Suggest related ideas shows up in this channel's material as a dedicated NotebookLM prompt — and, more interestingly, as a workaround. In the walkthrough of NotebookLM with Steven Johnson (Ep.22), the tool's notes and sources are separate contexts: you can focus the model on one or the other, but not both at once. Johnson's workaround for that constraint is the dedicated “suggest related ideas” prompt. In the live session, after the “pure oxygen” answer leads him back to his reading-notes notebook and he saves an Auguste Piccard quote — the sealed 1930s stratosphere gondola that relied on a pure oxygen supply — he selects that note as the focus context and runs the prompt against the NASA oral-history sources. It surfaces Gene Kranz on complacency about pure oxygen, Jonathan Yardley's oral history on atmospheric control, and redundancy gaps relevant to the Apollo 1 fire [^0cc0dd07-dc88-5d83-8cf7-2f6e48821796@2779]. The channel's step-by-step framing of the session names this move hunting for parallels: pin and curate what resonates into notes, then ask the sources to relate to a note. In that loop, suggest related ideas is the bridge between a curated note and the wider source pile — the mechanism that lets a documentary concept assembled from saved quotes (Piccard's gondola) reconnect with 200,000 words of unassigned NASA transcripts, ending in a generated opening script linking the two sealed, pure-oxygen capsule stories [^0cc0dd07-dc88-5d83-8cf7-2f6e48821796@52].