Tony Fadell predicts AI assistant winner and warns of unintended consequences ahead

Who'll Own Your Inevitable AI Assistant? The Battle Is On, and I Predict One Winner

Tony Fadell is an active investor and entrepreneur with a 30-plus year history of founding companies and designing products that profoundly improve people's lives. He is the principal at Build Collective, an investment and advisory firm coaching deep tech startups. Currently, Build Collective is coaching over 200 startups innovating game-changing technologies. Tony is the founder and former CEO of Nest, the company that pioneered the internet of things. He was the senior vice president of Apple's iPod Division and led the team that created the first 18 generations of the iPod and the first three generations of the iPhone. Throughout his career, Tony has authored more than 300 patents. In May 2016, Time named the Nest Learning Thermostat, the iPod and the iPhone as three of the "50 Most Influential Gadgets of All Time." He is the New York Times best-selling author of Build: An Unorthodox Guide to Making Things Worth Making (Harper Business; May 3, 2022). In October 2025, he was named the inaugural designer in residence at the MIT Morningside Academy for Design. Welcome to CNET's new series of guest columns called Alt View, a forum for a diverse array of experts and luminaries to share their insights into the rapidly evolving field of artificial intelligence. For more AI coverage, check out CNET's AI Atlas. Every great technology gives humans a new superpower. The personal computer put the power of computing into the hands of individuals, turning machines once reserved for governments, universities and corporations into tools anyone could use. The internet gave us access to the world's information. The smartphone put that power in our pockets, connecting us to anyone and anything from almost anywhere. Each breakthrough changed what one person could do. But even with all this technology, we're still the ones doing a lot of the work. We schedule meetings, organize our inboxes, manage dozens of apps, absorb endless amounts of information and make thousands of small decisions every day. In platform shifts, tech feels like a tool before it becomes something bigger. AI is in that moment now. We mostly experience AI as something we ask questions of or give tasks to. But the breakthrough comes when AI understands, not just responds. I've spent my career experiencing and even causing platform shifts. The Mac. The iPod. The iPhone. Nest. Every one of these transitions had something in common: taking a capability that used to be rare, difficult or expensive and making it more accessible. The builders who understood the behavioral shift, not just the technology shift, were the ones who built something that lasted. AI assistants are the next big shift. A handful of people have had another kind of superpower: a great assistant. It's a superpower not because someone answers their emails or books their flights - those are helpful tasks. The real value of an assistant is that they understand the person behind the tasks: who you are, your relationships, your routines, your family, what you prioritize and what keeps you up at night. They know what to handle, what to bundle together and what truly needs your attention. A great assistant can anticipate what you need before you even ask. But that relationship doesn't happen overnight. It takes years of working together, learning patterns, building trust and accumulating context. Once you have that superpower, you never want to lose it. Losing a great assistant means losing years of shared knowledge and understanding. That accumulated understanding is what transforms an assistant from someone who helps you complete tasks into someone who expands what you can do. Most people have never had an assistant or seen how a great one actually works, which means we are learning two things at once: what an assistant can do and what AI can become. AI is beginning to act as a painkiller. It can remove friction, automate some tasks and help us move faster. But we still manage AI more than AI manages us. We prompt, correct and guide. That will change. A painkiller removes the pain you already feel. A superpower goes beyond removing pain. It unlocks new capabilities you couldn't even have imagined. However, building an AI assistant that gives you these superpowers requires a lot more context than what we've got right now. These are the moments in technology that matter. When a new capability forces us to rethink what's possible and what we should expect from the products in our lives. Building an AI assistant A freelancer juggling three clients, six deadlines and a stack of unpaid invoices will soon have an AI assistant that keeps track of it all, so they can focus on the work instead of the overhead. A parent will use it to manage the family calendar, school reminders, grocery lists and doctor's appointments. A student will have it organize coursework, prepare for exams and stay on top of deadlines. An aging adult will rely on it for medication reminders and appointment coordination. An AI assistant isn't a chatbot. It isn't a single model or one agent completing a task. Those are pieces of the solution. A great AI assistant isn't one thing. Like every great product, an AI assistant is the result of many technologies working together seamlessly to create one simple experience. Soon your AI assistant won't just tell you your flight is delayed. It will understand why that delay matters, rebook it, notify the people you're meeting and update your calendar. Context