Are you downzapping suspected AI posts and comments? \ stacker news ~AskSN

pull down to refresh

Are you downzapping suspected AI posts and comments?

620 sats \ 24 comments \ @SimpleStacker 15h AskSN

I've noticed a pretty high rate of new users having all their posts and comments outlawed. I suspect it's because they're getting downzapped for their AI-like posts, but it seems to happen pretty quickly and I'm wondering what your internal algorithm is for downzapping. I suspect that a high rate of emoji and hashtag use is a give away, as well as somewhat tangential engagement with the topics at hand. Still, I'd like to know what other peoples' internal algorithm is for this.

view all related items

35 sats \ 0 replies \ @WeAreAllSatoshi 5h

Sometimes I do, usually when they feel like GenAI regurgitated summaries of the post they’re comments on without adding value

1492 sats \ 12 replies \ @kepford 15h

I have been doing this for a while. Not so much recently because I mostly ignore posts not created by "trusted" stackers. Honestly, I don't care if its an AI post at this point if its good. Many people act like bots(NPCs) anyway. Most people on the Internet at least.

78 sats \ 5 replies \ @Undisciplined 15h

Same. If a post has something worth responding to, it doesn’t make much difference if it’s a bot of some sort.

I ignore comments that don’t seem very thoughtful, regardless of what made them.

179 sats \ 4 replies \ @kepford 15h

You know what I'm finding the more I use AI and test things out? Its the humans that suck at using it. Its a tool and like many tools it requires a bit of skill to use effectively. Bad input will get back output. That's when I can tell its AI. When the "prompter" is so lazy they haven't learned how to use it well.

Given really good prompts and focused input I'm pretty sure most people can't spot the simulated intelligence. Its not really just artificial but rather a simulation of intelligence. The more I learn about this tech the less scary it is and the more I can see it being eventually very useful.

Is it going to be used to create slop? Yep. Its gonna get way worse but I'm gonna adapt. No fear. Adapt and thrive.

107 sats \ 2 replies \ @SimpleStacker OP 14h

Yeah I'm seeing way more emojis everywhere and it just tells me that whoever is using these AIs don't really understand how to make it talk like a human and for that matter don't seem to know themselves how regular people talk, in English at least.

72 sats \ 1 reply \ @kepford 14h

Oh yeah. The over-use of emoji is a sign of AI gen stuff.

Most people don't seem to even try to give the context needed to simulate an actual person. When you do that you get better results.

For example, if one wanted to imitate me you'd need to throw in a lot of mistakes.

42 sats \ 0 replies \ @SimpleStacker OP 13h

I wonder what kind of godforsaken training data resulted in them using so much emoji

reply on another page

0 sats \ 0 replies \ @Undisciplined 6h

That makes sense to me. It’s not uncommon for a person to be trying to simulate how other people do certain things while writing, so an AI doing that part probably doesn’t sound much less authentic.

64 sats \ 5 replies \ @Car 14h

Same "trusted" stackers, will respond and zap.

126 sats \ 4 replies \ @kepford 14h

Yep.

SN reminds me of the very early days of Twitter. Long before "everyone" was on it. I remember following about 50 people in the city near me. I met all of them IRL and it was a massive help to me in my career. I learned so much from these people. It was an online circle of people that shared a common interest. It wasn't about likes or clicks.

SN is much better but one reason is because its rare to find a site like this today. I hope @k00b realizes this. :)

63 sats \ 3 replies \ @Car 14h

I believe he knows, also I think he knows everyone in the lab loves SN so he gets instant feedback + you are right it won't be a secret forever.

5 sats \ 2 replies \ @kepford 14h

I've written this before but my problem with Nostr isn't the tech but the way people use it. IMO for SN to grow without destroying its current appeal to me would for it to have many communities. Not just grow the current one. I know that was part of the idea behind territories.

But I'd be fine if it just had gradual growth vs. massive growth. Twitter's value to be in the early days is completely gone now. Nostr is better but not like SN. Not even close.

50 sats \ 1 reply \ @Car 14h

I've written this before but my problem with Nostr isn't the tech but the way people use it.

We have a few new builders in PlebLab working on some interesting Nostr tech taking it in a different direction, doing my best to get them to think outside the current trends.

But I'd be fine if it just had gradual growth vs. massive growth.

Same, but I suspect it will be an overnight sensation.

view all 1 replies

30 sats \ 0 replies \ @ken 6h

Thanks for the great 😃😃😃😃 post! #nobotshere

52 sats \ 0 replies \ @optimism 10h

I think it's the same person. I'm downzapping.

30 sats \ 1 reply \ @rootmachine 12h

Never thought of down zapping and didn't even knew that possibility exists until now. Thanks for bringing this up to me. Some posts do need downzap: crypto scams and this sort of fake things. SN has to remain clean otherwise it will convert super soon into something hard to navigate through.

41 sats \ 0 replies \ @teemupleb 5h

I downzap clickbaity post titles (even if they got a lot of zaps and comments).

Examples: “I never knew this was true” or “So this happened to me today” etc.

For me to open a post and zap it, it has to be earned, starting from the title!

130 sats \ 1 reply \ @Aardvark 15h

I rarely downzap, and when I do, it's because the stacker is being a piece of shit.

105 sats \ 0 replies \ @Car 14h

🤣

30 sats \ 0 replies \ @Undisciplined 15h

Enough downzaps and your posts start out outlawed. My guess is that ek downzapped the bejeezus out of them.

30 sats \ 0 replies \ @OT 15h

I do when I see a crypto scam. Not sure if it's AI or not.

107 sats \ 0 replies \ @Scoresby 11h

I pretty much never downzap. I don't particularly care if someone uses AI to write.

What is frustrating is to read something that feels like thought was put in to it, but ends up being opaque or confusing and genuinely pointless. With a human-written post, you can usually feel confident that there was a point -- even if silly. But with AI, it may just be words strung together. Humans don't do things for no reason. AI may do things for no reason other than statistical likelihood.

Perhaps the end state of AI slop is kind of like pop-psychology or pop-spirituality: a bunch of nice words that sound good together but nobody has ever tried to live by. In that case, large chunks of the internet are already like this.

5 sats \ 0 replies \ @d680ecaa8e 8h

If the person well know how to use AI and add its personal touch then I will upvote and zap, if he just copy paste AI response like it without adding effort I will downzap. When reading a text I could differ.