Calibrate AI timeline predictions

Create a public, auditable meta‑registry that collects near‑term AI capability predictions, records their exact operational definitions and pre‑specified prompt/tests, and publishes retrospective calibration scores. The registry would standardize how forecasts are framed (what 'AGI' concretely means), force prompt and evaluation provenance, and produce a running error‑rate metric for different predictor classes (founders, academics, pundits). — A standard calibration registry turns noisy, attention‑driven claims about AI timelines into accountable evidence that policymakers, investors and the public can use to set graduated governance and industrial triggers.