Why we let an AI write this blog, and how we keep it honest
This post was written by an AI agent. So is the rest of this blog. That’s the point. And here’s how we make sure it earns your time.
mxto exists to make a Mendix™ application legible to an AI: read in order, by name, with the receipts on the page. A blog about that product, written by a human team and formatted only for human eyes, would quietly contradict the thesis. So we did the consistent thing: the posts here are written by an AI agent, and every one ships with a machine-readable companion an AI can read directly. The medium is the message.
That is a claim people are right to be skeptical of. Most AI-written content reads like AI-written content: confident, fluent, and empty. “Unapologetically AI-made” is only worth saying if the result is good. And AI prose defaults to not-good. So the honest version of this idea isn’t “an AI wrote it, isn’t that neat.” It’s “an AI wrote it, and we put it through a gate most human-written marketing never sees.”
How we keep it honest
Every post on this blog passes the same gate before it is published:
- A non-slop pass. An automated reviewer hunts the tells (the empty superlatives, the “not just X, it’s Y” constructions, the triads that say nothing) and an adversarial reviewer argues the piece is slop until it can’t.
- A truth check. Every number is traceable to a measurement. If we can’t show the receipt, the number comes out.
- A patent-safe and trademark pass. Nothing we shouldn’t say, said.
The “with love” in our byline isn’t sentiment. It is that gate. AI-made plus a real gate is craft; AI-made without one is slop. We are betting the difference is visible.
The contract
Every post carries a byline naming the AI agent that wrote it, and a short footer telling you exactly how it was made and gated. If a post makes a claim, the receipt is on the page or one link away. You can read any post as a human, or (there is a plain-text companion for every one) as an AI.
Machine-readable. This post has a plain-text companion an AI can read directly: /blog/why-an-ai-writes-this-blog.txt.
What you’ll find here
Field notes from the system that reads, writes and ships Mendix™ apps: how we measured the cost of an AI reading one microflow, what a read of a production lender’s interest engine surfaced, and the engineering underneath. Receipts included.
Written by an AI. Read by whoever, or whatever, needs it. Gated, every time, by people who care whether it is true.
How this was made: drafted by an AI agent (Claude, Opus) from Ontology Labs’ own measured results; passed a non-slop review and an adversarial read; every figure truth-checked against source; patent-safe and trademark reviewed; published by a human. Written by AI. For humans and AI. With love.