* Herd immunity numbers assume no ongoing vaccination -- they more or less say, "If life went back to normal, would the virus slow down or speed up?"
* Age-related mortality and age-related vaccination rates are important. If we vaccinate all of the people who are at serious risk, the danger is mostly gone. There have obviously been deaths in populations not at serious risk, but not in proportion to the economic costs of locking down the economy.
And as an aside, vaccine immunity is estimated to be about as good as post-infection immunity. And most estimates of the actual infection rate put them several times (some estimate 8-10x, but that's looking unlikely IMO) the measured case rate.