The picture produced by gpt only needs one sentence to get the author's meaning.

For the map produced by sd, sd needs to adjust many parameters and add reference.