Small-but-happy win:
— Sam Altman (@sama) November 14, 2025
If you tell ChatGPT not to use em-dashes in your custom instructions, it finally does what it's supposed to do!
Sam Altman says ChatGPT has finally learned to follow one oddly specific instruction: stop using em dashes. He called it a “small-but-happy win,” claiming that if users disable them in Custom Instructions, the model now listens.
And sure, some people tested it and got dash-free replies. But others jumped into Altman’s mentions with screenshots showing… well, dashes. A reminder that even the smallest AI fixes arrive with asterisks.
Some users also pointed out that ChatGPT was still hallucinating basic facts like naming the wrong US president, suggesting that stylistic tweaks aren’t exactly the biggest problem on the table.
OpenAI also rolled out Group Chats, basically WhatsApp groups with ChatGPT as a built-in mediator or idea-buddy. Fun update, but the em-dash saga shows the gap between AI polish and AI reliability.
