we have decent evidence (best is from the cruise ship and s korea) that the mortality rate is between 1-2%. call it 1.5 for this. we have 10k dead right now, but that's more an estimate of how many infected we had 3-4 weeks ago. my guess is that between then and now we've slowed transmission pretty substantially, so i'll guess we'll have 40k more dead in four weeks.
so my guess for the true number infected + infected and recovered in the us right now would be 50k divided by 0.015 = 3.33 m.
very wide error bars, obviously (and if i was to guess from other measures, i'd probably go under that), but not wide enough to get to 30 m or 100 m or any sort of number that would be implied for many of us to have already been infected.