On transitioning from Python to Go. Really good points that I've thought about as well—Go indeed makes concurrency seamless, and doesn't place any obstacles for parallelism.
I suppose the point is that concurrency must be supported natively by the language, and parallelism should be a runtime feature from the start. Both of these do seem like truisms if you think about it—any other model than M:N multiplexing has a bottleneck at one point or another.
I've been toying around the idea of trying to make a programming language again (and I'll probably find an excuse to do so as part of my uni curriculum at some point), hopefully a proper one this time, not some one-letter-Brainfuck-level-technically-Turing-complete-but-not-really-useful thing, so perhaps this could be a fair design cue to take.
(Via Chris Hodapp via Lobste.rs)