He tweeted 'who would like this' for a directory of indie hackers, read 50 likes as validation, spent four months building it, and never monetized it. Put payment in from day one.
He argues that asking if someone likes an idea is pointless unless you ask if they'll pay; waitlist signups and liked tweets don't count. Learned after five products that earned nothing.
Following levelsio: put up a landing page and a buy button for the idea. If people pay, it's validated; if not, you've avoided months of building the wrong thing. He suggests combining it with copying a validated idea and making one part better.
Decide what you're actually optimizing for; for most indie hackers it's freedom, and freedom comes from money. So validate early by asking people to pay before you build; you can refund them if it doesn't work out, but you can't get back months spent on something nobody wanted.