OpenAI has updated Advanced Voice Mode in ChatGPT to make conversations smoother and less interruptive. I use the voice mode quite regularly and can confirm that you had better not take a second to breathe, lest it thinks you’re done and starts responding.
Sometimes, it just starts responding mid-sentence, like humans do. For some reason, when it’s AI doing this, the levels of annoyance are off the charts.
Anyway, OpenAI says both free and paying users now experience fewer interruptions when pausing while speaking. Paying users also benefit from a more engaging, concise, and creative voice assistant.
That should help ChatGPT’s voice mode maintain its lead over Gemini Live, in my opinion. Gemini is unusable for me because of one major problem—it sometimes indicates that it’s listening when it is not.
You wouldn’t believe how many times I have waited for the chime to indicate it’s listening, then yapped for two minutes straight, only to be hit by a, “How can I help you today?”
To avoid that, you have to first greet it or say something to make sure it’s listening. However, even then, sometimes it just decides it has heard enough and starts responding, leaving you to yap into the void.
This has been my experience with Gemini Live on iOS, and I hope the experience is better on Google’s home turf—Android.
Anyway, the ChatGPT update comes as competition heats up in the AI voice assistant space, with emerging rivals like Sesame’s Maya and Miles and Amazon’s LLM-powered Alexa improving their natural-sounding AI. Not too sure if people are going to use them, though—for now, it appears to be a ChatGPT and Gemini two-headed monster.
Leave a Reply