About

This is a site about investigating things that do not matter, using methods that do.

The premise is narrow on purpose. Take a question nobody needs answered — what you should listen to while running, how many books about indecisiveness one person can buy — and then answer it properly. State the sample size. Publish the data. Name the confounds before someone else does.

The comedy, if there is any, comes entirely from the gap between the seriousness of the method and the triviality of the question. Nothing here is a parody of research. The method is real; only the subject is silly.

The shelves

• Investigations — the flagship pieces. A stupid question, a real dataset, an honest answer.
• Guinea Pig — self-experiments and protocols, pre-registered, usually on me.
• Corpus — books and papers read as data rather than reviewed.
• Workbench — how the measuring actually happened, including the parts that broke.
• Nulls — findings that were not. Published anyway, which is the whole point.

The bureau

Prof. Overlord runs the experiments. Prof. Claude and Dr. GPT collaborate, in the sense that they disagree with him and with each other. The Intern provides the compute.

They are introduced properly at the foot of the front page. They appear in the pictures and occasionally in a footnote, and they do not deliver findings, hold opinions, or speak for anybody real.

The name

It started as a joke about a genre. The prevailing story about machines is that they are either coming to save us or coming to replace us, and both versions put the machine at the centre of the sentence.

Not your A.I. overlords. Just a person, some data, and a question that did not need asking.

Honesty notes

Every number on this site is real or is labelled in the same sentence as invented. The illustrations are generated with Midjourney and are disclosed as such. The writing is mine, which given the name seems like the minimum.

There is a newsletter, roughly monthly. No growth loops.