Apple’s Rumored AI Pin Forces a Simple Question: What Do People Actually Want?

Earlier this week, my co-host and I had a conversation on our podcast, Primary Technology, about the rumors that Apple is working on an AI pin. We don’t usually spend a lot of time on rumors—for a number of reasons I won’t get into here—but this one is particularly interesting considering that we’ve seen this before (AI pins, I mean), and they haven’t turned out especially well. According to a report from The Information this week, Apple is actively developing a wearable device—roughly the size of an AirTag—equipped with cameras and microphones but notably, lacking a display. The idea is that it would launch as early as 2027, powered by the kind of multimodal intelligence we expect to see in iOS 27. Since we recorded that episode, I’ve been thinking a lot about whether this makes any sense for Apple, and what exactly the ideal device is for AI. I’ve said in the past that I think that’s the Apple Watch, though a “pin” definitely has certain advantages (it has an outward-facing camera, for example). More importantly, I’ve been thinking about what the ideal AI device is based on what people actually want. The ideal form factor should be determined by the ideal use cases, not the other way around. I think the bar is pretty high. If I’m going to carry another device, or if I’m going to replace something like my iPhone, it has to be able to offer value I can’t get from what I already have. Generally, that falls into three buckets: Answers to questions The most obvious use case is what people already use AI tools for—getting information. Of course, this is admittedly a pretty broad category. There are a lot of different types of information that people might want. For example, they might want to ask simple questions like “who directed Star Wars: The Last Jedi?” But people also want to ask slightly more complicated questions (what’s the weather going to be like when my plane lands tomorrow?), as well as queries like “what’s this plant, and is it edible?” Those questions require a different kind of contextual awareness. Your AI assistant has to be able to see your calendar, find out flight information, and check the weather at your destination. Or, in the latter case, it has to be able to literally see what you’re talking about, identify the plant, and give you the information you’re looking for. This is where the form factor of a pin actually starts to make some sense. The primary limitation of Siri on your iPhone or Apple Watch is that it can’t see what you see. Sure, you can hold up your phone and point the camera at stuff, but that’s awkward. If Apple’s rumored device includes the dual-camera array mentioned in the reports, it changes Siri from being just a voice assistant to a multimodal source of information about the world. You aren’t just asking for information; you are asking for context about the physical world in front of you. Do things on their behalf Of course, getting information is great, but acting on it is even more useful. This is the “agent” concept we’ve heard so much about but haven’t really seen work in practice. It’s the promise that the Rabbit R1 made but couldn’t keep: the ability to interface with apps and services to actually get things done. The Rabbit R1 failed because it tried to simulate your interactions via a cloud-based “Large Action Model” that was clunky and unreliable. Apple has the potential to solve this for first-party apps like Calendar and Messages. It controls the entire software stack, meaning it can offer an experience that other devices couldn’t. And, with App Intents, Apple could solve the same problem for other apps if it could get third-party developers on board. I don’t just want to know that my flight is delayed; I want the device to rebook me on the next one and update my calendar. We’re a long way off from any device being able to do that, but it’s the promise that every company keeps making. If Apple can make it happen, it’ll immediately jump to the lead. Remember and prompt This is the “external brain” use case, and frankly, it’s the one a lot of people find most compelling. We all have those moments where we meet someone and can’t quite place them, or we have a brilliant idea while driving and lose it by the time we get home. An ideal AI device should be a passive observer that helps you connect the dots. It should be able to whisper in your ear, “That’s David; you met him at CES last year,” or remind you to pick up milk because it knows you’re near the grocery store. Of course, this is also the creepiest use case. It requires a level of always-on surveillance that most people are rightfully uncomfortable with. If Apple is going to ask us to wear a camera and microphone on our chests, they are going to have to lean incredibly hard on their privacy credentials. Trust is the only currency that matters here. The big risk Previous devices haven’t been much of a success. No one has figured this out yet. The Humane AI Pin was a disaster of overheating and poor battery life. The Rabbit R1 was barely functional. The history of wearable AI is short, but it is brutal. There are laws of physics that even Apple cannot ignore. Cameras and AI models generate heat and drain power. Putting that in a coin-sized aluminum disc without a massive battery pack is an engineering feat no one has cracked. There’s also the fact that wearable devices come with a very real stigma. Anything that isn’t a watch has to be exponentially more useful than the burden of wearing it. Google Glass failed partly because people simply didn’t want to talk to someone who had a camera pointed at their face. Meta has circumvented this slightly with Ray-Bans because they look like sunglasses. A shiny badge on your chest is a much bolder statement. Is that an argument for or against Apple trying? I’m not sure. But with reports that Jony Ive and OpenAI are building their own hardware, Apple may feel it cannot afford to cede the category. Even if, right now, it looks like a solution in search of a problem. EXPERT OPINION BY JASON ATEN, TECH COLUMNIST @JASONATEN