Why aren't there any successful ML-family language based on BEAM?

Hello everyone.

I learn Elixir (and Erlang to the less extent) currently. One of the biggest pain points - dynamic typing. I am really tired of keeping in mind all the APIs, of catching subtle value errors during long-lived tasks execution (basically any server application). Given I had prior experience with Python I can say I am tired of dynamic typing.

I also have used Rust and Haskell. I’ve never been so calm and confident til coding - nor before, nor after.

Anyway OTP and BEAM both look like really genuine brilliants. I know they have their problems (like any other technology) but high-level architecture is beautiful and amazing.

All of these points lead to question - why aren’t there any successful ML-family language based on BEAM VM? I know there are such languages but all of them look abandoned and unusable. Maybe there are some unsolvable language design issues and they’re holding development of such languages?

Another one question - is there any need in such language at all?


See Alpaca and


Thanks for the link!