**Some judgments are verifiable, some aren’t.** Some judgments are verifiable like a prediction of how much rain will fall. Other judgments are not verifiable like how well did your employee do last year. It is obvious when a verifiable judgment is in accurate. If a judgement is not verifiable, it is not possible to systematically and unambiguously determine if it was accurate. > [!tldr] 👉 It may seem impossible to improve a non-verifiable judgment. But we can make improvements by focusing on the **judgment process.** [[Artificial Intelligence]] is easier when judgements it makes can be verified. # Examples ## Verifiable - How much rain will fall tomorrow? - What will the GDP of Mexico be next year? ## Non-verifiable - Which solution to a problem should we use? - What punishment should be selected for a crime? **** ## Source - [[Noise - A Flaw in Human Judgment]] ## Related