Heh i kind of agree here. Reasonable docs , easy to bootstrap and completely unaware of the underlying performance probs you are about to create bc it “just works”. Curious to your reasons though ?
Apparently the facebook team (and others in production I'm guessing) addresses some of this by only having a whitelisted set of queries in production. So you get the GraphQL flexibility during development, but a effectively a static set of APIs in production.
but GraphQL isn't a storage engine. or are you saying that in a few years we'll realize we've been using to solve a whole lot of things that really didnt need solving?