Photo-Illustration: Intelligencer; Photo: OpenAI
Last year, in a what’s-that-guy-been-up-to profile in the New York Times, legendary former Apple designer Jony Ive explained how he’d been filling his time. He’d founded a small design firm, LoveFrom, which was working on projects including a “Christie’s auction stand, Airbnb graphics and a Ferrari interior,” and expanded his real-estate holdings in his favorite neighborhood in San Francisco. He’d attempted to reinvent the jacket button for Moncler. He also revealed plans for a new “artificial intelligence device company that he was developing with OpenAI.”
This week, they made things official:
Sam & Jony introduce io pic.twitter.com/ej5K59kJq3
— OpenAI (@OpenAI) May 21, 2025
There isn’t a whole lot of information contained in this strange video, in which Altman and Ive profess to be friends, and demonstrate their friendship by describing each other in competitively superlative terms (“@grok summarise this, aint no body watching 10mins,” writes the top commenter on X). We learn that both Ive and Altman have distinct but equally cartoonish gaits. We find out from Ive that Altman is a “rare visionary.” We hear from Altman that Ive is “the deepest thinker over anyone [he’s] ever met.” We’re told by Altman that Ive’s design firm is the “densest collection of talent that I’ve ever heard of in one place and probably has ever existed in the world.” We learn that Ive believes his early AI projects are “the best work our team has ever done,” and from Altman that the prototype of Ive’s first device is “the coolest piece of technology that the world will have ever seen.”
The accompanying announcement, which is fittingly templated and illustrated like a wedding website, doesn’t give us much more, though it does hint at some of the procedural strangeness of the arrangement: “As io merges with OpenAI, Jony and LoveFrom will assume deep design and creative responsibilities across OpenAI and io,” the company says, glossing over an arrangement that, despite a $6.5 billion deal made up of OpenAI equity, leaves Ive and LoveFrom independent, even as they “take over design for all of OpenAI, including its software.” The intended message, however, is clearer: The AI guy is teaming up with the iPhone guy, and they’re trying to make the iPhone of AI.
What would that mean? Unclear. Start-ups and tech giants have been working on AI hardware for a while now, and the ideal form for an AI-centric device has yet to reveal itself. Dedicated AI gadgets like the Humane Pin and Rabbit R1 were dramatic failures, leaning too hard on AI models that too often failed to complete users’ requests or that testers simply didn’t find compelling. (Ive has defended against the comparison, calling them “poor products.” ) A few mishaps scattered among many acceptable interactions with ChatGPT are no big deal when you’re using a computer, in part because your expectations are lower and you have lots of available alternatives; a single failure with a voice-activated, screenless pin device purporting to see and hear as you do, and to act as a personal assistant, is a showstopper. Meta’s Ray-Ban glasses, which can execute some voice commands, identify images via the integrated camera, answer questions via chatbot, and handle some translation, make far more modest promises but actually work and have sold modestly well.
For now, the iPhone of AI is still the iPhone, a device with cameras and sensors and apps and an internet connection. Chances are that whatever OpenAI comes up with here won’t be fundamentally alien by the time it arrives. Customers are already surrounded by products that rely on technology that at one point was understood to be cutting edge “AI”: Smart speakers and screens like those from Amazon, Google, and Apple; earpods and headphones with integrated voice commands, like Airpods; smartwatches and fitness trackers that ambiently collect and analyze all sorts of data. The closest thing these announcements offer to a product description appears in the negative, as Altman outlines what he sees as the problem with AI interaction today:
If I wanted to ask Chat GPT something right now about something we had talked about earlier, Think about what would happen. I would, like, reach down. I would get on my laptop. I’d open it up. I’d launch a web browser.
I’d start typing. I’d have to, like, explain that thing. I would hit enter, and I would wait, and I would get a response. And that is at the limit of what the current tool of a laptop can do. But I think this technology deserves something much better.
The two things that stand out here are concern for form-factor — the clunky laptop – and ChatGPT’s need to be manually supplied with context. So, not a laptop, not and phone, not glasses, but probably listening all the time. A Humane Pin, but good? A phone-ish thing without a big screen? The WSJ reports that it could be positioned as a “third core device” that’s “capable of being fully aware of a user’s surroundings and life,” and “will be unobtrusive,” designed to “rest in one’s pocket or on one’s desk.” This sounds less like a gadget than a surveillance device, aligned with OpenAI’s recent pivot toward “memory” and customization, envisioning a world in which comprehensive interactions with AI are the norm, privacy as once understood is quaint, and ecosystem lock-in is total.
For Ive, this project is all but transparently about recapturing the glory of designing some of the most iconic objects in modern technological history, helping refine a series of genuinely transformative tools. “I have a growing sense that everything I’ve learned over the last thirty years has led me to this place and to this moment,” he says. “I am absolutely certain that we are literally on the brink of a new generation of technology that can make us our better selves.” He’s found a new project that also feels like a cause.
But there’s something similar going on for Altman, too. OpenAI’s ability to offer billions of dollars of equity to hire the iPhone guy follows from the success of ChatGPT, which is arguably still the only organic mega-scale product to come out of the current AI boom. Launched as a demo with low internal expectations, OpenAI’s general-purpose chatbot, deployed in web browsers and smartphone apps, ended up charting the course for the industry for the foreseeable future, incidentally proving the strange and complicated potency of the chatbot form. Better than anyone else, though, Altman is likely aware of its limits — as versatile as chatting is, it’s not an ideal interface for all computing, and like many other tech companies over the years OpenAI is surely anxious that its success depends on hardware and software contexts controlled by competitors — and with hundreds of billions of dollars in valuation to back up, he’s hoping he can repeat the trick: This time intentionally, and more durable form.
As Casey Newton writes, AI hardware is a “huge bet and a risky proposition.” ChatGPT was a runaway success, putting other companies on their back feet, leaving them in the difficult position of needing to convince users and customers to adopt features and tools they didn’t seek out or ask for themselves. (Would you like a chatbot in your Instagram? How about your search engine? What about your spreadsheets?)
Branching into AI hardware asks something new of customers. At launch, io’s mystery device will be held to the punishing standards of a new Apple product. Altman’s “coolest piece of technology” rhetoric suggests commensurate ambitions — Altman has told his staff that the device strategy could add a trillion dollars to the company’s valuation — but also high stakes. AI software tools like ChatGPT are versatile but slippery and hard to conceptualize; users like them and get utility from them, but are also fairly generous in their assessments of what they can and can’t do, in part because most of them aren’t paying to use them. Hardware is expensive, and customers are harsh. It’s difficult to know what people will buy, experimentation is slow and unforgiving, and, most importantly, there’s no hiding from a flop. But for OpenAI, which is one of the most valuable startups in the world, but which has massive operating costs and comparatively low (but rising) revenue, “huge” bets might be the only ones left.
