Of course they did but it's still going to suck without accuracy and consistency. Like manus uses so much time and credits to do super simple things it basically renders it unusable.

They're coming. I've seen the observability tools try to do this but I still have to tweak it. it's just time-consuming. Empromptu.ai is the closest to solving this problem. They are the only ones that have a library that you install in your to do system optimization, evals, for accuracy in real-time.

baml isn't great. You need to write it in their format and it still doesnt really solve the accuracy problem from humans interacting with your system that we're talking about here.