Some recent data points of interest: First, German researchers studied a hard hit town. High level takeaway: 14 percent of the town had antibodies, 2 percent had recent infection, .37 IFR estimate (mortality for all infected, not just confirmed cases. https://rp-online.de/nrw/staedte/kreis-heinsberg/kreis-heinsberg-bei-15-prozent-in-gangelt-coronavirus-infektion-nachgewiesen_aid-49993141
I haven't found a good English translation of the paper. Assuming it holds, would be both bad (lower community transition even in a hard hit area than we might hope) and good (lower mortality than worst estimates, though still far worse than seasonal flu).
Austrian study finds .32 percent of one sample with active infection. This suggests against broad exposure, but given that it's only testing active cases (not those who had it in the past), there are other plausible--however unlikely--interpretations. https://www.bloomberg.com/news/articles/2020-04-10/austrian-study-shows-coronavirus-cases-more-than-3-times-higher
I've yet to see a solid piece of evidence that the US or Europe has seen broad exposure. But still very early.
A good thread on the German study, by the indispensable @phl43 https://twitter.com/phl43/status/1248743065784741888