Most people ask it for answers.
Ask it for a fight instead.
Not the answer.
The other side of your answer.
You already have an idea.
You are half in love with it.
That is exactly when
you cannot see its holes.
So hand it over.
Tell the machine:
here is my plan,
now attack it.
Find the case I forgot.
Name the assumption I never questioned.
Tell me why a smart person
would do the opposite.
The model is a tireless sparring partner.
It never gets bored.
It never takes it personally.
It will argue the position
you are too invested to argue against yourself.
This is not about being right.
It is about being tested
before the world tests you
in production,
in the meeting,
in the place where it costs something.
A yes-machine makes you weaker.
It agrees,
it flatters,
it polishes the thing
you should have thrown away.
The machine that pushes back
makes you check your footing.
So do not ask it
to confirm what you think.
Ask it to break what you think,
and see what survives.
What survives the argument
is stronger than what you started with.
What breaks
you needed to lose anyway.
The danger is asking a leading question.
Ask "isn't my plan good?"
and it will find a way to say yes.
Ask "where does this fail?"
and it goes to work.
The words you choose
decide whether it flatters you
or sharpens you.
Choose sharp.
Use it to stress-test the idea,
not to bless it.
The point was never
to be told you are clever.
The point was to walk out
with an idea
that already went ten rounds
and is still standing.
Argue first.
Ship what wins.
– Serguey Asael Shinder
Top comments (0)