It took Facebook quite a while before they became profitable. The issue is with the lack of lock in. The "on demand" and service type startups are easier to substitute. It's just another challenge though. I could see Uber reach a point where they raise prices by 25% and I'd still pick them over another service with fewer drivers available.
Comparing to Facebook seems like comparing apples to oranges. The article is about selling physical goods and services for less than it costs to produce them, which is different than allowing people to access a service that you don't yet know how you're going to pay for.
A better analogy might be companies that manufacture video game consoles, where they almost always sell at a loss when a new console is first released, with the hope to make profit in the long term - but there you're locked in to some extent, as you pointed out.
