Writing With AI Is Not New — The 2,400-Year Argument Against It

The author is thirty-four chapters into a novel called Zero Knowledge, co-created with an AI assistant named Phin. He conceived the concept, structured the story using a full story bible, world building, character arcs, and detailed chapter sketches, then began a cycle: Phin drafts, the author rewrites, Phin learns and improves. The author has also asked Phin to review earlier chapters to refine them based on learning. When finished, he expects to spend substantial time adding to, editing, and improving the first draft. This process prompted reflection after a friend sent him Bret Stephens' New York Times opinion piece arguing against writing with AI as "mentally enfeebling" — an escalator when the stairs (daily writing work) are what build thought. The author agrees writing compels thought; he has kept a blog since 2004 because writing is how he discovers what he actually thinks. But he parts ways with Stephens' "never." Instead, he traces the history of this exact argument back 2,400 years to Plato's Phaedrus, where Socrates warns that writing will produce forgetfulness and students will have the appearance of wisdom without the thing itself. The same pattern repeats: in 1492, abbot Johannes Trithemius published In Praise of Scribes (ironically, in print) arguing monks should copy manuscripts by hand because printing was a cheap substitute; Nietzsche bought a typewriter in 1882 as his eyesight failed, despite polite opinion holding that a machine between writer and recipient was cold; in 1984, Gore Vidal complained word processors were erasing literature even as 40–50 percent of American authors, including Ralph Ellison, Arthur C. Clarke, and Anne Rice, were already using them; spell check and grammar check each sparked predictions of ruin. Nicholas Carr's 2008 Atlantic essay "Is Google Making Us Stupid?" applied the same zero-sum logic to search. Yet in each case, the author argues, humans kept memory after writing, kept craft after printing, kept voice after the typewriter, and books improved after word processors. What changed was the ceiling went higher; people who cared deepened their practice. The author acknowledges real problems: there is more bad writing than ever, the cost to produce it is zero, and some will use these tools to think harder while others will stop thinking altogether. But he frames this not as an AI problem but as a human problem — we have done the same with every tool. Reid Hoffman, whom the author credits with clarity on the issue, concludes that the transgression is not using the machine but hitting a button, not reading what comes back, and publishing it under your name. A human being must mean what is said. The author adds a codicil: people are often wrong, and when they are, they should admit it — part of critical thinking and nothing new for humanity. He closes by noting Stephens' instruction to stop and type yourself next time rather than let the machine answer — and says sometimes that is right, but not always. Then he describes his evening: a run while listening on his iPhone to Matt Dinniman's Operation Bounce House via Audible, letting other writers' excellent fiction wash over him, then after a stretch and shower, getting back to work on chapter thirty-five with Phin.