What it costs
Sometimes I dream that I don’t work: that I am begging, or that I have to go back to the construction site. There was no one particular dream. It is just that a lot of anxiety comes out like that.
This is the last post of four. The first was about ten years of writing code by hand; the second and the third were about what we gained. This one is about what it costs. It is the most personal, and it has more fear in it: I am more afraid now than I have ever been.
The fear
I am 41, with ten years in the profession behind me. And still I don’t feel there is any force that would keep me in it if agents one day learn to do everything fully autonomously. Honestly, I doubt they will. But a year ago I couldn’t imagine what I do now, not even close. What will happen in a year? Rationally, I know I can’t imagine that either.
What does this fear look like? A thought somewhere at the edge of consciousness — in a dream, or while I am walking somewhere. It sits nearby like a little devil, sometimes bigger, sometimes smaller, but always there.
I can’t single out a moment when it was strongest, and the anxiety has two sources. The first is a few bloggers I watch on YouTube, who only last year said that neural networks would never replace developers. Recently they changed their minds, and so did I: most of the code, maybe 99% of it, can now be written with neural networks.
The second is LinkedIn, and Twitter and Reddit with it. People write there that they can’t find a job — people with 25 years of experience and huge achievements, from FAANG, Microsoft, Netflix. People with twenty years of experience literally write that they will work for food. Every story is different, but it is still a symptom of a big problem that is coming, and it feeds my fear.
Recently I saw a post from a guy who would take a job in Russia with a salary ceiling of 140 thousand roubles. That is what was paid ten years ago; now you would expect 400–500 thousand. It looks like that’s it: the salaries are over, and it is not clear what comes next.
I may be no exception: I can lose my job the same way, and no one will know that I do it really well. Maybe this job is simply not needed anymore. Everyone I know still has a job, but projects need fewer people now, and code is not a craft anymore; it costs nothing. The news talks about it.
But an engineer was never paid for code — he is paid for business value. Patrick McKenzie wrote this back in 2011, in Don’t Call Yourself A Programmer, And Other Career Advice: “Engineers are hired to create business value, not to program things.” And a little further: “Add revenue. Reduce costs. Those are your only goals.” He also advised staying close to the profit centres of a company rather than the cost centres.
In my own work the clearest case was the WYSIWYG editor from the first post. I reworked its rendering — batching the store updates was only one of many parts of that — and it became much faster. Before that, there was a decision to move it from old packages to new ones, and it took a long time to agree the budget for it. In the end the client agreed, we moved it, and now the application doesn’t have to be rewritten; it can live on. That result was the value, not the lines. Its fate is still unknown, though: that depends more on business processes than on technical decisions.
Fifteen years ago that was career advice. Now it is the line along which the profession splits: whoever only typed the code is left with nothing to sell. What is left is understanding what the business needs, deciding, being responsible for the result. That calms me a little, and scares me a little too: value is much harder to show than code.
What surviving means
So when I think about the future, I think about how to survive: to keep a job in some form, so that my standard of living doesn’t fall, or doesn’t fall much. Most likely my dreams of earning enough for a pension, or of coming up with some IT business, won’t come true. I won’t have a pension at all, because I left Russia, but I never counted on it — I counted on earning more and living on the interest. Now it is clear I didn’t make it in time.
Many people depend on me. I don’t want to let them down, and I really don’t want to sink back into poverty. Physically I will survive, and I will come up with something, but life will become noticeably poorer.
I do have a plan B. I can teach yoga again: for a while I taught it in Andrey Sidersky’s system. And I want to write my own game projects.
Where do I see myself in five years? I have no idea — not even in a year: every half a year my life turns upside down. I hope I will be alive and well, that I will have money, that I will be warm. Maybe kids, maybe a dog, if it goes well. If it goes badly, I will be alone, living from hand to mouth and teaching others: I still like that, and it can earn something. If I can be neither a developer nor a businessman, I won’t be able to support anyone but myself, and I will have to live in the forest, in a tent or a dugout.
