Nice! So supervision trees are what differentiates the BEAM runtime from other runtimes in terms of process management? Does it automatically respawn crash processes, and can it (or some other abstraction) deal with pooling and some kind of auto-scaling?
