. So, before, there was such advice, and students who came to us without getting a job to study backend, I was like, well, like, well, okay, well, like, a bit in vain. And then we look at what indicator I highlight for myself, that it's very important, how urgently I need to get into the profession, because, naturally, look, the deeper you work on the foundation, the less of an imposter you will be when you come to the market. You will stand firmly on your feet, because you have a support in the form of a comprehensive understanding of how everything works. The question of your age, time, desire to get a job. If you have such impostor syndrome in your head and you're like, "Well, maybe I'm just a frontend developer, a form-shaper, not needed in the market." And this is your insecurity, just your fears, then to hell with backend, go get employed, yes? This is like postponing the main action for which you are studying. To become a specialist who is paid a salary, then to hell with this backend, let's go get employed. If it's a clear plan, a clear strategy, I'm like, I'll crunch all this like seeds. I understand that I will grab this, grab this, grab this, and then I'll go to the galley and row. And then, will I have the opportunity to develop, find time, or will I be putting out fires, as often happens at work? And this is part of the strategy. Then it's different. And you need to look at this. If you have some fears, that maybe, well, you know, maybe you need to be, maybe a more complete full-stack engineer, maybe. If you have such a thing, maybe, maybe, then no, not maybe, it's not possible. You've fixed one specialization, you move on to the second. Because, even, what are we talking about, guys? Even I, a frontend developer, I'm an Angular, React, or Vue developer, or am I everything at once? It's impossible, yes, in the nuances, everything at once, it's impossible. Again, with the help of AI, of course, the impossible is possible today, you know, fundamentally both, but in any case, the concepts are also peculiar, and you can only realize them from the height of a senior flight, so to speak. You can quickly jump into another framework, understand conceptually what it is based on, and you can manage a team through AI and configure a stable system for development. Only this way, but not from scratch, like, I'll study three frameworks now and be everywhere and nowhere, yes, this is called everywhere and nowhere. And you can't fit everything into your head at once, like, you'll fit frontend, you'll fit it, and then you'll go to backend. And frontend, it's, and in backend, oh, there's so much. Like, frontend will start to blur, and then you'll get lost. And what's good there, and what's better here, and what is the truth, in short. A good question about different programming languages and the meaning in different languages. Languages are different. First, it's competition, it's marketing, naturally, it's influence on the industry. Second, it's natural, well, I don't know what to prioritize. It's, of course, some different tasks are done with some languages.

more effectively. Well, generally, yes, like when you are again a software engineer, the differences between languages are very small, there are very small differences, of course, there is, for example, strict functional programming. This is my gap, which I professionally cannot close yet. Well, I don't see the point, yes, in my professional career, I don't understand, for example, why I need it, except to show off on YouTube, and in principle, this is also part of my work, and I need to do it, and therefore I will delve into functional programming. Not like in React, where there are pure functions, containers, immutability, but at the level of all these monadic, functorial, mathematical concepts. So, I just don't understand why I need it. I say: "Well, I'll go there," yes, just because, well, I need to close gaps. I like it when a gap is closed. You live with this pain for a long time, although you could go and spend just, I don't know, a month and study this thing. By the way, possibly, I will do this study in a YouTube format and invite guys who are knowledgeable in this. They will tell me something, I will ask questions, and I will understand, like, why functional, purely functional programming might be needed. Uh-huh. But one way or another, like, programming languages will not be. Well, of course, we, uh, will have to type fewer and fewer letters ourselves and remember that this language has this syntax, and that one has that. AI generated it, the compiler compiled it, so everything is fine. Tests passed, so everything is fine. And what syntax it is, what indentation it is, linters will do everything for us. Of course, of course. So. So, guys, if anything, please duplicate. I, it turns out, jumped down. Now, if I reach the end, I might scroll up. Although we will gradually start to wrap up. Uh, I haven't found a job as a front-end developer in 2 months and mastered back-end in Go. To be honest, I don't even want to return to front-end after this. It's kind of stuffy there, and it's much more interesting here, but harder, of course. >> How did you master back-end in Go and get a job? I'm entering the market in a week. Yes, I just didn't want to waste time waiting for interviews and decided to learn. It took 3 months. Ah, well, wait, you haven't left for work yet, right? Well, when you find a job, good luck, like. Well, first of all, yes, like you still, look, guys, MirLax, like, I'm generally so, honestly, so many people, even smart students we have, I'm amazed. I understand that guys will simply, damn, simply become cosmic engineers. My brains are somewhere over there, and