Earlier quoted context omitted.
That's called consulting, generally not really seen as a startup. I don't really get your point? You seem to be telling people without technical skills to go get into non-scalable business models?
The examples he gave are misleading; they tend to suggest consulting. But his point is right: There are applications that have been written for a small number of clients, but languish because getting that amount of success is something a developer can do on his own. In other words, it's tee'd up: There's a business case, a product and there are existing users. What remains needed is someone who can get additional use…
All one needs to do is go to freshmeat or ohloh to see loads of stuff that could be leveraged to fulfill a need and ultimately build a business. You know that it fulfilled at least one person's need, because it was created to scratch an itch. Sure, maybe they aren't polished or sellable out of the box, and many of them are infrastructure (rather than end-user-product) related, but that's where the work of building a business comes in.