Build a function of unknowns that gives probability of a “bad outcome”, plug in the ranges of the best data you can find for the unknowns.

Compare this to the probability of getting Covid (~100%) and having a “bad outcome”.

This let me confidently say things like, “getting vaxxed is 2-5x safer than not” and ending the argument (whatever the actual figures were when I got vaxxed).

The real answer isn’t technical though —- we need public health authorities to learn post-Internet PR. I’m cribbing Eric Weinstein here.