Another point which I rarely hear about in discussions of the iPhone is how it is fundamentally marketed differently. We have had an iPhone for many years now. It hasn't been the exact same hardware this whole time, but it has been an "iPhone" the whole time. Now think about other mobile phones. Ask someone what phone they have and it is either "LG", "Motorola", etc. Mobile phone models change so often than no one can remember what model they have currently, because there is a good chance that you can't even buy the model anymore. Apple has won the marketing game because by keeping the same name, they don't have to scrap all of their previous marketing whenever the model changes.
There's one big reason - latency.
Not all games are twitchy FPSs and racing games though; and not every element in even *those* games has to be calculated for instant feedback.
Let's say you have a flamethrower in a game. You need to be able to see where the flames are going, and where your enemies are going, 'immediately' so that you can get a good kill.
But all the indirect lighting from those flames bouncing around in the scene lagging behind by, say, 150ms would be perfectly acceptable.
e.g. offload the intensive bits and pieces to the cloud, rather than the entire thing.
Although the entire thing would be possible in due time as well - players are already playing with lag; lag induced from rendering off-site is minimal compared to the hops between you and the average game server. Bandwidth is a far bigger concern; especially when you consider more and more ISPs introducing caps.
As far as CPUs go.. oh, absolutely, they'll get ever more powerful especially after Larrabee. But the games get more demanding as well. So now you need a new CPU, but it uses a different socket, so you get a new motherboard, turns out your old cooling fan won't fit, beside.. the TDP of the new CPU is 120W, so you'll need a watercooling solution anyway, etc. etc.
There's a lot of reasons why 'gaming on the cloud' ( I do hate that term. ) can be a good thing (from the technical up above, to the energy efficiency, to technical support, to non-invasive anti-cheating constructions), just as there are down sides to it (potential for lag, slurps bandwidth, no resale options, game could be abandoned at any point and you'll have no recourse, etc.).
But to state it's doomed before it's even launched... hmm. Let's have them give it a shot, first?
What do all of these have in common?
Some sort of license to work, i.e. they have to prove in some way that they know what they are talking about. Why are you trying to hold someone in an entry-level job up to the same standard as a "professional"?
What ever you want is going to cost a little more than it is worth. -- The Second Law Of Thermodynamics