> And if those metrics aren’t tabulated for easy access, they probably should be.

I couldn't agree more. They're two different error rates for our society and measuring them accurately would help us go to where we should be on the curve.

Somebody should make a website visualizing the data we do have, perhaps with uncertainty bounds, and a recursive breakdown locale-by-locale… Nose goes!!

Edit: wow I bet this is a project that would be _way_ too difficult to vibe code with AI, with well documented data sources and what not. Sure would be a shame if somebody proved me wrong.