Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

SlopStop: Community-driven AI slop detection in Kagi Search

blog.kagi.com

Post reply on HN
Print view View original

171–180 of 271 posts

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#171

Post by throwaway2037 »

notepad0x90 wrote: post #144

I wish a smarter person would research or comment on this theory I have: Training a model to measure the entropy of human generated content vs LLM generated content might be the best approach to detecting LLM generated content. Consider the "will smith eating spaghetti test", if you compare the entropy (not similarity) between that and will smith actually eating spaghetti, I naively expect the main difference would b…

    > Consider the "will smith eating spaghetti test"
I thought this was a casual joke... then I Googled it. Yep, it's real: Consider the "will smith eating spaghetti test"

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#172

Post by sjs382 »

laacz wrote: post #60
Thrymr wrote: post #57

Earlier quoted context omitted.

Explicitly in the article, one of the headings is "AI slop is deceptive or low-value AI-generated content, created to manipulate ranking or attention rather than help the reader." So yes, they are proposing marking bad AI content (from the user's perspective), not all AI-generated content.

Which troubles me a bit, as 'bad' does not have same definition for everyone.

That's fine.

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#173

Post by input_sh »

estimator7292 wrote: post #90
input_sh wrote: post #48

Earlier quoted context omitted.

It's not the only thing the tool does, as they also publish that regurgitation publicly . You can see it, I can see it without even having a Kagi account. That makes it very much not an on-demand tool, it makes it something much worse than what what ChatGPT is doing (and being sued for by NYT in the process). > They provide attribution to the sources. It's listed under the headline "Sources" and is right below the sh…

> as I've demonstrated You have not, you've thrown a temper tantrum

Sure thing bud. Thank you for your well thought out counter-argument.

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#174

Post by raincole »

notepad0x90 wrote: post #144

I wish a smarter person would research or comment on this theory I have: Training a model to measure the entropy of human generated content vs LLM generated content might be the best approach to detecting LLM generated content. Consider the "will smith eating spaghetti test", if you compare the entropy (not similarity) between that and will smith actually eating spaghetti, I naively expect the main difference would b…

It might work for real photos vs AI-gen photos, but I really don't see how 'entropy' is so important when distinguish human-gen text from Ai-gen text.

I also don't see why AI can't be trained to fool this detection.

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#175

Post by raincole »

dvfjsdhgfv wrote: post #76

So we have two universes. One is pushing generated content up our throats - from social media to operating systems - and another universe where people actively decide not to have anything to do with it. I wonder where the obstinacy on the part of certain CEOs come from. It's clear that although such content does have its fans (mostly grouped in communities), people at large just hate arificially-generated content. We…

Are you implying Kagi is on the "nothing to do with LLM" side? Even Kagi uses LLMs to summarize news.

https://github.com/kagisearch/kite-public/issues/97

LLMs just make too much economic sense to be ignored.

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#176

Post by Marsymars »

senderista wrote: post #136

Isn't the scalable approach to ask AI to identify AI (and have a human review the results, but that's required no matter what)? I also doubt most people will be able to detect AI text generated with a non-default "voice" in the prompt.

> I also doubt most people will be able to detect AI text generated with a non-default "voice" in the prompt.

I'll grant you that if someone is careful with prompts they can generate text that's difficult to detect as AI, but it's easy to see that in practice, web results are still full of AI-generated slop where whoever is publishing it doesn't care about making it non-slop-like.

Second to that, much of what I read or search for isn't amenable to an AI summary... like I'm very often looking for facts about things, where trust in the source is of primary importance, so whether I can detect text as AI-generated or not doesn't matter, what matters is that there's an actual source willing to stake their reputation, either as an organization or an individual, on what's been written.

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#177

Post by zrobotics »

anal_reactor wrote: post #96
Kuiper wrote: post #88

Earlier quoted context omitted.

