Human creators do a thing not unlike LLM synthesis in creative tasks, but with the addition of taste.

Good designers absorb lots of good examples from the web of examples they're exposed to, selecting the subset they like (applying their taste) and then synthesizing that intuition into an answer to a given problem in front of them.

Not unlike what LLMs are doing.

Though LLMs don't have a taste criteria for what to absorb from.

Their sampling criterion is "things that humans decided to reproduce", which is aligned with a kind of wisdom of the crowd's generic taste for usefulness.

Things that are thought by other humans to be useful will tend to be kept around and that's what the LLMs sample from.