upvote
im thinking about making a hackernews but only for articles written before 2021 (or so).
reply
The underlying problem of distinguishing honest contributors from AI hustlers is quite hard, but worthwhile to solve properly.

Maybe there is a way to reduce incentives for AI hustlers to join a certain platform, while attracting honest contributors. But even honest contributors might have a bad day or a new project and suddenly they're out in promotion mode.

reply
Sadly I don’t believe ai slop is a solvable issue - we will never be able to filter out AI slop, because it is indistinguishable from human slop (and AI used right could be better than what that person could produce otherwise, eg fixing spelling and language).

I think the real problem is just trash content - content I could produce without AI in almost as quickly. The solution is probably more about determining if the content is interesting and well written, no matter who wrote it or how.

reply
Yes, it's a human-level problem. Everybody knows from school the people who did not help with the group work, and the ones who actually did their part and were nice to work with.

However are the "good ones" actually hanging out on social media? Or is it just the mediocre ones who are wasting time? "Learning new things" or something being interesting is also highly dependent on one's current knowledge level and skills.

The remainder of the content is ragebait and "discussing" with others so one can feel better about themselves. It is a really nice comfort zone where you don't have to leave your work chair in order to tackle other TODOs in your life while also not feeling like you wasted your afternoon.

I feel the most true version of HN are the "Show HN" posts where somebody did some actual work and is eager to get feedback and judgement from their honored peers. Unfortunately due to AI this has been drowned by sloppy humans as well.

An upvote by the biggest idiot one knows is worth the same as an upvote by a nobel prize winner.

reply
deleted
reply
Good catch!
reply