Hacker News Reader

I'd argue against this. From the things that I see removed from those I follow on instagram, I would never label as porn. And then they loose their entire account based on extremely puritan value judgements.

The whole argument being made here is about just those "edge cases."

What the Twitter post asked for was doing things in a "fair, reproducible, and representative way", rather than just being correct most of the time.

I suspect the vast majority of the time the misinformation removal is accurate as well. People tend to post the same dumb stuff over and over.

Even if that's true, you can still moderate in a fair, reproducible and "representative" way, because the goals are different. In the case of NSFW content, the filtering can be opt-in: "don't show me this stuff, because I'd rather not see it". The point of moderating "misinformation" is not to let anybody see it, because you don't want them to see it.