The discussion yesterday got me wondering. I am feeling more and more like I would like chalk to be more like polonius: something we can improve but that isn't directly blocking anything in rustc. I am therefore contemplating trying to create more "intermediate" goals to try and improve the rustc trait solver directly. I remember we went a bit down this road before, trying to decide e.g. how much of GATs we could do with the current framework, but I'm wondering if it makes sense to think about refactoring and improving the current trait solver in parallel with chalk. I am going to put some more thought into trying to make this more concrete.