What is the point of these LaTeX editors with a chatbot built in.
-
What is the point of these LaTeX editors with a chatbot built in. To help search for references? If you aren't familiar with a reference, you shouldn't be citing it (if it even exists). And if you haven't thought carefully about how it relates to what you're doing, IT IS NOT REALLY A CITATION.
-
What is the point of these LaTeX editors with a chatbot built in. To help search for references? If you aren't familiar with a reference, you shouldn't be citing it (if it even exists). And if you haven't thought carefully about how it relates to what you're doing, IT IS NOT REALLY A CITATION.
Checking grammar? Sure, okay, everyone can decide if that's a legitimate use case for them. Expressing something complex in LaTeX? Again, perfectly fine occasional use case as you familiarize yourself with LaTeX. Generating citations or descriptive text? NO, THAT IS THE PART A PERSON HAS TO DO.
-
Checking grammar? Sure, okay, everyone can decide if that's a legitimate use case for them. Expressing something complex in LaTeX? Again, perfectly fine occasional use case as you familiarize yourself with LaTeX. Generating citations or descriptive text? NO, THAT IS THE PART A PERSON HAS TO DO.
Look, I get being excited by the tech, I really do. But we know enough about its shortcomings that everyone should be appropriately skeptical about letting it intrude on the parts of the work that have to be done by humans to have any meaning.
-
Look, I get being excited by the tech, I really do. But we know enough about its shortcomings that everyone should be appropriately skeptical about letting it intrude on the parts of the work that have to be done by humans to have any meaning.
A citation isn't a pointer to another instance of a topic being mentioned. A citation is an intentional acknowledgement of existing work because you are building on it, or it provides context for what you are trying to do. That is something the author has to appreciate and then communicate.
-
A citation isn't a pointer to another instance of a topic being mentioned. A citation is an intentional acknowledgement of existing work because you are building on it, or it provides context for what you are trying to do. That is something the author has to appreciate and then communicate.
Imo the arXiv should impose an automatic ban the first time someone submits a preprint with a hallucinated reference in it. Zero tolerance policy. Same goes for peer-reviewed journals.
As soon as you identity a hallucinated reference in a manuscript, that means nothing the person does can be trusted.
-
A citation isn't a pointer to another instance of a topic being mentioned. A citation is an intentional acknowledgement of existing work because you are building on it, or it provides context for what you are trying to do. That is something the author has to appreciate and then communicate.
Thanks for saying this. I mean you loose so much context and history without it.
-
Imo the arXiv should impose an automatic ban the first time someone submits a preprint with a hallucinated reference in it. Zero tolerance policy. Same goes for peer-reviewed journals.
As soon as you identity a hallucinated reference in a manuscript, that means nothing the person does can be trusted.
@mcnees the pernicious thing is that (as per this WikiEdu post https://wikiedu.org/blog/2026/01/29/generative-ai-and-wikipedia-editing-what-we-learned-in-2025/) LLMs can often generate correct references to actual works…that don't say what they're cited as saying. That is an awful lot harder to catch.
Another thing I've seen in LLM-generated material is references that are very subtly wrong. Like the authors, journal, issue, year, etc., are correct, but the DOI URL goes to a different paper.
-
Imo the arXiv should impose an automatic ban the first time someone submits a preprint with a hallucinated reference in it. Zero tolerance policy. Same goes for peer-reviewed journals.
As soon as you identity a hallucinated reference in a manuscript, that means nothing the person does can be trusted.
@mcnees grammar checkers existed long before LLMs became a thing, same for syntax highlighting, so not even on those cases, I think
-
Imo the arXiv should impose an automatic ban the first time someone submits a preprint with a hallucinated reference in it. Zero tolerance policy. Same goes for peer-reviewed journals.
As soon as you identity a hallucinated reference in a manuscript, that means nothing the person does can be trusted.
@mcnees Even *using* the plagiarism machine should be grounds for academic misconduct.
-
What is the point of these LaTeX editors with a chatbot built in. To help search for references? If you aren't familiar with a reference, you shouldn't be citing it (if it even exists). And if you haven't thought carefully about how it relates to what you're doing, IT IS NOT REALLY A CITATION.
@mcnees a friend of mine found this use which I thought was pretty clever: he asked the chatbot to give him a diagram of dependencies for the Theorems in his paper, and the diagram was correct. The diagram was kind of messy, parts of it overlapped each other, but he could tweak that by hand and get it right. He says it probably saved him between half an hour and an hour.
-
undefined oblomov@sociale.network shared this topic