By skyforbes I've been messing around with image generation, I'm still really new, but I just found out about sending the prompt as a JSON code rather than plain text. It does seem to be working better for what little I've tried out so far (in particular it seems like it's much better at capturing things like facial features, for example), but I actually asked ChatGPT if there's any benefit to sending my prompt as a JSON and it said there isn't. Is that incorrect?