ChatGPT is not only a success of LLM technology, but also the birth of a new product category—arguably a paradigm shift in product design. It marks the transition from window-based products to freedom-oriented products (DOF, Degrees of Freedom). For decades, products have been designed around the “window” metaphor: from the DOS and Windows era, to Office software, to the age of the internet with web pages and mobile apps. With ChatGPT, and especially with LLM-driven products, a new dimension of freedom emerges—interaction powered by natural language. This gives users greater cognitive and interactive freedom. Yet this is only the beginning. Generative human–AI interaction points toward an evolution: more natural, more open, and defined by higher degrees of freedom. AI products are evolving from window-driven to freedom-driven.
The essence of freedom in human–computer interaction lies in embodied computation. The mouse and touchscreen represent embodied interactions through hand movements. In ChatGPT, the prompt becomes an embodied interaction of natural language—even of thought itself. But this is only the first stage of AI product design. Such embodied interactions will naturally extend to more degrees of freedom (or modalities), lowering users’ cognitive load and aligning human–machine interaction more closely with innate human characteristics. This is the product form that AGI will eventually take.
At the same time, because AI carries a degree of autonomy, it will gradually free users from constant screen dependency. Interaction will decouple into more asynchronous, human-like exchanges. Today, many people spend more than ten hours a day facing windows—screens and pages. In the AI era, the window is disappearing. It no longer serves as the workspace itself, but becomes more of an observation layer. Users are liberated from the alienation of “tools as windows,” while AI acts as their proxy to digital productivity.
Through embodied interaction, users will operate the most powerful AI systems.
This is the direction of our product practice and evolution at MOSS AI Office.