if someone dies in a traffic death while having stage four cancer then this would indeed not change year on year mortality (assuming the person dies of cancer).

This is an awful example though because the victim profile of car deaths is generally middle aged adults, who would not have died otherwise. If your average car driver was a stage four cancer patient traffic safety would indeed be less of an issue relatively speaking.

The risk profile of covid patients skews so heavily towards old age and existing preconditions that it is actually not trivial at all. It's immensely important to look at death rates in the context of general mortality, not a vacuum.

It may very well be the case that covid 'crams' the deaths of people who would have died anyway into a short timeframe and there'll be singificantly lower flu mortality and so on later this year.