I Let Claude Grill Me About My Own Startup Idea. It Was Brutal — and Right.

I had a startup idea I was fairly confident in. Confident enough that I almost skipped validating it and went straight to building. Instead, I ran it through `grill-me` first. What followed was twenty minutes of the plan getting taken apart, piece by piece, until the parts that survived were the only parts actually worth building. It was uncomfortable. It was also exactly what I needed.

The idea, before the interrogation

The pitch, as I described it initially, sounded complete: a clear target user, a clear problem, a rough plan for the first version. In a normal conversation, that description would have gotten a supportive response and I would have started building that same afternoon.

Where it actually started falling apart

The first hard question was simple and I did not have a real answer: "who has this problem badly enough to pay for a fix today, versus who would just say it sounds nice?" I had conflated interest with demand — the same appreciation-is-not-demand trap that kills most first products. It kept going from there: what happens to a user's data if they cancel, what the actual first week looks like for someone who signs up, what I would do if the one channel I was counting on for distribution did not work.

The moment it got genuinely useful

By question six or seven, I stopped defending the plan and started actually rewriting it in real time — not because I was being argued into submission, but because the plan that came out the other side actually answered the questions instead of hand-waving past them. The version I ended up building shared maybe half its assumptions with the version I walked in with.

What I would have built without it

Almost certainly the original version — confident, plausible-sounding, and missing the one question that would have surfaced two months in, at a much higher cost than twenty uncomfortable minutes up front. That is the trade the skill is actually offering: discomfort now, in exchange for not discovering the flaw after you have already spent the time.

If you have an idea you are fairly confident in, that confidence is exactly the reason to grill it before you build it. The ideas that survive being interrogated are the only ones worth the time anyway.