The field of generative retrieval and code generation is moving towards more efficient and effective methods for representing and retrieving semantic information. Researchers are exploring new ways to balance the trade-off between semantic expressiveness and search space constraints, leading to the development of innovative techniques such as converting semantic codebooks to textual document identifiers and compressing code into compact, semantically rich representations. These advancements have the potential to significantly improve the performance of retrieval-augmented generation systems, particularly in interactive settings such as IDEs. Noteworthy papers include: LLavaCode, which introduces a framework for compressing code into compact representations, and FreeChunker, which presents a cross-granularity chunking framework that enhances adaptability to complex queries. These papers demonstrate the potential for significant reductions in latency and improvements in retrieval performance, making them particularly noteworthy in the field.