3 ms·
The launch of ChatGPT polluted the world forever
- chistev 1y ago[flagged]
- Den_VR 1y agoSomeday maybe we’ll have a term similar to “low-background steel” for information and web content.
- etherlord 1y agohttps://blog.jgc.org/2025/06/low-background-steel-content-without-ai.html?m=1 https://blog.jgc.org/2025/06/low-background-steel-content-wi...
- ChrisArchitect 1y agoLarge discussion earlier this week: https://news.ycombinator.com/item?id=44239481 https://news.ycombinator.com/item?id=44239481
- willis936 1y agoThe root of it is deterioration in trust. Even before LLMs hit the scene there was suspicion of narrative manipulation by social media sites. ChatGPT only changed how popular this take is, but not its measure.
- cheschire 1y agoWhy did you paraphrase the article’s subtitle?
- Den_VR 1y agoI’m not sure if admitting I didn’t even open the article helps or harms my case.
- Eddy_Viscosity2 1y agoThis is a great analogy.
- happa 1y agoLLMs don't really need more training data than they already have. They just need to start using it more efficiently.
- _1tem 1y agoExactly. Smart humans work with far less training data and do better.
- ghusto 1y agoThe article keeps making it sound as if it's a problem for humans. e.g.: > Now here the date is more flexible, let's say 2022. But if you're collecting data before 2022 you're fairly confident that it has minimal, if any, contamination from generative AI. Everything before the date is 'safe, fine, clean,' everything after that is 'dirty.'" Though what it seems to actually mean is that it's a problem for (future) generative AI (the "genAI collapse"). To which I say;
- joshstrange 1y agoThis seems like a very badly written article that rambles on in random directions. It proposes incredibly dumb ideas to anyone with half a brain like water marking AI output. The most damning part for me is mentioning the Apple paper and the refute of the Apple paper, to my knowledge that paper had nothing to do with training on generated data. It was talking about reasoning models, but because they use the word “model collapse”, apparently, the author of this article decided to include it in, which just shows how they don’t know what they’re talking about (unless I’m completely misunderstanding the Apple paper).
- m4r1k 1y agoThis! And I’d add, it’s the Register–it has always had a very low bar.
- famahar 1y agolowbackgroundsteel.ai sounds really promising. I don't really care for it as a clean AI training source, but I'm interested in a curated internet where I know it's not diluted with generative content. I'm not sure what that would look like when it comes to social media. This AI era has made me return to reading physical books as a hobby and engaging with offline/non-anonymous online communities more. Confidence in authenticity is one of the most important things for me these days.
- iJohnDoe 1y agoI have no expertise in LLMs. I do think the article poses an interesting question. How do you get the models recent information without ingesting information that has been generated by AI. I’m sure it’s possible, but not without a certain level of uncertainty. Humanity now lives in a world where any text has most likely been influenced by AI, even if it’s by multiple degrees of separation.