AI by Hand ✍️

AI by Hand ✍️

Retrieval Footprint

What a document costs the window

Prof. Tom Yeh's avatar
Prof. Tom Yeh
Aug 19, 2026
∙ Paid

Library › Context Problems

  1. The Context Window

  2. Window Occupancy

  3. Growth per Turn

  4. Remaining Space

  5. Turn Budget

  6. Retrieval Footprint

  7. Retrieval per Turn

  8. Search Rounds

  9. Window Sizing

  10. Truncation

  11. Compaction

  12. The Compaction Bill

  13. Compacted Retrieval

  14. The Compaction Threshold

  15. Verbatim Tail

  16. Stateless API

  17. Retrieval Cost

  18. Context Tax

  19. Quadratic Cost

  20. Cost Forecast

Retrieval does not add a document to the conversation, it adds the document to the window. Whatever comes back from the store competes for the same cells as the system prompt, the question, and the answer. A retrieved document is split into fixed chunks, and every chunk takes a cell whether the model needs it or not, so a generous retrieval is quietly the largest thing in the window. [[rag-chunks-in-the-budget]] asks the next question: what happens when the chunks do not fit.

Paid members: the worksheet and its printable PDF are below ↓

This post is for paid subscribers

Already a paid subscriber? Sign in
© 2026 Tom Yeh · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture