The PC era is here to stay as long as the Internet lives. That is not to say however that it won't go through some physical changes along the way, in much the same way the QWERTY keyboard has been made more ergonomic, virtualized on touchscreens, and tucked away in tiny cellphone trays, but remains essentially the same.
Certainly other manifestations of the desktop's evolution, such as in this interactive exhibit at the DDR museum in Berlin will appear in mass production eventually.
Projected screens and keyboards that can be used on walls and other surfaces are also a given. As are hologram projection screens and keyboards, both 3D and flat.
But smartphones and other mobile devices will also have the features below. What makes a desktop remain a desktop is not these physical changes to it and its peripherals, but the fact that some data will remain local and therefore some processing and storage capability must also remain local. The computing process has remained the same for many years, and the desktop PC is its best incarnation.
Yes, I know that smartphones and laptops and other devices of today are tremendously more powerful than desktop computers of just a few short years ago. Even so, the desktop has continued to provide more processing and storage muscle than any of its mobile brethren.
Eventually the silicon revolution will end and Moore's law will collapse. It is at that point that we will entirely change how we compute. Some people predict we will change over to quantum computing, change from bits to qbits (or qubits), and do our processing on electrons rather than on chips. American theoretical physicist Michio Kaku predicts that molecular computing will see widespread adoption before quantum computing does.
No matter if we end up with molecular computing or quantum computing, we will radically disrupt the computing process -- processors, programming, storage and, yes devices -- all of what we use today will no longer fit into that new computing process and thus will all become obsolete. And then, the desktop will finally die.
And that, says Kaku, is as important to the world economy as it is to humanity's ability to compute. When we drive headlong into the silicon chip's dead end, devices will no longer sell at a fast pace simply because there will be no improvement in new models over old models. However, with the birth of a new computing process and all the new devices that it will spawn, the world economy will thrive.
And therein lies the answer to why hardware manufacturers are so eager to prematurely declare the PC dead: they need to sell more devices quickly before the silicon chip maxes out. And, they can already see that Moore's law will collapse entirely within the next 10 or so years. In other words, their desperation is showing.