Note / 003
Capsize / Product note
Decision systems
Break it before
you build it.
Capsize is a decision tool for finding the assumption your idea cannot survive without.
The problem
Most ideas do not fail because nobody worked hard enough. They fail because one important assumption stayed invisible until too much had already been built.
The expensive question is not
“Can we build it?”
It is “What has to be true for this to matter?”
What Capsize is
Bring it a product, service, offer, process, or venture. Capsize structures the bet, compares a founder-right world with a market-right world, checks the business story, and identifies the stage that carries the most weight.
Then it stops being an AI exercise and gives you one real-world probe to run next.
AI can make the argument. Evidence makes the decision.
The flow
01
State the idea.
What do you want to make, who is it for, and why should they choose it?
02
Build both worlds.
Capsize makes the strongest case for the idea and the strongest boring alternative.
03
Find the load-bearing stage.
It looks for the point where a small change has the biggest effect on the decision.
04
Run a real probe.
You leave with one external test and a clear artifact to bring back.
What it refuses to do
Capsize does not give you a “72% chance of success.” It does not call synthetic output validation. It does not turn confidence-looking numbers into proof.
No success score.
Reality gets the final vote.
Evidence changes the state
A fresh free run is synthetic. A reference or observation can be anecdotal. A registered real-world test can become probed evidence. The important part is that the evidence class is not whatever the model feels like calling it.
Capsize Pro keeps the project, the registered probe, the evidence that comes back, and the Reality Gate that can stop another internal rerun when the same missing external answer is still missing.
Payment buys continuity and governance. It does not buy a smarter answer.
When to use it
Before a build.
You are about to spend design or engineering time and want to know what deserves testing first.
Before a launch.
You have an offer or product but one uncertain stage could make the whole economics fall apart.
When you are looping.
You keep refining the same idea internally and need one outside answer before doing more work.
Notes / Subscribe
Get the next note.
No schedule. No content calendar. A note goes out only when there is something worth understanding.