Why Wall St. was disappointed by WWDC 26

Apple’s stock slid after WWDC 26, and the takes wrote themselves: not enough AI, too late, falling behind. I think the disappointment is real, but most of the diagnoses miss why.

AAPL, the week of WWDC 26

The problem wasn’t that Apple shipped too little AI. It’s that Apple shipped the wrong kind of AI — and in doing so revealed a worldview that’s quietly a generation behind.

The read-only frontier is over

For a decade, Apple has been the best in the world at one thing: making consumption feel exquisite. The Retina display, the App Store, Apple Music, the polished feed. The implicit model of the user has always been a reader — someone the device serves information to, beautifully.

So when Apple does AI, it does read-only AI. Summarize my notifications. Tell me what this email says. Surface the thing I was looking for. It’s all retrieval and condensation — pointed at you.

That was the frontier in 2023. It isn’t anymore.

Creation, writing, connection

The products people actually got excited about over the last two years all share a shape: they turn the user from a reader into a maker.

• You don’t ask the model for an answer — you co-write a draft, a function, a melody, a design.
• You don’t consume a feed — you generate something and put it into the world.
• The output isn’t private and disposable; it’s shared, remixed, responded to.

That’s the trifecta: create, write, connect. Generation plus distribution plus other people. It’s the difference between a smarter encyclopedia and a new kind of pen.

Apple, structurally, is allergic to two-thirds of that. Creation is messy and shows its seams — bad for a company whose brand is that nothing is ever janky. And connection means other people’s content flowing through your device, which Apple has spent fifteen years trying to keep at arm’s length behind privacy walls and walled gardens.

So we got a better reader. In 2026, a better reader is table stakes, not a keynote.

Why the market noticed

Markets aren’t reacting to feature counts; they’re pricing in where the value is migrating. And value is migrating to the layer where people make and share things — the creation layer — because that’s where engagement, lock-in, and network effects now live.

A company that owns the best hardware for consumption, but cedes the creation layer to whoever’s model and whoever’s social graph, is a company collecting rent on a shrinking idea of what a computer is for. That’s the thing the stock drop was really about. Not “less AI.” The wrong altitude.

What I’d want to see

The version of WWDC that wouldn’t have disappointed me looks less like “Siri, summarize this” and more like:

1. Creation primitives, system-wide. Not an app — an OS-level affordance to generate, edit, and iterate on text, images, and code anywhere you can type.
2. A connection story. Make it trivial to take the thing you made and send it, publish it, collaborate on it. Lean into other people instead of away.
3. The device as instrument, not appliance. Optimize for the person leaning forward and making something, not leaning back and being served.

Apple is the best in the world at the lean-back device. The frontier moved to lean-forward. That gap — not a model benchmark — is what WWDC 26 exposed.

Which, honestly, is also why I’m writing again. Reading the takes wasn’t enough. I wanted to make one.