You have to want it

In How to Earn a Billion Dollars, Paul Graham treats building what you want as a way to forecast demand: your own needs predict it, because whatever you and your friends adopt now is what everyone adopts in ten years. That role for wanting is real but secondary, because its main job is to supply the standard you build to, the working sense of what good looks like that decides whether what you ship is mediocre or great.

The gap between a good product and a great one is almost entirely in the last stretch, the unglamorous work of removing frictions that each seem too small to bother with. You notice those frictions only if they would annoy you yourself, and you keep removing them past the point where the product already ships only if you want the thing to exist for your own use.

The standard therefore cannot come from anyone who does not want the product, and market research is a weak substitute, because research tells you what a stranger reports wanting but cannot make you care whether the fortieth detail is right. A founder building for a market he finds boring can hire skill, raise money, and copy a proven idea, and still ship something mediocre, because wanting is the one input he cannot hire or buy.

This is also why PG tells young founders to build for themselves first: when you are your own user, the demand signal and the standard come from the same source, so the empathy he says startups run on is just the ordinary care you take over something you want to use yourself.