Howdy 👋🏾. After huge announcements from Google, Microsoft, Anthropic, and OpenAI, it was finally Apple’s turn up to bat, and, well, it was a bit underwhelming to say the least. On the AI front, as expected, Apple largely avoided the buzzwords like Siri but did make a few notable announcements:

👉 Apple’s Foundation Model Framework – a fancy name for its own local AI model that lets you tap into on-device generative AI. The benefits? It’s fast, private (everything stays on the device), and no token costs. But the real question: How good is this model? And for developers: Should you build around Apple’s ecosystem or stick with more universal cloud models that work on Windows, Android, and the web?

👉 Xcode finally gets AI-powered code assist tools – Apple’s lagging behind in developer productivity, while others have been turbocharging their tools with LLMs. (OpenAI spent billions acquiring Windsurf for this exact reason: OpenAI’s $3 billion Windsurf buy)

👉 Apple Intelligence – the suite of AI features got a few shoutouts, like recognizing photos and content from your camera. But let’s be real: these are novelty features compared to what Google Gemini is shipping. Last year, I predicted that Apple’s grammar tools at the OS level might Sherlock Grammarly. Instead, it’s been underwhelming, slow, uneven, and far from the “wow” we’ve come to expect.

Of course, the big thing we all hoped to hear was a new version of Siri with personal context, and in the first few minutes, Apple acknowledged that it wasn’t ready, basically confirming the recent piece by The Information showing just how behind Apple is in this AI race.

The big news was the launch of Liquid Glass, a new design language that heavily borrows from the look and feel created for the Apple Vision Pro device. It looks okay, but honestly feels more like a touch-up than a big shift in design philosophy. As many have mentioned, the idea of glass works in visionOS because it puts an emphasis on layering augmented reality with the real world, and the stacking looks nice, but could this be a hint at a bigger future push to a layered interface that pulls more of the real world into other products? Keep in mind that Apple usually reserves hardware announcements until the fall, so maybe we’ll see a more cohesive product when new hardware meets this new design language.

The one surprise I’m eager to try on my iPad is the slew of big changes to iPadOS. I joked with a friend that in many ways the addition of a windowing system, file menus, a mouse cursor, folders with custom icons, and so much more makes the iPad more like a Mac than ever. I’m not sure what this could mean for the future of these two platforms, but seeing these changes makes me wonder if maybe we might finally get a touchscreen Mac or maybe a loosening of the many restrictions on the iPad that keep it from doing the heavy lifting a Mac can do. These devices run on the same processors, and the differences between the two keep getting smaller.

The whole keynote clocks in at about an hour and a half, but for the first time in a long time, it was a bit underwhelming.

-jason

---

🔗 The Best in Tech This Week

🤖 Xcode Finally Gets AI Tools
I’m a big user of AI-powered coding tools, and until WWDC, Xcode was seriously behind. The speed these tools bring to development is massive. Any dev team not at least testing the waters is missing out big time. I haven’t tried Xcode’s ChatGPT integration yet, but Claude 4 is still my favorite. The lack of choice in which model you use feels like a miss.

💼 Windsurf Acquisition Rattles Anthropic
OpenAI’s $3 billion buy of Windsurf was barely announced before Anthropic started limiting support for Claude 4, my go-to dev model.

🔍 Google Zero is Here
I’ve been shouting about AEO, Answer Engine Optimization, aka Google Zero: the point where search traffic for many drops to zero. If your business is reliant on search or PPC, now’s the time to rethink your model.

💰 OpenAI claims to have hit $10B in annual revenue
Notable mention: OpenAI is printing cash.

---

🎤 The AI Roadshow: Workshops, Talks & Beyond

June 24, 2025 – WTCI AGILE: Building Earth’s Future From Space

July 10, 2025 – Delaware Division of Libraries AI Speaker Series

July 11, 2025 – Startup Grind AI Workshop

Aug. 11-16 – Black Is Tech Conference

---

P.S. Before you go…

If you haven’t heard of Trevor Rainbolt, he’s an absolute pro at a game called GeoGuessr. Basically, he looks at a single Google Maps street view image—anywhere in the world—and guesses the location in less than 0.1 seconds.

Well, guess what, ChatGPT is getting scary good at the same game. Without any special tools, it can pick up on subtle things like plants, house architecture, and street signs to guess where a photo was taken with frightening accuracy. And remember: this is the worst AI you’ll ever use. It’s only going to get crazier. Check this out: Watching GPT-4o guess a photo’s location.