"Discipline at the input creates discipline at the output" is the cleanest version of this argument I have read.
Pattern 3 is the one most people skip. They ask the model to confirm instead of disconfirm. The output sounds good, survives no scrutiny, and they wonder why it failed in the room that mattered.
Pattern 4 is the one with the highest stakes in your domain specifically. The leak you don't catch before release in a medical context is not a README problem.
Writing about the input discipline problem from a different angle at michaelklingebiel.substack.com. "Begin With the Question, Not the Answer" covers exactly why confirmation seeking is the default and how to break it.
Agreed. Our ideas are meant to be pressure tested. Discipline can give us uncomfortable outputs. This ensures that we arrive at an output reflective of reality
"Discipline at the input creates discipline at the output" is the cleanest version of this argument I have read.
Pattern 3 is the one most people skip. They ask the model to confirm instead of disconfirm. The output sounds good, survives no scrutiny, and they wonder why it failed in the room that mattered.
Pattern 4 is the one with the highest stakes in your domain specifically. The leak you don't catch before release in a medical context is not a README problem.
Writing about the input discipline problem from a different angle at michaelklingebiel.substack.com. "Begin With the Question, Not the Answer" covers exactly why confirmation seeking is the default and how to break it.
Agreed. Our ideas are meant to be pressure tested. Discipline can give us uncomfortable outputs. This ensures that we arrive at an output reflective of reality