About the publication
About
We publish about artificial intelligence for people who have to make decisions about it, not just opinions.
There is an enormous amount of writing about AI and startlingly little of it is useful a month later.
Most of it falls into one of two piles. The first is breathless: every release is a revolution, every benchmark a milestone, every limitation a temporary inconvenience. The second is dismissive: it is all autocomplete, it is all hype, nothing has changed.
Both piles are easy to write and neither helps you decide whether to put one of these systems anywhere near your work. So we started a third pile. Beyond the Prompt exists to describe these tools accurately — what they do, what they cost, where they break, and what genuinely changes when you build a habit around one.
We are independent. Nobody pays for coverage, no vendor sees a draft before you do, and when a tool we like does something indefensible we write that down too.
§ 01 — What we make
Three formats, one argument.
Articles
The written record. Essays, field guides and the occasional post-mortem, published when there is something to say rather than on a schedule.
Podcasts
The argued version. Some weeks it is the two of us taking a term apart; some weeks it is a practitioner explaining what actually happened when they shipped.
E-Learning
The taught version. The AI Fluency Programme turns everything above into structured practice, so you leave with a working system instead of a reading list.
§ 02 — Editorial standards
Four rules we hold ourselves to.
Say the limit out loud
Every capability claim has a boundary. A piece that describes what a system does without describing where it stops is marketing, and we do not publish it.
Show the working
Numbers get sources. Claims get citations you can actually open. When we cannot verify something, we say so in the sentence rather than in a footnote.
Assume the reader is busy, not simple
No talking down, no throat-clearing, no thousand words of context before the point. Respect is measured in paragraphs removed.
Correct in public
We get things wrong. When we do, the correction goes at the top of the piece and into the weekly letter, not into a quiet edit at midnight.
§ 03 — How it is made
We use the tools we write about.
And we show you whereIt would be strange to publish about AI-assisted work and hide how this publication is actually produced. So here is the whole pipeline.
Brief
A human decides what is worth writing and why now. This step has never been delegated.
Draft
A writer agent produces a structured first draft against the brief, with sources attached.
Edit
An editor pass cuts for clarity and voice. A human then cuts again, usually harder.
Verify
A fact-checking pass flags every claim that carries a number, a name, or a date.
Publish
A person reads the whole thing top to bottom and puts their name on it. Always.
Machines get us to a solid draft considerably faster than we managed alone. Everything that makes a piece worth reading — the argument, the cuts, the decision to publish — still happens after that, and a named human is accountable for all of it.
§ 04 — Get in touch
Corrections, arguments, and guests.
If we got something wrong, tell us and we will fix it in public. If you have shipped something real and can talk honestly about what broke, we would like you on the show.