alt Hacker NewsThis seems fine as a short-term solution, but human-only is no good as a long-term rule. The AIs will soon surpass human capability. Even in the present, I think some AI comments are already decent quality. It's just most of them aren't high quality yet.
And I'm worried banning AIs altogether will eventually lead to some form of prove-you-are-human verification to use the site, which will reduce anonymity. Even something seemingly benign like verifying email would mean many unverified accounts like my own will disappear.
And there is a legitimate use for LLM rewrite to counter identification by stylometry, so rewrite shouldn't be banned. I think we'll have to allow the AI stuff at some point, and make a system that incentivizes quality posts regardless of where they come from or how they're written.
Replies
> The AIs will soon surpass human capability.
The rule can be revised later.
> I'm worried banning AIs altogether will eventually lead to some form of prove-you-are-human verification to use the site, which will reduce anonymity.
Of all the sites on the Web to worry about this happening, HN is low risk. Oppose that change if it comes, not this one.
> And there is a legitimate use for LLM rewrite to counter identification by stylometry
Source for comment-level stylometry ever actually being someone's downfall, despite availing themselves to every other much more standard defense measure? Regardless, if your experimental means of deanonymizing yourself comes at the expense of the site's quality, it is probably not welcome.
"prove you are human verification" as in something like Sam Altman-backed World and The Orb [1]? Or maybe even the bead [2] (backed by me)
I don’t care to read a comment that nobody put their time in.