There is a huge audience for AI-generated content on YouTube, though admittedly many of them are oblivious to the fact that they are watching AI-generated content. Here are several examples of videos with 1 million views that people don't seem to realize are AI-generated: * https://www.youtube.com/watch?v=vxvTjrsNtxA * https://www.youtube.com/watch?v=KfDnMpuSYic These videos do have some editing which I believe was d…

Hot take but I don't care if the content I consume is AI-generated or not. First of all, while sometimes I need high-effort quality content, sometimes I want my brain to rest and then AI-generated slop is completely okay. He who didn't binge-watch garbage reality TV can cast the first stone. Second, just because something is AI-generated it doesn't automatically mean it's slop, just like human-generated content isn't…

I mean, that certainly is a hot take, but you are getting down voted without people responding why.

I can certainly understand just wanting filler content just for background noise, I had the history for sleep channel recommended to me via the algorithm because I do use those types of videos specifically to fall asleep to. However, and I don't know which video it was, but I clicked on a video, and within 5 minutes there were so many historical inaccuracies that I got annoyed enough to get out of bed and add the channel to my block list.

That's my main problem with most AI generated content, it's believable enough to pass a general plausibility filter but upon any level of examination it falls flat with hallucinations and mistruths. That channel should be my jam, I'm always looking for new recorded lectures or long form content specifically to fall asleep to. I'm definitely not a historian and I wouldn't even call myself a dilettante, so the level of inaccuracies was bad enough that even I caught it in a subject I'm not at all an expert in. You may think you are learning something, but the information quality is so bad that you are actively getting more misinformed on the topic from AI slop like that.

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#178

Post by Marsymars »

ehnto wrote: post #150
FridgeSeal wrote: post #140

Earlier quoted context omitted.

> He who didn't binge-watch garbage reality TV can cast the first stone Stand by then, because I have rocks and according to you, licence to throw them. You are free to watch all the slop you want. All I want is for your slop, to not be at the cost of all other media and content. Have a SlopTube, have SlopFlix, go for it! But do it in a way that is _separate_ and doesn’t inflict it on the rest of us, who would _like_…

Your later point is hard to convey to people who don't want to hear it. I don't want AI content, even if it is as good, or even if it were better. The human element IS the point, not an implementation detail. An AI song about sailing at sea is meaningless because I know the AI has never sailed at sea. This is a standard we hold humans to, authenticity is important even for human artists, why would we give AI a pass o…

> And I mean this earnestly, if an AI in a corporeal form really did go sailing, I might then be interested in its song about sailing.

Would you? That seems achievable with current technology, bolt a PC with a camera onto a sailing ship and prompt it to compose text based on some image recognition.

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#179

Post by Marsymars »

ehnto wrote: post #154
laacz wrote: post #38

Earlier quoted context omitted.

Should we care? It's a tool. If you can manage to make it look original, then what can we do about it? Eventually you won't be able to detect it.

Objectively we should care because the content is not the whole value proposition of a blog post. The authenticity and trust of validity of the content comes from your connection to the human that made it. I don't need to fact check a ride review from an author I trust, if they actually ride mountain bikes. An AI article about mountain bikes lacks that implicit trust and authenticity. The AI has never ridden a bike b…

Haven't we given some AI agents access to potentially motherboard-bricking commands yet?

Re: SlopStop: Community-driven AI slop detection in Kagi Search

#180

Post by Marsymars »

School-Cotton wrote: post #66
abnercoimbre wrote: post #37

Earlier quoted context omitted.

> Even now if you put an effort into prompting and context building, you can achieve 100% human like results. Are we personally comfortable with such an approach? For example, if you discover your favorite blogger doing this.

> Are we personally comfortable with such an approach? I am not, because it's anti-human. I am a human and therefore I care about the human perspective on things. I don't care if a robot is 100x better than a human at any task; I don't want to read its output. Same reason I'd rather watch a human grandmaster play chess than Stockfish.

There are umpteenth such analogies. Watching the world's strongest man lift a heavy thing is interesting. Watching an average crane lift something 100x heavier is not.
Post reply on HN