Building the foundational science of understanding human behavior.
Our mission
Our goal is to decipher what makes us human, starting with how we make decisions.
Every day, billions of people choose what to buy, whom to trust and where to go. Those decisions move entire economies, and we still can't explain them. We learned to model the weather and language; the human mind is still the frontier.
We listen to real people, follow them over time and build models of how we decide. It is an old question. For the first time, we have the tools to answer it.
Our research
Our team brings applied and theoretical research in machine learning and reinforcement learning to one problem: how people make decisions. We work on three fronts, and each one feeds the others.
Human decision models
Models of how individual people make choices, trained on data from real people and on reinforcement learning environments.
AI-led field research
An AI interviewer that talks with verified people by voice, at scale.
Intelligence
Agents that turn your past research into something you can query, and suggest what to ask next.
A model of how we decide
Language models learned to write by reading text.
We train models on decisions made by real people, to predict what someone will choose and understand why.
Team
Five years of research in AI and reinforcement learning, applied at leading institutions. The founders have led data acquisition for Fortune 500 companies, built ranking and calibration models for Meta's ads, and published machine learning research.
