Are generalist skills a new trend or a real necessity?

Over the past two decades, IT departments have been split into narrow specializations. In the case of web development, this meant frontend and backend. To put it very simply: the part of the application you can see, and the part you can only feel because it runs under the hood. Despite this, we can still find rebels who, defying a polarized IT industry, learned to build systems on both fronts (or rather, “ends”).

By working on both user interfaces and server-side systems, these developers gained the experience to build solutions spanning from the database all the way to the colorful buttons on a webpage-essentially becoming “One-Man Bands.” The market initially dubbed them Fullstacks, only to curse them shortly after. The worlds of front and backend evolved so rapidly that keeping up and remaining a top-tier expert in both would require an elastic 24-hour day and a lake of coffee. Instead, the community doubled down on narrow specializations and expert knowledge.

Today, the evolution of languages and frameworks is slowing down, and ideological debates over architecture are losing momentum. Due to-or thanks to-the AI boom, we are witnessing the blurring of lines between specializations, and we are once again valuing generalist skills that allow us to operate at every stage of software delivery. Since the code itself can now be written by artificial intelligence, a systemic view of the product, rather than just a slice of it, becomes the key to success.

Questions widely discussed today include: When will the lines blur between those who define what needs to be built and those who actually write the software? Are coding skills still necessary (and at what level)? Are the so-called Product Builders just the new Fullstacks, or are they the “future developers”?

Currently, IT professionals are either deepening their core programming skills or chasing frequent tool changes that should support their work rather than completely replace it. The way we use language models is shifting so fast that knowledge from just 2–3 months ago is often frustratingly outdated. Many shakeups still lie ahead, and one can’t help but wonder: what will happen when the market’s AI bubble bursts? Will the tools offered today (LLMs) remain as efficient and financially accessible?

Will Product Builders, whose primary skill is leveraging AI, be able to handle the constraints of advanced tool providers? Will a limited understanding of the programming paradigm itself be enough when working with less capable local models? Or will the market split permanently into generalists and specialists?

---

Paweł Nejczew

This blog post was translated by Gemini AI.
Link to original post