Per : when griping about bad metrics in fuzzy fields, too easy to go 180° & argue for anything. How can we *partially* use quant?
Conversation
Wrote on this a bit: blog.andymatuschak.org/post/159340765; & from : medium.com/quora-design/s. Still ?: how much to weight what when—how to sum.
2
7
The correct combination is necessarily unknowable—unknown unknowns & all—but we can do better than whim.
Kindling: meaningness.com/metablog/bonga