Marco (hate him or love him) nailed the criticism of this a few years back:
The joke of “next year will be the year of Linux on the
desktop” is almost as old as the internet, but it’s
true: desktop-Linux fans always say it’s “getting
better”, and there’s always a major distribution update
a few months away that’s about to be awesome. But it
never is. And it never will be, because the reasons why
desktop Linux isn’t awesome today will still hold
tomorrow: it’s still an extremely fragmented
development community for which the non-geek user
experience is one of the lowest priorities.
The fragmented community doesn't matter. At all, really. It just requires that one group (Canonical, Redhat, whoever) put their guns behind what they provide and try getting that into the market.
While having more developers would help the individual efforts, having other people rebuilding the same thing doesn't hurt the individual projects. It hurts the ecosystem to have excessive duplication (and I don't know why I can list 3 dozen window managers or file browsers but can't name one good non-linear video editor) but the success of any single distribution doesn't depend on everyone else getting behind it.
I think the comment is actually pretty accurate in relation to the non-geek - the problem has nothing to do with developer fragmentation. Consumers don't care about that. The reason desktop Linux doesn't take off is because Grandma doesn't walk into Best Buy and see an Ubuntu laptop, they see Windows. And every computer 99% of consumers will buy will be from a store, carrying Windows of some flavor.
They won't go out of their way looking for a competing OS, or how to install it. As soon as the thing boots into Windows, you have lost them, because they won't change anything. And you can't try teaching them, because Windows has won for them - the device works, hence nothing has to change. Internet Explorer shows emails, even if it is slow, I don't know what "security vulnerabilities" are, or why developers hate it (and why do I care? they already rewrite the site to work in 10 year old non-standards compliant web browsers anyway).
Linux will get widespread adoption the day it is the default. Because it honestly does work, when you take the time to tailor the hardware to it. System76 or a Dell Ubuntu laptop are as far as you need to go to see that in effect - they work out of the box because the engineers tailored the device to the OS. Like they all do with Windows already.
The problem with Ubuntu, though, is that it is too different. It doesn't look like Windows 7 (or, for a lot of people, XP still). Instantly, nobody likes it. They want the same thing, better. Not something different. They don't want change, or to have to learn or read. They want to read their kids facebook posts. And you can't blame them for that.
I've seen that article and I submit the Nexus 7 as a counter-example. (If anyone didn't read the link: that quote's making an analogy between desktop Linux and Android devices.) The Nexus 7 was widely agreed to be great on day one, even by certain die-hard Apple fans. (I forget if Marco was one of those.)
With that, Google now has a track record of achieving "great on day one". Who's to say they can't repeat that on the desktop? In fact, they may be well on their way with the new Samsung Chromebook, which seems to have received mostly praise: the biggest remaining criticisms are with the concept, not the execution.
The central point I agree with Marco on is the problem with the rhetoric of "X open source project will work soon or eventually". It's always a pipedream and not a project you can hold in your hand - sometimes because the person demonstrating it won't allow a hands-on experience. :)
EDIT: I resubmitted it, because it is as relevant then as it is now: http://news.ycombinator.com/item?id=5051152.