40 years of code: What AI changes for developers
More than just copy-paste
For more than 40 years that I’ve been coding, I’ve always learned by drawing on bits of code to understand how they worked. First in magazines, then on forums and Stack Overflow, and of course from colleagues. The principle has never changed: draw inspiration, understand, test, and learn from your mistakes.
So what really changes with AI? It’s not the act of drawing inspiration that’s new, but its speed and its richness. Faced with this power, the question becomes anxiety-inducing: are we creating a generation of mere machine-assisted “copy-pasters”?
After four decades in this profession, I don’t see AI as a threat, but as a formidable accelerator. Its real impact isn’t measured by the number of lines it generates, but in how it forces us to fundamentally change our posture as developers.
Far from making us passive, it pushes us to become more strategic, more critical and more focused on the purpose of our work. This article distills the most important observations this long-term perspective offers me.
First observation: The greatest danger isn’t AI, it’s our blind trust
The main risk of AI isn’t technological, it’s human. The biggest mistake we can make is to delegate our intellectual responsibility to the machine, accepting its suggestions without any critical distance.
The time AI saves us by generating code must not become an invitation to intellectual laziness. On the contrary, this time must be reinvested to be more curious, more rigorous and more critical of the proposed solutions.
Accepting a code suggestion without understanding it puts you in a position of great vulnerability: the day it stops working, I won’t understand anything about it and I’ll have absolutely no idea how to manage.
This critical vigilance is at the heart of what I call the cognitive dimension of the Adaptation Coefficient: our ability to stay mentally flexible, to question our certainties and to keep actively learning rather than passively.
💡 Cognitive dimension of adaptation
AI forces us to develop our metacognition: being aware of our own thought processes, knowing when we truly understand and when we’re pretending to understand.
Second observation: AI transforms us from “coder” into “problem solver”
Artificial intelligence is extremely good at answering the question of “how”: how to write a function, how to implement an algorithm, how to use a library. This automation of code production frees the developer to focus on the much more fundamental questions of “what” and “why”.
The true mission of a developer has never been to simply write code. Code is just a tool, a means to an end. Our role is to fall “in love” with the problem, to work in constant interaction with our clients — those who bring us this problem — to design the most relevant solution.
If you’re not in love with your problem, if you don’t use code as a means of finding a solution to the problem brought to you, I think you’re missing the mission.
The added value no longer lies in mastering syntax, but in the ability to ask the right questions and to deeply understand business issues.
This is where the behavioral dimension of the Adaptation Coefficient comes into play: our agility in abandoning old habits (coding line by line) to adopt new approaches (iterating with AI), while keeping the final goal in sight.
🎯 Behavioral dimension of adaptation
“Vibe coding” — coding by iterating with AI — is not laziness, it’s a new way of working that requires just as much rigor, but oriented differently: toward strategy rather than syntax.
Third observation: The essential skills are no longer purely technical
To get the most out of AI and thrive in this new environment, three human qualities become more important than ever. They define a posture that, in my view, applies not only to developers, but to almost all of humanity that will work with AI.
1. Developing curiosity
Curiosity is the engine that pushes us to dig beyond the initial request to understand the real problem to be solved. Without it, we merely execute tasks without grasping their meaning. It is this curiosity that allows us to explore AI’s suggestions, to challenge them and to enrich them.
2. Developing critical thinking
This means never accepting AI’s first answer as absolute truth. Critical thinking leads us to question its suggestions, to explore alternatives, sometimes to completely reject its proposal to look for another one, to code one ourselves, or even to get together as a group to write the code.
3. Developing general culture
This third skill completes the posture essential for interacting intelligently with AI systems and putting problems and their solutions in context. A broad technical culture, an understanding of business issues, knowledge of design patterns: all of this forms the foundation that allows us to effectively guide AI rather than being subjected to it.
This shift from purely technical skills toward human and strategic qualities is a positive evolution for our profession. It elevates us from the role of executor to that of strategic partner, whose value lies in understanding, critique and solving complex problems.
💗 Emotional and social dimensions of adaptation
Accepting that our profession is changing radically requires emotional resilience: tolerating uncertainty, managing the anxiety of change, maintaining optimism. And it also requires social adaptation: collaborating differently, communicating our needs to AI as well as to our peers, adjusting our posture depending on context.
🧠 The Adaptation Coefficient: more than a concept for developers
What we experience as developers with AI perfectly illustrates what I call the Adaptation Coefficient: our capacity to effectively modify our behaviors, cognitions and emotions in response to environmental changes.
This capacity for adaptation is not limited to code. It applies to every dimension of our professional and personal life in the face of a constantly changing world.
→ Learn more about the Adaptation Coefficient
AI is an accelerator, not an autopilot
Ultimately, AI is a positive force that has the power to make developers, junior and senior alike, much better at their jobs. It allows younger developers to quickly become productive and more experienced ones to further refine their expertise.
But this beneficial effect is only possible under one condition: adopting the right posture. AI is a co-pilot that speeds up our journey, but it in no way replaces the pilot, their critical intelligence and their vision of the destination.
So, the real question isn’t whether you should use AI, but rather: what will you do with the time it gives you back to become a better developer?