Here is how this mindset plays out in practice.
Starting a green field project looks great on a performance review,
so everyone wants to do that. It is how we end up with things like
the graveyard of abandoned Google products.
This is a problem I don’t know how to fix.
I feel like I’ve been partially inoculated to this process because of writing code. There’s something viscerally satisfying about creating PRs that remove more code than they add, or which only remove lines of code. I do feel it’s easier with code than it is with more bureaucratic systems, like the example the article used with removing unneeded programs at a university. We understand that removing code might have downstream impacts in the form of bugs, while removing systems involving human labor can have downstream impacts on people’s livelihood. One of them is more palatable than the other.
As has been noted though, there’s still a pressure in software to produce new work rather than clean up old stuff. Removing dead code or tests, or refactoring to remove redundancies is not considered sexy, and you’re unlikely to receive any serious merit for it on your performance review. Part of that is the difficulty in selling the value of it (how do you quantify the time and effort saved from something no longer being there?) and part of it is probably this mindset in action.