That's useful, thanks. Though an estimate of seroprevalence based on a blood bank survey (both of these are from the same group, using the same method) has issues with bias (in particular, see the number of re-weightings performed and the huge impact of these adjustments in Figure 2A of the second link), that makes me hesitant to believe any claim starting with "75% of the population was infected, therefore..."

In other words: if we're seeing a resurgence in cases now, the parsimonious answer is that the blood-bank survey estimate was wrong, not that the virus is escaping all immunity. My prior probability of this explanation is dramatically increased when I see that the seroprevalence data was adjusted upward by a factor of 2-3 to account for "seroreversion" (which is essentially guessing).

Yep, that quote is an opinion from one person. "Virus expert" or no, it is still speculation. The fact that said expert is confidently claiming that 75% number makes me doubt his opinion. We can't confidently estimate the proportion of people infected in the US or Europe. I don't know how you'd begin to claim such a thing in Brazil, where the access to testing and data reporting is quite poor.

> for millenia the most-effective method of weather forecasting was constancy: "Tomorrow will be just like today"

I disagree that this like predicting the weather. The weather has consistent seasonal patterns that give you ~90% (brown number) of the story (ironically, this virus is also probably strongly seasonal, but I digress).

Here, we're engaged in speculation that some "new strain" might be worse than the existing strain. It's the opposite of using long-term trends to predict the future. It's a bit like claiming that every new typhoon is going to be the worst one ever, because it's different than the last typhoon in some way.