Hard questions
answered quickly.
FUTURESEARCH
Better answers than any other AI system, in 10 minutes.
On questions where other AIs give up, FUTURESEARCH keeps researching. It does probabilistic inference, so it can always answer.
Search through FUTURESEARCH’s repository of questions about the future based on daily news.
Or write your own question. FUTURESEARCH will deeply research any topic.
1. Frame Your Question
2. Analogize the Past
FUTURESEARCH identifies the best historical analogues for current events, researches their outcomes, and builds them into statistical models.
3. Research the Present
FUTURESEARCH identifies the key actors in any situation, and generates profiles on them. It reads news, industry analysis, government reports, and foreign-language sources in real time.
4. Simulate the Future
FUTURESEARCH never refuses to answer. It always produces concrete dates, numbers, and probabilities, reflecting the inherent uncertainty in the question.
5. Publish a Report
Producing a report is as simple as selecting which FUTURESEARCH research, models, and deductions to include.
Get answers to questions other AI systems can’t handle
Unbiased
Other AIs perform worse on partisan questions. They are subject to the same biases humans are: gullibility, ideology, over- and under-confidence.
Deeply-Researched
Get answers citing broad evidence from worldwide sources. Draw from the FutureSearch Knowledge Base of statistical models and patterns from history.
Concrete
FutureSearch never refuses to answer. No matter how uncertain the question is, it will always give probabilities, dates, and numbers.
Why Trust FUTURESEARCH?
FUTURESEARCH confronts uncertainty head-on. It draws from the practice of modern judgmental forecasting to reason about the past, present, and future even when there is no clear answer.
The result? FUTURESEARCH outperforms most veteran human forecasters when put to the test in answering questions about how current events will unfold.
Prediction Market Profit
Profit on Manifold Markets, the highest volume public prediction market, on 50 important, liquid geopolitics questions about 2024
Forecast Accuracy
Expected accuracy on 55 key geopolitics questions about 2024with a high volume of expert forecasts across Metaculus, Infer, Good Judgment Open, and Manifold Markets
Meet the Team
-
Dan Schwarz
CO-FOUNDER & CEO
Stanford University, BS
Prior to founding FUTURESEARCH, Dan served as CTO at Metaculus, the pre-eminent public forecasting platform.
Dan was a Senior Software Engineer at Google and Waymo, and created Google’s currently running internal prediction market.
-
Lawrence Phillips
CO-FOUNDER & CTO
Imperial College London, MSc
Heriot Watt University, PhD Atomic/Molecular PhysicsPrior to founding FUTURESEARCH, Lawrence led the AI team at Metaculus.
Lawrence has held ML engineer and ML scientist positions at Cambridge Consultants, GTN, Jawbone Health, and GSK.
-
Peter Mühlbacher
FOUNDING RESEARCH SCIENTIST
Cambridge University, MSc
University of Warwick, PhD MathematicsPrior to FUTURESEARCH, Peter worked at UniCredit Bank Austria as a Risk Analyst, and at Metaculus as a Research Scientist.
-
Daniel Hnyk
FOUNDING SOFTWARE ENGINEER
Czech Technical University, MSc
Prior to joining FUTURESEARCH, Daniel was an ML/AI engineer at Wikidi/Semantic Visions as well as Director of Engineering at both GWI and Metaculus.
-
Peter Scoblic
ADVISOR
Harvard Business School, DBA
Peter brings a wealth of experience in the forecasting space as a Principal at Event Horizon Strategies, Senior Fellow for Future Security at New America, and Senior Associate at CSIS. He formerly served as Deputy Staff Director for the Senate Foreign Relations Committee and Director of Strategic Foresight at Metaculus.
Let’s work together.
Get in touch to schedule a demo of FUTURESEARCH.