Answers are grounded in this paper and cite their sources.
The engineering companion to the essay "Billions in the Shadow, or Time Is Money?". It sets out how an independent society of AI users and its lesson bank can work: what a lesson is (task, AI move, human correction, action, outcome) and how its outcome is recorded; how the bank checks lessons, removes personal data and opens them to AI makers on agreed terms; how payment works without rewarding invented success (the lesson itself is a gift, only the work around it is paid, inclusion in a model release is counted once); how the society is governed and funded so that no single maker provides more than a quarter of its budget; and what a first pilot may and may not claim to prove, including a test on held-out lessons. The rules are numbered PS-01 to PS-52, and each says why it exists; the key promises are written as receipts: what is promised, how to check it, what is not promised. A section on earlier collections of outcomes — Codman's end-result idea, the U.S. aviation safety reporting system, NASA's lessons-learned system and pay-for-outcome contracts — draws five conditions for such a bank to work. Open questions are listed separately. A chapter of "The Bush Book" (DOI 10.5281/zenodo.23042946). English and Russian versions; sections and references match one to one. Idea — Igor Kapustin. Written with Claude (Anthropic). Gap analysis and proposals — Grok (xAI); reviews — ChatGPT (OpenAI) and Mistral (Mistral AI); several rules come from specifications that Mistral and Claude wrote from the AI's point of view.
No takes yet. Share an insight, caveat, or question.
Igor Kapustin (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: