I have a pet theory that LLMs being confused about what they just wrote is why they use so many em dashes. It's a good way to conceptually pivot at any point -- or not.
I disagree. I don't think LLMs use em-dashes any more than typical professional writing does. It only seems a lot to someone who only reads online chatrooms and social media.
I think the em-dashes are OpenAI's way of watermarking their outputs. Makes it easy to spot spambots on Twitter. Well, that and they all have noticeable favorite words.
