This article never really tackles the fact that p(infected|vaccinated) != p(vaccinated|infected).
With high-vaccination rates in Israel, that matters a lot.
In fact, the scariest statistic it gives--that nearly 60% of serious hospitalizations are fully vaccinated) gives the p(vaccinated|infected), not p(infected|vaccinated).
Trivially, and to make the point, if 100% of people are vaccinated, then 100% of hospitalizations will be among the vaccinated--which tells you almost nothing about how likely it is to be hospitalized when vaccinated (beyond that its nonzero).
In fact, the scariest statistic it gives--that nearly 60% of serious hospitalizations are fully vaccinated) gives the p(vaccinated|infected), not p(infected|vaccinated).
Trivially, and to make the point, if 100% of people are vaccinated, then 100% of hospitalizations will be among the vaccinated--which tells you almost nothing about how likely it is to be hospitalized when vaccinated (beyond that its nonzero).
An embarrassing failure for science.org