Exactly, there's things you aren't shouldn't do with an llm. But generating helm charts, configs, action workflows, building specs and then implementing based on them? Simply a no-brainer.