The system I'm proposing isn't anything like reddit. Other than on /b/, most of the other boards on 4chan had their own rules, r9k had the robot that filtered speech and images.
The owners of an imageboard can use the system to FILTER the absolute worst, trashiest comments that don't add anything of value to the debate. That's no different than Wizchan having its 'Global Rules' which already includes a rule like "Do not submit low quality, inflammatory or off-topic posts."
It can also be used simply to present information to other users about the LLM's assessment of various dimensions to the comment. This is not the same as people voting up or down comments, as it wouldn't necessarily affect visibility of comments or where they appear. This thread would exist exactly as it does now, but perhaps to the left of the post number you might have a clickable\expandable element "Comment Rating", or perhaps in some cases where one factor was particularly pronounced, you could have something more prominent, for example your comment >>307727
could be labelled "This comment has appeals to belongingness".
Or, as I've already said, the user could simply be warned of that before posting, the moment you hit submit, your comment could have read "Warning: our system has rated your comment as moderately appealing to belongingness, please re-consider what you're posting", and you could perhaps edit it again before posting it for real. Maybe you would ignore it, but for others who are more considerate they might hold back and reconsider and change up their argument a bit. Up to the owners.
Seriously, in response to my comment, whether you agree or disagree with me, does this comment:
>Stop suggesting it for every practical problem you paid shill.
Really add that much to the conversation? Is it constructive or is it combative and utterly stupid? The LLM wouldn't necessarily be banning IDEAS or ARGUMENTS from being expressed, the author of this comment could have easily put foward a coherent argument for why they suspect I'm somehow in the employ of OpenAI or whatever.
The LLM could rate comments on a variety of factors. Here are some that I think are important:
-BePost too long. Click here to view the full text.