I think the question would be how specific the prompt would be in that scenario. Is the prompt having to effectively describe the new idea, or is the prompt describing a general problem space and goal while letting the LLM do all the heavy lifting of innovation?

Either way, my issue there is actually with the fact that the LLM can only predict what a human might say in response to the prompt andz as far as we know, don't have any way of logically thinking through the new solution or modelling reality to understand if the idea might actually work.

It's not a requirement to be useful tools, its a requirement to be an "analog for the language center of a brain". I personally use them as boilerplate-generator and as rubber ducks, but a neurotypical brain will create neologisms quite often (accross a life, not everyday ofc) when talking to people.