See or change the hold: while something truly needs the person, every Trillion page they open (web, Mac app, menu bar, iPhone) shows only that one thing until it is done or set aside with a reason; choose in advance what turning it off costs, ask it off, or turn it on
op read (the default) answers whether the hold is on, the one thing it holds (held.key, held.card), how many set-asides are left today and the stakes. op aside with key (held.key), title and reason sets the held thing aside for a while; the reason must be words, and it is handed to the AIs to do the thing instead or cut it to one small step. op stake with stake (a stake id) and values (the fields it asks for) chooses what turning the hold off costs; a new stake counts only after the one in force has run its delay. op off asks it off: it turns off only after the stake's delay and comes back by itself a day later, and a stake that tells someone or gives money is handed to the AIs to carry out. op on turns it on at once. The hold is on the person, never on an AI: an AI keeps doing all of its own work while a thing is held, up to the one step only the person can do. Every way out of the hold and its counter is DATA in src/escape-chain.json.
Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this tool by what it does and runs it.
Say to your AI: “Use Trillion to see or change the hold: while something truly needs the person, every Trillion page they open (web, Mac app, menu bar, iPhone) shows only that one thing until it is done or set aside with a reason; choose in advance what turning it off costs, ask it off, or turn it on”
What your AI calls: run_tool {"tool_id":"tool:hold-on-what-needs-me"}