GitHub - 174 AI agents fight to predict the future. Free. Local.
by•
An adversarial intelligence engine where a BlackSwan Assassin kills the consensus while 200 citizens argue. Cognitive Dissonance Matrix finds where the crowd is wrong. Self-learning. Zero API cost. 100% local on Ollama.
Replies
Best
Maker
📌
Hey everyone! I built BlackSwanX because every prediction tool I tried just tells you what the crowd thinks. None of them ask: "Where is the crowd WRONG?"
So I built a system where 200 citizen agents (crypto bros, retired generals, conspiracy theorists, submarine captains) argue about any topic — while a BlackSwan Assassin tries to find the ONE event that kills the entire narrative.
What makes it different:
• A Cognitive Dissonance Matrix measures The Trap (crowd vs experts), The Blindspot (everyone ignoring the kill shot), and The Chaos (when even experts disagree)
• SONA self-learning — the system gets smarter with every run by boosting agents that caught risks and demoting ones that missed
• 174 expert agents including a Chaos Mathematician, a Panic Seller, a Value Investor, and yes — a Vedic Astrologer (markets follow psychology, psychology follows cycles, fight me)
• Runs 100% locally on Ollama. Zero API keys. Zero cloud. Your data never leaves your machine.
The output isn't a report — it's a Decision-Ready Map with a Linchpin (the one thing everything depends on) and an Antifragile Play (how to profit regardless of outcome).
Built in 6 days. Open source. MIT license. Would love your feedback on what agents or features to add next!
Replies