las datasets get-documents <datasetId> <dst> should start downloading documents immediately instead of waiting for the entire dataset to be listed.