Back to glossary

A reading memory is the durable, queryable layer of what you have read — built automatically from bookmarks rather than authored as notes, organised as concept pages rather than time-ordered highlights.

Why it matters

The reading-to-writing ratio for knowledge workers has inverted in the last decade: reading is up roughly fivefold since 2015, authoring is flat. The tools we use to organise what we read — notes apps, bookmark managers, read-later apps — were designed when the ratio ran the other way. They optimise for the act of writing things down. A reading memory inverts the assumption: the unit of organisation is the article you saved, not the note you wrote about it.

The concept is adjacent to but distinct from three sibling categories. **Personal wikis** (Obsidian, Roam) require manual authoring. **Bookmark managers** (Raindrop, Pinboard) store URLs without synthesis. **Read-later apps** (Readwise Reader, Pocket) optimise for the act of reading itself. A reading memory composes the substrate beneath all three: capture is free, synthesis is automatic, retrieval is semantic.

The word *memory* is load-bearing. A library is a stack; an index is a lookup; a memory is something that returns the right thing without being asked. The 30th article you read on a subject teaches more than the first three combined — but only if the prior 29 are findable when the 30th asks them a question.

How Pith relates

Pith is built around the reading-memory concept. Bookmarks are the input substrate; per-concept wiki pages emerge automatically with citations back to the sources; semantic search retrieves by meaning, not keyword. You read the way you always did; Pith makes that reading remember itself.

See also

Last reviewed: 14 May 2026 · Licensed CC BY 4.0 · cite freely with attribution to Pith.