The AI Revolution Already Happened

If you ask people whether artificial intelligence is going to change the world, you'll get a wide range of answers. Some are convinced it will replace most jobs within a few years. Others think it's mostly hype. Many people seem to believe it's something that might matter someday, but not today.

I think they're all looking in the wrong direction.

If you work in tech, you know that the disruption is already here. The revolution didn't arrive all at once. There wasn't a singular moment where the world changed overnight. For everyday folks, it has been arriving in small increments, hidden inside products and services we already use. And the reason most people haven't noticed is the same reason they didn't notice the smartphone revolution until it was already complete — gradual change is almost impossible to see while you're living through it.

The smartphone didn't instantly transform society the day the first iPhone was released. At first it was a neat gadget. Then it became a great camera, a GPS, a boarding pass, a bank branch, a ride-hailing service, the way we ordered food, applied for jobs, communicated with friends, and navigated the world. Looking back, it feels obvious. Living through it felt gradual. Nobody announced the moment the smartphone became indispensable. It just did, incrementally, until one day the world it had replaced was distant in the rear-view mirror.

I think AI is following the same path.

A few years ago, most people started interacted with AI through chatbots that made obvious mistakes — often wrong, sometimes hilariously so. Many people tried them once and walked away with the impression that they were little more than toys. That impression is already outdated. The AI systems available today are dramatically better than the ones that existed a year ago, and those were dramatically better than the ones from the year before. The pace of improvement is unlike anything I've seen in my time working with technology.

I use AI almost every day because it's useful. It helps me edit and proofread articles and brainstorm ideas when I'm stuck. It helps me understand documentation and automate repetitive work. When I built the custom theme for this website, AI wrote, tested, debugged, fixed, validated and deployed all of the code. When I wanted a new feature, I put the developer documentation into an AI tool and simply asked it to implement what I needed. To my surprise, it did — in one shot. I still reviewed the code to make sure it was ok. But the amount of work accomplished in a few hours would have taken me weeks or months on my own.

That's the part people miss. The disruption isn't always about replacing people. Sometimes it's about allowing one person to accomplish what previously required several. Sometimes it's eliminating the tedious parts of a job, or lowering the barrier to entry for people who couldn't previously afford the expertise. And yes, sometimes it is replacing tasks that humans used to perform. All of those things are disruptive, and most of them are complicated — neither straightforwardly good nor straightforwardly bad, but genuinely mixed in ways that deserve honest engagement rather than either panic or cheerleading.

AI still gets things wrong with surprising confidence. There are legitimate concerns about privacy, misinformation, copyright, employment, and concentration of power. Those concerns deserve serious discussion. But it's equally important not to miss what's already happening because we're focused on the mistakes or waiting for some future arrival that has already begun.

This fall, many people who have never (or barely) touched ChatGPT or Claude or Gemini are going to get their first real glimpse of what that means — whether they're looking for it or not.

With iOS 27, Apple is preparing the most significant Siri upgrade in the assistant's history. The new implementation gives Siri awareness of your personal information, awareness of what's on your screen, and the ability to take actions across apps using natural language. The old Siri was largely a voice-controlled command system — set a timer, call Mom, what's the weather. It's useful in narrow circumstances, but limited and brittle. The new generation of AI assistants is moving toward something fundamentally different. Instead of memorizing commands, you'll increasingly be able to simply describe what you want. Ask it to find the restaurant your sister texted you about last month, locate the email with your passport information, or move the photos from last weekend into a new album and share it with your wife. These are the kinds of requests that historically required you to search through apps, messages, emails, and photos yourself. Apple is building systems that can understand the request, understand the context of your data, and take action on your behalf.

Whether Apple's implementation succeeds remains to be seen. But the direction is clear, and it represents something worth paying attention to: computers are slowly moving from tools that require exact instructions toward tools that understand intent. The smartphone changed how we interacted with our world. AI may change how that world interacts with us. That's a huge shift, and it's already underway.

The reason people didn't notice the smartphone revolution until it was complete is the same reason most people don't realize the AI revolution has already begun — gradual change doesn't feel like change while you're inside it. It feels like a slightly better app, a slightly more capable assistant, a slightly faster way to accomplish something you were already doing. And then one day you look up and the world has reorganized itself around something you barely noticed arriving.

I don't know exactly where this leads. Nobody does. But if your impression of AI was formed two years ago, you're already behind. And if you're waiting for the disruption to begin, you may have missed the fact that it already has.