> Even though the cutting edge is a lot less stable, it's by far the better investment, since the technology's lifespan is longer. Plus, generally a lot more pleasant to deal with.

I'd disagree on both counts there actually.

First, if something has already become a notable "legacy" technology then it's probably going to exist for a long while yet. There will still be people writing C++ in 20 years. I'd feel a lot less confident saying that about Node.js. When balancing the equation you have to factor in that most new technologies will never gain significant traction.

Second, there are different sorts of evolution around technologies. C++ or Java as a speciality now aren't inherently interesting, but they're disproportionately used in fields that I find interesting (information retrieval, real-time systems, signal processing, low-level programming, finance, big data, etc.). I find all of those more interesting than Rails in 2005. Much of that is personal preference, but again in older stacks the interesting things are the problem space, not the stack itself, while with newer stacks, there's more novelty value.