Scripting languages for scripts. There is definitely a place for them. It is all a matter of scope. Large programs that will be touched by many hands benefit from static checking to make sure all the pieces match up.

Personally, I like static type checking with type inference more than a lot of explicit specification of typing. It generally refactors better. I wonder how that would have come out in your example. Would having static, structural type checking have given you the flexibility that you needed?

There is also the issue that as soon as you start passing data between machines in serialized form, most of the type checking tools lose effectiveness. I need to learn more about GraphQL as it seems to address that to some extent.