I can’t speak for cancer specifically but a lot of research out there isn’t in easily parseable formats for people to sort through using traditional methods, and LLMs have been pretty helpful for adding to evidence of specific effects to be able to pare down the list of millions of compounds out there to a number that’s worth testing in the research I’ve been a part of. I’m about as skeptical as anyone can be of the use of LLMs in the medical field and try to avoid using them whenever possible but one of the only things they’re good for is parsing more text than any human is capable of in short amounts of time which is unfortunately useful for first passes in a lot of the work that goes into therapeutics development