Prediction markets are usually discussed through the lens of sports, politics, or headline-grabbing event contracts. Talking with FutureSearch.ai co-founder Dan Schwarz made it clear that another side is developing beneath all of that. For companies like FutureSearch, prediction markets are becoming raw material for something much bigger involving AI research, forecasting systems, and how people process information online.
Schwarz has one of the more interesting backgrounds in the space. Before launching FutureSearch.ai, he worked at Google and Waymo, helped build Google's internal prediction market, and later became Metaculus' CTO. That combination gives him a pretty unique perspective because he has spent years thinking about forecasting from both the human and machine-learning sides.
Exclusive Interview With Dan Schwarz
One thing that stood out during the interview was how naturally Schwarz talks about prediction markets as information systems instead of just financial products. That framing keeps coming up more across the industry lately. You hear similar language from Coinbase, Kalshi, and the forecasting communities. Which increasingly treat markets as live indicators of collective belief.
Schwarz repeatedly emphasized that AI can now play a role in filtering, researching, and forecasting large amounts of information at scale. That idea sits directly at the center of what FutureSearch.ai is building.
FutureSearch is Building AI Researchers, Not Just Forecast Tools
The company's website almost feels less like a traditional prediction market product and more like an AI research lab wrapped inside a forecasting engine. FutureSearch describes its platform as a system that enables users to deploy teams of AI researchers and forecasters to investigate complex questions in minutes.
That can include prediction market questions, though the scope goes much wider. The platform focuses heavily on large-scale forecasting, automated research agents, AI evaluation systems, and long-term probabilistic analysis. Much of the public work FutureSearch publishes focuses on testing how well AI systems can forecast real-world events.
Some of the company's research explores whether AI can outperform humans in forecasting tasks over time, while other projects focus on evaluating how models reason through uncertainty. You can tell the company views forecasting itself as a serious benchmark for intelligence.
The Industry is Starting to Blend
Another thing that became obvious while digging through FutureSearch's work is the significant overlap now between AI, finance, prediction markets, and information discovery. Those worlds are starting to blend fast. FutureSearch has published work on Polymarket forecasting, AI stock forecasting, benchmark systems for evaluating forecasting agents, and research pipelines that automatically generate thousands of forecast questions.
The company even openly discusses using AI systems to identify potential inefficiencies inside prediction markets. That probably would have sounded futuristic only a couple of years ago. Now, multiple companies are actively developing similar ideas simultaneously.
You are also seeing more crossover between forecasting communities and institutional finance. Some of FutureSearch's work focuses on long-term business forecasting and market analysis rather than just internet prediction markets. That shift matters because it shows that forecasting is increasingly treated as decision-making infrastructure rather than niche speculation.
Why Prediction Markets Matter to AI Companies
One of the more interesting themes from Schwarz's comments is that prediction markets create a measurable environment for testing intelligence itself. Forecasts resolve. Questions eventually get answered. Systems can be scored objectively over time. That makes forecasting unusually attractive for AI evaluation. Unlike many benchmark systems, prediction questions eventually yield real-world outcomes rather than synthetic test answers.
FutureSearch clearly sees value in that dynamic. Much of the company's public research focuses on scaling forecasting evaluation systems, creating benchmark environments, and measuring how AI agents improve over time. At the same time, Schwarz did not frame AI forecasting as replacing humans entirely. The conversation leaned more toward AI-assisted reasoning where research agents help narrow down information, identify promising questions, and support human judgment.
The Trade Handle Prediction Markets Take
The biggest takeaway from speaking with Dan Schwarz is that prediction markets are quietly evolving into something far broader than just event contracts. Underneath the surface, a growing ecosystem is emerging that treats forecasting as an intelligence layer tied to AI, finance, media, and research.
We also think companies like FutureSearch highlight where the industry could head next. The conversation is slowly shifting away from simply asking who wins an event. Moving toward understanding how information itself gets processed, analyzed, and priced. That may ultimately become one of the most important long-term stories in prediction markets.