Point is that we know Covid has very different risk profiles based on who you are. Saying "That means you were roughly twice as likely to die as you would have been if you were serving in the U.S. armed forces in Afghanistan throughout 2010" probably underestimates your risk if you live in a nursing home, but it vastly, vastly overestimates your risk if you're 20 and in good health.
I don't see that as an issue with the measure.
Moreover, average micromort increase in a population is a useful metric, in the same way GDP is a useful metric, even if your individual productivity is very different.
though i'm actually a big fan of this unit (e.g. as an expression of the relative risks of various activities), it's clear from the comments section that when people are presented with a simple number they tend to treat the underlying reality that it reflects as simple too. this is FAR from the only domain where people reduce complicated things to scalars, and then do atrocious, hideous, unspeakable things to those scalars (take standard deviations, %changes over time, and the like), and then make claims about reality based on the results of those mathematical perversions (i take particular aim at the social sciences here. sorry. they certainly don't have a monopoly on statistical innumeracy, but their domain is just more rife for this sort of abuse).