@nprofile1q... non of these people I speak of ever ran into this specific issue :) My guess is they don't even know that uuids are twice the size of bigints - not just in tables but in indices, foreign keys and hash joins. The probably don't know that the writes are slower and the index performance impacted by fragmentation. In the beginning of a project non of this matters but as the data accumulates suddenly the next bigger RDS instance is needed and then comes the "oops thats expensive" :)