Quoted from a trusted source that I have followed since early March:
"As I predicted 6 weeks ago and again last week, the CDC finally realized their infection fatality rate estimate of 0.4% (for symptomatic infections) and 0.25% for all infections including asymptomatic, which is the one everyone uses was too low and they've adjusted their "best guess" estimated overall IFR (including asymptomatic) to 0.65%, with an estimated range of 0.5-0.8%, which is very close to the 0.5-1.0% I've been predicting for months (since the NY antibody testing was done in early April, at least, revealing their IFR of about 1.1%, and this was before most epidemiologists were saying it - they're mostly estimating 0.5-1.0% now, as per the 2nd linked post below). I'd rather they were right, but they simply weren't factoring in the antibody testing data properly, IMO.
This is likely the last "inherent" IFR estimate, as we're now starting to see (I think) the impact of improved medical procedures and treatments on the IFR and we should expect the effective IFR to come down from here on out and especially once we have engineered antibodies and, of course, vaccines.
But the 0.5-1.0% IFR range provides a good "worst case" planning tool for death estimates should we not practice interventions (like masking/distancing and testing, tracing and isolating) or if treatments/vaccines don't pan out as planned. And that is what gives us the crazy high potential of 0.9MM-2.6MM US deaths eventually, if 55-80% become infected with an IFR of 0.5-1.0%."
This dude has been prescient on all things Covid throughout....
Coronavirus Disease 2019 (COVID-19)
Scientists Say New, Lower CDC Estimates For Severity Of COVID-19 Are Optimistic