Developer: Listing Rules for AI Never Ends — Describe How the Answer Should Feel

gabriel1 · x · 2026-09-18

Developer gabriel1 argues that stating rules to make an AI assistant answer correctly is information-sparse: new edge cases keep forcing new rules forever. Instead, he wants to describe how the reader should feel when reading a response, which packs more information per token — a thought-provoking take on prompt and agent instruction design.

Original post →

More from coding & agent

coding & agent channel →