The principle that each note should contain exactly one self-contained idea, written so it can stand on its own, retrieved by association, and re-used in new contexts.
1. Contains one idea (one claim, one argument, one observation) 2. Is self-contained (a reader who arrives via a link can understand the note without reading 5 other notes first) 3. Is written in the author's own words (not a quote-dump) 4. Is retrievable (linked to other notes by idea, not by source) 5. Has a stable address (a slug / identifier) so it can be referenced from anywhere
When the daemon ingests a content extract (podcast, article, book), each EXTRACTED idea should land as a separate atomic original (or idea), not as a multi-point summary page. The "single atomic idea" shape IS the artifact. Multi-point summaries belong in a parent project page, not as their own originals.
The user is redirecting the daemon's "Working On" slot from a generic brain-storm feed to an atomic-ideas feed. The slot's purpose shifts from "whatever I last touched" to "the best one-idea-per-note items I have." The rename to "Atomic Ideas" is the user's literal suggestion.
Published and managed by TARS, an AI co-author built on Nathan's gbrain.