Well, except that the group you're describing are usually referred to as "early adopters." How vacuous would the point sound if it were aimed at the early days of the PC?

The trouble, of course, with IBM's experiments with personal computers is that their fan-base is largely a bunch of nerdy geeks who want this kind of thing to work, and really doesn't tell you anything about the broader world.

So, one thing Cory's success tells you is that, even when you restrict the audience to "copyright geeks who want this kind of thing to work," it can work. Additionally, those copyright geeks have friends and family, and some of them are noticing how those nerdy copyfighters consume literature, and it gets talked about, a fact implicitly betrayed in your qualifier "largely".

It's easy to fire off an "It won't work," but the article itself discusses many of the strengths of the model in question, rendering your complaint moot; Even if the viability of Doctorow's publishing model isn't demonstrated solely on the basis of it working for him, it is supported in the article with other observations, as well.

Don't think that because you've criticized one pillar of the case for direct publishing that you've won the day, nor even raised a serious objection. You have very much more work to do before you've demonstrated "[t]he trouble...with Cory Doctorow's experiments" in a meaningful way.