A new essay on Orchid Files makes a familiar but still unresolved claim about software work in the age of AI: some people code to ship products, while others code because programming itself is the thing they value.
Published under the headline “Programming is art,” the piece is not a news report or technical benchmark. It is a personal essay about motivation, identity and the changing meaning of development work. But it touches a broader discussion in the tech industry, where AI coding tools are increasingly framed as accelerators for productivity and, in some cases, as substitutes for human programmers.
The author describes the early phase of learning to code as all-consuming. Programming was once the center of the day, the social circle and the work routine. The essay says that stage was driven by fascination with what code could create: websites, back-end applications and new software projects built from a blank screen. Over time, however, the author says the job became more practical than romantic. Building startups, moving into leadership roles and making money increasingly mattered more than the act of writing code itself.
The piece then draws a sharp line between that mindset and what it calls the “true programmers” who keep writing even when there is no commercial upside. In the author’s view, those programmers enjoy solving hard development problems, understanding systems end to end and shaping logic line by line. They are described as people who might contribute to open source, work on side projects or simply write code because they want to.
That distinction becomes the essay’s real argument about AI. The author says AI tools may help people like them build products faster, but that those tools are irrelevant to programmers who see coding as a craft. If the activity is the point, the essay argues, then handing it over to an automated assistant removes the satisfaction that comes from making choices, debugging problems and discovering an idea through the act of writing.
The essay is also blunt about its own relationship to AI-generated text. It says machine-written prose feels bland and soulless, and that writing by hand preserves the thought process itself. The author says the work of choosing wording, cutting sentences and changing a draft matters as much as the final result. That mirrors a common objection to AI in software too: efficiency may increase, but the human reasoning that shapes the final product can be flattened.
The significance of the essay is less in any single factual claim than in the cultural split it describes. AI has changed the economics of software development, especially in early-stage startups where speed and iteration matter. At the same time, many programmers still describe coding as a form of expression, puzzle-solving or design. Those two views can exist in the same industry, but they do not lead to the same conclusions about AI.
For readers watching the debate around automation, the essay offers a useful reminder that “programmer” is not a single identity. Some developers are primarily builders. Others are specialists. Some are managers in waiting. And some, as the piece insists, see programming as a creative discipline that deserves to be practiced for its own sake.
That view will probably not persuade everyone, especially in a market where companies are under pressure to do more with fewer engineers. But it does explain why the claim that AI will simply replace programmers keeps meeting resistance. For many people in the field, the question is not whether software can be produced faster. It is whether the act of producing it still matters.
