Introducing Breakthrough Markets at ACM Collective Intelligence 2020

At CI2020, Cassandra Labs co-founder Lorenzo Barberis presented a novel design for Breakthrough Markets—a hybrid prediction market that blends human and machine intelligence to accelerate scientific discovery.

Unlike replication markets, which validate existing results, Breakthrough Markets aim to identify promising combinations of ideas—mapping the relationships between concepts like a Bayesian network. By letting humans and AI-powered bots trade predictions about how scientific ideas connect, the system creates a real-time consensus about which interdisciplinary directions are most worth exploring.

Lorenzo walks through key examples—from Foldit to the Malaria Game—demonstrating how crowdsourcing non-expert input can outperform traditional research pipelines. The system doesn’t just prioritize what’s likely to be true; it prioritizes what’s worth testing next, helping teams coordinate toward breakthroughs with the highest potential payoff.