How Much Precision Is Really Enough? A Look at the Precision of Writing in an Age of Unlimited Judgement

Large language models have turned the supply of critique from something scarce and costly into something abundant, instant and almost free. As a result, the list of objections these models raise against any text is structurally endless and never reaches saturation. This article argues that the common question, ‘How much precision is enough?’, is wrongly framed. Contrary to popular assumption, precision and correctness are not actually in dispute. What genuinely needs to be settled is the degree of stringency: the effort a writer must spend to secure those two goals. Drawing on the sceptical tradition (the Agrippan trilemma), the philosophy of science (the Duhem–Quine thesis), the sociology of science (Collins’s experimenter’s regress), and cognitive economics, the article argues that stopping revision has always been a conventional decision, not the discovery of an objective point. The theoretical core of the article is this claim: the agnotology literature has already documented the demand for unlimited stringency as a recognised technique for producing ignorance, and what happens today is an emergent, agentless form of that same technique. The article closes with a practical four-question criterion and a distinction among five functional domains of writing.