[oleks]

Using an LLM to generate code is not an easily traceable and explainable process. Using a DSL to same ends is. PL research has yet to meet explainability in AI head on.