Even trusted people get things wrong every now and then. And he's not an expert. I suggest listening to experts in the field who are questioning the results. And a cursory look at the methodology shows that their sample size was not only relatively small, but also vulnerable to bias because only people volunteering to be tested were tested. This would likely impact the type of data we get. If it was truly a random sampling of the general population, I'd maybe give it more weight. But there is also the issue of the tests themselves which are also somewhat unreliable.
In any case, there are still questions about whether the presence of antibodies provides sufficient immunity to CV19. The WHO has been clear as mud on the issue.
And again, this had been discussed at length earlier in the thread and the reports were essentially debunked as incomplete and it was agreed more testing needed to be done.