The articles central premise is based on an assumption that 40% of the US population has already been infected. That seems very high to me, it would suggest 140M infections. If we work from the current deaths of 142K and assume a 0.25% fatality rate from all infections, it puts us in the range of 55M infections.
So not to say we wont see this taper off in current hot spots as certain high density areas get closer to herd immunity (and other areas are less at risk due to population density) - but I dont see any way we have had 140M infections to date.
RE: If we work from the current deaths of 142K
How accurate is that number?