Yes and no. Specialization is good for productivity on paper - but it breaks down horribly in practice. In programming, you may need to extend or re-write your framework or even web server's logic to scale your websites. In real-time trading systems and HPC programming, you need to have a good grasp of compiler and even kernel memory management to achieve low latency.

The analogy extends beyond programming to engineering/science in general. The field of molecular biology was formulated by quantum physicists interested in applying their experimental techniques in Biology. The new emerging field of Bioinformatics was conceived by statisticians and computer scientists to apply data mining techniques to genome data. Genome sequencing itself, is a blend of EE (taken from silicon waffer design) and molecular biology.

My opinion is that if everyone relegates to their own "tower," and aren't curious to drill deep into the bits of the other fields. There will only be incremental improvement to the existing towers, and no new towers. Hence the original author's point: kids are satisfied with interacting with technology on a superficial level (many in fact have a high level understanding of how cellphones work, cell phone gets signal from wireless tower), but they aren't curious enough to dig deep into the bits (program an Android game).