🧵 Untitled Thread
Anonymous at Fri, 15 Mar 2024 10:50:07 UTC No. 16078914
I would like to thank Zhang, Wu, Yang, Zhu and Liu for the laugh.
Special acknowledgement to the peer reviewers, The Journal of Surfaces and Interfaces, Elsevier and the China University of Geosciences.
Anonymous at Fri, 15 Mar 2024 11:22:04 UTC No. 16078940
6.2 Impact Factor.
Academia is dead.
Anonymous at Fri, 15 Mar 2024 11:38:32 UTC No. 16078957
>>16078914
Bro, what a disaster. I guess I understand trying to have an LLM write the paper if their English is terrible, but how was this not caught in review?
Anonymous at Fri, 15 Mar 2024 12:05:07 UTC No. 16078985
>>16078914
The link for those in need:
>https://www.sciencedirect.com/scie
Anonymous at Fri, 15 Mar 2024 12:11:48 UTC No. 16078998
>>16078957
LLMs won't help much becuase you still need people with contextual understanding (ie. possess excellent written english skills + adequate scientific knowledge at a high level).
I was reading a paper written by a Soviet mathematician and while his english is much better than picrel, it's obvious he's translating this thoughts word for word from Russian to English, so it's very weird to read.
Anonymous at Fri, 15 Mar 2024 12:57:52 UTC No. 16079051
>>16078914
>Elsevier
Elsevier a shit. Anyone else remember the Chaos, Solitons, and Fractals incident from a few years back? Always made me wonder just how many journals exist as just a little fiefdom for the editor to push their own papers.
https://www.nature.com/articles/456
Anonymous at Fri, 15 Mar 2024 13:37:08 UTC No. 16079089
>>16078998
LLMs won't help much if your goal is to actually write good and original papers. That's not what they were designed to do, and extrapolation is really not the strong suit of these ML methods (which are fundamentally regression/interpolation strategies but applied to generative problems).
For people who are just trying to get raw publication numbers and don't care about whether their work is any good, LLMs will do.