In this case I would read "check-boxing" as "reducing to a simple optimisation loop without considering the implications of what you're optimising for".
Granted, I'm doing some pretty heavy interpretational lifting there, but that optimisation loop is something I see as common to productivity, video games, social media, and many other obsessions of the nerdosphere. It feels good to give your optimiser a constant stream of feedback, even if trivial.
Of course, now our optimisers are being manipulated by much larger optimisers that run on engagement metrics rather than dopamine. Where did the intuition to build these myopic hyperoptimisers come from? Surely not the same people who grew up minmaxing character attributes or building spreadsheets for optimal daily task efficiency? When you gaze into the optimiser...
Weird that we managed to turn society into a GAN.