OpenAI CEO Sam Altman on Friday said ChatGPT will "finally" stop using em dashes when instructed, calling it a "small-but-happy-win". Several users reacted to Altman's update, with some claiming their instructions were ignored. "Told it to stop...it replied with an em dash. Peak betrayal," one user said, while another said, "Finally— we can escape — the sloppiest AI tell yet."