This Hacker News thread discusses the use of LLMs in biomedical publications, sparked by an article on the increasing prevalence of "excess vocabulary" in scientific writing. A key point is that many scientific papers are written by non-native English speakers who may be using LLMs to improve grammar and clarity. While this can lead to tidier text, concerns arise regarding correctness, nuance, and responsibility for errors. Some argue that LLMs, like "Author Services," may introduce inaccuracies that authors, especially those with limited English proficiency, might miss. The conversation touches on the impact of LLMs on writing style, noting a tendency towards business-like or generic prose. There's debate on whether LLMs improve or degrade the quality of scientific communication, with some fearing a "dumbing down" effect and a loss of individual voice. Others are more optimistic about LLMs as tools for overcoming language barriers, but emphasize the need for human oversight to ensure accuracy and avoid the proliferation of "weasel words" and stylistic tics.