I made a new update to my experimental learning app, Flashbang: we've now got LLM-graded spaced repetition review.
The video walks you through a quick demo.
I also wrote a post about it on Substack.
I've got some open source history textbooks I'm chopping and processing via Claude and GPT, which are then organized into structured atomic "learning legos" and coordinated into an SRS review system. Having a lot of fun!
Next up I'm going to experiment with visualizations and also explore the relationship between my course's content (which is going to be world history 1500-or-so to present) and the sources it draws from.
OpenStax is full, libre Creative Commons licensed, which makes me feel free to really experiment with LLMizing it and remixing. But at some point I think I'll need to introduce a compelling narrative shaped out of all its factual info.
If you'd like to see more, subscribe to the Substack at flashbangapp.substack.com.