I’d interperet the cease qualifier in goodhart’s law to require the measure to be good in the first place, but I think SLOC is a negative indicator regardless of whether it’s a target

I think there are parallels between Deutsch’s ideas about explanation and the concept of coincident vs inherent complexity