The problem is that the article misses is that there is an opportunity cost of specializing, specialize in the wrong thing and your career is over. Try finding a Cold Fusion job now that it has seen it's glory days, sure they are still around but if you loose your job, it may be a long road until you land in another job that fits your specialty. At a certain age many developers do specialize, many of them tend to sunset with a widely adopted technology. Take COBOL for example, if you participated in COBOL development you would meet a lot of specialist, because most of the people in that technology are past the point of choosing to specialize. Participating in the latest technology is always going to create a filtered bubble of generalist, those are the ones that have less opportunity loss in transitioning.
