Lots and lots of production code to train AI models.

Programming models have been sophisticated enough for some time to not be restricted to specific languages. It's rather like the difference between a junior developer (who knows how to use a specific language) and a mid developer (who knows who to program generally).

Thanks for creating Gleam!

Is there any research that supports what you are saying here? I agree it mostly works, but I’d expect more API hallucinations and test cycles at a minimum.

I'm not sure how much that matters. I suspect and anecdotally hear/see that compiler errors and docs are as/more important.

That mindset is a bit dated when models can pull up docs for a few years now. You can even have it write itself a little guide for lesser known languages within its own agents file.

The entire gleam language tour fits comfortably in a modern LLM context window. You're one web search away from "training" it yourself.