All guidesLearning Memory
Learning Memory
AI student: "Where did I see this?" — one search across everything you touched.
Half of studying is finding the thing again. Learning memory searches your own trail through a course — not the whole internet, not the whole course: the parts YOU touched.
What it is
One query fans out across five personal sources: your notebook notes, the lessons you have worked through (title and content), your tutor conversations, the quiz questions you got wrong — including the AI review explanations written for them — and your own posts in discussions.
Every hit shows a snippet around the match, when it happened, and a deep link straight back to the lesson (or the notebook panel) where it lives.
Why — the pedagogical rationale
Revisiting material in your own context is retrieval practice with better cues: the note you wrote, the question you asked, the answer you got wrong are all memory anchors you built yourself. A search that only spans the course genericizes those anchors away.
Searching your quiz misses also resurfaces the why — the AI review explanation names the misconception behind the wrong answer — turning a lookup into a micro study session.
How we achieve it
The search runs as five parallel lexical queries over only your own rows, with wildcard escaping so your query matches literally. Lexical matching is deliberate: the sources have no shared embedding space, and the goal is recall of where you saw something — not a chatbot’s opinion about it.
Where to find it
On the twin page (/twin), under Learning memory. Notebook hits jump straight into your notebook panel; everything else deep-links to the lesson.