What it solves
Students re-read their own work and feel certain they’ve understood. They haven’t — they’ve recognised. The gap between recognition and retrieval is one of the largest illusions in learning, and it widens dramatically in the AI era, where every topic feels familiar because an LLM just explained it fluently.
Ghost Map breaks the illusion visually.
How it works
After you reconstruct a concept on the canvas (Step 2 of the learning cycle) and answer the AI’s Socratic questions (Step 3), Ghost Map activates. Fluera generates an ideal concept map from the lecture material and overlays it — semi-transparent, aligned to your canvas — on top of your work.
- Green — concepts you wrote correctly, connections you drew right.
- Red — concepts the ideal map has that yours doesn’t. You never wrote them, or you wrote them wrong.
- Yellow — connections you drew that don’t match the ideal. A link you thought existed but doesn’t, or points in the wrong direction.
- Blue — connections the ideal has that you missed.
Tap a red or yellow node. You try to write the correction — by hand, on your canvas — before Fluera shows the ideal answer. Then the overlay reveals. The bigger the gap between your confidence and reality, the more durable the correction becomes.
The science behind it
In 2001, Brady Butterfield and Janet Metcalfe documented the hypercorrection effect. Counterintuitively, errors made with high confidence are corrected and retained more permanently than errors made with low confidence. The proposed mechanism: the surprise of being wrong when you were sure you were right triggers amygdala activation, which modulates hippocampal encoding. The correction gets a neurobiological highlight.
Ghost Map operationalises this. The confidence slider in Step 3 is not aesthetic — it primes the emotional contrast. A confidence-5 wrong answer, revealed by Ghost Map, leaves a far deeper trace than any number of passive re-readings.
What’s coming
- Concept-aware diff for mathematical equations (currently recognises shapes and text clusters).
- Per-domain Ghost Map voices — one tuned for STEM, one for humanities, one for law.
- Replay mode — step through how your canvas diverged from the ideal, move by move.