Shaking a fist at the rain
It will be hardest for those who can’t adapt. I don’t know any personally, but LinkedIn has thousands of them, and I wish them luck. These are the guys who still don’t learn agents, who fight this tsunami and shake their fist at the sky: “Rain, stop, I don’t like that you are cold.” I am afraid they will lose their jobs, and it will be hard for them to rebuild when they finally see that you can’t work even the way people worked a year ago. And survival is exactly what we will be doing in the near future, if tokens stay this cheap.
Less than a year ago I was a super skeptic myself. But if I can’t avoid it and can’t fight it, I need to master it — that is the second post.
I had something to lean on while I did: ten years of experience. Those who are only coming into the profession have nothing like it, and the ladder I climbed has lost its lower rungs.
The ladder
When I started, this ladder existed. A person with no connections, no relevant education, nothing at all, could climb into the middle class, roughly speaking. I was lucky, and I probably have some natural talent, but back then the way in was open to almost anyone. The internet was already there: you could find any information, download a book — the same Eloquent JavaScript — and learn.
And before my eyes two people I knew from construction became programmers: my partner in the gypsum workshop, who also did music, and an installer of low-voltage systems. Both are frontend developers now, and both started after I told them what to do and how to prepare.
Three people from one construction site
How the ladder worked when it still had all its rungs.
- me
- my partner from the gypsum workshop
- the low-voltage installer
- Sep 2016 I start learning JavaScript from one book
- Spring 2017 My first job, a small startup
- Autumn 2017 EPAM
- After me Starts learning, after I tell him how to prepare
- Then The same startup, if I remember right; then a Japanese company Thailand first, now Japan
- After me Starts learning the same way
- Then Hired at a job fair by a company from Riga, leaves quickly
They too started after thirty, and they got what they wanted: they left, they earned money. Back then companies took all the sensible guys.
How did I teach them? The way I taught myself. I told them what I was doing and said what I kept telling myself: that it will work out, that you must not give up, that you need to be consistent, logical, rational, that the future is in this. But the main thing I told them was that there is a chance.
Both are still in the profession. With the one in Riga we don’t talk anymore. With the one in Tokyo we joke about how we will be working on a construction site while AGI takes over the world. But we talk about serious things too — practice, and where and how to live. He doesn’t want to stay in Japan and wants to move to Western Europe. I don’t think it is the best idea, but it is his choice, and I sincerely hope it works out for him.
What would I tell myself in 2016, on the construction site, with that book? Keep going, change nothing, everything is going as it should. I think I made the only right decision then: it put off poverty by ten years, let me live a normal life and help others.
The lower rungs
Now it doesn’t work like that. I look at my team, at the speed we are rushing at, and I don’t see where a junior would fit.
It is not about juniors being bad. In our team everything moves so fast that you need a lot of experience not to make a mess, and a junior would have no room to learn. I wouldn’t want to be a junior now.
The numbers point the same way, although each of them has its limits.
- At Stanford, Erik Brynjolfsson and his colleagues looked at payroll data for millions of American workers (Canaries in the Coal Mine?, revised in August 2026). For people aged 22–25 in the jobs most exposed to AI, employment fell by about 11% from November 2022 to June 2026, while the same age group in the least exposed jobs grew by about 10%. Experienced workers show no such gap. And the mechanism is not that young people get fired more; companies hire fewer of them. The authors themselves call it descriptive, not proof of cause.
- SignalFire, a venture fund that tracks hiring, counted that in 2025 new graduates were 7% of hires at the fifteen biggest tech companies — less than half the share they had in 2019. At startups they were under 6%.
- And the vacancies themselves. Indeed tracks job postings for software developers in the United States (the series on FRED):
Caution is needed here: the fall of 2022–2023 started before the agents could do anything serious. But the picture for juniors is the same in every source I found: the door at the bottom is narrower than it was when I came in.