they are given this boom-boom info. They quickly chew it up, understand it, comprehend it. I comprehended concepts longer in my time. I still sometimes think: "Damn, I can't fall asleep because I can't understand why. Why? Although, of course, more often I understand that there is no answer why. And it's just because it's done this way and that's it. It's just because you can do it this way, and you can do it that way. But, guys, there are tons and tons of smart people, so the world is the same story, I think. Front-end then quickly went into Go for 3 months. Naturally, like, this is the story of 3% of people, out of those 100 people who are trying to get into the profession, who generally like it. 3%, 5% maximum. These are the ones who 1 2 3 5 boom-boom-boom-boom. And the majority are ordinary mortals who, like, need to, phew, strain themselves. Ah, yes, guys, continuation of React by Samurai without alternatives will be. Ah, it will be. That is, the presentation for this new lesson is ready. The videos will be a little shorter, but they will be, in short, I was already going to start releasing them this week because I dragged it out, dragged it out, dragged it out. I'm not going to drag it out, right? Ah, January is already ending. Well, I was going to release it at the end of the year, but, you see, I went into Prosobesim, into the platform, mm, into development, in particular, to play with React, to touch it, ah, to touch it directly, although through AI, yes, it can be called directly. And from this week I will, guys, try with all my might, the main thing is, you know, to take the first step, ah, to release content. I will release it, I will release it. I will try very hard, guys, I will try very hard, but possibly, I will allow myself not to say, like, I released it, so now at least twice a week. Okay, at least once a week, once a week. Oh, now twice a week. Once a week, twice a week, once a week. Twice, twice a week and so on. You need to get going. Mm, because I work an infinite number of hours. Simply time, yes, now in a day I don't even dance bachata anymore. I want to start dancing bachata. So, I have a return to the group training in February, so I say: desk, computer, sleep, desk, computer, sleep, sleep, desk, computer, sleep. So. Yes, there is such a story, guys, that experienced people are leaving the market, and they can't get in. Therefore, guys, to get in, to successfully pass the interview, go to Prosobesim and pump, no matter how scary the interview skill is. A service that I created personally for myself and for my team. We plan this year, I personally plan to find a Western American client with the English language. Now I will also have a question to close. And I made a service for myself and for my team, because, like, we developed internal products, like startups, this and that, and now we don't have money to develop, to develop for the sake of development. We need to earn money to continue to exist. Therefore, what will we do? We will work by hiring, yes, outsource, outstaff, or go our separate ways, or as an incubator team, depending on how it turns out. But in any case, we need the skill of how developers sell themselves. And we will download it through Prosobesim. I also put my team there. I felt awkward, the guys felt awkward, because when you haven't passed interviews for a long time, it's a struggle. No matter how great an expert you are, passing interviews is a real stress. And the task of Prosobesim is to make this stress go away, to make the fear completely devalued. The fear of interviews, yes, my God, it's just a normal thing. So, about your back-end OOP program? Well, you can say that it's OOP, because we have classes, we have inheritance, we have use cases, we have methods, ah, we have DDD. And this is all a bit about OOP. I can't say that it's strictly OOP paradigm programming, because, yes, any today is cross-paradigm, yes, programming, so a little OOP, a little functional, a little purity, a little mutability, and everyone is scrambling about what is better, where and how. And so, ah, we touch upon these concepts, because it's about classes, because it's about clean architecture, yes, you can't, for example, talk about architecture, about Dependency Injection, about patterns. like, outside, well, outside the context of OOP, because there is also encapsulation, and polymorphism, all this needs to be understood. And, by the way, we will, especially in the booster now, probably also one of the ways to strengthen the booster for back-end developers is to delve deeper into patterns, even separately consider enterprise patterns, perhaps even in theory. Ah, perhaps even we will insert some interesting, non-trivial, non-pop, complex patterns into each microservice, at least at the level. We showed how you did it, and you will comprehend it later, but you will have something to rely on, somewhere to look, to set the right vector. Yes, Timur seems to know functional programming, judging by his videos and reels, he writes some code, explains something. Yes, possibly we will talk to Timur on this topic. I think he will be interested. In this regard, he, well, again, not in production, I don't know, like, maybe we'll see, he'll get someone out in production. Well, I will also invite Timur as a guest. Big respect and admiration to Timur. Did you not think that AI is just another marketing campaign, nothing more? Its mass adoption is due to the fact that he, Vidi, wants to recoup his money, and that's all, who invested. Ah, well, Kirill, look, you could think about this if I, as a programmer, did not realize, did not accelerate my work tenfold with these tools. So, that it's just a marketing campaign, of