Read through the CHI '21 proceedings last week. My favorite paper was this very interesting AR tutor, which features in-context guides and multiple levels of scaffolding, adapted in realtime to user behavior.
PDF: andymatuschak.org/files/papers/H
Abstract/videos: dl.acm.org/doi/10.1145/34
Conversation
It does seem like AR can enable a much wider range of the in-context explanatory environments I describe here:
Quote Tweet
Not a board game manual next to the board game, but explanation *inside* the board game. Not a YouTube tutorial open next to Logic, but an explanatory medium *inside* Logic.
New "letter from the lab" on mass mediums which integrate doing with explanation: andymatuschak.org/doing-centric/