Can a shitpost be detected algorithmically?
Can we use CRM114 and other patterns in conjunction to detect shitposts?
Bill Yerazunis gave a great talk in 2010 at Black Hat about using a series grey filters to generate an outbound filter for classified materials in a language that he didn't speak: https://www.youtube.com/watch?v=JKB5CojW4AA
Can we do something similar?
As an aside, does anyone know how sophisticated sentiment analysis is currently?