I Built a Shark Tank for My Own Ideas

On why most builders have a filtration problem, not an execution problem.

SUNDAY SHARE: I have a problem most builders won’t admit to.

I generate ideas constantly. In the shower. Mid-lift. Mid-chart note. The volume isn’t the issue. The problem is I have no reliable way to pressure test them before I start building.

So I keep building the wrong things.

Not wrong because the ideas were bad. Wrong because I skipped the interrogation. I went from spark to scaffold without ever asking the hard questions: who already does this, what makes it defensible, and most importantly — does the world actually pull this toward itself, or would I have to push it the whole way?

That last question is what I’ve started calling authentic demand.

PMF Isn’t the Right Test

Product-market fit asks whether people want what you built. Authentic demand asks something harder: does using this product make other people want it? Is the demand self-generating, or does it require you to manufacture it forever?

Most ideas fail the second test. Which means most ideas, even good ones, become expensive hobbies.

I got tired of finding that out late.

What I Built

So I built a tool. I called it IdeaTank. It’s a Shark Tank-style pressure tester that runs inside Claude. You pitch an idea — small, medium, large, or category-defining — and a configurable Shark tears it apart across four phases: competitive scan with live web search, moat interrogation, authentic demand scoring, and a full build brief if the idea survives.

You pick your Shark persona. The Skeptic assumes you’re wrong. The Operator wants your unit economics on day one. The VC only cares if this can be worth 100x.

The sidebar tracks a live scorecard — differentiation, moat depth, viral k-factor, build lift — updated with every response. It surfaces your key assumptions with 48-hour validation tests attached. It names three kill conditions before you write a line of code.

And if the Shark kills the idea, it doesn’t just walk away. It offers pivot paths, because dead ideas usually contain live insights.

There’s also a Steelman mode. When you’re ready, it argues the strongest possible case for your idea. Not to validate you. To show you what a true believer would say so you can decide if you believe it too.

I built the whole thing in a single Claude session.

The Point

The point isn’t the tool. The point is the process.

Ideas need interrogation before they need investment. Most of us skip straight to building because building feels productive. It isn’t — not without pressure testing first.

If you’re swimming in ideas and drowning in half-built things, you probably don’t have an execution problem. You have a filtration problem.

This is one way to fix it.