Why does the same explanation land differently for different people?
Recipient design
People shape what they say for a particular listener, using assumptions about that listener’s knowledge and situation.
An everyday example
A beginner needs “the material the assistant can use right now”; a developer may need token limits and retrieval details.
Original illustration, not an observed participant case.
Where the idea stops
The concept comes from human conversation. Applying it to an AI interface describes the design of the explanation, not proof of human-like understanding.
Something to try
Explain one idea to a colleague new to AI, then to a builder. Keep the meaning constant while changing the example and detail.
Practise on a fictional case →Go deeper: the builder’s view
For builders: let people choose depth and correct assumptions about their needs. Technical vocabulary alone is a poor measure of ability.
Source and interpretation
This is a plain-language interpretation with an applied example; it is not a quotation or an endorsement from the source.
Sacks, Schegloff & Jefferson — A Simplest Systematics for the Organization of Turn-Taking for Conversation (1974) ↗Catalog v0.1 · checked 2026-10-29