Andy Matuschak

Design friends: I highly recommend @edelwax's new talk! [link] Years ago, Joe wrote a wonderful talk on the challenges of metrics called "Is anything worth maximizing?" This is the followup, presenting a design approach intended to escape HCD's goal orientation.

Was reminded recently of one of my favorite papers on social learning. It centers on a clever lie! Subjects study some material, then quiz either a peer or a bot about it. They learn much better when they think they're quizzing a peer. Secretly, in both cases, it's a bot!

Some tantalizing discussion of the instigation of Illustrator on p. 84, but I wish it were more than a few short sentences. Still, lots of interesting details here—thank you!

In case you haven't seen, there is this great 2018 talk from Chan and a couple excellent colleagues (miss you guys!):

[link] I found the Four Star books helpful for learning techniques (things to note when scanning a piece before beginning to play) but they must be supplemented with tons more at-level music—not enough in each book to pass each level IME.