The death of the 'PC' has been hyped quite a bit, but it seems to usually involve small parts of the computing experience changing for sections of the population. Here is a quick runthrough of the state of the 'PC' as I see it.
The definition that I have always associated with 'PC' is its distinction from centralized computing. How much of the user's experience is being computed in a 'private' processor and how much is being foisted onto a server somewhere to be processed. While talk of a thin client revolution has been around for a couple of decades now, it has failed to materialize. The web, however, has done quite a lot to slowly and steadily steal computation for some tasks, particularly for retail consumers. The amount that you can do with just a web browser is staggering, growing, and becoming steadily more popular. The other front where the PC is beginning to lose ground to servers is in high performance computing. There has always been a server component in this sector, but as distributed technologies improve, the balance between PC and servers is shifting back toward servers, although it will probably be a while before the right disruptive technology comes about to replace it.
The biggest obstacle to centralized computing over personal computing is network connections. Fast enough, reliable connections are simply too expensive (if available at all) for moderately intensive computing at this time, but as networks continue to improve, this is a limitation which is going to continue to fall away over time.
Another definitional aspect for 'PC' seems to be terminal interfaces with a keyboard, mouse, and monitor, which I will address in order. Keyboards, by leaps and bounds, are the fastest and least error prone way of inputting language into a computer. Voice input is still clunky at best and thumb keyboards are slow and often rely on error-ridden auto-completion. There may be a few exceptional people who aren't frustrated trying to write long-form using alternative input, but those are definitely the exception. The disruption for this technology doesn't appear to be anywhere even close to the horizon.
Mice, similarly, have proven time and time again to be the best device for interacting with a two-dimensional field. Here we have gaming to thank for case studies of attempts to use other devices, only to come back to a good mouse for the best interaction. The continuing dominance of mice, however, isn't as great as keyboards, as touch-screen technology could provide better interactivity for some very common usage scenarios. The problems that touch-screens have, are still numerous, however, including slow response times, imprecision, and difficulty in indicating different types of interactions (e.g. right-clicking). The biggest obstacle to touch-screens is that they are ergonomic nightmares when used with a keyboard, with users having to have two completely different orientations for keyboard and pointing work. Touchpads frequently encroach on traditional mice in this space to allow for compact laptops without having to manage accessories and it is from laptops that mice have had their greatest competition.
Monitors provide large screens, not uncommonly in multiples. Interacting with large amounts of data or viewing entertainment usually means a big monitor. Setting aside the ubiquity of monitors in living-rooms where we call them televisions, spreadsheets, one of the most ubiquitous--if perhaps controversial--applications used in offices just don't work well on small screens. Monitors can also ride the coattails of full-size keyboards.
Probably most telling is that even in the I, Crigley article, he says that the I/O problem that mobile devices have is likely to be at least partially addressed by docking stations, which are the traditional terminal.
The 'PC' could also be partially defined as the traditional desktop operating system. In this respect, MacOS, Gnome, KDE, Windows, and many others are all on the same boat. While 'death' is probably a vast overstatement, this is probably the most threatened of the 'PC' components and it has less to do with the merits of the platform and more to do with the business of computing and your average consumer. Apple has stolen a lot of mindshare with its iOS products, at least partially because they provide a managed computing experience. Derisively called a 'walled garden', Apple controls distribution of all applications to your device and promises that your device won't get foweled up by malicious software. It also provides software developers an entire distribution infrastructure in which to sell small inexpensive programs. The success of this model is influencing other OS developers and has drawn a huge portion of the newest generation of developers toward thinking in this model. If you are a young developer these days, chances are you are either developing for web or mobile and ignoring desktop software. The consequences of this particular turn of events are difficult to predict and may end up blowing over, but they also might change the desktop OS for a long time to come.
Finally, there is the distinction between stationary and mobile devices, with 'PC's being seen as stationary. Here, the stationary PC has become more specialized and niche oriented. People who require as much computing power they can get or fast network connections will still have a demand for stationary computers. Devices connected to multimedia systems (e.g. PVRs and gaming consoles) have little need to be mobile and can be made less expensively for that purpose. Businesses which require employees to be chained to a desk also will frequently not be able to justify the costs and compromises of mobility. Those types of niches aside, the stationary PC is dead. Retail consumers buy mobile phone devices and laptops for these purposes and as such the retail stationary PC has long had its heart cut out and served up to the competition. This ultimately means that if you want to buy a stationary PC, you are in a marketplace which is losing investment, meaning higher prices for products which meet needs less specifically.
In conclusion, PCs still have their place and for many of those things it is irreplaceable for the forseeable future. The big change is that there are people with very few computing needs for whom mobility is the primary deciding factor. When your communications are