There is something, not novel but persistent, that has led me to ponder: what future awaits the programming profession? Are we on the cusp of a new era for this vocation?
In my observations, even students in elementary and middle school are no longer focusing on the fundamentals of programming when first introduced to coding. Instead, they immediately turn to artificial intelligence. While AI offers remarkable capabilities, it cannot replace core programming skills. Many children and adolescents today cannot write a single line of code. It is troubling that a twelve- or fifteen-year-old can become a “millionaire developer” by merely conceptualizing an app idea, feeding it into an AI tool, and accepting the finished product on the other side. The emerging generation, I believe, are not programmers; at best, they are idea generators.
Educational institutions bear much of the responsibility for this shift. I do not wish to impede technological progress—it is the natural course of our times—but as schools, colleges, and universities increasingly rely on AI to handle coding assignments, the foundational discipline of programming risks disappearance. Is it appropriate for individuals to adopt the title “programmer” when they perform no programming whatsoever?
If blame is to be assigned, it rests squarely with our educational system. By normalizing AI-generated code in their curricula, higher education institutions are eroding the very essence of programming. I have discussed this concern with two colleagues—both holding multiple master’s degrees in AI and IT. One specializes in artificial intelligence technology, and the other serves as a global lecturer and researcher in quantum computing and programming. They argue that this transformation is an inevitable and positive aspect of our field’s evolution. I respect their viewpoint and do not intend to obstruct global progress, but I maintain that we must preserve the critical thinking and technical rigor that define true programming expertise.
My own experiments with AI-assisted development have revealed its current limitations. AI systems frequently introduce errors, exhibit limited contextual memory, and demand constant re-orientation. Far from accelerating projects wholesale, these tools can decelerate progress and introduce new challenges. While AI may soon mature enough to undertake substantial development tasks, it has not yet reached the point where it can reliably manage an entire project or even significant portions of one. Anyone who believes that mastering AI prompts alone constitutes programming proficiency has, in my view, misunderstood the discipline.
I invite the community to weigh in: how do you perceive AI’s role in software development? Do you integrate AI into your workflows—partially or fully—and what have been your experiences and outcomes?