Every great technology gives humans a new superpower. The personal computer put computing power into people’s hands, turning machines once reserved for governments and universities into tools anyone could use. The web gave us access to the world’s information. The smartphone put that power in our pockets, connecting us to anyone from almost anywhere.
Each breakthrough changed what one person could do. But even with all this technology, we’re still the ones doing most of the work. We schedule meetings, manage our inboxes, juggle dozens of apps, consume endless information, and make thousands of tiny decisions every single day.
In platform shifts, tech feels like a tool before it becomes something bigger. AI is in that moment right 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 those transitions had something in common: taking a capability that used to be rare, expensive, or difficult 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 us have had another kind of superpower: a great assistant. It’s a superpower not because someone answers your emails or books your 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’ll 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’re capable of.
Most of us 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 removes friction, automates some tasks, and helps us go 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 might not have even imagined. But building an AI assistant that gives you these superpowers requires a lot more context than what we have 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 demand 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 tests, and stay on top of deadlines. An aging adult will depend on it for medication reminders and appointment coordination.
An AI assistant is not a chatbot. It’s not a single model or one agent completing a task. These are pieces of the solution. A great AI assistant is the end result of many technologies working together seamlessly to create one simple experience. Soon your AI assistant might not 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 when you are 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 learn your preferences, workflows, habits, relationships, routines, and past conversations over time.
The right interaction and persona build trust. Every assistant develops a communication style, tone, and patterns that feel natural to the person using it.
Expertise gives 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 of us still don’t know we need an AI assistant. That’s not unusual for a new platform. In the early days of personal computing, most people didn’t realize why they’d ever want 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. Once you learned on Apple, Apple became second nature and trusted.
ChatGPT reached 100 million users faster than any technology in web history. People are already living with AI, but we’re still in the early phases. The next shift moves this tool from occasional chatbot to a trusted assistant working constantly 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 is the critical shift.
The more useful an assistant becomes, the more you depend on it. The more you depend on it, the more memory accumulates and trust builds. Eventually it weaves itself into your life, helps you manage life, and even enhances life.
The iPhone had a similar path. It started as a phone you could use to check email. Then it became your wallet, your navigation system, your camera, your music player, your messaging platform, and finally something most of us can’t leave home without. We’re still in the email-checking phase of AI assistants. There’s a lot more to unlock.
The Fight 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, expertise, 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 instance. 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. Expertise allows it to act. The interaction builds trust. Reflection helps it connect patterns and anticipate what you’ll need next. The model is the brain that processes information and generates responses. But a brain alone is not enough. Without the rest of the system, every interaction starts from zero.
Microsoft, Google, Apple, Anthropic, OpenAI, Perplexity, and xAI are all competing to have 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 run different AI brains — in the cloud or on devices themselves — whichever are cheapest, fastest, or offer 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. Every 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 few questions a day. It’s constantly helping. And that requires constant compute. If every interaction must go to the cloud, the cost of running a truly personal AI assistant starts looking a lot like paying for a human assistant — which is out of reach for most users.
The only way this scales is to put more intelligence closer to you.
Think about your home security camera. You don’t need 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 shouldn’t be any different. On-device 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.
The future won’t be everything in the cloud or everything on device. The architecture will be distributed. That’s how you make the economics 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 rushing, 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 user 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 on 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 is 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 humans, is always available, is endlessly patient, is never judgmental — that’s a truly powerful tool. Perhaps 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 need to ask that question now with AI.
When we chose the iPhone over an Android device, or a Mac over a PC, we were picking a phone or a laptop. This time we’re picking something that will know how we think, work, and live. That has never happened before. Humans are at the center of this platform shift. Make it matter.
This article is part of CNET’s series of expert 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.
