AI;DR: Every New Post Now Points You at One a Person Wrote

By: on Oct 18, 2026
A black and silver fountain pen resting on a sheet of paper covered in handwritten cursive

Scroll up. There's a green box above the photo on this post. That's the new thing.

Every new post on this blog now opens with a line called AI;DR. It's a play on TL;DR. It says the post was written with AI assistance, and then it points you at one article on the same topic that a person wrote. If you'd rather read that, go read that. No hard feelings.

Takes about two minutes to explain. Here's the whole thing.

What Was Already Here

Every post on this site already ends with a footer saying it was made with human and AI-assisted workflows. That footer links to my standards for AI-assisted blogging and to the post about how the workflow runs. The short version of those two: I bring the material, the model helps with the shape, and I'm responsible for every claim.

The footer is honest. It's also at the bottom, and it only talks about me.

AI;DR moves the disclosure to the top, and it gives you somewhere to go.

What the Box Does

  • It gives a reader a real alternative. Some people don't want to read AI-assisted prose, full stop. That's a fair preference. A link to a good human-written piece respects it more than a disclaimer does.
  • It sends credit and traffic to someone who wrote the thing well. Most of my posts stand on somebody's earlier work. The box names one of them where you'll actually see it.
  • It keeps me honest about what my post adds. If I can't say what my post has that the linked article doesn't, I probably shouldn't publish mine.

The Rules

These live in the blog's instructions, so every draft gets checked against them:

  1. The article has to exist, and the title, author and URL have to be checked by opening the page. Never from memory.
  2. Same topic. Not "AI in general" bolted onto a post about NixOS.
  3. A person clearly wrote it.
  4. It's worth reading even if it disagrees with my post. Disagreeing is fine. Sometimes it's better.
  5. No good match? Leave the box out. A weak pick is worse than none.
  6. New posts only. I'm not going back through the archive to bolt one onto every old post.

It's one block of frontmatter. The layout renders the box:

aidr:
  title: "Friends Don't Let Friends Use Ollama"
  url: https://sleepingrobots.com/dreams/stop-using-ollama/
  author: Zetaphor

Sometimes one article isn't enough. Make it a list and the box joins them into one sentence. Two is usual, three is the ceiling:

aidr:
  - title: "AI;DR (AI; Didn't Read)"
    url: https://www.rickmanelius.com/p/aidr-ai-didnt-read
    author: Rick Manelius
  - title: "Why A.I. Isn't Going to Make Art"
    url: https://www.newyorker.com/culture/the-weekend-essay/why-ai-isnt-going-to-make-art
    author: Ted Chiang

Three Real Picks

Dude, Just Use llama-swap points at Zetaphor's "Friends Don't Let Friends Use Ollama". That essay already recommends llama-swap. It made the case before I did, so it gets the top slot on my post.

The tripwire post points at Geoffrey Huntley's post about fixing things with a Nix overlay. His framing, that you can change any tool on the box with an overlay, is where that post starts.

The Disco Elysium voices post links more than one piece in its box. The game's voices and Robert Kurvitz's novel Sacred and Terrible Air (no official English translation) are each worth a read.

An agent can find candidates. Picking the one at the top is my job.

This Post's Box

This one points at two. The first is Rick Manelius's "AI;DR (AI; Didn't Read)", the post that took the term to the top of Hacker News (he credits the coinage to a tweet by @seclilc). His policy is blunt: if you didn't care enough to review and edit the AI output, he won't bother reading it. That's the reader this box is for, so his post goes first.

The second is Ted Chiang's "Why A.I. Isn't Going to Make Art" in The New Yorker. It argues that a model can't make the many small choices that writing is made of, which makes Chiang the strongest case against a post like this one. That's rule 4.

Header photo by Álvaro Serrano on Unsplash.

Content on this blog was created using human and AI-assisted workflows described in my standards and workflow posts. Original ideas and editorial decisions by Justin Quaintance.