I wouldn't say that it's excellent now. It still has a long way to go. Small example, try enabling captions in Google's meet. Those are awful. Don't get a lot of words right, if you have an accent it's even worse. And distinguishing between different people talking? Forget about it...
You can use a steno mask so your voice doesn't disturbs others but also so that the voices from coworkers and other background noise doesn't disturb the speech-to-text software.
It would look funky though, and I've read from people having problems with their voice after an 8 hour day speaking software.