Just to clarify, I only use UUID v4 myself, for the reasons you specified against using UUID v1.
That said, I do not disagree with any of the points you made. My perspective with such things is that I like to get a solution out the door as quickly as possible and iterate on that. If I was developing software where I needed random values for session/transaction tracking, as soon as I saw Node-UUID used the crypto library, I probably would have gone with it.
Now if the project instead involved say, trillions of records, that’s the point where I might consider using a more compact number scheme if there were significant performance savings to be had.