I mostly don't, but there is still a disconnect here.
Who wants to not know the sort of group they belong to?
Censoring things that are not acceptable at all is different from censoring (or reordering) things that are determined to be sub-optimal for a particular person to see, that come from members of the group who are not going to be banned.
So if you go to a listing of groups, without some active curation, how do you prevent it from being that kind of content, or hate groups, or other undesirable content? The standard gets real blurry real fast.
As I contemplate the subject, I think I draw the line between a set of human-understandable rules, enforced by some combination of human and machine partners, versus a set of goals that drive a ruthless optimization process that is opaque.
