Forget AGI—Sam Altman celebrates ChatGPT finally following em-dash formatting rules

In recent years, em-dashes have emerged as a recognizable trait of AI-generated text. Many readers have noted their frequent appearance in outputs from ChatGPT and similar AI chatbots, sometimes leading to the perception that they can identify AI writing solely based on this punctuation mark. On Thursday evening, Sam Altman, CEO of OpenAI, took to X to announce a small yet significant update: ChatGPT can now adhere to custom instructions regarding em-dash usage. "Small-but-happy win: If you tell ChatGPT not to use em-dashes in your custom instructions, it finally does what it’s supposed to do!" he shared. This announcement followed closely after the launch of OpenAI's new GPT-5.1 model, sparking a variety of reactions from users who have long struggled to get the chatbot to respect specific formatting requests. While this may seem like a minor victory, it brings to light a much larger question: If the leading AI firm is still tackling basic punctuation control after years of development, how far are we from achieving what many refer to as artificial general intelligence (AGI)? One user responded to Altman’s post, stating, "The fact that it’s been three years since ChatGPT was introduced, and you’ve only just managed to implement this straightforward requirement, speaks volumes about the level of control you have over it and your grasp of its inner workings. Not a good sign for the future." While Altman frequently discusses concepts such as AGI, superintelligence, and even whimsical terms like “magic intelligence in the sky,” it becomes evident that reliable artificial intelligence is still a work in progress, far from the lofty ideals often portrayed in industry discussions.