This is a very provocative paper, but it's missing several key details. 1) We still have no idea how many virions are needed to lead to infection. Thus, even if talking generates 1000 droplet nuclei the infectious dose might be so high that the risk of transmission is small. https://twitter.com/carlzimmer/status/1260652736783069184

2) What are the effects of wearing a mask on droplet production? Could most of the droplets be stopped by any simple barrier mask?

This thread by @firefoxx66 makes case risk exists from droplets even at a distance, but I believe same data make it clear that this kind of spreading is still low probability. Attack rates would be much higher if a few droplets hanging in air were enough. https://twitter.com/firefoxx66/status/1260905937910587392