Of course large, important projects require a certain amount of funding to succeed. I'm not sure where this is really disputed, I don't see the actual argument the author is railing against. Fedora Linux may not generate revenue directly, but it is an important component to the development of RHEL, which is the core component of Red Hat's highly successful business model.
There are many, many, many organizations of all types all over the world that use different economic models to fund various different interests. I sing in a choir. We're a non-profit organization and raise money in a variety of ways to pay for music, a professional director, various instrument ensembles, venues to perform in, and people to record the concerts. The choir itself is auditioned, but all volunteer. We spend a few hours per week for fun because it enriches our lives, and as a result people get to come enjoy our concerts. Last spring we performed Mendelssohn's Elijah at Sanders Theater in Harvard Square. Unlikely to happen, but in theory the recording of that concert could become popular, and for a trivial cost via file sharing networks that performance could be listened to by 1 million people all over the world. Demand for those digital copies on that scale might average $0.001/per copy. Not worth it to try and charge for it anyway.
But we aren't a business. There are a few people associated with the choir who are compensated for their work. They are highly trained and skilled and put in more than the typical 3 hours per week. The rest of us are not paid and in fact pay dues. Is it really that different from the way an OSS project might work? 2-3 key maintainers and a hundred other contributors averaging 3-ish hours per week?
According to their website, WCVB-TV supports 484 non-profit organizations in the Boston area. http://cf.thebostonchannel.com/bos/sh/charity/front_charity....
There are many valid ways to fund software projects. Some of those projects turn out to be extremely successful and popular, and in many cases the Open Source nature of the product actually adds real value to end-users. Sure, there is some bad software, but I've used lots of bad software in my life, both free and not-free. The fact that people might be working without the compensation they might deserve in a perfectly balanced economic world is not really a big concern. That's life. It doesn't just apply to software, and the reverse is also true. It is most assuredly the case that some large companies are paying some programmers far more than they theoretically deserve for the software that company sells.