We built the feature everybody asks for, then made it refuse in 4 of 7 states

We built the feature everybody asks for — tell me what to build first — and then made it refuse in most of the states a user can be in. Here is the reasoning and what it cost.

The problem: what should I build first is the most answerable-looking question in the process. Ask any chatbot and you get a reasonable feature list in twenty seconds. It is also a plan for building the wrong thing efficiently, if nobody has established that anyone wants the thing.

So the analysis checks a precondition. It reads the latest verdict the idea already carries, and because that map covers the whole verdict set plus the no-verdict case, the states are countable: 7, of which 4 refuse and 3 produce a scope.

The refused set includes the state EVERY idea is in on the day you have it — nobody has evaluated it — and the one where the evidence came back inconclusive. Inconclusive is not a soft yes, and I had to resist writing it as one.

The part I would defend hardest: the refusal returns an empty scope and a paragraph, not a smaller scope and not a scope with a warning on top. A tool that hedges has produced the scope, and you will build from it, because the scope is the concrete artefact and the caveat is one sentence.

When it does open, two of the mandatory fields are the ones nobody asks a chatbot for. A condition for quitting — a scope with no stated point at which you abandon it is a wish with tickets, and the useful thing about writing it is that you do it before you are emotionally invested. And an explicit not-building list, because the features you cut are the ones you re-add in week three, one reasonable decision at a time.

Also enforced at the parser: every must-have feature has to name the job it serves, or the output does not parse. That single rule removes most of a typical first scope, because features that exist because a competitor has them have nothing to write there.

And the build window is capped at 4 weeks. Not as ambition — as a constraint that forces the cut you keep avoiding. A first version you cannot finish inside that is the product, and you are building on faith again.

One honest edge we publish rather than imply: the contract asks for more than one non-goal and the invariant only rejects an empty list, so a scope with exactly one passes. The intent is stricter than the enforcement.

The gate and the field lists: https://whittleos.com/guides/what-to-build-first-mvp-scope