Update: It misses every 4--6 first words on a 64GB Mac 3 Ultra Max.
You have speak, wait, then it'll catch you. But it's not worth it. it's not accurate. there's little context. The setup is weird and you can't select a different mic. It's not polished. It's a beta, likely using open source transcription models so nothing unique. Willow voice is a 10 if this is a 1 honestly. Maybe it works on Windows. IDK. Mac. Steer clear. I have 3 macs, m2,m3.m4 = no luck. reducing review to a single star given the supposed upgrades were amazing. My baseball bat to the keyboard dreams are long off with this.

Original review:
It's a pretty good tool, but definitely slow. I'm writing this with Willow Voice, which is part of another deal here on my Mac, and it's significantly faster. Willow Voice has no delays, inserts text quickly, and it's actually usable. The importance of voice transcription and one's productivity can't be overstated. Even paying a monthly fee for the speed advantages of something like Willow Voice is easily worth it, versus VoiceDash at this point. Now I see the team responds and is working hard, so I'm just going to have to keep it at two stars and keep my tier two subscription to see if they improve it. But for now, on my powerful Mac, it's not usable. And when I say "not usable," it works. But the point of these tools is to work fast, not wait around.

[edit in response to your very kind reply to my original note. I typically only talk for about 10 to 20 seconds, maybe 30. I never really dictate anything long whatsoever, because I'm using it for coding and fast interaction with my computer. I look forward to seeing if you can really kick this into gear. And while I appreciate that AI is involved, somehow your competitors have figured out how to provide smart text editing on the fly as well, with zero speed delays. So there must be some OS layer thing that you can do to make it better. ]