I am baffled any time I’m on facebook how horrid it is. Loads insanely slowly, won’t accept picture uploads at random, randomly doesn’t let you share to people in your friends group (it will just hide people for no reason). Such a terrible website. Sadly I have to stay on it for friends and music reasons (there is nothing else anyone I know uses) but man, it’s crazy how badly they want you to install their spyware app so you don’t use the browser version.
GraphQL has all the right combination of abstracting a really hard problem, ignoring the hard details, and giving you enough toggles to pretend you can solve the worst problems that makes hyping it really, really easy.
I’m mostly into it for the strong typing, self-documenting nature of it. In my own GraphQL APIs I’ve done a pretty great job of avoiding common pitfalls.
I’m a Ruby on Rails developer currently developing a service that’s basically ripped out of another Ruby on Rails app and the legacy data is just crazy bad — a lot of it has to do with poor validation but it’s understandably easy to get to that point in a dynamic language like Ruby if you’re not careful.
I also manage a REST JSON:API and it’s just so bulky and horrible to deal with. The tooling is barely there and it’s way overly complicated compared to GraphQL — the concept of “only query what you need” is fantastic.