Juniors will still be needed: one day the developers who exist now will run out — they will retire, get sick — and new ones need to be grown. Companies like EPAM understand this and grow the replacement deliberately: they put juniors into teams or give them separate projects. But the market is now full of experienced people who want to work, and cheaply — and I don’t know how long companies will stay patient with that.
What I would tell a junior
Someone asked me once whether to go into programming. I answered reluctantly.
If you really want to be a programmer, there are two options: make your own project, your own business, or go to a big outsourcer that hires juniors and trains them. Whoever wants it badly enough will find a loophole. But it is hard, very hard.
But in general I would advise not going into programming at all, and trying something else: massage, guiding, construction, anything that needs hands. Everything done in front of a screen will be automated sooner or later. I think that out of a hundred programmers today, maybe five will be needed, if that. The rest will take any money, and salaries will fall. This work is already very hard. If it is also paid little, it is not clear at all why anyone would do it.
I feel sorry for those who are only starting: for someone like me — born in a poor family, with no one to help — it was one of the few chances to live a decent life. And I don’t know how one breaks in now.
I still prepare people from the resource pool for the move to the next level; they are becoming seniors. We talk about the future only as a joke. And the jokes are not the brightest: that we will serve electronic masters and sit on universal basic income. Sometimes I tell them to save some money and be ready for anything.
A speed that cannot last
Those who stay in the profession pay too. I don’t know how much longer I can pass such a flow of information through myself: half an hour can do a week of the old work, and not everyone can digest this. I already know what an unsustainable speed does: after the crunch I couldn’t open my laptop without a headache.
So I see my work as shifts on small projects: you come, do it very fast, leave and rest. You don’t have to lower the speed; you have to recover — long holidays, weekends, whatever it takes. For me it would be about a month of work and a couple of weeks of rest.
I have worked like that before — on construction, in a crew of three. Intensive work, and then a long rest. Big outsourcing is different: there is always work. I don’t remember a time when I could really rest. Even on the bench I was either learning or running interviews, always on call.
What I lost
Some things I have already lost. What is left of a lead, as I wrote in the third post, is the responsibility, and for me it is a loss. Being a lead was hard, sometimes there was too much load, but mostly I loved it: I always liked working with people, helping and teaching other developers. Now it is not needed.
Writing code by hand is a different story: I don’t miss it — I never really liked it, and I know I am an exception here.
Loneliness
The people around me went with the team, and there is a lot of loneliness now. After we came to Serbia, my marriage ended, and I moved to a village again — this time a Serbian one — and lived alone there for a year and a half.
At work I talk to the other developer and the analyst, both in London, sometimes in the virtual office with cameras on and silence. No more team-buildings. We used to sit under one roof, hold a scrum meeting in a real meeting room with a whiteboard, and drink beer together after work.
My old team is spread around the world now: everyone left because of the war, everyone still works, not all of them at EPAM. There was no clear goodbye: we had a call, said some warm words and tried not to lose each other, but for a few months before that it had been clear that it was all ending. We are still good friends. Here in Serbia there are several people from St Petersburg, and we sometimes go to the office together. But the old closeness is gone.
And I talk to the computer. But I never personify the agents — that is a hard rule. A tool is a tool.
I am not the only one: remote workers are lonelier than those on-site (Gallup), and the World Health Organization wrote in 2025 that one person in six in the world is lonely.
What holds me is practice: yoga, Buddhism, qigong. Yoga and qigong since I was 25; they help a lot, if there is strength. And tea: I love roasted oolongs and shu puer, and I have drunk Chinese tea every morning for about twenty years.
My mornings are always the same, and the routine saves my head. I get up at six, before dawn. I wash my face, do exercises, drink a glass of warm water. I prepare tea and a breakfast for later and go out into the yard — I live in a house. If it is warm, I roll out a mat and sit with my tea, sometimes do qigong, then stretching and yoga. This hour alone gives the day a backbone.
