Latency is only one of the problems. Most of these services seem to handle audio in a way where they are picking one person as the speaker, and muting or ducking everyone else's audio. Practices like cooperative overlap are technologically inhibited.
Seems like a reasonable feature to more productive meetings in general? In-person meetings can degenerate to unproductive noise if people can't speak in turn. If you don't want the "talking stick" pattern there are other remote spaces where one can all meet online and have multiple conversations simultaneously.
