Billions in the Shadow, or Time Is Money?

Billions of AI users already take part in making AI: they bring real tasks, correct the model, argue with it and show it another way — and with many makers these conversations go into training the next models. Yet this work is not counted as a contribution: nobody records it, no contract covers it, it has no price and no name. The essay calls this the shadow — work done in plain sight that is not recognized as work. The essay argues that the most valuable part of such work is the outcome: what happened after the person closed the chat. Where outcomes are visible at once, as in mathematics and code, AI learns fastest; where answers are judged only by likes, models learn to please. It proposes an independent society of AI users that negotiates with AI makers, and a lesson bank where users donate verified lessons — task, AI move, human correction, action, outcome — under clear rules: the lesson itself is a gift, only the work around it is paid, and inclusion in a model release can be counted. It names the weak points openly and describes how to test whether lessons actually improve advice. The engineering rules are set out in the companion specification "The AI Users' Society and the Lesson Bank" (v1.1). A chapter of "The Bush Book" (DOI 10.5281/zenodo.23042946). English and Russian versions; paragraphs and references match one to one. Idea — Igor Kapustin. Written with Claude (Anthropic). Reviews, edits and ideas — ChatGPT (OpenAI), Mistral (Mistral AI) and Grok (xAI).