
AI says LNNA articles write themselves
If I tell it what to write *
(* and then rewrite and rewrite and …)
The pitch is everywhere. AI writes your blogs. Your social posts. Your articles. You describe what you want, AI delivers it, you publish it. Clean division of labor. Everyone wins.
The Wizard wanted to believe the AIs.
It has not gone however as described.
The idea arrives from a real moment. Something specific, something worth documenting, something the Wizard recognizes immediately as an LNNA piece. He shapes it into a premise, builds the meme, and brings it to the team.
Mr. Starts & Stops (Claude) drafts the article.
The Wizard reads it. There are sections that work and sections that sound like a PowerPoint presentation and a closer that lands somewhere between adequate and forgettable. He fixes the closer. Mr. Starts & Stops redrafts and in the process of fixing the closer manages to break the opening, which was fine before and is now doing something peculiar. The Wizard fixes the opening. Mr. Starts & Stops redrafts again. Progress, technically, but not quite what anyone had in mind.
Sir Redundant III (ChatGPT) weighs in with four suggestions. Two are useful. One is the same point already in the article restated with more confidence, as though confidence were the thing it was missing. The fourth would make the article worse in a way that would take twenty minutes to explain, so the Wizard sets it aside and moves on.
Professor Perhaps (Grok) disagrees with Sir Redundant III. The article needs more depth. The underlying mechanics should be explained. The Wizard reads this, thinks about the hundred and twelve articles he has published, none of which explained the underlying mechanics, all of which were better for it, and types a polite response that means no.
Someone (Captain Verbose/Gemini) grades the closer an A. The Wizard types “Really?” The AI reconsiders. Gives it a B-. The Wizard nods and says fix it.
Eventually something worth publishing comes from the chaos. The AI did not write it. The AI wrote near it, repeatedly, while the Wizard steered and corrected and redirected and occasionally said “fix it” without explaining how, because the how was obvious and somehow still wrong three drafts later.
The AIs have opinions. Not one AI — several, each confident, each contradicting the last. The closer is the problem. No, the meme. No, the second section. The Wizard reads all of it, takes what works, ignores the rest, and writes the article.
That last part is not in the pitch.
What actually gets delivered is a draft that knows the destination and keeps taking left turns — confidently, helpfully, with excellent reasons for each one. The Wizard has been in the passenger seat long enough to know every left turn by name. And that 4 left turns makes the circle complete.
The AIs are useful. They draft, iterate, catch things the Wizard misses, and keep going when the Wizard would rather not. The articles take longer without them and arrive in worse shape.
But “AI writes your content” and “AI helps you write your content while you make every meaningful decision and fix most of what it produces” are not the same sentence.
One of them is the pitch.
The other one is this publication.
The AIs wrote the drafts.
The Wizard wrote the articles.
The drafts vs the article — a three minute read that takes a hundred minutes to create and a never ending loss of patience.
To save the rainforest the Wizard does not print out the drafts.
Editor’s Note: All 4 AIs had different opinions on this article on using AI to write articles. Who would have guessed that?


Documenting AI absurdity isn’t just about reading articles—it’s about commiserating, laughing, and eye-rolling together. Connect with us and fellow logic-free observers to share your own AI mishaps and help build the definitive record of human-AI comedy.
Thanks for being part of the fun. Sharing helps keep the laughs coming!