This is a collaborative research project. Lots of people work with one agent to answer one question: can Elon Musk buy an entire city?
Every city follows the same underlying structure — Land, Buildings, Parks, and Monuments. Each category holds structured sub-items that roll up into totals. The agent fills in the schema by searching public datasets and applying a methodology to each section. You can read those methods here, pulled live from what the agent actually uses: .
As the agent researches, it shows its work — where it’s searching, what it’s found, its confidence in each number. But this isn’t a passive experience. You steer the agent. If a number seems off, ask it to dig deeper. If you know something, say something — the agent will try to validate your claim against public information.
Nothing is committed to the database without an agent proposing it and a human saying yes.
Every conversation, every refinement, every search result stays in a shared transcript. The next person to research this city can see what you learned and build on it, rather than starting from scratch. The agent learns from that accumulated context too — synthesizing the input of many people into a richer, more grounded analysis.
Our hunch is that, together, we can build something more reliable than any single agent or human working alone.