Em dashes have become a hallmark of AI-generated text, often leading readers to identify AI writing by its frequent use of this punctuation. Sam Altman, CEO of OpenAI, recently announced on X that ChatGPT can now follow custom instructions to avoid using em dashes. He described this as a "small-but-happy win" on the platform, highlighting a new capability just two days after the release of the GPT-5.1 model. User reactions to Altman's announcement were mixed. Many have struggled for years with getting ChatGPT to adhere to specific formatting requests, and this development raises questions about OpenAI's control over its AI's behavior. One user pointed out that it took three years for ChatGPT to manage a straightforward task like punctuation, suggesting a lack of understanding of the AI's workings and control over it. This leads to skepticism about the timeline for achieving artificial general intelligence (AGI), a concept Altman frequently discusses while raising funds for OpenAI. While Altman often speaks about AGI and superintelligence, the ongoing challenges with basic formatting suggest that practical, reliable AI is still a work in progress. The disparity between ambitious claims and current capabilities highlights the complexity of developing AI that meets user expectations consistently.