It is pretty astounding that a lot of the "new" things we are doing like functional programming are as old as OOP which itself is older than I am. That isn't necessarily a bad thing it's just fascinating. The big thing a new thing needs to do to displace the old is to be so much substantially better than the previous for it to make sense for there to be a shift. Perhaps we will one day get there. Meanwhile since concurrency is hard his beef with IPC seemed interestingly in the opposite direction of where things seem to be going. It seems more and more systems are going towards things like Isolates/workers/etc. models rather than direct multithreading in order to help eliminate entire classes of bugs. Meanwhile that is essentially harkening back to Erlang and how it did these things starting almost 40 years ago.

i dont see how posix has *anything* to do with the problems that actually exist and become more pressing in the hw/sw interface space. adding a posix interface to a system does not magically transform its architecture and other interface to something worse. also, calling flash drives super reliable must be some kind of joke...