I don't disagree with you - I just wanted to point out that "see this caught 300m$ error therefore it's worth it, and it also validates other regulations" (which is how I read OP and a common line of reasoning I see in public debates) is dangerously wrong because it ignores costs completely.