Reminds me of Soundex, where words can be encoded to a shorter number of characters:

https://en.m.wikipedia.org/wiki/Soundex

Maybe someday AI can generate novel media encodings. The rest of the blanks are filled in via inference client-side.