There seems to be really a resurgence of ML; I found the first part (ML-centric) of proglang course on Coursera [https://www.coursera.org/course/proglang] quite enjoyable.
From a quick skim the approach taken in extending the powerful static type system of ML in a dynamic/open programming context is quite elegant [see Alice ML Quick Tour > packages and following sections]: it builds the notion of packing/serialisation/distribution (both code and data) taking advantage of the ML module system.
alice is probably my favourite language i’ve never used - i keep thinking that if i ever need to write that sort of concurrent/distributed app i’ll give it a serious look, but my work code is all c++ and my personal projects tend to be more single-user cli/desktop apps for which ocaml is a better option. i was quite sad to see that it had died the last time i checked; nice to see chris double’s effort to revive it.
I think with a great remoting solution, Rust could fill the same design space as Alice ML in part by being heavily influenced by both ML and Erlang. It already does copying or change-of-ownership when doing intraprocess threading.