Today I had a 20-minute conversation with ChatGPT while driving. I didn’t plan for it to go that long, but somewhere in the middle I realized I was looking at a feature that could change how I work.
The voice mode itself has gotten much better. ChatGPT now runs on GPT-Live-1, and if you’ve used the older voice modes, you’ll notice the difference right away. It sounds more natural, the pauses are shorter, and the conversation actually flows. It feels a little strange to say that about software, but it was smooth.
The real surprise was that voice mode can use the apps I’ve already connected to ChatGPT. While I drove, it pulled up my tasks in Capacities, and I added and updated items just by talking. It also went through my calendar and saved notes from the conversation so I could look at them later.
The one real limitation: when it’s calling a tool, it goes completely silent. In a car, with road noise, I couldn’t tell whether it was working or had dropped. A small sound or a quick “one sec, checking” would fix that.
With tools like Dots, Muse and GrokBot out there, I know this kind of workflow is going to become normal. But a voice that’s this good, combined with access to my own data, felt like a real step up. If you’ve already connected your apps to ChatGPT, try it on your next drive. Start with your to-do list and see how far you get before you reach for your phone.