The traditional operating system (OS) was designed to manage hardware resources and isolate applications from one another, a necessity in an era where software was created by expert strangers and distributed as fixed-function packages. However, the rapid advancement of AI is fundamentally altering this model. By enabling users to generate their own applications using natural language, AI is blurring the lines between frontend and backend programming, as well as between developers and end-users. In this emerging paradigm, the rigid partitioning of applications becomes less relevant when most software is created by the user or their immediate circle to solve idiosyncratic, real-world problems. This shift suggests that future computing platforms may move away from the 1970s-era design of running deterministic, pre-packaged applications. Instead, the focus is shifting toward devices that facilitate the creation of malleable, task-specific software on demand, potentially rendering the current concept of the OS as a mere gatekeeper of fixed applications obsolete.