100% wrong my friend. Your biased comment doesn’t take into consideration who’s writing, what I’ve been doing in the past few years, and my writing process. But I appreciate you coming here and give me some audience.
Ok, that was mostly intended as a bit tongue-in-cheek, but it was probably a bit too harsh/rash, I mean there are obviously lots of text styles and/or conditions that will lead to text that sounds like an AI could have written it (e.g. most of LinkedIn - even before ChatGPT existed - sounded like AI). And there are also a million different shades between completely AI generated and completely human written (e.g. "hey here is my rough text, please spellcheck and add some jokes").
Honestly I mostly don't care either way, I prefer to sort by good article vs. bad, not by AI-generated vs human-made (which would be quite hard anyway) I just found it quite funny and ironic in this particular case :P
That being said, this article still **sounds** like AI generated to me. So not saying it is/isn't (also basically impossible to prove or disprove), just saying it sounds like it, as some honest feedback. I think the main point of the article, that for humans just "being human" will become really valuable is true not just for training AI but also for other humans. So "authentic" human-made text will become more valuable, but the flip side of that is imo that anything that sounds AI will be automatically devalued (independent of if it is AI generated or not).
If you haven't yet, give NotebookLM a try, especially the podcast generation, honestly it's my favourite AI tool right now and by far the best at generating interesting text imo, but after listening/reading a few of the podcasts/texts you can kinda hear the style/patterns and they're very close to the style here (and some patterns match up very closely too, some examples from the first few paragraphs: "We're about to dive into the wild world of ...", "Picture this: ...", "a concept so mind-bending, ...", all things that you'll hear quite often with NotebookLM). Like I said, I don't wanna say that you're using that here, but after listening to a few podcasts, maybe you'll see what I mean by it sounds very much like that model.
AlphaGo Zero trained on itself to become the baddest ass Go player on the planet. Reinforcement learning. So Im thinking we can figure this one out?
Super cool word Ouroboros; when I first saw it I thought it was a spoof on some type of bro
An online article talking about model collapse from consuming lots of ai generated content online that sounds obviously ai generated…
I think I could even guess the model (NotebookLM?) :D
100% wrong my friend. Your biased comment doesn’t take into consideration who’s writing, what I’ve been doing in the past few years, and my writing process. But I appreciate you coming here and give me some audience.
Ok, that was mostly intended as a bit tongue-in-cheek, but it was probably a bit too harsh/rash, I mean there are obviously lots of text styles and/or conditions that will lead to text that sounds like an AI could have written it (e.g. most of LinkedIn - even before ChatGPT existed - sounded like AI). And there are also a million different shades between completely AI generated and completely human written (e.g. "hey here is my rough text, please spellcheck and add some jokes").
Honestly I mostly don't care either way, I prefer to sort by good article vs. bad, not by AI-generated vs human-made (which would be quite hard anyway) I just found it quite funny and ironic in this particular case :P
That being said, this article still **sounds** like AI generated to me. So not saying it is/isn't (also basically impossible to prove or disprove), just saying it sounds like it, as some honest feedback. I think the main point of the article, that for humans just "being human" will become really valuable is true not just for training AI but also for other humans. So "authentic" human-made text will become more valuable, but the flip side of that is imo that anything that sounds AI will be automatically devalued (independent of if it is AI generated or not).
If you haven't yet, give NotebookLM a try, especially the podcast generation, honestly it's my favourite AI tool right now and by far the best at generating interesting text imo, but after listening/reading a few of the podcasts/texts you can kinda hear the style/patterns and they're very close to the style here (and some patterns match up very closely too, some examples from the first few paragraphs: "We're about to dive into the wild world of ...", "Picture this: ...", "a concept so mind-bending, ...", all things that you'll hear quite often with NotebookLM). Like I said, I don't wanna say that you're using that here, but after listening to a few podcasts, maybe you'll see what I mean by it sounds very much like that model.