The problem these kinds of tools are trying to solve is the one thing video conferencing can't do, which is the ability to communicate using the semantics of 3d space. In a meeting room, I can direct an audience to different parts of a room for different conceptual "spaces". For example, I might use one part of a room to collect and display engineering drawing, another for prototypes, another for a Kanban, etc.
After a solid year of literally thousands of teleconference calls, it's clear to me and my team at least that this spatial component is important in real-life meetings and is noticeable when it's not there. When we all come into the office and hit a meeting room together, the ad-hoc ability to use a whiteboard (often together and collaboratively), or sit together in groups around the conference table, it all has important value.
So far I've messed with Spatial.io and a couple other remote training applications, and while they're "ok", they're fussy to use in practice. Most importantly they don't really have a smooth way of integrating work artifacts from your computer into the experience, and the whiteboarding frankly stinks. Lots of comments here also talk about how this kind of meeting eliminates subtle facial cues and can be kind of uncomfortable for the wearer. I also totally agree with this.
But I'm also aware that this is a technology in its absolute infancy, and I'm expecting it to get better at some point.