Can agree with this. I have had some funny hallucinations, though, where it mixes up Elixir keywords with totally unrelated to programming things and then acts very confident that said things are in Elixir. So you just need to be a bit skeptical and double-check it. But that goes with the LLM territory anyway.
Can we say that being a functional language makes it easier for AI to understand and write cleaner code? Basically, there are no mutations of input data and I guess somehow that makes it easier for the LLMs to understand and keep a track of the input and avoid weird and fancy mutations down the call stack.
YMMV, but I think it writes fine unit tests, but really sub par functional or end to end tests that need to check business logic. I think that's just a hard case for LLMS and not an elixir issue though.
