Interesting newsletter on “open-ish-ness” in ML, and more broadly:
Reminds me, too, of @brianwdavis’s post: