> The easiest way to understand 100,000 concurrent threads is by making all 100,000 threads do the same thing and execute the same code.
If all your threads are off doing different things at different times, and executing different sets of code, you might have problems understanding your algorithm.
Its a restriction for sure, but it does lead to far simpler "thinking". There's still a very rich space of parallel algorithms that can be explored within this restricted space.
In effect: you probably *don't* want to write general-purpose "desynchronized" code unless you really have to. Keeping all 100,000 threads synchronized and in roughly the same spot of code really makes your brain think a lot easier.
