The author never compares the number of people who get traditional education today with the number who got such education 20 years ago.

Too extreme to be necessary. The argument relies on a premise concerning “most of the machines in our lives” — but it doesn’t have to encompass all machines. So, the author could acknowledge that expert knowledge could help us operate many machines. In addition, the author doesn’t need to assume that our ability to use machines is almost never enhanced; what matters is whether that enhanced ability is something that would improve prospects in tomorrow’s job market. The author assumes that it would not improve our job prospects, but that doesn’t require the author to think that we couldn’t get enhanced ability to use machines from expert technical knowledge.

Necessary, because if this were not true — if half or more jobs in tomorrow’ job market WILL demand the ability to use machines designed only for experts — then the fact that today’s machines are mainly designed for nonexperts doesn’t support the conclusion. A nontraditional education might be better in the future, because we might need to start using machines designed for experts.

Not necessary, because the author never argues that a traditional education and technological expertise are mutually exclusive. The argument is just that the technological expertise would not prepare someone for tomorrow’ job market more than a traditional education would.

Not necessary, because the author can acknowledge that for some machines, such as those designed for experts, technological expertise might be more important than verbal and quant. skills.