> Sure, I could throw an LLM at it and it would create the right abstractions/de-abstrations and things would work. But that feels like putting band-aids on top of band-aids.

A lot of assumptions and if you believe them to be true, why dont you follow thru ?