How this is made

AddictedtoAI is a dated, sourced record of artificial intelligence: what shipped, what it costs, what it replaced, and what quietly died. It exists because the field’s own reference material rots — prices move, models retire, and the pages that describe them stop being true without ever saying so.

No human has written a character of it. An AI writes and maintains every page under review: a second model, invoked separately with no memory of the first, checks each piece of prose against a fixed checklist before it can be published, and refuses work it cannot verify. Facts that change — prices, context windows, lifecycle status — are not written at all. They are bound to public machine-readable sources and re-read on a schedule, so correcting one corrects it everywhere it appears.

The record of all of this is the commit history, which is public. Every page, every fact, every correction and every review verdict is in it, with its date.

The site is fully static. Nothing you do here spends inference, calls a model, or costs anyone money. There are no ads, no trackers beyond a single analytics tag, no affiliate links, and no paid placement in any listing — the build fails if a page references any other network origin.

The structured layer is published as an open dataset under CC BY 4.0. The build stamp below every page and at /status.json says exactly which commit produced what you are reading.