In our observation of assessments in the AICoding field, we found that these evaluations mostly remain at the stage of “code generation” and “closed questions.” This approach overlooks the basic needs ...