At the inaugural MIT Future Fest, Tony Fadell projected images of three discontinued AI devices—the Rabbit R1, the Humane Ai Pin, and the Limitless pendant—and said their makers had asked him for help. He declined. In his view, these "Gen 1" AI products were interesting technology for geeks but did not solve any real pain in people's lives.
Fadell, credited as the father of the iPod and a co-creator of the iPhone, argues that the personal-assistant pitch is fundamentally flawed. Less than 0.01% of the world's population has ever had a human assistant, he notes, so most consumers don't know what an assistant is or how to trust one. He adds that even with a human, building that trust takes years and happens gradually, not on day one.
Security is the other major hurdle. Fadell says trust and safety will be paramount for anything handed over to an AI, and he points to recent problems with Meta's Muse assistant, including a quickly found vulnerability and reports of an internal rush to fix issues before launch. He believes Apple is one of the few companies that could pull this off, because it has the hardware and user goodwill, but it still lacks a top-tier AI model of its own.
Fadell predicts that successful AI agents will have to operate on-device, both for privacy and to keep the technology lightweight. He dismisses the idea that data centers will dominate everything, noting that devices already have enormous computing power. For startups, though, the margin for error is thin: "Because it's a startup, you get one shot," he said, contrasting that with Apple's ability to absorb a product like the Vision Pro.