This will be nice when LLMs will have ubiquitous network access. And you can prompt them in the paper to push the prompt and other details to a specific endpoint :D.