[repolfx]

Depends what you mean by "a forum like HN". Before HN the main geek watering hole was Slashdot, which famously never censored content and fought strongly against attempts to force it to do so.

Slashdot also had a rather sophisticated moderation and scoring system, that allowed spam (hot grits etc) to be downranked and appear auto-collapsed, whilst longer form content was upvoted and expanded by default - even if it was a reply to negative ranked content.

You may feel a personal preference for HN, or not, but they were essentially the same from the perspective of any lawyer.

In other words HN could easily keep its distinctive feel without ever banning or erasing anything, just by implementing sufficient controls that let users see what they want to see: in fact it already does via options like showdead.

[potatoz2]

The idea that it's OK to implement an algorithm that takes some user's input to censor posts but it's not OK to do so directly seems like it skirts the central question.

If I have upvotes and downvotes, weighted the way I like, and I limit who can get an account in the first place, I can probably achieve the speech outcome I want relatively easily "without" human intervention.