GPT-4o and the End of the Required Screen

Originally published in Chinese on May 14, 2024. OpenAI's latest launch event introduced GPT-4o, a model with voice interaction that is free to all users. Here are the elements I think matter. A warm-up for something bigger I think this was only a warm-up release, a product tied to model technology. A bigger technology event should follow soon. What GPT-4o is worth GPT-4o is impressive. Beyond the technical system and the stability of its performance, the more important point is that OpenAI provides this feature free to all users as a basic function. That suggests a larger ambition: upgrading the interaction between people and machines. The previous core value was text interaction between people and machines. Text is a form of interaction that people have long known and accepted, and before graphical interfaces appeared it was the stable, accepted state. Now text interaction has been lifted into voice, which is an upgrade to a basic capability in how people and machines interact. Its biggest meaning is for future ha