The Limits of Rubrics

Worth saying explicitly: a rubric captures the articulable part of taste, not all of it. There's a part of evaluating work that can't be written down. It's the judgment call in a borderline case. It's the sense that something is off even though it meets all the criteria. It's context-specific—what's excellent for this organization or this moment might be different than what's excellent in another context. It's the part of taste that is genuinely personal and earned.

This is actually reassuring. It's why you still need human directors even in an AI-driven world. The rubric can tell you whether work meets basic standards. But it can't tell you whether basic standards are good enough for what you're trying to do. It can't weigh tradeoffs that your specific situation requires. It can't recognize excellence that doesn't fit the pattern. That's still judgment.

But the rubric captures enough. It makes standards visible. It makes feedback teachable. It makes evaluation consistent. It takes some of the burden off intuition. And that matters.