We are not all mechanics. In fact, the knowledge of how to make a simple repair like fixing a puncture in a tire is much less prevalent now than it was forty years ago. Nevermind the knowledge of how to troubleshoot a vehicle that won't start.
The consequence of this is that knowledge of the mechanics of the complex system that is a modern vehicle is typically beyond any one person, and certainly well beyond the majority of its users.
Today there is leverage in far flung industries for the individual who has specialized knowledge of his field AND knowledge of how to apply software development to that field. But these opportunities will not last long. Any software that can increase profits in an industry will attract competition from other software developers. That software will be improved and refined by software specialists. It will (and has) evolved to a systemic complexity that equals the modern vehicle.
Software is a tool. The civil engineer who uses AutoCAD to design your road does not need to know how to code his own AutoCAD.
Although industry standard tools with scripting environments often have active communities of people developing sophisticated tools on top of them. Encase is one example; the Canadian Mounties write quite a bit. Splunk is another. Niches may asymptotically outrun good software.
That's a little too good to be true, and the bakery example stroke me as very ill-suited.
I enjoyed the article, though, and I wish it were true that clueless incumbents would die rapidly and developers would get infinite leverage.
What I think is true is that developers will get a little more leverage while clueless incumbents will take forever to disappear.
(Edit: in the long run, everyone will be a developer. There will still be good and bad developers, but there won't be non-developers professionals, just like today there are not many people with a job who can't read or write.)