I think the author may have stumbled over his analogy a bit; rather than suggesting "Learn to sing, or else you'll be at the mercy of an asshole", it's more along the lines of "Learn to sing, so you know what to look for in a good singer". Big Idea People (henceforth referred to as BIPs) are not necessarily a bad thing -- sometimes they do genuinely come up with something good that would benefit the market -- but BIPs have an issue: they tend to have absolutely no concept of what is required to execute their idea.
People on here are imagining CEOs trying to do their jobs and having a good chuckle, but they're missing the point. Your boss, unless he himself used to be a lowly programmer, isn't going to be executing his ideas in any form that could be marketable. I don't believe the author was intentionally arguing this; rather, I think the more important point deals with being able to bring their vision down to a realistic level -- less "My boss is coding our new Android app" and more "My boss now understands why we can't duplicate the functionality of Google in two weeks time". The more he or she codes, the more they begin to understand the work involved.
I've never really considered myself much of a programmer, but having learned to code, I can respect the word involved and that a simple line count doesn't tell the whole story. The same basic principle applies here, too. Not only can it shoot down unrealistic ideas but it can keep them from getting proposed in the first place; they die at his desk, never having left the room, because he figured out long before opening his mouth that, under the current circumstances, it was unreasonable. At the same time, knowing what is possible can potentially give him new ideas about the direction of a project that not only work, but actually might make sense. This is not a bad thing.
Ultimately, there's only so much you can do and there will be limits to what the person in charge of your project knows. Be happy; you're not redundant.