AI has truly changed the economics of programming. A significant portion of code is now produced faster, cheaper, and by fewer people. From this, businesses drew a very clear conclusion: if programming has become easier, it means fewer programmers are needed.

At the initial stage, this logic works. The old volume of development is indeed becoming cheaper. A single strong specialist with AI is capable of doing more than before, and the market naturally stops paying previous money for the previous volume of work. This mechanism has already been described in the concept «When a specialist starts doing more, it is natural for the market to pay less for the previous volume of work».

But then an interpretation error occurs.

For business, a programmer long looked like a person who writes code. AI has learned to write code — therefore, it seems logical that the programmer themselves becomes less needed.

The problem is that code production was only the most visible part of technical work.

AI on its own does not decide what system a company even makes sense to build. It does not understand the economics of a specific business, does not automatically determine the boundaries of automation, does not know where a network of agents is needed, where a single tool is sufficient, where ordinary deterministic code is required, and where it is better not to automate anything at all.

It does not organize a reliable corporate information loop on its own, does not decide which data can be trusted, does not design sustainable memory, does not build integrations with existing systems, and does not determine where autonomy is useful and where it is dangerous.

That is, after the cheapening of execution, the importance of other work sharply increases: designing, connecting, and managing the new computing environment.

Business has not yet learned to buy this work

It took companies years to realize why they needed programmers at all.

At first, computerization was perceived as an auxiliary function. Later, it turned out that software determines the speed of operations, the quality of processes, competitiveness, and the very structure of business. Programmers became scarce specialists and, for a long period, almost an elite of the labor market.

Now a new technological leap has occurred, and business has once again found itself at a similar point.

The tool already exists. AI is capable of a great deal. But the organizational model for its use has not yet taken shape.

A manager can buy ChatGPT, Claude, Gemini, or corporate access to agents. But having a tool does not answer the main question: how to restructure the company's work around it so that it creates economic results, rather than just pretty answers.

This is precisely why a paradoxical picture is emerging right now: business is cutting the people who are closest to understanding the new computing environment, and at the same time trying to figure out on its own how to apply this environment.

This is not necessarily a long-term strategy. Rather, it is a transitional phase when the old demand has already collapsed, while the new one has not yet received a normal name, budget, and place in the organizational structure.

Programmers also do not see the new market right away

However, it would be a mistake to explain the situation solely by the short-sightedness of business.

Most programmers have worked for decades within an understandable model:

business formulates a task → programmer implements it.

They have gotten very good at answering the question "how to program this?", but much less often have had to answer the questions:

• what even makes sense to program now;
• where a new shortage has emerged;
• what new function can be sold to the market;
• what has become valuable after code became cheaper;
• how the very economics of technical work have changed.

When business said "we need less programming," many specialists heard only the first part: "we need less of you."

But the market may not return for the old work.

It may start buying a completely different technical function.

New demand is forming around the management of AI systems

We are not talking about the profession of a "prompt engineer." That is too narrow and temporary a description.

The new value is broader: the ability to turn AI capabilities into a truly working system for a specific business.

Today, this already includes:

• searching for and verifying high-quality information;
• creating specialized AI consultants;
• programming networks of agents;
• building memory and context;
• connecting models with the company's internal systems;
• process automation;
• designing information loops;
• choosing the boundary between human, AI, and ordinary code;
• quality control and error handling;
• translating AI capabilities into an economic model understandable to business.

In a few years, the specific set of technologies will change. But the economic function will remain: someone needs to turn the new computing environment into the working infrastructure of a business.

This logic is directly linked to the HVQ concept «Natural language is the new programming language, and AI is the new programming environment».

If AI is indeed a new programming environment, then the emergence of specialists who know how to work within it multifold better than the average user is almost inevitable.

The accessibility of a language does not cancel out differences in qualification. Python has been free for decades, but it didn't make every entrepreneur a programmer. Excel is available to millions of people, but it didn't make professional financial analysts obsolete. Linux is open-source, but it didn't make systems engineers unnecessary.

The same thing is happening with AI.

Why this is especially profitable for the solopreneur

The new market will not necessarily be built around large full-time teams.

A company may not need such a specialist full-time. It may need to:

• design the system once;
• integrate AI into a few processes;
• build an information loop;
• set up agents and integrations;
• train employees;
• periodically review the architecture and improve it.

One strong specialist is capable of performing this function for multiple companies at once.

What looks like a reduction in the number of vacancies to a hired programmer may look to a solopreneur like the emergence of a market for fractional technical capacity.

This is directly related to the concept «AI turns a single person into an organization»: a strong specialist gains the ability to perform work that previously required a team.

And with the concept «Abandoning the old market can open a new one»: if the previous service is getting cheaper, trying at all costs to hold onto the old model may be less profitable than an early transition to a new category of demand.

The market is currently between two forms of demand

This best explains the current situation.

The old form of demand is already fading:

"we need a person who writes a lot of code."

The new one has not yet become familiar:

"we need a person who understands how to turn AI, agents, data, and processes into a working system."

Therefore, the current market is easily mistaken for the final collapse of the value of programmers.

But perhaps this is merely the gap between two market models.

Historically, such transitions are especially interesting for solopreneurs. Large companies are slow to change job titles, budgets, and organizational structures. An independent specialist can spot a new function earlier, master it, and start selling it before the market agrees on what exactly it should be called.

The main opportunity here is not in waiting for the return of the old demand for code.

It is in seeing what technical problem business has not yet learned to formulate, but will soon be forced to solve.

Today, management of the new AI environment is increasingly becoming such a problem: information, agents, context, automation, and the boundary between machine and human decision-making.

Demand for traditional programming may already be dropping at a time when demand for new technical work simply does not yet know how to be properly named, bought, and measured.

It is precisely in this gap that the next big market for technical solopreneurs may now be forming.