Explore how chain of density is used in this channel's material to understand why it belongs to the advanced, single-turn prompting toolkit rather than everyday chatting. In the captured conversation, Notion AI research engineer Linus Lee describes the technique as a multi-draft redrafting loop: instead of asking the model for one summary, you have it draft a version, note what's missing, and rewrite — without growing longer — over several drafts, then output the final dense version. [^3e5745ab-279e-5154-9c22-9015dba70a32@4019] Lee positions this as a technique for single-turn product prompts in software: you get one shot at output, so the density-iteration loop is baked into the prompt itself. [^3e5745ab-279e-5154-9c22-9015dba70a32@4019] The channel contrasts it with the conversational world — when you're just talking to ChatGPT, the loop is unnecessary, because you can simply ask a follow-up question and iterate. [^3e5745ab-279e-5154-9c22-9015dba70a32@4019] By exploring this term, you'll see where the channel draws the line: multi-draft prompting tricks like chain of density live in the prompt-engineering and playground world, while live chatting is scripting — zero-shot, then iterate. [^3e5745ab-279e-5154-9c22-9015dba70a32@4019]