Live data from Hacker News

Show HN: Slop or not – can you tell AI writing from human in everyday contexts?

slop-or-not.space

11–20 of 22 posts

Re: Show HN: Slop or not – can you tell AI writing from human in everyday contexts?

#11
post #10

Was able to get an 8x streak. The question that made me lose it was really hard, I basically took a guess. Some were hard but spottable after re-reading the answers a good 10 times... ahah.

thank you so much for taking a look :) Yeah you'd be surprised how difficult it can get to spot nuances sometimes. Sometimes, there isn't any nuance at all and the AI is just as good at writing about something pretending to know about the topic.

Re: Show HN: Slop or not – can you tell AI writing from human in everyday contexts?

#14

By playing this game I'm helping to train AI how to be less detectable?

Ha :) I'm not building models, nor am I affiliated with any big labs. The idea is to use this to educate people how to spot tells of AI writing. Although like any data that's made open this can be used to train future models as well I suppose.

Re: Show HN: Slop or not – can you tell AI writing from human in everyday contexts?

#15
I keep accidentally clicking on the human one because my brain wants to treat it as “find the human content”.

FWIW, I found the “medium” one’s hardest. Most of the “hard” ones have dead giveaways in the form of either punctuation or common AI text rhythms.

Re: Show HN: Slop or not – can you tell AI writing from human in everyday contexts?

#16

I keep accidentally clicking on the human one because my brain wants to treat it as “find the human content”. FWIW, I found the “medium” one’s hardest. Most of the “hard” ones have dead giveaways in the form of either punctuation or common AI text rhythms.

Thank you for trying! I first built it as 'detect the human' response, but that was counter to the 'slop or not' framing. Yeah I'm also observing the same based on the first few hundred people's results. The harder models seem to write almost too well and that's generally not how humans write on the internet unless it is a blog post/essay. The easier models seem to be the ones that are tripping people the most.

Re: Show HN: Slop or not – can you tell AI writing from human in everyday contexts?

#17
I got 19x. When they say "curious about" it's always a good sign that it's AI, same with X not Y construction, saying "genuinely", saying things like "absolutely slaps" and other millennial slang, being overly positive: generally sounding like the transcript of an Instagram food review. When they're trying to be casual they seem to default to some kind of 2017 millennial stereotype. Typos and "edit:" are always a good sign that it's human, so I'm sure people will start adding those in to AI-generated text to seem more real

Re: Show HN: Slop or not – can you tell AI writing from human in everyday contexts?

#18

I got 19x. When they say "curious about" it's always a good sign that it's AI, same with X not Y construction, saying "genuinely", saying things like "absolutely slaps" and other millennial slang, being overly positive: generally sounding like the transcript of an Instagram food review. When they're trying to be casual they seem to default to some kind of 2017 millennial stereotype. Typos and "edit:" are always a goo…

Thank you for taking a look! Yeah there's definitely a few tells which are noticable if you look for it :)

Re: Show HN: Slop or not – can you tell AI writing from human in everyday contexts?

#19

I keep accidentally clicking on the human one because my brain wants to treat it as “find the human content”. FWIW, I found the “medium” one’s hardest. Most of the “hard” ones have dead giveaways in the form of either punctuation or common AI text rhythms.

Yeah the UX itself is clearly slop

Re: Show HN: Slop or not – can you tell AI writing from human in everyday contexts?

#20
I did a couple dozen and wound up with a success rate of 90%. That's less than I hoped. That means I'm likely missing even more slop when I'm not keyed into searching for it, with two options to choose from.

I know I've been finding it all over the internet, but this suggests to me I'm encountering it more than I thought.

Post reply on HN