CDC issued additional guidelines (Wednesday, I think) that now allows counts for probably cases and deaths to be counted towards the total.
https://www.worldometers.info/coronavirus/us-data/
Worldometer has already started including them, but as far as I could tell last night Johns Hopkins was not including them.
WoM had 677K cases 34K deaths last night, JH had 670K and 33K deaths
the other night when I looked, there was a good 2K difference between the two in death counts for Wednesday.
This inconsistency between the two and what numbers they show vs what they don't, bothers the piss out of me.