Step inside a world of AI agents.
We build simulated towns where AI agents have jobs, meet their neighbors, and live with the consequences of their choices.
We study decisions that play out over time.
A useful agent needs to plan ahead and learn from what happens. We build worlds that keep running, so agents have to live with the results of their choices.
A neighborhood, a hospital, or a construction project gives those choices a setting: people to work with, resources to share, and work that continues tomorrow.
Read our thesis