It would be interesting to actually model the portfolio effect somehow. If you had 100 people with say n entrepreneurs, m social workers, p investigative journalists, q frontline first respondents, r soldiers, s homeless people etc, what does the net civ risk profile look like?