Live data from Hacker News

Hacker News, Sans AI

elijahpotter.dev

31–40 of 112 posts

Re: Hacker News, Sans AI

#31

Interesting this keeps moving up the front page when the site is inaccessible because it's hosted on a baked potato.

73% of users vote prior to reading TFA, according to this research. (I am sometimes guilty of this myself) We live in a world being dimished by confirmation bias, but this isn't a new thing. Those who wrote/approved the headlines always had more power than those who wrote the articles. https://www.researchgate.net/publication/315096490_Consumers... edit: disclaimer, no hate on TFA. Just responding to the comment.

This submission should be studied by people holding clipboards. It needs a follow-up:

"How I made it to the top of HN with zero content beyond a catchy title"

It further proves the key to getting your stuff on HN is not to post interesting content, it's to post something that sounds interesting.

Re: Hacker News, Sans AI

#32

Earlier quoted context omitted.

I'm curious to find out if they're using a model to detect AI content. I'd be more amused than disappointed. But, alas, potato.

It's not an LLM. Why use big program when regex do trick?

You can get the best of both worlds with a small embedding model

Re: Hacker News, Sans AI

#35

Earlier quoted context omitted.

73% of users vote prior to reading TFA, according to this research. (I am sometimes guilty of this myself) We live in a world being dimished by confirmation bias, but this isn't a new thing. Those who wrote/approved the headlines always had more power than those who wrote the articles. https://www.researchgate.net/publication/315096490_Consumers... edit: disclaimer, no hate on TFA. Just responding to the comment.

This submission should be studied by people holding clipboards. It needs a follow-up: "How I made it to the top of HN with zero content beyond a catchy title" It further proves the key to getting your stuff on HN is not to post interesting content, it's to post something that sounds interesting.

D1 hater right here.

Re: Hacker News, Sans AI

#36
post #18

I built (well, vibe-coded) a version of this a while back that runs against the Hacker News API, it's static HTML on GitHub Pages so I don't have to run a server for it: https://tools.simonwillison.net/hacker-news-filtered

How often do you use it? Immediately I started thinking how nice it would be to use natural language to have LLMs generate a deterministic filter for stories matching content I DO care about, filtered from New. Instead of filtering it out.

I don't use it myself. I built it for a Hacker News thread where someone else was saying "I wish I could browse Hacker News without all the AI stuff".

Re: Hacker News, Sans AI

#37
post #18

I built (well, vibe-coded) a version of this a while back that runs against the Hacker News API, it's static HTML on GitHub Pages so I don't have to run a server for it: https://tools.simonwillison.net/hacker-news-filtered

interesting but i rarely use any HN reader

there's just something about this UI and its consistency

I also don't mind all the AI related news

If anything I just wish they had a mute/block button. its not fun when somebody is stalking you replying to every comment you make.

Re: Hacker News, Sans AI

#38

Earlier quoted context omitted.

I got a partial load and what it looks like it does is just search each submission for a list of key words and discards any that hits, so it would discard this submission.

I would argue that in a small way, this post _is_ about AI, so it wouldn't be a false positive.

I was not suggesting that it would be a false positive, I was suggesting that this will filter out many submissions that would be of interest to those that want less AI on HN. This would flag a blog that has nothing to do with AI if some random person mentioned AI in the comments of that blog post, right?

Re: Hacker News, Sans AI

#40
post #25

The filter doesn't appear to be perfect, one of the top posts right now is "My Agent Skill for Test-Driven Development"

I guess they could run the titles through a (flash) LLM to catch ones were keyword detection won't cut it.

Have to love the irony. :)

Post reply on HN