Reproduce “A Simulation Testbed for Self-Governing Human-Machine Institutions” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Callum Berger, Jeremy Pitt, Stefan Sarkadi
Megabike provides a simulator for studying whether AI groups can make, follow, and revise rules while sharing resources. Five riders coordinate steering and effort, pursue lootboxes, and face a threat; the simulator enforces rules and records actions. Across 40 all-LLM sessions, no rule-reopening vote passed, and sanctions varied by mechanism; findings are limited to this small, all-machine study.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.