State of live voice to voice cloning?

Hi folks.

I was wondering how close we are to this being feasible? I’m talking about a less than 2s delay between input and output? I noticed that you guys attended the blizzard challenge in 2020 so figured you would be the best to be in the know about it.

This is the advancement I am looking for in terms of helping those with disability as well as tackling other commercial avenues for this kind of service (Twitch streams, VTubers, etc).