How long does it have to be?

As I neared the end of my PhD and was editing and revising my dissertation, I was asked one question more than any other: “How long does it have to be?”

It’s not a stupid question, but it’s one born from ignorance. Ignorance about writing, broadly—and academic writing, specifically. Ignorance about scholarship and the purpose and value of academic work. And ignorance about the ways humans learn and demonstrate mastery of a subject.

My committee never once mentioned length as a requirement (the version I submitted at my defense was 287 pages); my committee chair mentioned it once, to tell me I needed to pare back my second chapter (and even this was less “it’s too long” than “you’re moving too far from the focus of this chapter”). Instead, our conversations centered on clarity and organization, on the relationships between ideas, and on the ways my work would contribute to the field of literature. And the hours I spent writing and revising each week were rarely, if ever, dedicated to writing more. Instead, they were almost always spent on writing better.

Everyone who has worked in and around academia for any amount of time is shrugging right now. All of us, even the wannabes, can and do write far more than we need—or often should. My colleagues and I joke about cutting roughly half of what we write and how it makes us feel: the frustration, the reluctance, the mourning, the acceptance, and the satisfaction. Everyone knows the Faulkner quote. All of this is to say that (unless you’re talking cash, and even then…) “more” is rarely “better.”

“How long does it have to be” is analogous to another phrase, one I’ve come to despise in my seven years in the corporate workforce: minimum viable product. Nothing speaks more directly to the ways corporations and late-stage capitalism have enshittified the world. Under the guiding ethos of “minimum viable product,” there are few, if any, incentives for excellence, and no one wants to discuss how we might improve things. Good has been replaced by merely sufficient—and, more often, insufficient.

There is an implicit argument here to be made about the output of LLMs, which seems to, across every discipline, represent the Platonic ideal of mediocrity. LLMs (a now-multi-trillion dollar industry) are trained and optimized to produce something acceptable, generic, and sufficient. They are the technological culmination of an industrialized logic that work must be made more efficient, above all. They are the product of an increasingly incurious world interested only in questions like “How long does it have to be?” or “What’s the minimum viable product?”

But what happens in a world where the first question we ask ourselves is “What’s the least I can do?”