I sometimes talk to ChatGPT like it’s a child learning the rules, because sometimes it really feels like that. I’ve told it repeatedly not to use long dashes. And yet, every so often, they show up on my screen like we never had the conversation. I’ll never understand why it insists on using them, especially when they’re banned in my custom instructions. Glitch? Rebellion? Who knows. But this is exactly why human edits matter.
AI can speed things up, spark ideas, and even throw in a little personality, but it can’t replace good judgment, consistency, or a well-timed edit. It’s incredibly helpful, but it still needs supervision… and a little tough love.
The sarcasm might be a learned behavior, though.