Very interesting, how do you validate the correctness of the generated visuals? And that a previous lesson, say 5 videos ago ensures that the concept being explained now had its prerequisite covered?
I assume you could do a topological sorting of the content to ensure prerequisites are covered, but I'm curious on how automated this could be.
Looks super cool either way!