← sources

Future of NLP for Computational Social Science

source: reddit (reddit:LanguageTechnology)published 2026-11-03 08:43ingested 2026-11-04 02:40

retained text

hi I’m a social science researcher interested in NLP. for the past few years I thought this was the state of the art method for social science researchers but now with general LLMs being amazingly good at annotations, is there any need in using more traditional methods? or are they still relevant due to cost, accuracy or reproducibility? If we have all the money in the world to use general llms, what Is the reason to use and learn other methods for information retrieval, text analysis, annotation, sentiment analysis etc? very confused and lost! 

document fetched 2026-11-04 00:38 · canonical https://reddit.com/r/LanguageTechnology/comments/1xmnscq/future_of_nlp_for_computational_social_science