The community member is asking for methods to speed up the "synthesize" process, noting that the LLM (Large Language Model) calls within "synthesize" are the most time-consuming part. In the comments, another community member suggests two ways to potentially speed it up: getting a faster LLM or decreasing the amount of data being sent to the LLM (by adjusting the chunk size or top-k parameters).
