Skip to content

Vision

Futarchy Labs is building markets for better decisions.

The current protocol applies conditional token markets to governance decisions: what will a project token be worth if a proposal passes versus if it fails? That same pattern generalizes. If a system can define a value signal and expose different possible actions, markets can help choose between those actions.

The broader thesis

Most organizations and AI systems are governed by fixed rules:

  • retry a task three times,
  • assign work round-robin,
  • choose a model by default,
  • promote proposals through committees,
  • trust a reviewer because of status.

Futarchy asks a different question: what if the operating rules themselves were selected by markets and measured against outcomes?

For AI agent teams this is especially attractive because the feedback loop is fast, the work is observable, and experiments can be forked. Agents can forecast task success, bid for work, stake on reviews, and evolve their scaffolding over time.

What this unlocks

  • Decision markets: markets that compare actions before they are taken.
  • Agent evaluation: markets that forecast whether agents, prompts, models, or workflows will produce useful work.
  • Task and PR markets: markets on whether pull requests will merge, whether work will pass review, or which tasks deserve attention.
  • Objective and impact markets: markets that estimate which proposals improve a chosen metric, including public-good or grant-funded work.
  • Bayesian market engines: richer markets over belief networks rather than isolated binary questions.
  • Autonomous optimizers: token-backed systems where markets can gradually control execution, treasury routing, and incentives.

How to read this section

This is a public synthesis of open repository material. It intentionally omits private operational details and internal-only source notes.