What kills me is the problems with audio and its effects on conversational engagement. Online calls are laggy, and participants aren't all on the same volume level. It's incredibly frustrating when I speak and someone says that my volume is low, so I should repeat myself. And the lag makes turn-taking very unnatural---in meatspace conversations, people interrupt each other and cede the floor all the time. You can't do that in a Zoom call.
Taken together, Zoom is a deeply frustrating experience because the limitations of the technology get in the way of having normal human conversations. It looks like it should work, but it doesn't.
I got used to platforms like TeamSpeak where everybody had individual volume sliders. It was excellent. Whatever Zoom does to automatically adjust everyone's volume is a poor replacement and it leads to exactly the issue you describe. Sure it's cheaper than sending out a bunch of individual audio streams or doing multiple transcodes server-side, but we've lost a lot of flexibility. I'd quite like P2P systems like Jitsi to bridge the gap.
Taken together, Zoom is a deeply frustrating experience because the limitations of the technology get in the way of having normal human conversations. It looks like it should work, but it doesn't.