I think there is an element of truth to this that some people probably do not like to admit ...
Up until the release of the Nintendo DS and Wii the primary motivating factor for people to "upgrade" from one console generation to the next console generation has been the processing power of the new systems; we see a lot of this continue today with people like MikeB talking about how the Cell processor is 20 times as powerful as a 4 year old G4 processor when performing the fast fourier transform without any consideration of what that means or why that would be important.
This focus on technology can not survive for much longer being that within 5 years for the same (inflation adjusted) price of a PSP or XBox 360 you will be able to buy a laptop that is able to provide 3D graphics at a similar quality to what a cutting-edge console will be able to provide.
Why consoles and portable gaming systems will survive is that their hardware is dedicated to gaming which means that their user-interfaces can be perfectly suited to providing a better gameplay experience than any lap-top can provide.
The problem is, there are major economic incentives against what he is saying.
1) Convergence, or, one device to rule them all. History has shown time and time again that for consumer electronics, we do not want one device that does everything. Sure, some people will and thus there will always be a niche for those devices, but they have never been something the populace wants. We already have the be-all end-all of convergence; the limit of convergence technology is the computer.
And if we've learned anything about computers (or other general-purpose devices) it is that you cannot control distribution with them. The more your device is billed to do everything, the more people will want to get around the limits you impose on them. Take the PSP - a device billed to do everything, and people don't buy games. Or UMD. Or pretty much any part of the ecosystem Sony set up. Thats because when you are selling a device that is billed as being able to do anything, and then you artificially stop people from doing things, you destroy consumer confidence and lead them to alternative paths. Then when you take a just-as-easily cracked device like the DS, which is only billed as a gaming machine, you still get the piracy -- but you also sell software and cases and pretty much every part of the ecosystem.
This pattern can be seen time and time again in consumer electronics. Jailbreaking iPhones and iPods is a booming business in and of itself - while people happily shell out $3 for ringtones on lesser phones. It's human nature; the more you *think* you can do, the more you *want* to do -- despite whether or not others want you to do it.
And these are just economic incentives against convergence. This doesn't even get into more features necessitating more size (which consumers don't like) or increasing cost (which consumers don't like) or consumers asking why their movie player also plays games when they don't have any desire to game.
Don't get me wrong - there is now and always will be a market for convergence devices. That market is NOT the mass market. This has been shown to us time and time again.
2) The 'Dumb Terminal' argument (which the author mistakingly calls 'cloud computing,' which is slightly different). This is the argument that all you really need is a net connection, and everything else can be streamed to you over a network. This came up all the time right after the days of the mainframe computer, where people would connect to a mainframe computer via a terminal which was basically a monitor and a keyboard. As microcomputers got big, it was constantly assumed that computing would eventually return to the mainframe model, as it is easier to support one complex computer (or cluster, or whatever). So you could have all your data on the central model, and pull it up on demand, and your personal computer would be cheap because it would basically be a connection to this other computer.
Once again, history has shown us that this is not acceptable to the masses. EVERYBODY does something on computers that they would not want to put in control of a third party, even if it would probably be safer -- whether thats illegitimate file sharing, or just emailing personal information to a family member. It may be safer in the hands of someone devoted to keeping it safe, but that doesn't remove the psychological reaction to giving your sensitive data to someone you don't know.
This gets even worse when it comes to media. We are so used to owning our media, and doing whatever we want with it. In a digital world, this means ripping it to our hard drive, putting it on our iPods, putting it on Youtube, watching it on our computers, or whatever. Rental systems are one thing, but when we think about what we 'own', we want to 'have' it. We don't want it streamed to us over the internet, or have anyone tell us when, how, or why we can use it. People do not want devices that try to limit this, and you can be damn sure that a centralized solution like that would not let me use DeCSS or AnyDVD to rip my DVDs (all legally purchased, thank you very much) or Blu-Rays (still not rip-able enough for me to buy, but eventually...).
This point meshes with #1... if I am buying computing, I want to compute regardless of what it is... but you can be damn sure a content distributor is not going to allow me to get content from other channels if it is up to them.
Sorry for the length, I expect many "TL/DR" to follow this... but don't think that this guy's argument holds any weight. We've been hearing it for decades, and its no truer now than it was then.