Connecting ideas across fields to build compositional foundations for collective intelligence
I study how groups of agents — human, AI, or mixed — can coordinate better. My work sits at the intersection of multi-agent systems, AI safety, and democratic mechanism design. Rather than reducing everything to one framework, I try to connect perspectives from game theory, active inference, computational social science, and category theory to find where they compose. I'm building simulation infrastructure to test whether coordination mechanisms actually work before deploying them in the real world.
Approach
A quick way to explain what I do: if individual agents are trees and game theory is about how trees interact, I'm looking at forest-level properties. How do interaction networks look? What are the ecosystem-level effects within specific niches like democracies and markets? How can we protect these systems? I assume someone already solved my problem in a different field — so I read across disciplines looking for patterns that repeat across contexts. I take play seriously as a learning mechanism, and I try to make the implicit explicit.
I'm outgoing when something interests me, otherwise I sit and read books. If I say something that makes you wonder if it was irony, assume it is. I'm a good listener, but if you bring up something I've thought about a lot, I will go on a five-minute rant. Known for these rants in certain circles.
Metal music, strategy games, meditation, nature, sci-fi & fantasy, applied math, history, info hazards, and literally any subfield of science except thermodynamics.
