I agree with your description of what the real, practical problem is, but I'd like to hear your opinion about solutions. Personally, I do think a viable answer

I agree with your description of what the real, practical problem is, but I'd like to hear your opinion about solutions. Personally, I do think a viable answer could involve rule-makers/enforcers making subjective judgements on sometimes unclear issues (even if that group must be small to feasibly operate). The practice is necessarily common because that's the nature of many of our social systems. We introduce as many abuse-curbing objective elements as possible, but in the end there will always be some amount of human judgement. If the balance tips one way, the system loses effectiveness, and if it tips the other way (the "slippery slope"), it becomes a new problem. Personally, in the example given, I don't think it would be impossible to find a balance that allows the mean-spirited speculation about why the first lady wasn't around, but disallows the borderline calls to violence and direct calls to hatred of people whose children have been murdered. Again, I agree that the problem is "how", but I absolutely am not pessimistic about it. I think we can use content moderation, despite its subjectivity, to find a balance - certainly not a perfect balance, but a better one than exists today.