course not. That there is manipulation, PR, investments, recouping money, various cunning speculations, when they buy shares from each other, transfer money, so there is definitely smoke around, but not fire. There is fire, and I, as a specialist, as an engineer, understand how much this tool will push the entire engineering industry forward. Therefore, mm, the only thing, like, I have thoughts that, possibly, all these AIs will be more expensive. Now, literally hundreds of billions of dollars a year are burned so that we, conditionally, well, in principle, use all these tools for free. That is, I speed up work tenfold, and I pay for the tools $100. Well, $150, $100-150. So, roughly speaking, conditionally, let's say, my senior salary, for example, is $5,000. And I work tenfold faster. Like $50,000, and I pay only $150 for it. Like, damn it, man. What's the question? Of course, of course, of course, like. But again, it's understandable that tenfold. I don't know what's happening here simultaneously, like everyone understands this, more is demanded from developers. Therefore, in fact, this is the topic, that we are just the proletariat. And, well, well, the tool is fantastic, of course, here. Ah, and here, just start doing it, guys. Like. Yes, Kirill, I admit that, ah, such a question is related to the fact that you haven't started, look, super, leave the stream with this important, main advice of your whole life. If for you programming with AI is part of GPT in your window somewhere, deep, whatever, it means that you don't even imagine what programming with AI is. AI should be immediately in your project. cursor, code, Gemini, Codex, whatever. Take from this list and try. It's a struggle, it's unclear, start, and in a month or two or five you'll be amazed, and you'll understand that life is divided into before and after. Analogies with construction. Current houses are built at the speed of light. Yes, only the quality is so terrible that the demand for ordinary builders has only increased. Who can fix it after these shoddy builders? Boom-boom-burum-boom. Well, maybe, like, maybe. Well, of course, look, what's the thing, like, ah, the tool is available to everyone, both experienced and inexperienced. And it's the same with construction, like, yes. Well, again, look, what's the thing. Well, there is demand for houses. There is demand for houses. There are many of them. Real estate is also some kind of bubble. I'm looking at Batumi, like, I also don't understand why it works like that, but I just don't understand. Uh, people buy, build, like, and why build well, like, there is no demand for it. There is demand to build and sell, yes? Well, that's how the market works. Build, sell, build, sell. The same is true in programming. Feature and sell, feature and sell, feature and sell. So there must be a compromise, cost, price, quality. That's why, like, you need to approach this maturely. We don't need everything to be perfect, high-quality. That's not the point. Ah, well, that's how the market works now. Ah, oh, duplicating. Dimych, thank you for duplicating. For a front-end developer, the minimum task is now either to go into back-end or into UX design. The minimum task is to accelerate front-end development tenfold with the help of AI. Yes, that's the minimum task, I would say. And then, having accelerated tenfold, go more boldly and okay, I shouldn't be afraid to change something in a Docker container. I'll figure out this Nginx or something. Aha, I need to configure gzip there. Okay, I'll figure it out. Aha, I need to change the endpoint. Before, like, how can I change the endpoint? I'm not a back-end developer, and now ChatGPT, Cursor, Gemini. Where is the endpoint in the back-end that I'm sending this to? Where is it in the back-end? Oh, great. And where do I add something here so that something is saved differently? Oh, cool. Well, wait, let me study it. Explain it to me. Done. Something like that, yes. To fully switch to back-end, well, that's also an engineering endeavor, a meaningful one, it makes sense to study back-end, to switch, not to switch to back-end, but to study back-end, like, to know back-end. But if you don't like back-end, and you like front-end more, you really like front-end, then engineering in front-end also exists. There are also complex things there, there are complex front-ends that cannot be done with AI, not at all. Well, take, for example, what I showed you, where I was drawing. AI won't do it. It won't do it, because there are no code examples. It won't do it effectively, maintainably, scalably. It will do it somehow. There are complex front-ends, there are complex projects, even forms, buttons are complex in front-end, when there are very many of them, when it's a complex composite application, like, so you can do something in front-end too. Are there discounts for former students? Yes, guys, if a front-end developer, a graduate, comes to us for back-end, there's something with a discount, if Oh, let me tell you about the discount. Look, I mentioned the price increase, but I wanted to mention it and forgot. Look, what I'm going to tell you now. You'll be amazed, like. Everything for you, so that it's just a pleasure for you, so that you just fly into space. Look, the next intake for back-end. The current intake, I think, has a higher price. The next intake will have an even higher price. But for the next intake, personally from Odymych, tell the sales manager, you will be promoted to the JUNIOR level here. That is, now you are entering the platform, you are triple, guys. The platform is free. All you need to do is register now and start passing interviews. There is front-end, there is. If you want