3 ms·
I think it's just an AI-generated simplification, sucks that it made it to the front page. The subject matter is interesting, I would have loved to have read so
by robbie-c 10mo ago
I think it's just an AI-generated simplification, sucks that it made it to the front page. The subject matter is interesting, I would have loved to have read something written by an expert!
- fabian2k 10mo agoI would assume so, but I didn't see any smoking guns in the text itself. But I'm also not familiar with the newest models here and their quirks.
- D-Machine 10mo agoSee my point above (https://news.ycombinator.com/item?id=46271980 https://news.ycombinator.com/item?id=46271980) for smoking guns. There are some pretty basic and grievous factual errors re: GPUs being used when in fact TPUs are used, and completely false claims about physical models not being huge parts of AlphaFold development and even architecture.
- fabian2k 10mo agoThose errors don't seem AI-specific to me, they could easily be made by humans.
- D-Machine 10mo agoTrue, it is the style of the post that reveals obvious overuse of AI. The errors could well be made by a human, especially since a trivial visit to Wikipedia or one of the original papers will show most of what is being said here re: the actual deep models to be wrong. This is more likely the error of a human than an AI. EDIT: Ugh, it is late. I mean, if you used e.g. ChatGPT-5.X with extended thinking and search, it would not make these grievous errors. However, ChatGPT without search and the default style, produces junk basically indistinguishable from this kind of post. So, for me, the smoking gun is that not even the most basic due diligence (reading Wikipedia or looking at the actual papers) has been done, and, given the length and style of the post, this is effectively a smoking gun for (cheap, free-version) AI use. But, more importantly, it is indistinguishable in quality from AI slop, and so garbage regardless.