🗣️ Say hello to a less-annoying AI voice — finally.
OpenAI just rolled out some major tweaks to its Advanced Voice Mode — you know, that real-time chatting feature in ChatGPT that’s supposed to feel like talking to a human but kinda doesn’t? Well, it’s getting a glow-up. Manuka Stratta, who works on post-training at OpenAI (aka one of the folks fine-tuning the personality behind that synthetic voice), dropped the news earlier this week via a social media clip. And yes, it’s all about making ChatGPT act a little more like a patient friend and a little less like a needy coworker.
The update zeroes in on a common gripe: those awkward moments when you pause for a breath or to think — and the AI just bulldozes over your sentence. That’ll happen way less now. 🧘♂️
Free-tier users will notice the voice assistant actually lets you speak uninterrupted. And if you’re one of the fancy folks subscribed to Plus, Pro, or any of OpenAI’s premium flavors (Teams, Edu, Business), you’re getting the deluxe treatment. Fewer interruptions + a more charming voice assistant that’s sharper, snappier, and just plain nicer to talk to.
A rep from OpenAI told TechCrunch that this more advanced voice mode delivers answers that are not just clearer and more direct, but also more imaginative — more “you’d think this was written by a real person” kind of energy.
Why now? Well, the pressure is on. Voice AI is the new battleground, and OpenAI’s got rivals breathing down its digital neck. Remember Sesame — the buzzy startup from Oculus cofounder Brendan Iribe? Its AI bots Maya and Miles are trending for sounding freakishly like humans. Amazon’s also ramping things up, teasing an Alexa revamp powered by large language models. Competition, clearly, is not sleeping.
But if voice is the future of human-AI connection, OpenAI’s making sure it doesn’t sound like a conference call from 2009.