Multi-agent AI forecasting that tracks calibration against real-world outcomes and Polymarket. Deploy a council of distinct AI agents to debate events, eliminate bias, and audit the trail.
The base rate analyst opens with a frequency. The contrarian sells consensus short. The verdict, after three rounds, is a number neither would have produced alone.
A reader's $47k decision, debated in full. Five rounds, two concessions, one vote against the math.
How the council aggregates eight rankings into one resolution — and why ties still resolve cleanly.