221–230 of 271 posts
Isn't "detecting slop" an identical problem to "improving generative AI models"? Like if you can do one surely you can then use that to train an AI model to generate less slop.
I notice a distinction made in the docs for image, video, and "web page" slop. Will there be a way to aggressively categorize filter web page slop separately from the other two? There's an uncomfortable amount of authors, even posted on this forum, who write insightful posts that (at least from what I can tell) aren't AI slop, but for some reason they decide to header it with a generated image. While I find that dist…
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.
Matt also shared insights about the other signals we use for this evaluation here https://news.ycombinator.com/item?id=45920720
And we are still exploring other factors,
1/ is the reported content ai-generated?
2/ is most content in that domain ai-generated (+ other domain-level signals) ==> we are here
3/ is it unreviewed? (no human accountability, no sources, ...)
4/ is it mindlessly produced? (objective errors, wrong information, poor judgement, ...)
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"
HN could use some of this. It'd be nice if there was a safe having from the equivalent of high grade junk mail.
Earlier quoted context omitted.
> I wonder where the obstinacy on the part of certain CEOs come from. I can tell you: their board, mostly. Few of whom ever used LLMs seriousl. But they react to wall street and that signal was clear in the last few years
"Completely detached from reality" we used to call it. But where is the money coming from? Is it because we abolished the idea of competition, they never suffer negative impacts of bad decisions any more?
Earlier quoted context omitted.
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…
> An AI song about sailing at sea is meaningless because I know the AI has never sailed at sea. Human singers often sing about topics they have no authentic experience with. Some pop singers exclusively sing songs written by other people. You're entitled to dislike AI music, but I think your attempt to justify this dislike doesn't hold up to scrutiny.
Regardless of any of that, I could also say that I don't like AI music because I prefer my artists to have hot showers and it's somewhat none of your business, respectfully.
Earlier quoted context omitted.
> An AI song about sailing at sea is meaningless because I know the AI has never sailed at sea. Human singers often sing about topics they have no authentic experience with. Some pop singers exclusively sing songs written by other people. You're entitled to dislike AI music, but I think your attempt to justify this dislike doesn't hold up to scrutiny.
HackerNews holds AI to much higher standards than humans. In other news, water is wet.