Meanwhile, that army of "commodity" programmers will be out there, energetically growing the market for your services by bringing lots and lots of new PHP-powered sites online. Every blogger who installs Wordpress, and every school and church and magazine and music label that goes online with a basic Drupal site, is a potential future customer.
Of course, just because PHP work pays decently and is easy to find doesn't mean you'll enjoy it. There's a lot of pain involved in working with PHP, particularly other people's PHP. (I ameliorate the pain somewhat by sticking strictly to Drupal, which has higher coding standards and a better defined, more modular structure than the average PHP newbie's code.) But the work is certainly out there.
It's easy to see that an interference from the attribute of one good to the attribute of another is usually not sound. For example, just because bricks are commodities doesn not mean everybody is able to build proper houses.
Also, if we follow his assumptions and logic, any language that would have won the popularity contest, would have been a "looser" by his definition. For nearly everybody would code in the most popular language, therefore reducing prices for programming jobs in that language.
But, of course, this logic is also flawed. For it assumes that supply of programmers increases while demand remains more or less constant, which consequently leads to decreasing prices for programming jobs. However, the popularity of a language may also increase demand for programming jobs. Depending on the relative effects, (average) prices may decrease or increase.
I don't know why the author thinks PHP coding is somehow different from every other kind of programming, and that there aren't huge differences between programmers (we're talking orders of magnitude) as far as speed of output and code quality go.