[oasisbob]

There needs to be a meter for the amount of AI writing in blogposts. Same physics, same climb, same afternoon fog.

[salahadawi]

I've actually been tracking the front page of HN for AI posts for a while now: https://www.salahadawi.com/hacker-news-ai-detector

This post evaluates to 99% AI generated.

[jsnell]

Nice! Have you considered doing a Show HN for that?

That's valuable in at least three different ways: public education, showing that most of the articles are still human-written which can be easy to forget about sometimes, and as an easy way to cross-validate my intuition when flagging something as AI-generated without having to manually run Pangram.

I despair a little bit about how many HN voters either seem to want to read slop or don't understand when they're reading it. This post is obviously AI generated from the first paragraph on, and still has 480 votes.

[salahadawi]

I've thought about it, will probably fix a few things and then do it

[salahadawi]

Seems someone at HN didn't like my post, tried posting in Show HN but was immediately shadowbanned

[JonathanRaines]

Pangram claims to be super accurate. Do you know how? I understood detection is a hard problem.

[salahadawi]

Their technical report (arxiv.org/abs/2402.14873) goes into it.

They train a classifier on human text paired with AI-generated versions of the same documents. Dataset scale is comparable to LLM pretraining sets.

[naet]

This reads to me as AI generated. Apparently it's still good enough to the general audience to be the #1 post on HN right now though. Which is honestly a troubling signal for the state of the world...

[LordDragonfang]

Yeah, I clocked it as AI the moment I read "Here is the uncomfortable part". FWIW, Pangram detects it as 100% AI-generated.

https://www.pangram.com/history/c410d4b4-abfd-4ca0-b52d-db0d...

[stavros]

Yeah, it's really tiring reading Claude's output all day, every day. Nowadays I yearn for a different style.