For our purposes, anyone building on LLMs and tinkering with the prompts to use.