to come to front-end with us with a $100 discount, well, we are increasing the price by $100, and you can, roughly speaking, annul this increase by proving to us that you are, like, with a fighting spirit, with full this, with readiness for self-sacrifice and self-discipline. And you will raise your level of training to Junior 1. And then you come to the manager and say: "Here is my account." Share your screen, show, here is my account, here I am. I am on Junior 1. Give me a $100 discount. And you will get a $100 discount. This is the offer for the next intake. And to compensate for the increase a little. That is, there is no task to make the increase infinite and just, like, the task is some kind of selection, when a set of some kind. Uh, I really don't want, guys, to drag in those who, well, who might be better to dissuade for later. This is a very difficult topic on the one hand, like, dude, well, damn, it's very difficult for me, ah, ah, dude really wants to. And we, in fact, guys also gave me examples of some of our students who came to us at the zero level and from scratch, yes, from scratch. I was against teaching from scratch in the incubator. A terrible opponent, like, well, how to teach from scratch? I, well, all the time the thesis is, a person should learn from scratch themselves. This is a criterion of professional suitability, that if you can't learn from scratch yourself, you can't learn from scratch yourself. Well, then, well, like, at the same time, of course, I don't want to learn from scratch myself, I want to learn with a master, with a team. I need motivation, I need a roadmap, I need support, I need a company, I need professionals nearby. I want to understand that I am now self-learning in the right direction, so I need some confidence. But nevertheless, I was an opponent, but our guys are moving to the main course and got jobs recently. So, we have a relatively new direction, from scratch. And guys got jobs. The team gives me examples: "Look, this one, this one, this one, like, they came to learn from scratch." Well, and I again started to grumble, I said: "Well, damn, we need to take fewer guys, those who are boom-boom-boom-boom-boom. Zeroes, like, no, it's hard to work with them." And they start giving me a counterargument: "Look, boom-boom-boom, everything is great." I say: "Damn, well, uh, well, yes, like that." And, in short, and therefore there are people who are just obsessed, disciplined, adult, mature. And here the question of emotions should not be, that some Leo. Well, I hope there are no emotions left for anyone, because, in principle, already, well, like, it's generally cringe. It used to be cringe, like, when they shouted that it was easy, easy, easier than simple, uh, and now, like, well, you'll just be pelted with tomatoes if you even hint that getting into IT is easy. Ah, thank you for making free quality content. You are giving a gift to students. From the soul, guys. Glad to try and the samurai's path will continue with this goal, so that you can get a reference point and improve and enjoy. Mm, as if the hiring market has degraded. As if the hiring market has degraded, because it is very difficult to get to those interviews. Resumes are written and checked, and HR doesn't understand what they are doing, except for setting filters. There is such a feeling, guys, well, the hiring market has significantly shrunk, accordingly. Mm, there are many candidates. Naturally, it is necessary to optimize the work with these candidates who apply. Naturally, this is also, and naturally, candidates also apply and, to do something. Naturally, filters, markups, naturally, work goes in both directions. Well, like, it's a natural process, yes? This is a natural process. If, again, there were, well, there were simply not enough developers before, and, like, and Hire took everyone, and everyone said: "Oh, great hire, you applied, you were invited for an interview, you didn't apply, you were still invited for an interview." And now, like, well, there are not so many developers needed, yes, at the moment. And, accordingly, it seems that hiring is broken. As if it's always been like this, where there is no balance, supply and demand, so you need to pump, pump, pump, pump. Pass interviews. We are now going to invite our partners, not our partners, to our Prosobesim platform, so that they come from the platform, take people, interview you, find them there, because on the platform with interviews, the idea is that there will be no random people among you on the platform, because you need to have a certain character to go and interview even with those like you, to interview and be interviewed, and, secondly, the content there is so complex that you need to prepare for it. So, this is definitely not about random people. And I think this will be interesting for some companies, so that they come and take you from there. Follow the news. Here. Follow the news. You can sign up, pay for the back-end course now, and take it later. Write to Alexander, please, to the managers. I don't handle these issues. I don't really like this scheme, because, like, to sign up, pay for the course, and then, like, not go to it, and then all these banking transactions, chargebacks, refunds, it's all quite a hassle. Well, in short, please write to the team. I think we've had such cases for sure. And, possibly, this will work. Didn't you think about making a course on system design for front-end for middle plus developers? How can one improve in this? Oh, well, system design, guys, it's not about front-end, it's more, it's more, if we talk about system design, there is system design that is asked in interviews. I think you see a lot of this on