This is an important point:

> For example, LLMs can aid in improving readability and grammar, which might be particularly useful to those for which English is not their first language.

I don't know whether this has been empirically confirmed, but I have the strong belief that a manuscript with poor grammar, by a non-native English speaker, has a much higher probability of being rejected than the same manuscript but copyedited by something like Grammerly or a SOTA LLM.

Ideally writing style should matter much less than the quality of the research, but reviewers are not just influenced by objective criteria but, unconsciously, also by vibes, which includes things like writing style and even formatting.

▲

zahlman 2 days ago | parent [-]

Meanwhile, more and more posts on discussion forums etc. is clearly "copyedited" by these tools and the result is quite grating for the regulars.

▲

cubefox 2 days ago | parent [-]

Probably less grating though than broken English. (Copyediting is also different from pure LLM replies which don't involve editing anything.)

▲

precompute 2 days ago | parent [-]

Broken English still has its charm and brings the structure of the writer's native language to the fore, which makes it relatively easier to parse and glean intentions from than polished LLM-speak.

	▲	cubefox 2 days ago \| parent [-]
		That might be true, but I think it's false. Or more precisely, I think manuscripts with broken English have statistically a higher probability of being rejected than ones that are copyedited with AI.