I think that's not quite the point in the article. The idea is, in my reading, that we've built lazily on castles of sand for so long that sometimes we think it

I think that's not quite the point in the article. The idea is, in my reading, that we've built lazily on castles of sand for so long that sometimes we think it makes sense to throw away things we shouldn't, and other times we obsessively wrap/rewrap/paper over things we should throw away. What falls into each category is obviously debatable, but the author seems to be critiquing the methodology we use to make those decision--debatable or not, people aren't debating it so much as they're taking the shortest and often laziest path without prioritizing the right things (efficiency, consistency). Even with our priorities in order, there will still be contentious, hard choices (to deprecate so-and-so or not; to sacrifice a capability for consistency of interface or not), but the author's point is that our priorities are not in order in the first place, so the decisions we make end up being arbitrary at best, and harmful/driven by bad motivations at worst.