OpenAI has mounted ChatGPT’s notorious ’em sprint’ obsession (considerably)

OpenAI CEO Sam Altman shared in a social media publish that the corporate has now mounted ChatGPT’s overuse of the “em sprint,” which is the extra-long hyphen that’s generally seen in AI-generated textual content.

Previously, ChatGPT was overzealous in its use of the em sprint, to the purpose the place it’d proceed to incorporate them even when customers requested it to not. Now, with the repair, a person can instruct ChatGPT to not use em dashes and it’ll respect the instruction.

“Small-but-happy win,” Altman writes in his publish. “Should you inform ChatGPT to not use em-dashes in your customized directions, it lastly does what it’s alleged to do!” ChatGPT will proceed to make use of em dashes per traditional if no instruction is given. That half hasn’t modified.

The AI chatbot’s fondness for the em sprint has been so prevalent in its responses that many have (misguidedly) begun to see it as a surefire signal {that a} textual content was generated utilizing an AI software.