swarm dynamics lab

distributional safety

What stays safe when many AI agents act together? Over three days we read the field, ran a swarm of experiments, and wrote down what broke.

Oct 1 - 3, 2026
01the questions We read the field and turned it into testable hypotheses. enter -> 02swarm research lab Agents ran the experiments on a fleet of servers, through the night. enter -> 03the findings What held, what broke, and the films that show it. enter ->