I see tablets just like another "flavour" of a computing device.
First there were mainframes (e.g IBM-360). Then minicomputers (e.g. PDP-11).
Middle of the 70's saw two new classes of computers: personal (Apple) and super (Cray-1 and CDC).
Ever since, mainframes are few years from being dead. But it never happened and it looks like the rumors about mainframe's death will die first.
The 80's brought workstations (Apollo, Sun Micro) and the other flavour of personal computer - IBM PC.
Now, 40 years later, the minicomputer and workstation markets are taken over by PC.
I believe this happened because of Intel (variety and low cost of CPUs) and Microsoft (Windows).
Every single attempt to beat Intel and prevent them from taking over neighbor markets failed.
Also, every attempt to create a new CPU niche with sustainable demand (VIA, Transmeta) failed.
The 90's brought Linux.
It has beaten most of the OSs (Windows included) in everything from embedded to supercomputers.
With one exception: desktop.
It looks like the only way to beat Microsoft on the desktop is to kill desktop as a class.
At least this is what Apple decided to try, I believe.
There are few factors that make this idea not as ridiculous as it would sound even 5 years ago.
More's law. Nobody needs the latest and greatest. Half the way to the top does everything just fine.
Better Windows
)). Outsourcing (a PhD in India costs less than a high school dropout in the US).
And just like the share of farmers went from 50% some 100 years ago to 2% today, creative jobs are leaving NA.
This is where the tablets come in.
Moore's law made ARM competitive for whatever tasks are remaining of interest to NA populace (consumption, mostly).
Unless this planet starts outsourcing to Mars or something, I think somebody would have to do something outside of "consumption".
That task you can't accomplish without a keyboard, i.e. computer. Hence, the later won't die any time soon.
It might happen in the US. But not in the world.
Diogen.