The OP suggested a dynamic positioning of the minibuffer. That would be a significant change in the UI and as Eli points out would present many aspects that would need to be worked out.
A little over a year ago I broached a more modest change in this thread:
There I raised the notion of (optionally) moving the modeline to the top of each window and positioning the minibuffer to the top of the frame.
Screen technology keeps, evolving enabling ever larger screens with ever more pixels. At the time that I initiated that thread I owned a 30" (diagonal) 2560x1600 monitor. Since that time my employer has provided me a 43" 3840x2160 monster (HxW = 26"x38"). This is a thing of beauty but sadly a miserable experience when emacs tries to manage the whole screen as a single frame. Yes, I can have many full height windows side by side showing me an unprecedented amount from each buffer. And I can cope with the 38" screen width by collecting the two or three windows of most current interest side by side at the center of the frame. But _very_ often buffer content fails to fill its window. This means that I cannot scroll that content to the middle of the screen. This leaves buffer content often displayed nearly 2 feet removed from its mode line and the minibuffer.
Trying to work in this configuration is essentially impossible. My current compromise is to waste 2/3 of the screen's pixels, creating a 1/3 height, full width frame.
Interestingly, I feel much less disoriented using the full screen height with a tiling window manager (awesome) displaying full height web pages and GUI desktop productivity apps. My conclusion is that some aspects of the emacs UI (rooted at least partially in considerations of repainting early glass TTYs) could stand to be reassessed. For instance, positioning the modeline at the top of our windows would align emacs with the now nearly universal UI idiom that positions a titlebar at the top of a window.
/john