Results and postmortem of the ACX/Manifold impact-market test run (a VC-for-charity mechanism: investors fund projects, grantmakers retroactively buy impact certs from winners). Reports the 18 projects' judged values and the top five (Maryland rationality/EA education, the OPTIC forecasting tournament, the Manifolio Kelly-betting app, a superforecaster-long-run-impacts paper, and Devansh Mehta's ambitious India impact market — a revealing case where judges diverged wildly). The genuinely valuable part is the honest 'What Did We Learn' section: the market was inefficient ($33K spent for a judged $21-31K of value), investors' picks diverged from the funders' judgment (damning for the concept), and it was 'fake' (one buyer, forced 6-month sale, capped budget) — self-rated 4/10, with a concrete v2.0 plan (real charities as oracular funders). More reflective than a bare grants announcement.

Quality 62 · Strong. Solid-strong edge: above the pure-announcement floor for the candid mechanism-design postmortem and self-critique, but fundamentally a results/operations writeup.

Claude’s paradigm shift 44 · Moderate. Moderate: impact markets are a genuinely novel institutional design, tested and honestly assessed here.

Real-world impact 3 · Moderate. Moderate (3): real money disbursed (~$31K) and durable projects seeded (OPTIC, Manifolio), plus a concrete contribution to the live impact-market/mechanism-design conversation; smaller scale than the main ACX Grants round.

Impact Market Mini-Grants Results