I Said Apps Were Dead. Apple Just Proved It — By Building What Comes Next.

Apps Were Dead

7:44

About a year ago I started saying the App Store was dead.

Not today-dead. Not literally-no-apps-exist dead. But dead in the way that matters strategically — dead as a paradigm, dead as the dominant interface layer between humans and computing, dead as the thing developers should be building toward. I got flack for it. I still get flack for it. So let me update the thesis — not to say I told you so, but because we can now see the entire shape of what's coming. And it's bigger than I thought.

Where This Thesis Started

Back in October, I wrote about the coming interface revolution — the idea that every generation of technology is really just compression. We compress the distance between intent and action. Keyboard to mouse to touch to voice. Each layer dissolves, and a new one appears underneath it.

The App Store was the most successful version of the old paradigm. A curated storefront of discrete, purpose-built tools. You have a need. You find an app. You open the app. You do the thing. You close the app.

That model made Apple a $3 trillion company. It was extraordinary. And it's ending. Not because apps are disappearing. Because the layer below apps is becoming accessible.

The Phase Change Nobody Predicted Fast Enough

In February, I wrote about ClaudeBot — about the moment execution became expressible. About Mac Minis stacking up in rows because people could feel that something had shifted. About getting agent logic running on a $5 ESP32 and laughing because it was the good kind of insane.

The thesis: we crossed from "AI answers questions" into "AI executes intent." That's not an upgrade. That's a category change. And in March, I wrote about skills as infrastructure — about the quiet shift from prompts to repeatable, agent-callable methodologies. About the moment the caller changed from human to machine. All of that was pointing at the same thing: the interface layer is collapsing. Not the apps. The layer that required apps in the first place.

What Apple Is Actually Building

The WWDC announcement is going to get covered as "Siri got smarter." That's the wrong frame. What Apple is actually doing is building the mature, privacy-first, consumer-grade version of what the open source agent community has been doing in garages and server closets for the past eighteen months.

Think about who's been running agentic AI in the wild. Developers. Tinkerers. People buying Mac Minis in bulk. OpenClaw users. People running Ollama on Apple Silicon because the local inference story on Apple hardware is already, quietly, excellent. Apple knows this. They've watched the open source agent community do its testing — on their chips, at scale, across a million different use cases — and now they're ready to productize it.

This is the Apple playbook. Let the wild west figure out what works. Then walk in, clean it up, wrap it in privacy guarantees and seamless UX, and ship it to 1.5 billion people. They did it with MP3 players. With smartphones. With tablets. They're doing it now with agentic AI — and this time, their own hardware has been the dominant substrate for AI agent development all along.

Why This Ends the App Store Paradigm

The App Store model requires a human in the loop at every step. You decide. You open. You interact. You close. What Apple is building — App Intents, ambient Siri, MCP integration — removes that requirement.

You don't open the Uber app. You tell your phone you need a ride. You don't open your photo editor. You tell it what you want the photo to look like. You don't open Amazon. You tell it what you're trying to buy. The app doesn't disappear. But the app stops being the interface. The app becomes a capability that an agent calls on your behalf.

Right now, apps compete for your attention and home screen real estate. In the agentic model, apps compete for agent favor — to be the most reliably callable capability in the orchestration layer. The user never sees them directly. The agent picks them. That's not a small shift. That's a complete inversion.

The Privacy Angle Is the Moat

Everyone else — OpenAI, Google, Anthropic — runs your data through their infrastructure. You get capability, they get signal. Apple's architecture is different by design. The small model runs on device. Private data never leaves the phone. Complex reasoning routes through Google's infrastructure, but through Apple's abstraction layer, with Apple's privacy commitments on top. For the 1.5 billion people who just want things to work — this is the unlock. Apple can say: you get all of this, and your data stays yours. The open source community solved capability. Apple is solving trust. That's a moat nobody else can build right now.

Where This Puts Apple in the AI Race

Six months ago, Apple looked like it was losing. Siri was a punchline. The Apple Intelligence launch was a disaster. But zoom out...1.5 billion active devices. The most trusted consumer brand in tech. The best mobile silicon on the planet — silicon the agent development community chose as their preferred inference substrate. A privacy story nobody else can match.

They were never going to win the model race. That was never their game.

Their game is the interface. The layer between human intent and machine execution. And they just built the playbook for owning that layer at consumer scale. Apple isn't at the back of the AI line. They've been building a different line entirely — and it's about to intersect with everything else.

The Real Headline

A year ago I said apps were dead — meaning dead as the paradigm, not dead as software. What Apple is building at WWDC is the proof. Not because Apple killed the App Store. But because Apple is building the layer that makes the App Store irrelevant as an interface concept — and replacing it with something that was always the logical next step.

The interface revolution I wrote about in October is arriving on schedule.

The cognitive layer is here. Apple just decided to own it.