It depends entirely on your audience.
If you're dealing with a hundred people in an office, then the vast majority can probably deal with a high-density UI. Because they are there for work and the UI - and more importantly the concepts underneath that UI and the mental model it requires - will be related to the stuff they already know.
But if you're dealing with an audience of millions who are trying to pass the time, or using apps or parts of the system that they rarely interact with, or have no real interest in the device, then a high-density UI will be extremely intimidating. They have no mental model of what is being represented on-screen - and if it's casual usage, they have no desire to spend the energy to learn that mental model. What they will want is the device to have two big buttons saying "do this" or "do that". Which is why the iPhone's grid of icons endures - "I want a bus ticket", "I want the camera", "I want to watch a video".
Personally I think the hardest part of UX design is ensuring the model presented on-screen matches the model that the user is expecting.
Which has two implications - firstly, it's entirely audience dependent and secondly, the further the UX model and the actual underlying model diverge, the more cracks will appear while people are using it - so you have to design from the bottom up for that given audience.