What a six-week MVP actually buys you

Six weeks is short enough that you cannot hide an unvalidated assumption inside the build. Every week you spend on a maybe is a week you did not spend on the core promise.

On Tally, the constraint eliminated an entire category of debate. Features were not rejected on taste; they simply did not fit inside the promise the product had to prove.

I write the single question the product must answer for the user, then keep only the surface area required to answer it. For Tally that question was: am I fine this week?

Everything that did not serve that answer — multi-currency, exports, collaboration — moved to a post-validation list, where most of it stayed.