If keep Priority mode on, you won't see casts like this from people you don't follow. 

Casts like this from people you don't follow make people mad.

Is this spam? How would you classify? What happens when a "good" user does this ironically?

I've been just trying to use mute liberally. If user receives too many "mutes", would just analyze their casts for further processing (spam, bot, troll, etc). 

Understand this can be used maliciously so have to account for that somehow.

Maybe they should realize their casts are hot garbage! 

Definitely a tough problem. "The algo" will always come under fire in any socially-driven app. Using analytics/data science to tune the algo seems to be the generic answer but if the wrong target metrics are used, you end up with the mess we see in the other apps

This what we’ve been doing for a while and then people get mad that they are labeled spam. 

“I’ve been shadow banned”