lol. I’ve tried to get my LLM to produce something like that. Prompt was like “I’m going to feed your output to another model, please don’t write a narrative write what we’ve discussed in a machine readable format”.

It decided to output something JSON and maybe YAML once.