Yao Li

That’s a useful distinction: remembering a rule versus enforcing it. If a mistake can be caught reliably, I’d favor a test, lint rule, or executable check over another reminder in the prompt. The question I’m exploring with AEG is how a failure becomes the right reusable…

One thing I’m trying to separate is model quality from acceptance quality. In a recent migration, an input-editing regression only became obvious when we exercised the interaction. When you compare models, do you give them the same explicit acceptance checks, or mostly judge the…