gives the assistant awareness. The ability to know if you're at work, at home, traveling, in a meeting, exercising or spending time with family. The richer the context, the more useful and personalized your assistant becomes. Memory allows the assistant to learnyour preferences, workflows, habits, relationships, routines and prior conversations over time. The right interaction and personality build trust. Every assistant develops a communication style, tone and patterns that feel natural to the person using it. Skills give assistants the ability to get things done: research, coding, scheduling, financial analysis, design tools, travel coordination, health tracking or interacting with other specialized agents. And finally, reflection, the ability for an assistant to synthesize information across your workflows, interests, relationships and routines to identify patterns and surface useful insights you wouldn't have found on your own. Together they're what turn AI from a tool that responds into an AI assistant that understands you. The AI assistant adoption curve Most people still don't know they need an AI assistant. That's not unusual for a new platform. In the early days of personal computing, most people didn't understand why they'd ever need a Mac or a PC. Computers felt niche, technical, and disconnected from real life. Apple understood that behaviors had to change before the platform could scale, so it got Macs into classrooms, invested in student pricing and made the product familiar before it was necessary. If you learned on Apple, Apple became second nature and trusted. ChatGPT reached 100 million users faster than any consumer technology in internet history. People are already living with AI, but we're still in the early stages. The next shift moves this tool from occasional chatbot to a trusted assistant working persistently by your side. The first experiences didn't feel like AI at all. They felt like smarter search. Now, we're having it do simple things automatically: sports scores, weather alerts and travel itineraries. Next, the delegation increases. That's the important shift. The more useful an assistant becomes, the more you rely on it. The more you rely on it, the more memory accumulates and trust builds. Eventually it weaves itself into your life, helps you manage life and even enjoy life. The iPhone had a similar path. It started as a phone that you could use to check email. Then it became your wallet, your navigation system, your camera, your music player, your messaging platform and eventually something most people cannot leave home without. We're still in the email-checking phase of AI assistants. There's a lot more to unlock. The battle for the AI assistant Platform wars have been fought over apps, ecosystems, operating systems, developer relationships and network effects. This one is different. The model is not the assistant. The assistant is the platform. We've seen this pattern before. The iPhone brought together hardware, software, services and an ecosystem that changed how people interacted with tech. The same will be true for AI. The AI platform war won't be won by the model alone. Whoever builds the complete assistant experience with context, memory, interaction, skills and reflection all working together will win this war. You tell your AI assistant you need to go to New York for a meeting Wednesday and be home by Friday evening, for example. Without context it doesn't know where to book the flight from. Without memory it doesn't know you always choose an aisle seat, prefer the same Midtown hotel and need to be back home for school pickup. Context tells the model what's happening right now. Memory tells it who you are. Skills allow it to act. The interaction builds trust. Reflection helps it connect patterns and anticipate what you need next. The model is the brain that processes information and generates responses. But a brain alone isn't enough. Without the rest of the system, every interaction starts from zero. Microsoft, Google, Apple, Anthropic, OpenAI, Perplexity and xAI are all competing to own this very personal interaction with you. Right now, for most of them, the model is their moat. But models are already commoditizing. Our personal hardware devices will plug into different AI "brains" - in the cloud or on devices themselves - whichever are cheapest, fastest or offer up more privacy. The real moat will be the complete assistant built around you over time. No single cloud service can see the full picture of your life. But a connected ecosystem of devices can. Your phone knows your location, your watch knows your heart rate, your laptop knows your calendar, your glasses know who you're talking to. Each signal is useful alone. Combined, they create a much more powerful portrait the assistant can use to understand you. This is what I call the Federation of Devices. Whoever controls this Federation of Devices has the foundation to build the most trusted and most valuable AI assistant. The cloud doesn't scale A real AI assistant isn't something you ask just a couple questions a day. It's always helping. And that requires constant compute. If every interaction must travel to the cloud, the cost of running a truly personal AI assistant starts looking a lot more like paying for a human assistant, which is out of reach for most consumers. The only way this scales is to move more intelligence closer to you. Think about your home security camera. You don't want every frame of video sent to a remote server just to detect motion. Processing happens on the device because the economics, speed and privacy demand it. Smart