We should have some kind of dashboard tracking how many people die from terrorism, car crashes, heart disease, COVID, etc., and how much time and money we spend trying to prevent these deaths. We are just so terrible at evaluating risk, for some reason it is OK to waste everyone's time with over the top security to stop a tiny portion of terrorism deaths, while still being OK with high speed limits that kill thousand…
People are just fine at evaluating risk. It's just that the prioritization of the population is different than the prioritization of HN/Reddit/whatever. "People" aren't bad at stuff. On average, they're about average at it, which is tautologically true. People don't wanna die in a car crash or from covid but they want to die at the hands of others even less.
That is absolutely not true, at least not in the way anyone means.
Humans are "terrible" at remembering numbers. The average-ability human has average ability, obviously, that is tautological. But as a whole, humans are not good at remembering numbers.
A reasonable question is, what does "good" even mean? Well that depends on context. We are worse at remember numbers than computers are. We are worse at remember numbers than the same human, but with pen and paper. We are worse than some animals.
Obviously, there is no objective "good" or "bad". We are better than some animals, etc.
But when someone says that humans are terrible at evaluating risk, they usually imply that we are bad at it as compared to risk evaluation done with real calculations, and that is internally consistent. And that is something that IS somewhat objective, and most humans fail as compared to this objective standard (e.g., lots of contradictions in risk decisions).