4 ms·

Wait, really? I read this entire article, quite enjoyed it(in fact I sent it to 3 of my friends already, as it's a topic that's very dear to me) and I hasn't ev

Wait, really? I read this entire article, quite enjoyed it(in fact I sent it to 3 of my friends already, as it's a topic that's very dear to me) and I hasn't even crossed my mind at all that it might be AI generated. Are you sure it's AI generated? If yes, how?

throwaway290 11mo ago
I don't think I'm as hyper sensitive to llm writing as some others here but this triggered me very quickly. I started scanning and selectively reading after the intro which was just completely soulless. After I reached the end of section 1 I closed tab. It's cumbersome. Things like "This essay examines two questions." Bold is used wrong. Uniform style wall of text all the way, with identical size sections and everything. Section conclusion not reflecting its title. Looking closely I noticed a couple of other tells but I won't share them in case they are reading.
barrkel 11mo ago
It's got that AI smell. The word choice. Antithesis ("not this, but that"). Persuasion words. There's meat in it, it's not pure slop, but it was definitely fed through the slop factory. And as some have mentioned, the facts are a bit dubious.
observationist 11mo ago
100% sure. It's low information repackaged bland pap with nothing to add with regard to perspective or novel information, meticulously referenced like a research paper, with the formatting and grammatical quirks of AI, the simultaneously broad, yet weirdly limited vocabulary, and so on. I'm sure if I spent more time I'd be able to articulate exactly which patterns I'm recognizing, but it's very recognizable. There are also similar patterns in previous content from that substack. Some effort has been put in to it, there aren't any emdashes, but it's slop. It's like a "ChatGPT, what's an interesting topic to write about that's relevant to region XYZ" "ok, thanks, write a detailed, interesting essay, but edit it to avoid all the AI tell-tales like em dashes." and so on until you get this article. If it gets on the HN front page, it gets propagated and an insane level of visibility all over the internet - say, 100 million people see it over the course of 24 hours. If they charge $5 for an annual subscription and 1% (or .1% or even .01%) are so impressed with the content that they sign up, that's a lot of money. I can't find any other possible reason for this showing up, it's nothing to do with tech or AI or silicon valley or the usual weird eclectic stuff that gets discussed, lol. Kudos to the author for knowing enough about how things work to make such an elegantly targeted campaign, I guess.

throwaway290 11mo ago

I don't think I'm as hyper sensitive to llm writing as some others here but this triggered me very quickly. I started scanning and selectively reading after the intro which was just completely soulless. After I reached the end of section 1 I closed tab. It's cumbersome. Things like "This essay examines two questions." Bold is used wrong. Uniform style wall of text all the way, with identical size sections and everything. Section conclusion not reflecting its title. Looking closely I noticed a couple of other tells but I won't share them in case they are reading.

barrkel 11mo ago

It's got that AI smell. The word choice. Antithesis ("not this, but that"). Persuasion words. There's meat in it, it's not pure slop, but it was definitely fed through the slop factory. And as some have mentioned, the facts are a bit dubious.

observationist 11mo ago

100% sure. It's low information repackaged bland pap with nothing to add with regard to perspective or novel information, meticulously referenced like a research paper, with the formatting and grammatical quirks of AI, the simultaneously broad, yet weirdly limited vocabulary, and so on. I'm sure if I spent more time I'd be able to articulate exactly which patterns I'm recognizing, but it's very recognizable. There are also similar patterns in previous content from that substack. Some effort has been put in to it, there aren't any emdashes, but it's slop. It's like a "ChatGPT, what's an interesting topic to write about that's relevant to region XYZ" "ok, thanks, write a detailed, interesting essay, but edit it to avoid all the AI tell-tales like em dashes." and so on until you get this article. If it gets on the HN front page, it gets propagated and an insane level of visibility all over the internet - say, 100 million people see it over the course of 24 hours. If they charge $5 for an annual subscription and 1% (or .1% or even .01%) are so impressed with the content that they sign up, that's a lot of money. I can't find any other possible reason for this showing up, it's nothing to do with tech or AI or silicon valley or the usual weird eclectic stuff that gets discussed, lol. Kudos to the author for knowing enough about how things work to make such an elegantly targeted campaign, I guess.