One day the labs will actually train models to de-slop and refactor a growing codebase... One day...

You'd think it would have been something they did a year ago, but here we are. Still.