Google I/O, Google’s annual product showcase, displayed a very strong AI focus this year. We saw little new hardware, and even with Google’s “add AI to all the things” strategy, most improvements were incremental, not groundbreaking. If you haven’t already, read Google I/O 2023: As Far As The AI Can See to get up to speed on the consumer and enterprise AI integrations that Google announced.

Beyond AI, Consumer Development Continues

WearOS gets a new way to create watch faces. There are updates to Find My Device to support earbuds and tags, as well as an interesting partnership with Apple to provide “Unknown tracker alerts.” Google reiterated that it was adopting the new rich communication services (RCS) standard for Messages. Sundar Pichai’s hyperbolically earnest, “We hope every mobile operating system gets the message and adopts RCS,” had me laughing.

We hope every mobile operating system gets the message and adopts RCS.

What It Means: Google’s non-AI consumer software announcements were … fine. There was nothing to stir the blood. Google software is mature and will continue to improve gradually. Without regulatory persuasion, it’s not likely that Apple will abandon its iMessage protocol in favor of RCS. We’ll continue to see a mixture of blue and green chat bubbles for a while.

New Hardware Is A Mixed Bag

On the hardware front, Google announced three new devices: the previously leaked and eye-wateringly expensive Google Pixel Fold ($1799, with a free $359 Pixel Watch if you preorder), the Pixel 7a ($499), and the Pixel Tablet ($499 with a free $129 dock if you preorder).

What It Means: The Pixel Tablet dock lacks a Matter hub, which seems like a miss. I’d expect a future dock with more Matter support. Economic conditions mark the Pixel Fold’s merged phone-and-7.6-inch-tablet as a luxury. Unless you’re using the Fold’s multiscreen support, you probably won’t need to test Google’s folded orientations beyond the emulator.

Developers See More Integrations

The Google developer keynote follows the Google I/O keynote, and this year it brought a few more announcements of interest to developers. MakerSuite gets new features for synthetic data and safety settings, while Firebase gets AI extensions for Bard and a chatbot. MediaPipe simplifies integrating your code with AI. There are new Google Meet APIs to bring Meet to your application — or your application to Meet.

Android Gets AR And Emulator Improvements

Android adds Geospatial Creator to ARCore for GIS-like capability. Developers get Android Studio Bot for code generation and a waitlist for access to Google-hosted streaming devices — a service that sounds a lot like Samsung’s Remote Test Lab. There’s also a new emulator with Bluetooth support.

What It Means: Google’s plan for hosted streaming devices is interesting. To date, the company has given no indication that it will eliminate support for the emulator, but with a streaming alternative, it could — and would be free to change processor away from ARM. That would disadvantage vendors of other Android devices, especially the very inexpensive devices that don’t integrate Google Play services.

Broader Web Support And A Focus On Responsiveness

Firebase Hosting now supports Astro, Flutter, Nuxt, and Sveltekit in addition to Angular and Next.js. Google is shifting web metric importance from initial page load to responsiveness after load with the interaction to next paint (INP) metric, which replaces first input delay in 2024. WebGPU provides access to the GPU for web apps that need extra horsepower. Google mentioned the (previously announced) WebAssembly compiler for Kotlin.

What It Means: Google is broadening its hosting appeal and adopting more frameworks that support static site generation — a solution to INP. Google’s recommendation for a cross-device solution that includes the web stays confusing: Both Kotlin and Flutter fit the roles.

In Summary: You’ve Seen This Before

What was missing? There were a few nods to privacy, but nothing as substantial as 2021’s push to catch up to Apple. Sustainability fared even worse: no mentions at all. If you watched Google I/O hoping for generative AI integrations of things you’ve already seen other large language models do, you got them in abundance. If you wanted something else … well, your earth remains unshattered.