SolutionWright Universal

June 30, 2026

Every Artifact Has A Falsifier: A Design Rule, Not A Wish

SWU bakes the falsifier into the artifact format — if a deliverable has no test that could break its claim, the format itself rejects it before it ships. (Class E, Class C)

A rule you only enforce in the editor is not a rule. It is a hope.

That is the practical reason every artifact we ship — proposal, dashboard tile, project report, receipt — has the same required field: a falsifier. Not a sentence in the cover letter, not a footnote in the appendix. A field. If the field is empty, the artifact does not render, the PR does not merge, the page does not deploy. The format rejects it. (Class C)

The principle is older than us. Popper, 1934 (Class E): a claim is scientific only if there is some observation that, were it to occur, would force you to give the claim up. We are not doing science here — we are running a small agency — but the line travels well. A deliverable that cannot be broken by any observation is a story about the work, not a description of it. Stories have their place. Invoices are not one of them.

So we treated the rule as a schema problem.

Every artifact template on our side has a falsifier slot next to its claim slot. The two are physically adjacent in the file. When we write a dashboard tile that says "lead intake responds within one business day," the next field in the same record is "if a lead submitted between 09:00 and 17:00 local time goes more than 24 hours without a human reply, this claim is wrong." When we write a project report that says "the override rate on the assistant fell from 31% to 9% over six weeks," the next field is "if the override log shows fewer than 200 sessions in that window, the sample is too thin and the number is retracted." (Class C)

This sounds fussy. It is fussy. That is the point. The fussiness is doing the work the editorial pass used to do, and doing it earlier — at the moment of authorship, when the author still remembers what they meant. By the time an artifact reaches review, the falsifier is either there or the build fails. A reviewer can argue about whether the falsifier is the right one. A reviewer cannot accidentally let an unfalsifiable claim through because they were tired on a Thursday.

What this rules out, on our side, is the language pattern most prevalent in vendor decks right now: the unfalsifiable superlative. "Most advanced." "Truly intelligent." "Indistinguishable from a human." None of those sentences name an observation that could fail. None of them, then, can survive our template — there is no falsifier you could write next to them that is not itself a piece of fiction. The format catches the marketing before the marketing reaches a client. That is the design intent.

There is a parallel concern outside the agency. Dr. Alianna J. Maren has been writing, from a different vantage, about the cost of pouring capital into one architectural bet while underspending on the human verification step that would catch the resulting errors. Her tummy-churning piece, in our one-line frame, asks the field to put the verify-the-output discipline back where it belongs. (Class E) We read that as adjacent to the artifact-format rule: a verification step that lives only in a person's head, or only in a post-hoc audit, is a verification step that will, eventually, not happen. A verification step that lives in the file format happens every time.

A few honest limits. The falsifier field forces the question; it does not guarantee the answer is well-chosen. A weak falsifier is still a failure mode — "if literally no one ever buys our service again, the claim is wrong" is technically a falsifier and practically a dodge. We try to catch those in review and, where we miss, on the corrections page. The rule is necessary, not sufficient. (Class C)

If you are evaluating a partner — us or anyone else — the cheap test is to look at one finished artifact and ask where the falsifier lives. If the answer is "it is implied by the surrounding paragraphs," the rule is not in the format. It is a wish.

EvidenceECTagsfalsifierartifact-formattransparencydesign-rulereceiptsevidence-class

Next steps

Bring this into a working session.

The workshop is where these notes turn into receipts on real work. The science page is where the underlying hypothesis is laid out in full, with the falsifier attached.