On July 24, OpenAI unveiled an upgraded version of the ChatGPT desktop app, which now incorporates Codex technology and introduces a cutting-edge ChatGPT Voice mode, powered by the GPT-Live model. This latest iteration allows users to initiate, review, and refine a variety of tasks—ranging from casual chatting to office work and programming—through intuitive voice commands, fostering seamless multitasking collaboration. As a full-duplex voice model, GPT-Live is capable of listening and speaking simultaneously, thereby facilitating conversations that closely mimic genuine human interaction. Users can effortlessly kickstart multiple workflows within a single conversation and maintain dialogue even as the AI agent diligently carries out tasks in the background. For instance, users can verbally direct ChatGPT to concurrently check for calendar clashes, sift through flight change details, and compile meeting minutes.