YouTube. System design, in my opinion, is about designing systems. It's about, like, a full-stack engineer, let's say, about just a software engineer. Look, what's the thing. I announced, remember, I announced under some stream. The link is previous, the link is there. You can fill out the form. My team and I are thinking about it. That is, we are not just throwing it out, we are gradually thinking about it. The course was announced there. Let me show you. Like, that was the idea. There was a post on this topic in the Telegram channel. I'll send it now. Kama Sutra. Kama Sutra. I'll find it and send it to you. You can look. And for those who are interested, and for those who are interested, leave an application, because, in general, on this very topic we will copy message. Now I can open it here. Interesting, I can't. Ah, I can. We are preparing a month-long marathon for current developers with the team. We will conduct this once a year. The idea was that either you are a front-end developer or a back-end developer, you come here, but most likely we will focus on front-end developers, because it is front-end developers who feel that there is something to tell and how to turn them from form-shapers into this scheme, this diagram, into engineering. Ah, so there will be system design, full-stack, agents, MCP servers, decomposition, production, DevOps architecture, that is, ready-made projects. By the way, they are in production, yes, by design, they will all go into production. Possibly, we will buy some of the projects you create during this marathon. Possibly, you will stay to support some projects, and we will possibly agree with you, we will sell them, and we will take you, like, for a 10% share, and you will, like, well, you can support it as a pet project, and at the same time, if there is monetization, you will receive some payments. I'll send the link to the chat, take a look. Mm, well, if we launch quickly, it might happen at the end of February, at the beginning, at the end of February, such a rather thin story. Well, it all depends on the price, how we prepare. If we prepare very deeply, the price will be one. If we prepare less, the price will be different. Therefore, possibly, we will do the first launch, less polished, but still driving. Ah, formulate, yes, well, in short, stay tuned, guys. So, yes, there are thoughts in this direction. Ah, Dimych, and you yourself? Oh, good question, because, ah, I'll tell you now, ah, why it's a good question, I think. Mm, Dimych, and you yourself, how deeply do you dig into the AI topic? Is there an interest in building some systems around agents or is conditional cloud code enough? Ah, there isn't much idea to build anything. Why? I'll tell you. Because I think, ah, well, I am an expert in engineering, and I understand that I just need to strengthen my expertise, and chasing trends, hypes, like, is not for me. So, to stand aside and not look there, not touch it, to be an old man, a fart, who says: "This is evil, we need to do it the old-fashioned way." No, so, I, like, use these tools to strengthen my knowledge. I don't rush, I got acquainted with the basic things that are there, especially, well, there aren't many of them, how to connect these agents, do RAG, and build an NN chain of interactions. So, it's not difficult to get bogged down in this as a niche. There was a hype. Everyone was like: "So, we need to automate." And now it's all subsided, it seems. Including because you do something today, automate it expensively, right? Why expensively? Because, well, like, look, I am a specialist, I can do it for you expensively. And in fact, what can you do? Well, you'll figure it out. It's labor-intensive now to figure it out. Labor-intensive. And, in principle, I think it's pointless, because tomorrow there will be tools, like, monopolist tools in this market, which will do it cheaper and faster. And you don't need to waste time figuring out these things. And then, like, the thing is that real adult AI is a very complex topic, and you need to dedicate your life to it, and just, like, stir up topics, like, these agents here and there. Well, it sounds cool when it's sold as a course. Let's teach you to make such a project and you'll make your SaaS service and earn millions of dollars. It sells well. Ah, I don't see much practical use in this, oh, I don't see much practicality, because much of it now

Everything is changing so dynamically. And what can be done there with a team of five people, something like a small project, a startup, some integration, automation, the day after tomorrow will have millions in investments, billions in investments. Perhaps, from the point of view of starting some venture for investment, it might be worth it, perhaps it would be worth it, and perhaps it's not too late yet, to jump in here. Maybe, or maybe it's already too late. I don't know. But again, within my expertise, in short, guys, I'm like, uh, slow and steady wins the race. Foundation, stability, but movement in some direction, uh, also, yes, there's a bit of inertia. It's difficult to just start these things. Those who are younger, of course, who are getting into this movement, it seems like a good option to me. Uh, so now those who studied to be frontend developers must switch to studying to be backend developers, because there's no point in studying to be a frontend developer, since AI will replace us. Look, nobody has to do anything. Frontend, backend must be chosen strictly based on your internal state. You can develop very deeply in frontend engineering, very deeply. Mass frontend, when frontend developers were hunted, it seems like that won't happen in the near future. Frontend with the help