I'm with you, but I don't see a clear path from "here" (illusion of the same memory) to "there" (each processor has its own memory, I guess). It's really hard to reason about latency, synchronization and communication costs in the abstract.
In a way I see the various cloud computing services as the organically grown (i.e. haphazard and clunky) answer to this problem.
One of the reasons it is so hard is because the discipline of thinking about those things isn’t baked into the language like it should be. You can’t look at a piece of code in a typical high-level language and reason about synchronization overhead because those semantics aren’t expressible in the code. We need languages that support expression of effects and coeffects, which would enable that kind of high-level reasoning about tradeoffs.