This is kind of a double edged sword though.

Say the virus mutates so that only 50% of its features are recognizable.

You'd think that the fact that your cells can recognize this is good, and it is to some extent, but, it also prevents new cells from learning about the virus(in efforts to prevent autoimmunity in analogous scenarios with non malicious objects; our immune system has to walk the fine line between being too permissive and allowing infections, and being too restrictive and murdering its own host through nasty autoimmune diseases).

So, it mounts a 50% response, which is often good enough, and more than makes up for the more limited response with speed.

But what happens when the % recognizable drops to like 3%? Some percentage of cells will recognize the 3%, and be able to mount a weak response, but more critically, block other cells from developing new responses. This limited response will slow down a proper response and itself will be likely ineffective.

Its known as "original antigenic sin", if you want to read more.

I hope COVID-19 isn't reading this.

I'm pretty sure covid-19 doesn't read Hacker News.

Not that surprising, since an immune system would be very likely to encounter a slight mutation or "variant" of the virus in the future, after the virus spread through others in the population.

But this obviously isn't the case with seasonal viruses such as influenza or coronavirus.