You don't need to systematically test people in order to assign their death as a flu death.

We have many different ways of counting deaths. We either count the people who had a test; or we count death certificates; or we do surveillance to see what respiratory illnesses are circulating and how many people are searching google for flu-like symptoms and how many people are attending primary care and ED for flu-like symptoms.

(and which ever method we pick, if we pick the same method for covid-19 we see many more deaths)

The whole point of the article is to argue that those flu deaths are "wildly inflated statistical estimates" and not real flu deaths in the same way that Covid-19 deaths - where we are systematically testing increasingly wide-ranging groups of people who are hospitalized or even just in care homes - are real Covid-19 deaths. That the only apples-to-apples comparison is not to statistically compensate for the lack of flu testing at all and only compare tested, confirmed cases.

It's obvious just from comparing the numbers in the article to the ones in the comment I'm replying to that lack of testing has a much, much bigger effect on confirmed flu numbers than confirmed Covid-19 ones - confirmed Covid-19 deaths are at 2/3 to 1/2 of excess deaths, whereas confirmed flu is at 1/4 to 1/8th of estimated flu deaths, and the flu estimate does more work to exclude unrelated deaths than just counting raw excess deaths would. The supposedly apples-to-apples comparisons in the article looks to be even less directly comparable than the apples-to-oranges one it's rejecting.