WWDC 2026 turns optimism into real Siri momentum

After a keynote that leaned hard into Apple Intelligence and safety, developers left WWDC 2026 with a different kind of energy. In the days that followed, the message became clearer: Siri AI felt genuinely usable—shipping in the first beta—while Apple’s broade

For several days at Apple’s campus during WWDC 2026, the mood didn’t feel like the usual mix of hype and hesitation. Developers I spoke with sounded energized in a way that cut through the quiet skepticism that has followed Apple for the past two years of underwhelming AI progress.

The keynote morning started like it always does—badge in hand. media shuttled into a dedicated area within Apple Park—but it carried a sharper edge of confidence. There was even a small, physical reminder of the moment: the swag. A tote, a water bottle, randomized emoji stickers, and pins. The throwbacks stood out. including nods to Clarus the dogcow and the Apple Jolly Roger flag. plus an official Little Finder Guy (Fin) pin.

Then came the shift that made the day feel bigger than the usual tech event. Before the announcements, Craig Federighi took the stage just before the 10 a.m. PST kickoff. It was common for a top executive to speak first. but this time it was different: Tim Cook’s final conference as CEO. Federighi delivered a heartfelt speech about Cook’s tenure and how important WWDC always was to Tim. Cook was invited onstage, met with a raucous standing ovation.

Cook spoke for a few minutes, thanking everyone for being there and expressing what the developer community meant to him. As he teared up, the crowd applauded again—warm, loud, and unmistakably real—before Cook headed backstage and the pre-recorded dub dub video began to play.

In the audience, the questions were practical and pointed. Would Siri 2.0 be in the beta? Would it ship this year?

That keynote didn’t follow the familiar rhythm of platform-by-platform updates. Instead. Apple leaned heavily into safety first. then moved into general improvements across platforms. and finally went deep on Apple Intelligence. The result was a strange feeling of compression: fewer headline feature moments up front. but a sense that a lot of work was being tucked into the background.

A “huge word wall” loomed over the presentation—over 260 individual features listed. The Apple Intelligence demos also moved slowly enough to be unsettling, at least at first. It felt intentional, though, like Apple wanted the audience to notice the difference between live reality and polished theater.

By the time the keynote ended, the air in the room had shifted toward something more concrete. Siri AI wasn’t just presented as a concept. It was real, and it was shipping in the first beta. There was also the promise that many details would be uncovered in the days and weeks ahead—something attendees could already feel they would be able to test.

Outside the campus, the work continued. The Apple Park Visitor Center—temporarily converted into a media filing area—became the place where the beta download had to happen fast. As soon as the update landed, release notes, screenshots, and whatever small breadcrumbs could be found started to matter.

What followed wasn’t a single feature moment. It was a flood.

After the keynote. there was time for a tech talk hosted by Craig Federighi and the team. where they walked through Apple’s new third-generation Apple Foundation Models and their Google Gemini collaboration. Demo briefings filled in more gaps, and testing the beta personally turned curiosity into something closer to trust.

The biggest takeaway so far is blunt: Siri AI felt like the real deal. Not a simple skin over Gemini. Not something you use out of obligation. It was shockingly good—and it pushed the way many people treat Siri at baseline.

I had gotten used to avoiding Siri entirely because it had felt unreliable. But during testing, that habit started to loosen. Siri didn’t just handle requests; it helped complete them in a way that felt natural.

At one point. I wanted to save someone’s current address from Find My as their work location on their contact card. I started trying to copy the address manually, then realized Siri could do the job instead. I asked Siri to add that person’s current location as their work address to their contact. Siri did it, only asking me to confirm the change to the contact card.

Later, there was another kind of test—timing and context. A local artist was doing a special new drop at 8 p.m. the next day, and the artist had teased it on X. I asked Siri: “Look up this artist’s website and remind me to check out their new drop when it’s released with a link to the store. ” and Siri used the info from the tweet to correctly make that reminder.

Neither example was flashy in the way big AI announcements sometimes are. They were useful. That’s the point—and the feeling that it can keep getting better is what lingered after both moments.

Outside Siri, there was a different kind of progress across devices, built around speed.

Apple TV updated to tvOS 27 didn’t come with a long list of headline features. but Apple called out specific performance improvements: faster app launches. more responsive Control Center. quicker music playback. and faster AirPlay. The performance gains also extended beyond the demo stage. The Vision Pro, for instance, boots much faster than it did in the initial release.

And the update pattern followed through into the smaller practical tools. tvOS 27 adds keyboard pasteboard suggestions and Apple Watch gestures. It also brings keyword support in Photos and 4K camera compatibility in Apple Home.

Safari and developer workflows were where the energy turned from personal utility to creative possibility. Getting to try extension creation in Safari felt like a tangible preview of how quickly developers can build for the modern platform. I told Safari to create an extension to act as a recipe manager with a clean UI with blue accents. a rating system. comments. and a category for each recipe. It offered options in the App Store, then wrote the code using Apple’s Safari extension APIs.

The extension didn’t just generate and vanish. It came back from Private Cloud Compute (PCC) a few moments later, ready to use. In under a minute, I created a custom extension that would automatically sync across my devices.

There was also an Xcode demo that helped create an entire pin trading app via Vibe coding. It included camera-based pin identification, animations, and AI tools.

For many people, these improvements will sound different from the kind of features that grab attention online. Speed, battery, and performance improvements aren’t always the stuff of viral headlines. But using iOS 27 and the other updates felt different—less like a bulky download with a handful of standout toys. and more like a device that immediately behaves better.

The message wasn’t that everything was perfect. Several months remain to improve and polish the experience, and that reality hung in the air even as people tested. Still, the developers at Apple Park were equally excited. There was a clear belief that many of them could tap into these AI models on their own.

For now. the beta experience is doing its job: making Siri AI feel usable instead of experimental. and making the rest of the platform feel quicker. sharper. and more personal. By the end of the campus day, the optimism didn’t feel like a promise. It felt like something people were already trying—and, for the first time in a while, actually liking.

WWDC 2026 Apple Intelligence Siri AI iOS 27 tvOS 27 Apple Foundation Models Google Gemini collaboration Private Cloud Compute Safari extension APIs Xcode demo Vision Pro boot speed