The line inverts the usual sequence. Most founders begin with a solution hunt — scanning markets, chasing trends, asking what would scale. Graham is naming what breaks first.

Ideas that begin as ideas sound like ideas. They arrive fully formed, packaged in deck language, optimised for pitch rooms. The problem is they solve nothing you have actually felt. So when the first customer pushes back, when the MVP sits unused, when the pivot question arrives, you have no ground. You were never the user. You were guessing about users.

The problems you have yourself come with free validation. You know the pain is real because you are paying the cost right now. The invoicing tool you wish existed, the coordination overhead you cannot automate, the five-step workflow you repeat daily and hate — these are not theoretical. They have budgets attached. If you would pay to fix it, someone else probably would too.

Most startup advice begins at the business model. Graham begins earlier. The idea is not the start. The problem you cannot ignore is the start. The rest is just deciding whether to fix it only for yourself or fix it for everyone still stuck where you were.

So today: write down the three recurring problems you complained about this month. Not the big strategic ones. The small irritating ones you keep working around. One of them is probably worth six months.