What your describing sounds like the wantrepreneur process and it usually doesn't end well because you will likely have no existing experience in the business domain you are targeting, unless your target market happens to be developer tooling*.
Most successful startups are made by people who already know the problem they are trying to solve because they've experienced it first hand. They have the industry contacts to quickly find early customers and their search for "product-market fit" is usually about whether clients will pay enough to make the startup worthwhile rather than "do clients even want what I'm building."
Ideas are a dime a dozen. Competent engineers are a dime a dozen (relatively). Distribution is what builds startups and that requires industry experience and contacts, or a cofounder that can carry the domain side of the business (in which case they will be the ones filtering ideas).
* I'm assuming you're an engineer. Otherwise you're an idea guy without a clue on how to create viable ideas.
I agree that firsthand domain experience and distribution are huge advantages, and most successful companies start there.
I’m less interested in replacing that path and more in understanding how people outside a domain avoid fooling themselves when evaluating ideas they see online.
Out of curiosity — have you ever seen someone without deep domain roots succeed by validating their way in, or do you think that path is mostly a dead end?