AI, Software
Arbiter placed 28th globally in Reasoning Architecture
Part of Arbiter
My solo entry to the Open Agent Hackathon 2026 finished 28th globally in the Reasoning Architecture track. You can run its console here.
The case study explains how Arbiter’s agents gather evidence and why a versioned policy, not a model, makes each decision. You can also run its console: pick a vendor and watch the agents reach a ruling.
