I initially thought the same, but using even the current gen avatars for multiplayer on the Quest 2 feels great and is surprisingly convincing. The speech to mouth movement detection looks natural and combined with the body language being accurately represented with reverse kinematics I felt like I was colocating very realistically.