> You're assuming the current ensemble of commonly used software stacks is the most optimal there is.

No, I'm assuming that the level of practical complexity has a much lower bound than people seem to be trying to service, and that while that bound will grow with time it doesn't grow at the rate the available “solutions” do.