COVID-19 Outbreak Information Updates (Reboot) [over 150.000,000 US cases (est.), 6,422,520 US hospitilizations, 1,148,691 US deaths.]

Is that three weeks or four?

If three weeks, that's an annual rate of 65K - If four weeks, that's a rate of 49K. That would represent a fairly bad flu year, especially the 65K, but I think yes, that would be much closer to flu.

Of course, May isn't typically a bad flu month so if the idea is Covid might begin to behave like the flu, you would have to compare those 3,800 deaths to any year for flu, at the same time of year. If the actual average flu deaths for three or four weeks of May into June is much lower than the mean (total/12) monthly average, then it's a different story.

Hard to say, still not enough data, I don't think.
It looks like the most accurate date and backed by Johns Hopkins, 17MAY2022 looks like the day we crossed the 1M threshold. So less than a month and probably still comfortably ahead of the flu in annual deaths.