IT stopped being a shortcut to the good life, says viral post on Habr

A post published on Habr, a Russian-language technology blog and community platform, on November 1, 2026, by a developer who goes by 1endstick went viral in the Russian tech community: more than 64,000 views in 30 days and 385 comments as of this writing, according to the original text. The author says he built HR products for an American big tech company and now implements AI agents at companies, which he says gives him a view from both sides of the table.

The post's central thesis flips the debate that dominates the conversation about AI and IT jobs. Instead of asking whether AI will replace programmers, it asks what happens if it does not replace them. The answer, it argues, is worse than mass layoffs: the profession stops being rare and becomes just another one, subject to the same salary compression as any ordinary job market.

What Sustained High Salaries

The argument starts with an uncomfortable question: what exactly was the market paying so well for? The standard answer is scarcity: it was hard to find someone who could solve the problem without a developer. But the post points to a second half of the answer that is rarely said out loud: the opacity of the work itself.

A task finished on Tuesday would often stay "in progress" on the board until Friday, because the manager had no way to verify how long a task actually took. Story points worked like a ritual in which both sides agreed not to get into detail, and the phrase "that part is complicated" shut down any further questioning. In that gap between what a task actually took and what was reported, there was room for a gym session during lunch, a second remote job, and meetings taken from the beach.

The post cites Stack Overflow's annual developer survey: in 2025, only 24.5% said they were happy at work, and another 47% chose the "complacent" option (something like "it's fine, don't rock the boat"). In other words, even those living this comfortable arrangement didn't necessarily like the work itself, they liked the package around it.

When the Industry Itself Inflated the Bubble

With the path into IT well mapped out, an entire industry grew up around it: "how to get into IT" courses, ready-made roadmaps, interview simulators, and, in Russia's case, even state incentives such as exemption from military service and easier mortgage credit for people working in the sector. A survey by Habr Career, Habr's career and job platform, cited in the post shows the result: 29% of IT professionals admitted to inflating their experience on their resume, and 88% of that group still passed their probation period anyway.

This created, according to the author, a parallel market of services to get through the profession's selection process without necessarily mastering it. The word "developer" stopped identifying someone who likes computers and started identifying someone who likes remote work. As long as the labor shortage lasted, no one cared about this dilution.

AI's Real Role in This Story

The reasoning error the post points to is linking two claims that don't actually follow from each other: that if AI makes developers stronger, then they're worth more money. The tool was rolled out to everyone at the same time: you got better than your own self from last year, but not necessarily better than the rest of the market.

The post cites data Anthropic itself released about its internal usage: non-technical employees went from 2 sessions per month with coding agents to 90; managers started making twice as many commits; and the share of junior developers among the company's technical hires fell from 27% to 9%.

The post's more specific argument is about where a project's context used to live. Until recently, all the knowledge needed to understand "why that part is complicated" stayed in developers' heads, out of AI's reach, and that worked as a kind of layoff insurance. But for an agent to work well, that knowledge has to leave people's heads and turn into documentation, detailed tickets, and meeting transcripts. And everything that becomes text for the agent to read, the manager (or the manager's agent) can read too.

The Precedent of Other Automated Professions

The post draws on an MIT study of 40 years of automation to argue that AI's effect depends on which part of the job it takes over. The three cases cited:

Taxi drivers weren't replaced, quite the opposite: more people ended up doing the job. What disappeared was the market value of one specific skill: knowing the city by heart. And that scarcity, once broken by GPS, never came back.

Where This Leaves Programmers

The post's hypothesis is that the profession will split into two extremes, like the accountant and stock clerk cases happening at the same time. At the top, fewer people, more expensive, accountable for the whole result from start to finish. At the bottom, lots of people delivering something "close enough" with an agent's help. The comfortable arrangement used to sit in the middle: doing your part well, earning like a factory director, closing the laptop at 6 p.m., and it's exactly that middle range that risks emptying out.

The post cites a figure from the Russian job site hh.ru: today there are 19 resumes competing for each junior-level opening, against only 2.5 resumes per senior-level opening, an imbalance that illustrates how entry into the profession got more competitive while the top remains scarce.

"But I'm Senior, That's a Junior's Problem"

The post rebuts this objection with a recent case on Habr itself: a senior developer, with ten years of experience and an open source project with 10,000 stars, was turned down in hiring processes at large companies, a post that reportedly got more than 300 comments. The person's experience didn't change; what changed was the filter that turns experience into a job offer, and that filter stopped recognizing their profile.

The author also reports that an Anthropic manager wrote that, on his team, recent hires have been outperforming people who probably code better than they do, and that this manager cancelled the technical coding interview stage of the hiring process. The post's reading: defending your own competence and remaining competent have sometimes become two different things.

The people who went into IT to avoid asking for a raise, avoid going to the office, and avoid asking for leave to live in another country will end up exactly where they ran from, except the job posting will say Senior Developer.

1endstick, author of the post on HabrТо есть люди, которые бежали в айти, чтобы не просить прибавку, не ездить в офис и не отпрашиваться пожить в другой стране, придут ровно туда, откуда бежали, только в вакансии будет написано Senior Developer.

What's Still Speculation

The author himself flags the text as a hypothesis, not a finding: he acknowledges that the cooldown in the IT market also has other causes, such as the broader economic slowdown and a rebound in jobs in other sectors that started paying better. He also notes that Anthropic estimated that even non-technical employees would need half a work week just to read what the agents produced for them, a review cost that the "infinite productivity" narrative tends to ignore.

The post ends by promoting an online conference organized by the author's group, ai-dev.live, scheduled for November 11 and 12, 2026, with talks on career assessment in the age of agents and reviewing AI-generated code.

What This Means for Programmers in Brazil

The scenario described in the post is Russian, but the underlying mechanics aren't local. The boom in bootcamps and "how to get into tech" tracks also happened in the Brazilian market over the last decade, driven by the same shortage of senior talent. And the adoption of coding agents like Claude Code, Cursor, and GitHub Copilot at Brazilian companies follows the same logic the post points to: the tool becomes available to everyone at the same time, which reduces the relative advantage of those who already knew how to code well.

If the post's hypothesis holds true here, the most likely effect isn't mass layoffs of Brazilian developers, but a compression of the mid-career range: hiring processes less dependent on traditional technical tests and more dependent on portfolio, network, and documented delivery, exactly the kind of record that, according to the post, AI agents make mandatory for working well within a team.