Ordinary flu killed some 80k americans last year. Covid-19 seems to be ~20x as deadly as the ordinary flu, and spreading at least as fast if not much more. So if it becomes endemic, embedded in the general population, just as flu is now - then the next flu season will have 1.6 million dead americans. Most people would consider that this is a large and undesireable result; that worrying about such a result is reasonable, and that it would be imperative to try and prevent it from happening if we can.

If everyone in the world were suddenly infected today and the onset would be, say, five days, the world would be in a mayhem in a week. The current fatality rate in countries such as Japan that doesn't have overloaded medical system, may seem low, but: 1) it's low because the system isn't overloaded. When overloaded, it starts to rise as the patients with heavy symptoms just die. 2) at a stage of a epidemic, it affects infra, which has feedback loop effects. 3) (for the sake of an argument) two percent of humanity is still 156 million people. A death sentence for 156 million people? That's harsh.

Indeed, overloaded medical system = high mortality.

To give a concrete example: recently there was a video report in Chinese state TV showing a heroic tag-team (doctors+nurses) doing 6 intubations per hour. Note: intubation is practically the last stage before you end up in a body bag.

That's a crazy number of intubations/hour for a single team. In such circumstances, you're guaranteed to get less-than-perfect recovery rate due to cross-infection.

Slow rate of infection is key to low mortality.

Same for car accidents, cancers, wars, why should we care right ? it's only an extremely small handful of people who are affected.

What is a "small handful" of people? One percent? One tenth of a percent? 0.1% of a billion is a million.