From my experience teaching research for the past 9 years, being able to verbalise your thoughts and write precisely and succinctly are things you have to learn through practice. Using LLMs to shortcut these before you learn how to do them proper you'd just never get these skills
More specifically to that discussion: The ability to "tame" fleeting thoughts and ideas in a format that won't let you "lie" with internal inconsistencies, or with half-formed ideas is itself I think at least one of the major goals of the scientific endeavour (and is therefore a cornerstone in scientific education).View quoted post on X
About the use of LLMs there - I'm not sure we can afford not to use those to aid in these processes, but it's very clear where the scientific thinking of the author wasn't formed enough.
In those cases I think, our main goal is to teach (and learn together) younger researchers to identify themselves when ideas are half formed, when understanding isn't complete etc.
In many ways writing is one of the highest forms of thinking. And if as DFW wrote the goal of higher education is to teach you how to think - then we must let students write more.
