A: Since only incremental changes are processed (via streams + merges), you avoid re-processing the whole table on every run. That reduces compute and data transfer. Also, less storage/io overhead for frequent full loads. (Implicitly discussed in the blog through the stream + merge pattern.)