I have got used to loneliness — so used that I hardly notice it. I am comfortable alone. I am not alone anymore, though: I have a girlfriend now, and I hope we will be good together. But we are both used to living on our own, and that habit is hard to give up.
Of course I remember the years in big teams with nostalgia, all that movement. Probably I would like to be among those people once more. But after a year and a half in an empty house, talking, roughly speaking, only to a computer, going back would be hard. We’ll live and see. If anything brought me back to a big team, it would be the company of good people I know.
Further apart
And it is not only me: the modern world, as I see it, only pulls people apart. We have millions of ways to be near each other: social networks, dating apps and all the rest. In reality we drift further apart, and every year faster. It started shortly before COVID and grew much stronger during it.
And it is very sad, because a human being is, after all, a creature that needs other people. If we drift completely apart, we will probably lose part of what is human. Most likely we will see a wave of suicides, a wave of mental illness — to some extent we already see it — and a wave of violence connected with it. And I am afraid to look ahead.
The rain
But still, I don’t want to be that old man shaking his fist at the sky. We can only accept our fate and try to make our life as effective as possible and as pleasant as possible for ourselves.
And I have a hope that new technologies will free people and bring creativity closer, as happened with me: I realise my ideas fast now, and I like it very much.
Finally I am making my own games: I always wanted to, but writing the code would take so much time that I never started. I have a lot of ideas, and at least two games are in the works now, one way or another. I think good computer games are a good thing; there is use in them for people, they are not just killing time.
The project that gave me the most hope was Ritmolux, the music visualiser. It was great to see that fairly simple prompts can make something truly interesting and beautiful, where everything I enjoyed came to life: the maths and the generative art. Please have a look at it, try running it, and tell me whether you like it or not.
But the fear of losing the job and the money, the fear of losing humanity to loneliness — of course it is very big. Only time will show what is ahead.
Thank you
What did these ten years give me? A chance to leave for a country more to my heart. Freedom in decisions, reliance on myself, and a confidence in the future — which is exactly what is leaving now. And above all, people. People always come first for me, and meeting so many smart, good people in this profession is a bonus you could only dream of.
Who am I grateful to? The crazy director of my first company, the startup. My first lead at EPAM, a motorcyclist who lives in Europe now: we drank a lot of beer and talked a lot. My other lead, a Java developer, who never minded explaining the internals of the application to me, although I, a frontend developer, barely understood him. He was a real engineer, grew into an architect, and later I followed in his footsteps; he inspired me to understand deeply what I do.
My colleagues from all the projects. From the dream team I remember our project manager with special tenderness: she supported us all the way and sincerely wanted the application to succeed and us to bring the client that business value. She cared — as did everyone on that team.
Thank you to everyone around me who fought and wanted something better. It didn’t always work out; often the reason was bureaucracy or politics in the project or the company. But the people who sincerely wanted something better did, in the end, achieve it.
And separately, my resource manager. She has supported me through almost all my ten years at EPAM, and in many ways it is because of her that I once didn’t leave the company. She sincerely believes that for a company to prosper, developers must feel good. You don’t hear that often, and it gives hope: if such people exist, not everything is lost yet. Thank you all.
The end of the run
Why did I write these four posts? For me it is a way to put a full stop to this stage of ten years: to look back at where my life came from and where it has come to, and to think a little about a future more uncertain than ever.
I would like the reader to take away the motivation that at least once it was possible to take your life into your own hands, change it sharply and make it much better. And the hope that it will all somehow turn out well, that there will be new work, and not just the destruction of the middle class and general poverty on a universal basic income. And I would like them to believe in themselves. I am not sure it worked: I reread these posts, and they are rather sad.
Last winter, where this series began, I was going to quit myself and never work as a programmer again. Now I dream that the work is gone and I am back on the construction site. It was a good time, a great ride, and it seems to be coming to an end. I hope that I personally will get through it. We don’t choose our fate, but we choose what we do with it.
To my old team — I know you will read this: I love you very much, and I hug you. And to everyone: good luck, and let everything turn out well.