There Are More Programmers Now, Even Though AI Promised to Replace Them Within a Year

The forecast looked unambiguous: six months, a year at most, and machines would be writing the code. About three years have passed. According to the summaries Habr cites, more people are paid to write code, not fewer. Job openings remain, but their requirements have been rewritten.

The available excerpt of the original contains no tables or statistics, so what follows covers the story itself and the context in which it unfolded.

What Was Promised and Who Said It

The mood of 2023–2025 is easy to reconstruct. Executives at large companies told investors they would not be hiring new engineers. In December 2024, Marc Benioff said Salesforce did not plan to expand its developer headcount in 2025. In May 2025, Anthropic CEO Dario Amodei warned that AI could eliminate up to half of entry-level office positions within one to five years. Coding schools reacted quickly: in their ads, Python courses gave way to prompt courses. In the comments under such news stories, people advised becoming a welder, since a neural network can't hold a torch.

As often happens with loud predictions, the details decided everything. Part of the work did get automated: draft code, tests, documentation, routine fixes. But a developer is responsible for more than a set of lines. They define the task, verify the result, answer for security and maintenance, and sometimes argue with the client about what actually needs to be built.

Why There Are More People

Economists call this the Jevons paradox: when a resource gets cheaper, people consume more of it. If code costs less, businesses order more code. Automation of internal processes that used to be postponed because of cost becomes profitable. Demand appears for integrations, agents, dashboards and scripts, and someone has to build, check and maintain all of it.

The second effect is described in the article itself: the word "programmer" has spread into roles where, not long ago, the only tools were email and a spreadsheet with a sales plan. An analyst, marketer or logistics specialist now writes scripts with the help of an AI assistant. Formally they are not developers, but they perform the same function. In such cases, employment statistics depend on who is counted as a programmer.

What can be checked against open data:

• Timeline of the original forecast: from "six months" to "a year"; about three years have passed.
• Trend: according to the summaries, the number of people paid to write code has grown.
• Job market: the volume has held, the requirements have changed.
• Geographic coverage: the original's announcement gives no figures by country.
• For comparison: the US Bureau of Labor Statistics projected employment growth of about 17% for software developers for 2023–2033, noticeably above the average for the economy.

What Developers and Hiring Managers Should Do

Developers should assume that what is valued is not typing speed but the ability to take a task all the way to production. That means reading other people's code and generated code, finding bugs in it, understanding architecture, and answering for the consequences. Juniors have it hardest: the routine tasks on which they used to gain experience are now more often handed to an assistant. That is why a portfolio of real projects and a grasp of the fundamentals work better than a certificate for completing a prompt course.

Hiring managers would do well to revisit their interviews. Tasks that test memorization of syntax measure what a model can do. Reviewing someone else's pull request, debugging a failing service and discussing design trade-offs show what a person can do.

And for everyone else: if your job involves spreadsheets and repetitive operations, basic programming skills with an AI assistant are becoming a working tool, just as Excel once did. The developer profession has not disappeared. It has become broader and dissolved into other specialties.