Orthogonal to escaping, the choice is what characters to use for unit separator and record separator.
If the data are for machines only, then for me the choice of characters doesn't matter. If the data are potentially for reading or editing, such as by a programmer, then my choice is to prefer typically-visible characters over typically-invisible characters and/or zero-width characters (e.g. ASV a.k.a. DEL a.k.a. ASCII 30 & 31).
My choice of USV is thus because U+241F and U+241E are visible, and also in Unicode they are semantically meaningful.
I'm still not sure what the value is over CSV, which also has visible delimiters. It's true that you have to establish/enforce a specific convention around escaping and quoting, since CSV has historical variation here. But it would make more sense to me to encourage any particular consistent handling of CSV, rather than yet another entirely new separator. At least some tools already support CSV, whereas nothing currently supports USV, as far as I know.
Orthogonal to escaping, the choice is what characters to use for unit separator and record separator.
If the data are for machines only, then for me the choice of characters doesn't matter. If the data are potentially for reading or editing, such as by a programmer, then my choice is to prefer typically-visible characters over typically-invisible characters and/or zero-width characters (e.g. ASV a.k.a. DEL a.k.a. ASCII 30 & 31).
My choice of USV is thus because U+241F and U+241E are visible, and also in Unicode they are semantically meaningful.