It is mind-bending to think about the immediate future of modern computer, forget about what's going to happen in a hundred years. I, for one, hope there's still a humanity left to ask this questions in, say, 50 years. For the meanwhile, let's just imagine we are not under constant threat of an all-out nuclear war.
One of the main points of the article which I find particularily interesting is the need for clutter in a language's process of evolution. Numeric types being an example for LISP's dreadful, despicable, and messy transformation into Clojure. Clutter is wasteful, but it has a reason for existence, even if it's not totally useful at first.
Another important concept mentioned is the fact that Moore's Law is bound to stagnate by its very nature, which means programmers and hardware designers alike need to work around the problems arousing from this process. This event is one of the cornerstone reasons for the development of a universal programming language that shall endure the test of time. Considering the different elements involved in Language Evolution (clutter, necessity, data types, etc.) this process may prove to be chaotic. If we add the fact that we don't know how the future (or its hardware) looks like, the creation of a universal hundred-year language seems near impossible.
As many a developer of Silicon Valley hopes, whatever the outcome, it better make the world a better place.