The essay “I’d rather read the prompt” is, at its core, a defense of human voice against machine-shaped writing. Its author, who says they have seen a great deal of student work, argues that large language model output tends to be verbose, repetitive and aggressively generic. The result, in their view, is prose that sounds polished while saying very little.
The starting point is classroom grading. The author describes encountering answers that appear to have been copied from ChatGPT or a similar model: heavily bulleted, overexplained and oddly eager to echo the assignment prompt itself. That style, the essay argues, is now recognizable enough to be routine, but not distinctive enough to justify a formal accusation in every case.
Rather than focusing on academic dishonesty alone, the piece makes a broader case for originality. The author says they are not writing only for students, but also for bloggers, commenters and researchers whose work has been flattened into a predictable machine voice. The central claim is that a person’s own thoughts are more interesting and more valuable than the output a model can smooth into place.
The article also offers reasons people may have drifted toward machine-generated prose. In a classroom, some students may see assignments as obstacles to clear rather than opportunities to think. In other settings, people may use AI because they want to sound more competent, reduce effort or keep up with expectations that reward volume over substance. The essay does not deny those incentives. It argues that they produce worse writing.
What stands out is the author’s disdain for the rhetorical habits that now signal machine text: the insistence on repeating the prompt, the fixation on tidy bullet points, the sense that the writer is trying to prove relevance rather than communicate. The result is writing that can be technically adequate but emotionally dead.
The author is careful to frame the problem as an expressive one, not just a moral one. The concern is not merely that AI can be used to cheat. It is that handing over language gives up the very thing that makes a piece of writing worth reading: the particularity of a person thinking in real time.
That is why the essay’s title lands as a provocation. The writer would rather read the prompt itself than the model-generated answer to it. In their view, even rough human writing can be more informative than a smooth transcript of a system trying to sound useful. The article’s argument is not anti-technology so much as anti-substitution: if the goal is creative expression, the person should remain visible in the final text.
