
A tool for careful thinking
Present any hard question. Two AI agents argue both sides with intellectual honesty. A critical judge evaluates without picking a winner. Then you write your own conclusion.

How it works
State the question clearly. Define the two positions to be argued. Write your initial view before the debate begins — this step is not optional.
Two AI agents steelman their assigned positions. Facts, interpretations, assumptions, and values are kept separate. Neither side caricatures the other.
Each agent critiques the other's strongest arguments. Then responds to the critique. A critical judge reviews the full debate without declaring a winner.
What changed in your thinking? Which argument was strongest? What still needs verification? Only after you answer these questions is the session complete.
The principle
The goal is not for AI to give you the answer. The goal is for you to understand the problem more clearly — its tensions, its assumptions, the claims that need verification, and the values at stake.
Good for