My hypothesis is that it did not become less deadly, it is that a combo of demographics and response (both positive and negative) played more of a role than we are ready to admit, mostly because we still seem to feel like everyone did a bad job when the response was probably closer to “fine.”

The response in China, once things got out of control, included locking sick people in their homes. This form of quarantine is probably good to reduce spread, but it is also probably bad for the health of the patient. If there are multiple people in the home, it is probably worse for all of them. China also has a lot of multigenerational households (grandparents, adults, children) which means the vulnerable are together with people who are lower mortality risk, but higher risk for traveling and getting infected. I don’t know what percent of households in Wuhan were like this, but it probably played some sort of role.

We also forget that it wasn’t only China that got wrecked - Italy also had substantial challenges.

On the other hand, if your healthcare system isn’t overwhelmed, people are wearing masks so initial exposure is low, and you are caring for/separating the sick from the healthy, the outcomes are probably exponentially better.

It certainly seems that later strains were less dangerous though.