Building is the fun part and the expensive part, which is exactly why I force myself to prove an idea is wanted before I let myself start on it.
The idea I am most tempted to build first is always the one I have not validated, because building is the fun part and validating is not. Every product instinct I have pulls towards opening an editor before I have asked anyone whether they want the thing I am about to make, and I have learned not to trust that instinct, because it optimises for what I enjoy doing rather than for what turns out to matter. I notice the same instinct in almost everyone I have worked with who is good at building — the skill that makes you fast at construction is a different skill from the one that tells you whether construction is warranted yet, and the two rarely arrive in the same person by default.
Validation, in the way I actually practise it, is not a survey and it is not a landing page with a fake button, though I have used both. It is finding out whether the problem I think exists is one somebody is already spending time or money trying to solve badly. If nobody is doing that — if the problem only exists in my head, as a thing that would be nice to have rather than a thing anyone is currently in pain over — that is the strongest signal I know of that I am about to build something nobody needed.
The place I go looking is not other people’s opinions about my idea, which are close to worthless, but their existing behaviour, which is not. What are they already paying for that is adjacent to this. What workaround have they already built, badly, out of spreadsheets and browser tabs and habit. A workaround is evidence in a way that a compliment is not — nobody builds a bad workaround for a problem they do not actually have.
I hold this standard especially hard against my own product ideas, because they are the ones I am most likely to talk myself past it on. It is easy to apply rigour to someone else’s pitch and remarkably easy to wave it through for my own, on the theory that I am different, that I can see something they cannot. I can, occasionally. Far more often I am simply the person least able to be objective about whether the thing in my head is actually wanted. The tell that I am doing this is when my case for an idea starts relying on my own taste rather than on anyone else’s demonstrated behaviour — taste is a fine reason to prefer one solution over another once the problem is established, and a poor reason to decide the problem exists at all.
None of this is an argument against speed — once the validation is real, I want to move as fast as anything else I have ever built, and building with an agent has made that part faster than it has ever been. That speed is exactly why the discipline matters more now rather than less: when construction was slow its cost was its own filter, and now that a working version of almost anything is a weekend away, the only filter left is whether anybody wanted it. It is an argument against skipping the one step that is cheap to do properly and expensive to skip: an afternoon of finding out whether the problem is real costs an afternoon; a month of building the wrong answer to a real problem, or the right answer to a problem nobody has, costs the month and the credibility to try again soon after.
The version of this I keep having to relearn is that validating an idea and being excited about it are unrelated facts, and excitement is not evidence. I have shipped things I was certain about that nobody used, and I have shipped things I was lukewarm on that people paid for almost immediately because the workaround they were already living with was worse. My excitement was never the variable that decided the outcome either way.
What I actually watch for now, before I let myself start building, is whether I can point at a specific existing behaviour rather than a specific enthusiastic reaction. Enthusiasm is cheap to produce and cheap to receive, in a conversation or a comment thread, and it tells me almost nothing about what somebody will actually do when the ask involves their own time or their own money instead of mine.