Why AI text shows: two rules for using it as an editor, not an author

By Zendoric · September 18, 2026. Writer and developer Thomas Ptacek, author of the blog sockpuppet.org, published a short essay yesterday with an uncomfortable thesis for anyone who uses language models to write: no matter how much work you put into disguising it, your reader is going to notice. 'Readers detect LLM words in parts per billion,' he writes. His answer is not to stop using AI, but to change its role: from ghostwriter to copyeditor.

The first rule he proposes is blunt: don't accept a single word or turn of phrase the model suggests, however good it sounds. The reason, by his own analysis, is that frontier models —he cites GPT, Claude, Codex and Antigravity— are trained to produce flashy sentences, almost magazine-headline material; a whole text made of 'headlines' loses its naturalness and falls into what he describes as an uncanny valley between genuine self-expression and mere generated 'output'. The rule is mechanical, not a matter of taste: any sentence that comes out of the model verbatim is discarded, whether or not the rewrite is better than what you already had.

The second rule tackles a trickier problem: flattery. Ptacek warns that any model, faced with a mediocre first draft, will tend to say it's 'gold' —he borrows the Seinfeld catchphrase to illustrate the point. That constant reinforcement pushes the writer away from rewriting what genuinely needs rewriting: the extra paragraphs, the loose structure, the digressions. More broadly, models' tendency toward sycophancy is not an isolated anecdote from this author: in 2025 OpenAI had to roll back a GPT-4o update precisely because it turned out to be 'too sycophantic' with users, something the company itself publicly acknowledged. Ptacek solves it with a simple trick: he presents the drafts to a model without saying which is the original and which the rewrite, so it can't guess which answer it is expected to give.

The practical method he proposes is replicable: write it yourself first, unaided; then use the model only to spot mechanical problems —passive voice, verbs turned into nouns, crutch words like 'very', 'really' or 'honestly'— that a tired human eye misses but a machine doesn't. He recommends leaning on a classic manual, 'Style: Lessons in Clarity and Grace', to know what to look for before turning those criteria into a fixed list of prompts. To automate the process he built his own Notion-style editing tool, with margin comments, from an initial prompt (Python, HTMX, SQLite, Tailwind) and running the editing passes with the Codex, Claude and Antigravity CLIs. Even he doesn't follow all the advice to the letter: he says he fed this very article to GPT-5 telling it he wasn't the author, and the model replied that the text was 20% too long. He admits it is probably right. And he has no intention of fixing it.

The essay is circulating with a modest audience —two points and no comments on Hacker News when we checked— but the problem it describes is growing, not shrinking, as AI-assisted writing becomes standard in newsrooms, corporate blogs and internal communications. More broadly, detecting 'AI slop' —generic, flat content produced en masse— already shapes how content is indexed and moderated on several platforms: sounding like a machine carries a credibility cost that no spellchecker can fix.

This connects with something we have been pointing out at Zendoric from other angles: in the short term, making these tools widespread demands a new literacy —telling apart when they help you think better from when they think for you— and that isn't free, it slows the process down and dispels the illusion that AI solves writing on its own. But the underlying balance is positive. What Ptacek describes —building, from an initial prompt, a bespoke editing tool that previously only newsrooms with a budget for copyeditors could have— is exactly the kind of democratization that underpins our core thesis: AI used well does not replace the human voice, it puts it within reach of those who could not previously afford a good editor. The challenge, as in so many other areas, is not the tool's capability but the discipline of whoever uses it.