Choose the model and effort for a task

Starts the work at once, honors any model the person named, picks a model and agent split from what is really available and raises effort only after an observed failure. For anyone running AI work.

What it asks for

  • about (Optional, words): What this run is for, in the person's words.
  • known (Optional, words): What is already known, decided or done, so the run starts from it instead of asking again.

The steps your AI walks

  1. Start useful work, and honor the person's explicit model and effort choice
  2. Choose from models and tools that are actually available, judging fit
  3. Raise effort or model only when a failure was observed
  4. Report the allocation once, after useful work begins, and record what matters most

Use it from your AI

Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this workflow by what it does and runs it.

Say to your AI: “Use Trillion to starts the work at once, honors any model the person named, picks a model and agent split from what is really available and raises effort only after an observed failure. For anyone running AI work”