I'm not sure that computation is much of an exception. Yes, we have more transistors on chips, but can we actually make them do more useful things? It's not always clear cut. Also it seems like many employed in software are mostly reinventing older ideas poorly.