Basically he theorizes that the reason Apple will price SL as low as $29 (a hundred bucks lower than normal) is because the iPhone app store has shown that the cheaper something is, the more money you make.
I'm glad he is making the point. I've bitched before that, for example, selling ebooks or audiobooks for $30 or $50 is idiocy. Much cheaper would be better for both seller and buyer. I've also experienced this in my own business.
However, I must add that Pogue perhaps over-simplifies a little, he makes it sound like an infinite street. Obviously that's not true. There's a limited number of potential customer no matter how cheap you make something, so there will be a point where profits fall if the price falls further.
The theory is that there is the classic statistical bell curve: the top of the bell curve is the optimal price, where profit is highest, so that's the price you choose.
The exceptions are that you might want to price it lower if you're trying to expand your market share quickly at the expense of some profit. Or you might want to price it higher, so as to position your product as a high-end product, again sacrificing some immediate profits for future good-will.
Update:TidBITS actually ran an article some weeks ago, they hoped that Apple would pursue exactly this cheap pricing with Snow Leopard. Simply because it has few new features, and it would not sell otherwise, but it's necessary for the future of the platform that it sell well. I agree, and it's good that Apple thinks so too.