What do you think is lacking in OpenAI voice?
I wish they didn't take out the "off the screen" outputs from the previous voice mode. I miss being able to ask it to output an image or codeblock when I wanted to come back to something later.
Just the latency and naturalness of the conversation. Then again, I haven’t tried it for a month, so perhaps it is time to try again.
The random shrieks and noises that come up are something else. Those are wild when you encounter them.