The field of natural language processing is witnessing significant developments in text embedding and preprocessing. Researchers are exploring new methods to evaluate and improve the performance of text embedding models, including comparing human and model performance to identify areas where models succeed and fail. Large language models are being leveraged to perform various preprocessing tasks, such as stopword removal, lemmatization, and stemming, with promising results. Additionally, unsupervised pipelines using large language models are being developed to automate corpus annotation, enabling faster and more accurate analysis of large datasets. These advancements have the potential to improve the efficiency and accuracy of natural language processing tasks, including clinical text eligibility classification and summarization. Noteworthy papers include: HUME, which introduces a framework for measuring human performance on text embedding tasks, and Investigating Large Language Models' Linguistic Abilities for Text Preprocessing, which demonstrates the effectiveness of large language models in replicating traditional preprocessing methods. A large-scale, unsupervised pipeline for automatic corpus annotation using LLMs also shows great potential for automating data preparation tasks at scale.