This view only makes sense if you believe one language is enough. One language that has it all figured out, one language that is the right solution for any task.
Even if such a language existed (be it Lisp or not), which is unrealistic, there is still a second aspect to the whole thing: programmers' brains. Different languages map differently to different kinds of brains, and that's a very good reason to explore multiple concepts in itself.
The Split is the natural way to explore a large search space using different strategies. It happens all over nature, and it's the way human-powered research works as well. Monocultures don't perform nearly as well. Yes, there is some effort wasted, but you can rarely ever come to an objective consensus where this waste actually occurs - except, typically, in hindsight.
Also, there is quite a bit of cross-pollination of ideas happening among different environments and languages, so advantageous traits do get passed around beyond their initial ecosystem.
I imagine cross compilers and common runtimes will get more popular, too, now that we have the processing power - and that's also a good strategy do deduplicate effort.
The only thing that really strikes me as a bad strategy is militantly advocating one programming language as the master race. You're not even doing the "pure" in-community a favor by denying that concepts from the outside world might occasionally be useful. Instead members are reduced to armchair criticism along the line of "yeah, nice library, but you didn't write it in Lisp, so at the end of the day nothing at all was achieved."
Every language has its use, but really we ARE wasting time by splitting the worldwide community of programmers in different languages.
Look at how most programmers love one language and advocate it's the best there is. That's called a split.
Aren't we supposed to not reinvent the wheel?