A 60-second self-audit

“The pen is mightier
than the sword.”

Not because the pen is stronger. Because it acts through the mind.

That was the deal with every tool before AI: the tool helped, but you still had to think.

AI changes the deal. It can sharpen your judgment, or it can take over the very capacity you think you are improving. And it will not warn you when the shift happens.

This takes 60 seconds. It does not grade you. It shows whether your AI use last week made you sharper, or more replaceable.

Most people guess wrong.

1  /  2  ·  Your gut
First, no thinking. How did last week with AI feel?
The gut read on the week overall, not the audit. That comes next.
Bad Rough Meh Okay Strong Great
Okay
Did the job. Nothing special.
2  /  2  ·  Last week, actually
How often did each of these happen?
Seven days. Be honest. The gap only shows if you are.
0 of 9 answered
Where your week landed

Two things about last week, on two axes. Up the side: how it felt. Across the bottom: how much judgment you actually put in. Your week is the dot. Which corner it lands in is the read.

Mirror felt strong · barely engaged Sharpener felt strong · fully engaged Idle felt flat · barely engaged Grind felt flat · fully engaged How it felt Great Bad Judgment you put in Low High you
The gap. The distance between how the week felt and how much judgment you actually put in. Not a grade. A direction. The one to watch is a gap that reopens week after week.
 

Why those two words. A mirror reflects you back. It shows you as you are and asks nothing of you. Useful for checking the fit of a dress, useless for making the dress fit better. A sharpener changes the blade. It costs friction and leaves you better than it found you. The same AI does either. Which one you got is set by what you brought, not by the tool.

How this is scored
Nine questions, two hidden halves. Five ask whether you engaged: pushed back, reframed your own question, argued the other side, verified before relying, caught something the AI made up. Four ask whether you offloaded: used an answer unchecked, shipped it unedited, stopped at "good enough," leaned on a conclusion you couldn't re-derive. Judgment (across the bottom) is engaged minus offloaded. Your gut slider sets the height. That's the whole machine, no black box.

Jim Collins wrote that good is the enemy of great: settling for good is what keeps you from something better. The AI version is quieter: a good week is the one you never examine. Rough weeks make you ask what went wrong. Good weeks feel like enough, so you don't look. And "don't look" is the one condition the mirror needs.

Here's the part that isn't about you. You just measured this by hand, on one person, for one week. Now scale it: a team, a quarter, ten thousand decisions where no one is watching the gap and every "good enough" compounds.

At that size it stops being a habit worth minding and becomes something an organization has to govern, and can only govern if it can measure. That measurement has a name: capacity-degradation risk. Building AI that keeps the gap closed is a discipline: capacity-preserving AI. What you felt in sixty seconds is the doorway. The rest is the building behind it.

See the framework this comes from → WHEEL
One catch

Noticing this feels like progress. It isn't. Not yet. The insight was today; the test is next week. Have this realization and still skip the check, and the realization was just a nicer place to not do the thing. So the only real question: will you run it again in seven days?