So much of the stuff I have done here has been regarding Probes, but this might be one of the best ones yet.
The important idea is that the main mechanism behind what makes anki effective as a spaced repitition system is how it already leverages the telematry of how you rate cards you see to gauge your understanding, and uses a Neural Network approach for this for the FSRS algorithm. Along with other things like Self Configuration For Chess, could serve as a great testing ground for long term efficay of Neuropharmacology / Utilizing Cognition Enhancement Effectively
Has two signals: immediate working memory evaluation for seeing how quickly one learns new/errored cards during a session for a day, and then one for long term memory formation in how cards reviewed/accumulated over a few days are immediately recalled without prior practice for that session.
Also a good implementation of Consuming and Choosing Content Effectively, especially when you can use coding agents to make cards as you go as defined by the Spaced Repetition note.