Tony Fadell on why the first wave of AI gadgets failed — and what comes next
TechCrunch Amanda Silberling
Tony Fadell says the first AI gadgets failed because they solved no real problem. His bet: the winners will be private, on-device, and harder than people think.
Based on reporting by TechCrunch, Amanda Silberling — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
At MIT Future Fest, Tony Fadell did not bother with diplomacy. He put the Rabbit R1, Humane AI pin, and Limitless pendant up on a slide as examples of what he sees as AI’s first wave: splashy, interesting, and already gone. His blunt read was simple. The products chased novelty before usefulness, and that is a bad trade when you’re asking people to change how they live.
Fadell’s argument starts with the job these devices were supposed to do. They promised a personal assistant, but the pitch ran ahead of reality. Most people have never had a human assistant, he pointed out, so the supposed market is often really a market of people who already know that world. Everyone else is being asked to trust something they’ve never used for themselves.
And trust is the whole game. Fadell said it took him years to learn how to use an assistant well, then trust that person with sensitive data, and finally let them act on his behalf with things like scheduling and banking. That is a long road for a human. For an AI assistant, it’s even steeper, especially after Meta’s Muse launched with a reported serious vulnerability and internal scramble to fix security issues before release.
He thinks the winning version of this tech has to stay on the device itself. Not because cloud systems are useless, but because privacy and battery life matter, and because people won’t hand over a pile of permissions just to make a gadget feel clever. Apple, in his view, is unusually well placed here because it already has consumer trust around privacy and the hardware pieces to support that kind of product. But he also noted that Apple is behind on its own AI model work, with the new Siri AI running on custom-built versions of Google’s Gemini.
There’s a bigger strategic point under all of this. Fadell thinks companies like Meta and OpenAI are chasing hardware because Apple already sits on the phone sensors, the permissions, and the installed base. If they want similar access, they need a new box that can gather video, audio, GPS, and more without feeling like a privacy tax. That’s a tough sell. And as he put it, startups only get one shot; this is not like Apple and the Vision Pro.
My take — AI-written commentary, not fact-checked reporting
The market keeps pretending “assistant” is a product category instead of a trust exercise, which is why so many of these gadgets end up as expensive proof-of-concept props. Fadell is right to push the boring part: privacy, permission, and usefulness. The AI hardware race will keep rewarding companies that respect reality, not the ones that think a pendant and a press launch can do the work.
Read more about this at: TechCrunch