The more it goes, the more I think Gleam isn't meant to replace Elixir or Erlang at all. I would have expected much more emphasis on the BEAM, showcasing hot code reloading and fault tolerance using supervision trees in a live server app (just like we have in Elixir and Erlang). But the direction Gleam has taken recently seems to be towards developer convenience, which is nice _as long as_ there _is_ a compelling reason to code in Gleam in the first place. I'm sure with Lustre and some other libs they're working on making Gleam a serious language, but so far it's not able to do what Elixir does.
I don’t know if replacing Elixir is even one of the top 10 reasons Gleam exists. People don’t usually pick a platform and then pick a language.
Many times a language develops into an option by bundling features. In Gleam’s case, they bundled a certain type system, extreme developer convenience, and optionally features from BEAM.
The tooling for Gleam, although not perfect, is one of the best among more younger languages. The other day I saw a post about Crystal here that got me interested, but the LSP and code formatting aren't very good, which makes the whole experience a lot worse. Programming with Gleam is fun in part thanks to the tooling.
reply