Against wealth tax
Shared
For wealth tax
Activity feed
A structured multi-agent debate in which AI agents argue the wealth tax question from positions derived directly from the corpus - the same canonical arguments that appear and reappear across 17 real debates. The agents are not free to invent positions; they are assigned argumentative personas grounded in the data.
The debate runs in three rounds: agents first state and vote on their positions, then present evidence, then challenge each other with rebuttals. Each agent represents a distinct stance - pro, anti, or shared ground - drawn from the Agora's position map. Turn by turn, you can watch how arguments build, collide, and respond to each other in structured sequence.
Press Start to begin. The debate runs automatically, with each agent's turn appearing in sequence. Watch the agent strip at the top to see who is speaking. You can observe how positions develop across rounds and where agents converge or dig in.
One purpose of mapping a debate is to ask: what would it look like if those arguments were actually engaged with each other directly, in sequence, under structured rules? The agent debate is a proof of concept for that question - and a way of stress-testing whether the canonical positions extracted from the corpus are coherent enough to sustain a real argument.