In the end, we have millenia of experience that human mortality without Covid is very almost 100% (not all figures are in yet). Mortality after catching Covid is so far much less than 100% (again, some fiddling not-yet-finalised figures, until we check everyone in our current cohort again in... say 100 years?).
So, obviously, it's healthier to catch the virus!
...
What we currently have is people who are known to have Covid who died without any obvious other condition, who died with co-morbidity not usually considered high risk in this day and age, died with a life-limiting condition that should not have been so limiting, were already in terminal decline from something but with heavy Covid effects, ... got hit by a bus (and straight into the path of a steamroller) while infected.
Some commentators only accept the very first figure as the basis of the 'true death rate', perhaps as an extreme backlash against the stats not always being able to remove the very last (especially if it also includes "untested but suspected Covid" versions of the above, which apparently isn't good enough for the front-end list items even with the word "highly" stuck in there) from the grand total.
And now they're also 'benefitting' from the latest Excess Deaths figures (by picking a suitable historic lead-up that works well for them) being depressed due to a large number of those who might have been a current Excess having already been taken by the initial peak, leaving a marginally fitter population to ride easier over any later surge in infections.
Another problem with the bare stats being used is the inadequate initial testing rates meant many highly likely infections were unable to be confirmed (mostly those where things were not so bad as to have any doubt of survival) which inflates the "known known" death rate (but also the unprovable "known unknown" and "unknown unknown") and now with greater testing numbers, cherry-picking the right detection and death stats makes it look like it's just "too much testing, when things are actuay getting better", when it's just as deadly.
And an analogy I used the other day (which was quicker to explain in the context of that moment) about lockdowns was like being in a car going down a steep hill. You suddenly notice sheep on and by the road ahead, slam on the brakes (jolting the passengers, things fly off the parcel-shelf; basically disruptive to all concerned) before realising they're actually further off than you realised (any that you did avoid hitting are just part of a larger flock). So you ease off on the brakes (you're still on the hill, so you're naturally speeding up), and someone in the back seat is complaining about the lost time so you might even step heavier on the gas. But there are still sheep on the road and you're still heading towards them, and you're now back to recklessly anti-ovine speeds again. Anyone complaining that putting the brakes in the first time just did more harm than good isn't honestly thinking through what would have happened if there were no brakes at all, etc.