This is not really a contradiction, IMHO. I published about systems 15 yrs ago and experienced the crisis in systems research. My personal take is, that the academic community failed to set new standards and research methods for computing system. (Maybe they cannot even exist in our short lived publish or perish world) I still review for conferences that still have systems (like ieee percom), but it is really difficult to reach consensus on system papers in comparison to AI or HCI stuff, that is also published.