I also think if you read the rest of what they're saying, it's probably is true that we will only get Linux GUIs which are competitive with Windows or Mac when people are willing to fund their professional development. OSS works exceptionally well for lower level components, or for components targeted at developers, because everyone involved benefits. For a professional developer a contribution acts as a good mechanism to raise your profile within the community and improve your ability to earn a decent salary elsewhere. For a company, running open source projects means they can solicit contributions and improve their profile; they contribute to other projects because it gives them a base to be able to build their products. We don't expect or find that OSS developers of high profile projects are starving in garrets, they are in fact leaders of their communities, highly prized and in-demand.
But these mechanisms do not have the same force when the consumer of the software is a general user, not a developer in the same field. i.e. when you're writing software in C++ which is then going to be used by graphic designers or artists, or by a casual user. In those case, the direct reputational benefits just aren't there, and that's why those sort of projects often struggle to take off. And on top of that those projects also involve a lot of painstaking and arguably boring work to get everything polished to a fine sheen. Krita for instance is excellent, but it has found a way to fund development through yearly crowdfunding. I think the same applies to Linux GUI's, that ultimately you need some pot of money able to support developers, which moves around according to the desires of informed consumers. In my opinion, that's what it will take for desktop Linux to become a serious mass market alternative to Mac and Windows.
