I consider myself fortunate to have learned how to code in a pre-AI era—a time when there were fewer abstractions, and I was motivated to figure out how things work under the hood. Now, with AI coding assistants and tools like ChatGPT, I can’t help but feel that the ancient art of problem-solving is slowly being lost. I look around and see what might be the rise of the perpetual junior developer: those who have never truly needed to wrestle with the underlying mechanics of software development.
Yet, as I write this, I realize that this sentiment is hardly new. Every generation worries that the next generation is “losing the recipes” or forsaking deeply held skills. Maybe this is just another chapter in the ongoing evolution of our collective consciousness—a natural part of technological and societal progress. Still, I can’t shake my unease about what I perceive to be happening in the context of software engineering.
When I reflect on how reliant I’ve become on AI to write and correct my code, I feel a twinge of concern that my own problem-solving abilities are beginning to atrophy. This is not merely a theoretical worry: I’m noticing how easy it is for me to reach for an AI assistant rather than push through a problem on my own. If I’m experiencing this, I suspect that junior developers who have never worked without AI tools must be facing an even more profound challenge. After all, if you can just ask a coding assistant to solve a problem, there’s little incentive to learn how or why things work.
In the past few years—ever since AI and large language models (LLMs) became widely available—we’ve heard a great deal about increased productivity and cost savings for businesses. What we don’t hear as much about, however, is the trade-off: what happens when we outsource our brains?
The answer may not be a dire apocalyptic scenario, but we should still pay attention to the long-term implications. Are we raising a generation of “perpetual juniors”—developers who remain in a state of perpetual dependency on AI tools, never learning to stand on their own? Or is this simply the natural next step in the evolution of software development, where problem-solving shifts from manual to AI-assisted?
I don’t claim to have all the answers, but I believe this discussion is critical. Perhaps the best approach is to remain aware of this dynamic and find ways to balance the convenience of AI with intentional learning and critical thinking. In doing so, we can preserve the spirit of innovation that fueled earlier generations of programmers, while also benefiting from the remarkable tools at our disposal.
Author’s Note: This blog post is driven by my own observations and concerns regarding skill atrophy in an era of rapid technological advancement. I hope it sparks thoughtful reflection on how we can nurture true expertise in a world increasingly influenced by AI.