alt Hacker NewsDoing it by hand now. Besides the site itself, I have spent most of my time building out the backend automation. I imagine I can manage the first 10k users myself and opening it up slow enough that a community itself develops and will be self-policing.
I got back on instagram recently and the first thing I see when I login now is content from repost accounts and it's like, do I really want to see this? Is this what happens when we optimize for engagement blindly? A hypothesis I have is that if we make rules in the beginning this will be easier to enforce at scale. See an account that's just reposting stuff -> report it -> review
I hope this isn't coming off as snarky or anything, I am honestly interested in the project...
So you're thinking something like Twitter(X)'s community notes, but instead of flagging misinformation, it will be used to 'community flag' an account as AI? Maybe it shadow bans, maybe it brings it to moderation queue, whatever...
I've seen a few of these 'NO LLM ALLOWED' sites pop up now, but any countermeasure to LLM use is easily subverted. I think the only real way to keep the LLMs out is, as you seem to agree, some sort of user-reported system (But that seems open to fraud....difficult problem to solve!)