Covert social signaling is constantly evolving though; trolls and other people actively look for ways to circumvent defenses, bypassing existing content filters or user filters. You can see this on Twitter; there's a busy secondary market in buying existing accounts (with various levels of age, quality, verification) and then naming and using them for trolling or influence campaigns. Dive into a contentious topic and note how many of the toxic participants have default usernames, very low follower counts, and (if you're really interested) atypical activity histories.
This is not a comment on the source article, whose arguments seemed largely reasonable.
These aren't just theories, these are how extremist ideologies get spread around the world, from ISIS to anti-government militias in the US.
I think most people are quite agreeable to filters (personal blocks and mutes, etc.) to deal with speech we find unpleasant. What we don't want is for those filters to provide cover for extremist recruiters.
obviously we all have a vested interest in ensuring citizens are educated and informed
People do not like words that do not confirm their own bias.