of AI, even if they are hunted, they will still demand: "Come on, make me 100 landing pages, like on Tilda." But here, come on, frontend, generate 100 applications for me quickly with the help of AI per month, because you take it, generate it, and it all works by itself, and the cost of error is small. Here. Next. A frontend developer is obliged, of course, to know about the backend and learn backend additionally. I think so, obliged. Always, yes, an important point, I've always said this. A senior frontend developer should be a junior backend developer. And vice versa, a senior backend developer should be a junior frontend developer. And this, as if, is amplified in the age of AI. That is, a full-fledged full-stack, who knows, will you be in the profession in 10 years, yes, a full-fledged full-stack at the beginning of the path. No, I'm a frontend developer, but I can do backend. No, I'm a backend developer, but if something needs to be done on the frontend, I can. For example, Artem, our DevOps, was initially a frontend sysadmin, then frontend, then backend, then DevOps. And now he's like a DevOps backend. I won't do the task, but if it's really necessary, okay, he opens the AI, generates, and everything is fine. But like, no, yes, it's not for me, I don't want to, I don't like it, I won't. I, for example, like it, and I'm interested. But, and also, yes, the question about the depth of AI is also such a point that so much is appearing there, that an important point, I had a friend, she's more about business, a product manager, she has her own company there too, they, well, in short, it's an IT company, but not about development. And she had a conversation once, just when this AI boom appeared. She was like: "Damn, I can't keep up." She had this like distress, not distress, but like anxiety. That there's so much, and I can't understand and learn it all. Oh my God, how will I fall behind, fall behind, fall behind. No, like, these are YouTubers who are engaged in it, earn money from it, they review these tools. Their job is to review, but to understand all this and worry, no. You, as a specialist, find at least one tool for yourself, implement it, and use it. Well, then find a second tool. And, perhaps, in reality, out of ten tools, you actually need one or two tools. Well, learn to speed yourself up with them, yes, to truly speed yourself up by orders of magnitude. In programming, well, we don't need 10 tools there, I think, of course. Although here I also want to discuss this topic with ITboroda, invite him to the podcast. I also want to continue podcasts now. And ITboroda is developing his community, a club, on AI. Well, he has a maximally comprehensive vision there, what's happening in AI from the perspective of a developer and AI. I'll talk to him, invite him to me, if he comes to my channel, we'll gladly chat. Here. M. Dimych, hello. Graduated 3 years ago. Worked at Obitech due to health problems. Dropped out of development for half a year. Now I can't get back into it. True, already on Naknau. Anton writes: "Uh, like, bro, well, contact Max, knock on Max's door, tell him your situation. I think Max will help you get settled." Uh, well, like, write to Maxim, you have the contact. Maxim, yes, you should have the contact. Here, write, bro, we'll help. Uh, oh, damn, I recently heard some podcast on this topic about Liberman, that they are doing something with distributed PCs, user crypto computations. I heard something, I didn't understand, I was listening to something before bed. Accordingly, something was loaded into my head. I was like: "Something is happening and that's it." And I didn't go back to it. They're up to something, guys, you need to understand that Liberman are super business guys. They're up to something that can be sold expensively, into which a lot of investment can be attracted. And what's about money, about hype, so accordingly, something distributed, something like that they will be up to. Whether it works out or not. One way or another, sooner or later, they will succeed, or others will. So, respect to those crazy people who still subscribe to these things, do projects that are like, let's fly to Mars, like Elon Musk. Dmitry, you also watched your Samurai path, the first course, I understand. Thank you very much for it. I got into IT in 2023. I use neural networks sometimes, but I perceive it as a tool, nothing more. A new Google, so to speak. As a tool. Everything is correct, as a tool. But Google is still about search, and here there are huge volumes of code generation. And I assume, Denis, that you probably don't use this deeply, uh, cursor and cloud-code. Accordingly, there is no large-scale code generation, no analysis of the codebase, in an instant. The codebase that would have taken a month to understand, to look at, a week, in an hour, yes, you can get the necessary slice from the perspective that interests you now. From this perspective, I need to look at the system from a financial point of view, how it's structured, come on, give me documentation from a financial point of view. I told the students today that this, that is, documentation doesn't make sense, because documentation is about everything, and you can get documentation in the moment in the slice you need. So try it, guys. Yes, because it's not just a search engine. A neural network is a real assistant, it's a code generator, it's a code reviewer, it's a powerful helper that is nearby. What's the point of rejoicing and clapping your hands that you've increased developer performance tenfold, if now they will demand tenfold more for the same salary, but the responsibility is also X10? Well, like that. Well, not responsibility