The total developer, or the age of versatility

In April 2025, Justin Searls published a widely discussed essay on the end of the specialist developer and the rise of the total developer: someone who codes, deploys, writes, collaborates with AI, and navigates multiple stacks. Here is my translated commentary.

The central argument

For Searls, the effect of generative AI on software work is not the disappearance of the developer. It is the disappearance of internal boundaries. Front end, back end, devops, technical writing, product analysis: those silos were already fragile; AI makes them obsolete.

What follows is my adapted translation of his essay: it is Searls speaking.

The world of software development is in the middle of a deep shift. Generative AI made a thunderous entrance, shook up the job market, and rendered certain skills obsolete. At a breaking point like this one, there are winners and losers. From where I stand, total developers — minds that combine technical skill with a sharp sense of product — are the big winners of this new era.

This is not a hunch; it is an observation. Over the past few months, I have watched developers with even a minimal understanding of the business burn through to-do lists at a dizzying pace. Code-generation tools recently crossed a major capability threshold. I experienced it myself: in two days, I completed the equivalent of two months of work.

How? By holding a clear vision for the product, maintaining rigorous technical standards, and letting AI do the rest. If you can fit critical thinking, good taste, and solid technical chops into one and the same head, these tools can multiply your productivity. Split those roles apart — separate the person who thinks the product from the person who codes it — and productivity collapses. The cost of communication simply becomes too high.

The end of an era: not everyone can win

And yet most of the companies and developers I see around me seem headed straight for failure. For decades, we were taught that “product” and “engineering” roles had to be strictly separated. Suggesting that a single person could drive both design and technical execution would sound absurd to many. Even companies that have understood what is at stake struggle to recruit these rare profiles. Their outdated job descriptions and salary bands are simply not built for such chameleons.

There is a sense of urgency in the air. A few months ago, the best developers played the violin. Today, they conduct the entire orchestra.

What I learned at Google

This subject has fascinated me since the start of my career, so forgive me if I indulge in a certain satisfaction when I think back to this anecdote.

In 2007, while I was still in college, I passed a phone screen with Google that earned me an all-expenses-paid trip to the fabled Googleplex. I flunked the interview miserably. I remember a group session with a renowned engineer, introduced as the inventor of BigTable. At one point he said: “one of Google’s great strengths is that the engineering track and the product track are completely separate.”

Having always wanted to combine my technical and non-technical skills, I felt personally challenged. Unable to keep my mouth shut, I raised my hand and asked: “But what if you’re a hybrid profile? What if you believe it’s essential for everyone to engage with both technology and product?”

He looked me straight in the eyes and told me I was not cut out for Google. The recruiter broke the awkward silence by walking us to the cafeteria. She suggested I try the famous ice cream sandwiches. My appetite, for some reason I could not quite explain, was gone.

Since then, more and more companies have adopted this “dual-ladder” system imported from Silicon Valley: the technical people on one side, the “idea people” on the other.

What separates people

Back to winners and losers. Some have gone “AI-first” and thrown out everything they knew. Others denounce AI as a passing fad, a bubble about to burst. It has become a subject I broach with caution, the way you would ask someone about their politics. For months, I wondered why it was so hard to guess a developer’s opinion on AI.

That is when I remembered that day at Google. And I realized that the developers I know who have embraced AI tend to be more creative, more results-oriented, and blessed with good product taste. The detractors, by contrast, are more likely to code for the sake of coding, to wait for crystal-clear instructions, or to want their work to fit into a routine. The former feel liberated by these tools; the latter feel threatened by them.

Today, when I look at who is doing well and who is struggling, the willingness to be versatile is the best predictor of success.

A glimmer of hope

If you have read this far and the message has not resonated, maybe it has stirred up your fears about AI. Or maybe you think I am talking nonsense. Either way, whether your organization is not built for this new era or you do not yet see yourself as a versatile developer, this section is for you.

To leaders: hire a good agency

My goal is to find a new expression to describe this transformation, but we have always had a word for these total developers: consultant.

Not because consultants are geniuses. But because, as I learned at Google, for these developers to do their best work, they need to exist outside classic organizational structures. It is no surprise, then, that my favorite consultants were among the first to adopt AI. Not because it is a trend, but because their natural disposition is perfectly suited to these new tools. They are the first to see AI’s potential to improve how we build software.

To developers: congratulations on your new job

I know many developers are frightened by these changes. Yes, AI is being used as an excuse for layoffs. Yes, the way the foundation models were trained was questionable. Yes, the hustle bros are spouting nonsense. All of that is true, and yet it changes nothing. Your job as you knew it is gone.

If you want to keep getting paid, you may have been told to “move up the value chain.” Let me put it more plainly: figure out how your employer makes money and position yourself directly between the company’s bank account and your customers’ credit card details. The more sentences it takes to explain how your work makes money for your employer, the lower down the value chain you are and the more worried you should be. There is no other option than to step out of your comfort zone.

Take these new tools seriously. At first, you will recoil in horror. You will find that they are genuinely bad at replacing you. That they make mistakes constantly. Your new job begins with learning to master them. You will gradually work out how to extract something that resembles what you would have done yourself. Once past that hump, the challenge becomes scaling it up.

Your new mission is to create maximum impact. Do not wait to be told what to do; identify the opportunities. Your manager will appreciate that you are looking for ways to make a bigger impact. Listen for the major challenges the business is facing and volunteer to be part of the solution.

All of this would already have been good advice ten years ago. It is not rocket science; it is just deeply uncomfortable for a lot of people.

What I would add

Searls’ analysis applies to developers, but it extends far beyond them. Support functions - communications, HR, legal, finance - are going through the same shift. The specialist becomes an equipped coordinator. That is exactly the terrain of AI Activation Sprints.