By revealed preference, modern academia is about churning incremental papers in order to obtain tenure and research grants.

While i don't disagree that there's a problem with publishing for the sake of publishing, we really do need the incremental papers. GPS (the evergreen science example) is the result of incremental papers starting in the 40s with grants for atomic clocks

Yeah, those greedy guys just want to (1) keep their (underpaid) jobs and (2) have a bit more money to increase pay for their (underpaid) students. :) As with software engineers, and every other job, some academics are crappy and some are amazing. Let's not paint everybody with the same brush. As for incremental, well, the AI you're all so fond of was incremental academic research for a couple decades before it was this sudden world changing thing.