X10. Well, responsibility is always responsibility, guys. And salary is always salary, yes? That is, you do, like, well, well, if you can do 10 times more, do it, yes, the question is, like, again, work. Nobody cancels work. An eight-hour workday, like, nobody cancels it. Those who run blogs on Instagram and talk about how you can live happily, work 2 hours a day and do nothing, they cancel it. Uh, and in general, like, this is what we see on the surface, and it attracts us, this, like, lifestyle. Of course, it's attractive, it attracts me. I would like to have the opportunity not to work, but to work exclusively with what I, well, just absolutely for fun, just for creativity. No, like, work is routine, it is like that, I can't imagine how one can talk about professionals who are free from this routine, this responsibility. I can't imagine at all. That is, yes, like, not everyone should be such a pro, not everyone should worship work, enjoy work, but I like work, like the craft, I like creating products that people need, that people use. I'm interested in understanding engineering things, and I'm glad that before I had to spend time on, uh, writing function syntax, and now I can spend time connecting one system to another, because typing letters is no longer necessary, and I can focus on more, well, engineering things, like how to connect what, at first glance, you can't connect, for example. And, accordingly, hence, there is such joy. That is, I'm not afraid of work. Work, like, I think, is still necessary for all of us. For most people. It is necessary for us. I'm a frontend developer. 4 years of experience. In Sobatikle, after backend tech, I've been burnt out for a year. Now I'm returning to interviews. I've never rested from work for so long. I'm restoring the Samurai path connection, no alternatives. Cool. And how are you returning to interviews, go to our platform. Uh, try it. It's cool, like, it's normal to be burnt out. Uh, it's also normal. M, burnout is also normal, like. What else? Burnout, vitamin deficiency, fatigue, midlife crises, junior, senior, transitional years, like, all this is something that needs to be treated calmly, and not overdone, so that it turns into some kind of depression, worked through somehow. But many people exaggerate their problems, like, and, accordingly, on this basis, they get depressed a lot, as if life should be super easy. For some reason, it seems to me, it seems to me, my most subjective opinion based on my observations, that people would just spend less time on Instagram and YouTube, I mean entertaining Instagram, YouTube, uh, watch less of these fitness handsome guys and girls, these ones who make smoothies in the morning and open their MacBooks an hour before that, after doing yoga, standing in some pose, and then go to a party. And then, well, I have such friends here who are very much in the image of a successful lifestyle from the outside. And there are even acquaintances whose lifestyle is not an Instagram picture, but it's really a part, and it's a mix of an Instagram picture and a lifestyle. That is, it's like a desire to live like that, like, and they succeed, they have their own business, all that. Well, then you need to understand what path was taken to open this business, the suffering, the responsibility, blah-blah-blah-blah. That is, it's not done as easily as it might seem. And it seems to me that people look at this and think, well, why does it seem to me, I also had moments then that I look at something, and it seems that everyone is a thousand times more expert than me, a thousand times happier, a thousand times richer than me. And I'm sitting here with my small business: "Well, I should have flown somewhere like Elon Musk already." But in reality, if you turn off all the fakes from social networks, normal people will remain. Normal people, and they have ordinary lives, and those who have great achievements, they just work their asses off, 24/7, and they are not on social networks. And they are not on social networks. And it seems to me that if you remove social networks, all burnout will disappear. The same thing, for example, take a programmer. Burnout. If, let's say, you look at conscious mercantilism and read that, like, damn, I'm here at work, and I could be a wolf who howls and earns 10 times more and is happy there and all that, and then you start, and I'm such a [ __ ] and something is wrong with me and it starts, it goes on and on. Although in many activities, yes, conscious mercantilism is respected, because after all, it's also like on the side of, well, like, the proletariat, a little bit, of course, with some distortions in some scheme, like, and let's not have win-win schemes, but schemes where, like, uh, I'm more for the win-win model, again, that, well, like, you should try to be more honest, truthful, professional, uh, but nevertheless, in short, so, accordingly, I mean that, of course, we have YouTube images. What tech images are there? You just don't see them, guys. You, if you're in the profession, how many programmers, 99% of programmers, they are not on YouTube, not on Instagram. 