> Historically, IDEs optimized for a tight inner loop: open files → edit → build → debug → repeat. The “death” argument is that this loop is no longer the domin

> Historically, IDEs optimized for a tight inner loop: open files → edit → build → debug → repeat. The “death” argument is that this loop is no longer the dominant unit of productivity once agents can execute most of it autonomously. And the argument is rendered moot because this loop was never considered the unit of productivity. There was always a whole planning and design phase that was conducted before this one. And for some tasks, it was not a loop, just the sequence edit → build sequence. The article seems like written by someone who has never used an IDE professionally, just watch someone using it for 20 seconds through a window.