Early in my career I started out just writing code for every problem. Now I loathe the prospect of writing a line of code unless I’m sure it’s not a waste of time. It can be exploratory code, but I need to know in advance exactly what questions I want it to answer, so even if I throw it away it definitively tells me a path I should not go down.

I think this is a pretty common career progression. Coding can be fun and rewarding but I’ve written hundreds of thousands - possibly millions - of lines of code in a dozen languages and there is very little novelty left for me. The challenges of team organization, project planning, and personnel issues are more attractive right now.