It's mild for non-vaccinated South Africans as well.

The same still applies as always: Vaccinate the risk groups.

SA had very high numbers of previous infections, so there was a ton of pre-existing immunity.

Which some would see as a partial argument against lockdowns.

Comparing death rates in Australia, which actually did pretty major lockdowns, to South Africa[1], I don't know that this is a great argument against them. South Africa definitely paid for that infection-based immunity in human lives.

[1] https://ourworldindata.org/explorers/coronavirus-data-explor...

Of course heavy and complete lockdowns are effective (at delaying the spread of the virus) - if no one is in contact with anyone else there's no way for the virus to spread. A more apt comparison would be made between countries/states with slight or "normal/average" lockdown to ones without, but that's very difficult to do in practice. And then measure the financial and societal impacts (which is also difficult/impossible).

So if everyone gets Omicron, we will have a ton of pre-existing immunity?

We already have a great deal of pre-existing immunity. The big question we're answering right now is whether vaccine-based immunity (which we have more of in AU, UK, and the US) is as robust as immunity from infection (main source of immunity in SA). It's still too early to say with certainty, but it's looking like the vaccines are working if you're not immunocompromised.

Seems that we don’t have much choice in the matter so we will find out pretty soon. That wouldn’t be the worst outcome.

According the the CDC, natural immunity is not be named or acknowledged.

Yet the unvaccinated are still overwhelmingly more likely to wind up requiring hospital care than the vaccinated.

The seronaive are. The unvaccinated with a prior infection are unlikely to require hospital care: https://www.nejm.org/doi/full/10.1056/NEJMc2108120