As much as I see the argument that the author makes, it is difficult to agree with them. Briefly, the rates of course completion in an online (MOOC) setting is far lower than in a classroom setting. Second and probably more important, IMO, is that classrooms are a space of mental growth for the students. If the students do not benefit as much from classrooms as they should, then it is a problem of either the curriculum or the lecturer. The same set of students are likely to perform worse in an online setting, not better.
The fallacy in the argument, am afraid, lies in defining the role of a teacher or a professor. They are not lecture giving machines. I have benefited from, and delivered lectures that cater to a broader range of interests within the classroom. The job of the professor or lecturer is one of eliciting an interest from the broader group. This means that lecturing is an active process than passively spouting our mind. An active process that demands the complete participation of both the teacher and the taught. This would mean foraying into unforseen topics to build an example in real time. Such diversions to illustrate concepts is not possible in online courses.
Last but not least: humans are still social animals, and they benefit from live interaction with their mentors. It would be a sad day indeed, when mere projections on the wall would be the way to teach. When that day arrives (it is not an if) the privileged will still want their wards to be taught by a person.