It reads well and it is still wrong. The gap is not the AI. It is how you asked.
I am Paul Gerber. I work with people and teams in Tokyo on getting more out of the AI they already use.
Two routes to the same three scores.
Quick. Nine questions about one task you just did.
This is my thought process, running in the open. Tell me where it gets you wrong.
Nothing you paste leaves this page. The reading is decoded in your browser and stored on your device.
Run it in a chat that already has real work in it. An empty chat has nothing to read.
Three things are always going on. Neither pole is better.
Fourteen moves, two at a time, by how much the work matters.
Capture the gist. Nothing more from you.
Comprehensive and technical. You answer its questions.
Laid out for people. You read it.
The levels stack. Get it right adds to get it down, and get it ready adds again.
So you never decide what to drop. You only decide how far up to go. Three moves, then nine, then all fourteen.
Two exceptions. Word cap and Escalation belong to get it down, because get it right wants the questions and detail they hold back.
Each test asks how far the work travels past the moment you made it. Any one of them can raise the level on its own.
Scored on their own public stance. Tap a point for the case.
Tap a point to see who they are and why they land there.
Reading it. Each person and company is scored on their own, against fixed poles.
Not charity. Fulfilled people are more productive and stay longer.