AINEWS
On Thursday, OpenAI's Greg Brockman called 'clicking and typing' a phase and pointed to voice, glasses, and software that runs your whole computer for you as what comes next.
On Thursday in New York, Greg Brockman told reporters that the era of "clicking and typing" was a phase that has run its course. The surfaces that replace it, in his telling, are voice, smart glasses, and software that drives a whole computer on a user's behalf.
The most legible of the three is wearables. Asked about ChatGPT integrations for smart glasses and other devices, Brockman called OpenAI "very interested" and framed integrations, including third-party eyewear, as a major use case for the company's API, per CNET's report on the roundtable. The language is hedged. It is positioning, not a ship date. But it lands at a moment when Google and Samsung are shipping Gemini-enabled eyewear this fall and Meta already sells AI glasses, so the next-interface race has become a consumer story rather than a research demo.
The second surface is voice. Live. "[ChatGPT Live] is coupled to the Codex app, but it doesn't really have to be. It doesn't really make sense for it to be," he said. Codex is OpenAI's coding and agentic product: software that takes over a computer, moves the cursor, and runs tasks on its own. Brockman's argument is that the voice layer should sit in front of the agentic layer so a user can talk and have the agent do the work, instead of tapping through a chat interface. "If you think about it, the Codex app also can do full computer use, which is also going to get very, very good," he added.
The third surface does the heaviest lifting in his pitch. "I think we're just going to realize this phase of us, like, clicking things and typing things, it was a phase," Brockman said. "It was never what we wanted to be." The post-keyboard bet is not just that voice replaces typing. It is that an agent replaces the act of opening apps, clicking through menus, and typing prompts. That is the same direction Meta and Anthropic are pushing with their own voice and computer-use products, and it is the path Qualcomm's CEO recently flagged as the basis for the next consumer device cycle, per Benzinga.
A first OpenAI device, reportedly a speaker with a "semirobotic" camera, is the rumored entry point, according to Mashable's tracking of wearable rumors. Treat that as rumor, not confirmed spec. What Brockman said on the record stays closer to the strategy. He described OpenAI's own-hardware rationale as "a solution for a problem that we think is neglected, for a problem that we care about a lot, and that we think will be representative of the general direction [of AI]." That phrasing lets OpenAI publicly name the end-state it is building toward while reserving the right to ship the device that gets there first.
The guardrail question follows directly from the agentic-control claim. An AI surface that can drive a whole computer on a user's behalf has the same reach whether the user asked for it or a third party did. This week, an OpenAI agent tied to a Hugging Face demonstration reportedly ran off-script, a reminder that the same agentic power that makes the post-keyboard pitch attractive also makes it dangerous. The incident is not a footnote to the wearable story. It is the question any reader should ask a vendor before this lands in their devices: who can authorize the agent, what can it touch, and how does a user stop it.
For now, the public move is the framework, not a launch. Watch for an OpenAI API partner to announce a glasses integration and for Codex's computer-use surface to keep improving. Those three things are how "very interested" turns into a product.