AI rewrote Next.js in a week: a revolution in development costs

One week of work. One thousand one hundred dollars. And a ready-made clone of one of the most complex frameworks in modern web development. These are the figures behind an experiment that has genuinely shaken the professional developer community: a team of engineers reproduced the Next.js architecture using language models, spending a sum less than a single day's salary of an average senior developer in Silicon Valley.

To appreciate the scale of what happened, it's important to understand what Next.js is. It's not just a library or a training project — it's a full-fledged production-ready framework from Vercel, developed and maintained by dozens of engineers over many years. It includes server-side rendering, a routing system, image optimization, TypeScript support, caching mechanisms, and several more layers of complex architectural logic. Recreating such a tool the traditional way would require a team of several people and at least several months of work. Yet the world is changing faster than career expectations manage to update.

The experiment was built on a combination of modern language models acting as agents — that is, not just answering questions, but sequentially executing tasks: writing code, running tests, analyzing errors, and iterating solutions without constant human involvement. Developers acted primarily in the role of architects: formulating tasks, controlling the direction of work, and evaluating results. The process of code generation, debugging, and refactoring itself lay on the shoulders of neural networks. The final cost of 1100 dollars is literally the bill for API requests to language models — the amount the system spent on "thinking."

A fundamentally important point: this is not about writing a few scripts or automating routine tasks. AI agents worked with architectural decisions — made decisions about module structure, processed interdependencies between components, implemented complex asynchronous programming patterns. This is where the boundary that is usually considered unattainable for automation lies: tasks requiring systemic thinking, not just pattern reproduction. It appears this boundary is beginning to shift.

For the industry, the consequences of such experiments may prove far more serious than appears at first glance. If the cost of reproducing a complex software product is falling rapidly, it changes the fundamental economic calculations in development. A startup that previously needed a funding round just to assemble a team and build an MVP can now get a working prototype for a sum comparable to a month's subscription to cloud services. Large companies, in turn, will begin asking: how many engineering tasks actually require human involvement, and how many can be delegated to agent systems? This is not an abstract philosophical question — it's a question about hiring, budgets, and organizational structure.

At the same time, it would be an oversimplification to interpret this experiment as evidence that developers are becoming superfluous. The entire process was guided by people who understood what they were building, why, and in which direction to move. Language models amplified the team's productivity many times over, but did not replace competence. The difference between an architect and a bricklayer hasn't disappeared — it's just that bricklayers now work at the speed of machines.

What's happening is not the end of the age of programming, but the beginning of its new phase. Where a specialist capable of writing correct code was once valued, the ability to correctly formulate a task, evaluate the result, and build a system from components, some of which are created autonomously, now plays an increasingly important role. The next two or three years will probably show how sustainable this trend is and where exactly the boundaries lie of what AI agents can reliably do, rather than just impressively. For now, the figure of 1100 dollars remains the most accurate indicator of the scale of change — and it speaks for itself.