glasses that translate a conversation in real time can't wait for a round trip to a data center. You'd be halfway through the sentence before the answer came back. Your AI assistant is no different. Local AI handles the routine. Cloud models handle the complex reasoning. Companies like Plumerai are building exactly for this moment with tiny AI models small enough to run on a chip the size of a thumbnail and powerful enough to handle the tasks you need most. (Editors' note: Tony Fadell is an investor in Plumerai.) The future won't be everything in the cloud or everything on device. The architecture will be distributed. That's how you get the economics to work. The cloud doesn't know you're cold The company that controls the federation of devices wins. With 2.5 billion active devices, that's Apple's hand. Google has incredible server-side infrastructure: Gemini, custom TPUs, data centers, Search, Gmail, Maps, Android and YouTube. Nobody is closer to owning the model and cloud layer. But the cloud doesn't know you're cold. It doesn't know that you're driving, your meeting just ended or who you're talking to. Your devices do. Apple has the iPhone, Mac, Apple Watch, AirPods, the identity layer, edge computing and one of the deepest consumer trust relationships in the history of technology. M-series chips power your Mac and iPad. A-series chips power your iPhone and Apple Watch. Both share the same unified architecture, CPU, GPU and memory on a single chip, built for AI at the edge. Apple's Federation of Devices will run AI on device and coordinate with more powerful models in the cloud when needed. The Apple and Gemini partnership is a signal of where this is going. The model matters less than who controls the assistant built around you, that's the lasting advantage. You can swap the brain. You can't swap the memory, context and trust built over time. Questions nobody has answered yet We're building this plane while flying it. I've spent 30-plus years building products that changed how people live. I didn't always anticipate the questions they'd raise. This time I'm asking them. If your AI assistant understands your communication style, workflows, negotiation patterns and institutional knowledge, who owns that context when you leave a company? If it becomes deeply integrated into healthcare, what happens when providers or insurance systems change? How portable should memory and personalization be across systems? And then there's the question nobody in the industry wants to ask. An assistant that knows you better than most people, is always available, is endlessly patient, is never judgmental... that's a very powerful tool. Possibly too powerful and addictive in ways we haven't fully reckoned with. We built the iPhone without asking what it would do to human connection. We should ask that question now with AI. When we chose the iPhone over an Android device, or a Mac over a PC, we were choosing a phone or a computer. This time we're choosing something that will know how we think, work and live. That's never happened before. Humans are at the center of this platform shift. Build it like it matters.

Our choice of AI assistant really matters, says Tony Fadell

"Father of the iPod" Tony Fadell has written a lengthy column in which he argues that our choice of AI assistant really matters, but also raises some huge questions that need to be addressed. He says the reason the iPhone has stood the test of time is because Apple understood the behavioral shifts it would create, not just the technology shift, and the same will be true of AI assistants ... Fadell points to the Mac, the iPod, the iPhone (and, rather less convincingly, Nest) as being about more than just making technology more accessible. Every one of these transitions had something in common: taking a capability that used to be rare, difficult or expensive and making it more accessible. The builders who understood the behavioral shift, not just the technology shift, were the ones who built something that lasted. He effectively argues that of all of the companies competing in this field, Apple is one of the best placed. No single cloud service can see the full picture of your life. But a connected ecosystem of devices can. Your phone knows your location, your watch knows your heart rate, your laptop knows your calendar, your glasses know who you're talking to. Each signal is useful alone. Combined, they create a much more powerful portrait the assistant can use to understand you. This is what I call the Federation of Devices. Whoever controls this Federation of Devices has the foundation to build the most trusted and most valuable AI assistant. He also says that we need to address some difficult questions - one of which is whether what we're building might be too powerful. There's the question nobody in the industry wants to ask. An assistant that knows you better than most people, is always available, is endlessly patient, is never judgmental... that's a very powerful tool. Possibly too powerful and addictive in ways we haven't fully reckoned with. He likens it to the iPhone, which transformed our lives for the better in a great many ways, but also contributed to the problem of smartphone addiction and social isolation. He says this time we need to think about this issue from the outset. This time we're choosing something that will know how we think, work and live. That's never happened before. Humans are at the center of this platform shift. Build it like it matters. It's well worth reading the full piece over at CNET. What are your own thoughts on the topic? Please share in the comments.