Ask HN: Is there a need for YAML in post-LLM world?

[throwaw12]

Curious to know your opinion about YAML and other DSL created to make workflows/processes less software engineering (imperative) and more configuration (declarative).

Take a look at K8s for example, lots and lots of YAML configurations, which was used for human readability, but with LLMs do we still need them?

Wouldn't it be easier with LLMs to say: expose this service with 10 pods and enable only internal routing - which then updates couple lines of Python code (pulumi, aws cdk and etc,.) instead of generating 5 yaml files each with at least 40 lines of YAML and also try to reuse same attributes/labels/tags

[falcor84]

I would argue that it's more important than ever for effective engineering change management. With how easy it is have AI make changes to a system, I'm reminded of this tweet from DevOps Borat:

> To make error is human. To propagate error to all server in automatic way is #devops.

I'm not sure if YAML specifically is the best format, but we absolutely need something that offers a single readable source of truth for people to manage and reason about changes.