1% on YouTube, on Instagram will beautifully tell you that there is some kind of professional happiness, which is like this and like that. And if you don't look at this nonsense, then focus, well, on your own life. Well, that's also not always the case. Sometimes psychologically in the head, yes, hormonally, something changes, and then it needs to be treated, yes, for certain conditions like burnout. To be treated, literally treated. And if I connect Opus on a VPS, tell it: "Write me such a project or fix something in the current one." And it does everything, launches it in Docker QBD itself, and it works. How to find motivation if I connect it? Well, don't completely give the task to this AI, or figure out what it did, and try to close the task not by having the task closed. Everything works and works. But how to do it better? How to do decomposition? How to make it more architectural? Is everything okay with security? Is everything okay with performance? So learn new things for yourself. AI, like, it doesn't just do something for you, but it also turns out that this year it's better to study to be a backend developer, to study, right, to be interesting to the market later. It's better to study for what is more interesting, guys. Frontend/backend only out of interest. First and foremost, professionals are interesting to the market. You can become a professional where you are willing to spend time even beyond normal hours. If you're not interested in backend and are interested in frontend, you won't have the strength. And vice versa, that's how it is. I always say: "Knowledge should serve people." Yes, Reels make you doubt yourself 100%. I've noticed it, I watch it, but I try to do it in small doses. Dmitry, have there been cases in your practice of transitioning from frontend to blockchain development? How optimally? Initially, frontend with elements of 3D, then predominantly smart contracts, pulling up backend. I can't say how optimally, guys. Exactly, Yur, there are such transitions. Many of our students went into production, studying, whoever is interested in crypto, they go. That is, there is crypto, there is just another node that works according to its own rules. This is in the form of a backend, this is in the form of a frontend, yes, this extension in the browser, I don't remember, I'm not a crypto expert, and I haven't delved into it, like, but it's quite a good thing if the topic is interesting, it's quite a good thing, you should follow your interest. The topic is in demand, blockchain, I think, will not go anywhere and will continue to develop, uh, and applications for it will be found. Therefore, here, here I am studying to be a frontend developer, should I drop my frontend studies and switch to backend to be in demand in the market? Well, I've already answered that. I just connected. Why? And it will replace. Well, guys, you can rewatch that later. Study frontend, become an engineer. I think that soon there will be a rise again, like before. The main thing is not to give up. We've already had the second layoff in the company this year. Salaries are being cut, but we're not discouraged, we're not giving up, guys, so that you don't get cut. And if you do get cut, so that you can easily find a job, practice interviews, on our platform too. Everything is free. The link is in the description of the video. I talked about this. Everything here, guys, you're awesome. Work, work, work, work. Turbulent times. What tomorrow holds, it's unclear. They say that 2026 might be even more difficult, even more problematic. You see what's happening in politics, in economies. Some wild shit is happening. So we're working hard for the future. Everything will settle down, yes? After winter comes summer, and then we'll exhale. Those who work hard, then we'll exhale, so that it's not like the grasshopper sang its song and didn't even have time to look back. Winter is still coming, yes, there are thoughts that it will be even worse, it will be more difficult. Therefore, we are developing professionalism. If you don't know how to develop it yourself, come to Frontend Developers, come to our Dev Club. Although we don't even have a Dev Club right now. Well, in the Dev Club, you get a subscription to our middle programs, our course on the platform. Study new materials from any authors you trust. Study new materials, practice interviews on our platform. Practice, if you're afraid to go to real interviews. Visit real interviews sometimes too. Sign up for our Frontend Developers marathon, which we will hold at the end of February, beginning of March. Here's the system. Well, I've talked about this too. You can go study backend, you can study in a continuous mode, that's also fine. My friend Batumsky, a bachata dancer, my colleague too, also a beginner bachata instructor, like me, signed up for our backend. Not a b, not a c. A cool guy, he does everything well. Well, breaks, breaks. Came back, got lazy, came back, got lazy, you know, well, he's already a developer. It's hard to find a good partner, but in any case, he learns about architecture, something, backends, databases, MongoDB, SQL, in short. In any case. That is, you are strengthening yourself, yes, fundamentally strengthening yourself. And tomorrow, when they say: "Okay, who among us, uh, okay, is completely clueless about backend." And I'm clueless about backend. I can, look, with the help of AI, make a full-stack. Uh, okay, you're staying, and you're clueless. No, well, I'll learn. No, when you learn, it's all over, goodbye. That is, for the future, yes, you need to do it for the future. This whole concept of my incubator now, the concept. For the future. For the future, because tomorrow is unpredictable. But when it's unpredictable, I'll definitely exhale and go on some kind of long vacation and will also be programming. Well, like, for fun, for kicks. Here. But not for kicks now. Now we're trying to choose the most necessary, juicy things on the market, which are also fundamentally important. Not just some fleeting trend, but fundamentally important. And it will definitely pay off in the long run. If it's difficult alone, then ask questions and let's move forward together. That's all, guys, I hugged everyone, lifted them up. I'm going to rest a bit today. Ja.