Which is what your sibling comment mentioned. The author does touch on that approach but never touches on the drawbacks, some of which I mentioned in my other comment and some which are directly in conflict with the benefits of UUIDs. My point is merely that there are other details that a MSSQL user should take into consideration before simply following the author's suggestion. A novice MSSQL user might be better off with auto incrementing bigints in a way a novice user of a different RDBMS might not.
I think article was pretty focused on PG. :) unlike MSSQL it does not cluster data on an index (you can run cluster command but it will not be preserved as data is inserted updated) so that concern is fairly specific to MSSQL.
