"I mean, partly. We know that sticking to outdoors, wearing masks, etc. helps a lot. So we should try to avoid situations such as offices with central AC, indoo

"I mean, partly. We know that sticking to outdoors, wearing masks, etc. helps a lot. So we should try to avoid situations such as offices with central AC, indoor restaurants, indoor gyms, etc." We don't "know" any of this, except in the sense that it gets repeated a lot, uncritically. The actual published data is mainly anecdotal and/or ambiguous. Consider restaurants -- South Korea still doesn't consider them to be high-risk facilities, and they're largely open: https://en.yna.co.kr/view/AEN20200819005400320 https://en.yna.co.kr/view/AEN20200819005400320 That case, in particular, is enlightening about the quality of the data: they blame the transmission on aerosols, but can't rule out droplet or surface contamination, either. It's a judgment call. Most of these stories are like this. The famous "Chinese restaurant" case-study that is cited all over the place to justify shutting down restaurants presents a grand total of nine infected (even though 73 "close contacts" in the restaurant were not infected), of which three may have possibly been infected at home: https://wwwnc.cdc.gov/eid/article/26/7/20-0764_article https://wwwnc.cdc.gov/eid/article/26/7/20-0764_article Then you have studies like this, which suggest a fairly low attack rate (3.5%), even if you're sitting right next to an infected passenger on a train, for hours: https://academic.oup.com/cid/advance-article/doi/10.1093/cid/ciaa1057/5877944 https://academic.oup.com/cid/advance-article/doi/10.1093/cid... The authors characterize this risk as "high"...but is it, really? I mean, clearly, being in a sealed tube next to someone for hours increases the risk, but how far are we willing to go for a risk that can be characterized in fractions of a percent for face-to-face interactions that aren't measured in hours? I feel like people don't know these numbers, and most people would be shocked if they did.