On December 6th , we announced Gemini, our most capable AI model yet!
It’s been optimized for three different sizes: Ultra, Pro and Nano. Gemini Nano is our most efficient model built for on-device tasks, running directly on mobile silicon which opens support for offline, privacy-sensitive inference.
We announced that you will soon be able to use the Google AI Edge SDK to leverage Gemini Nano through Android AICore. It is currently rolling out to Pixel 8 Pro with more devices and silicon partners to be announced in the coming months.
AICore is currently being used to summarize content in the Recorder app, even when the phone’s network connection is offline.
If you are interested in building apps using Gemini Nano and Google AI Edge SDK, please sign up for our Early Access Program.
The new ML Kit Subject Segmentation API, now in beta, lets you easily separate multiple subjects from the background in a picture. The API takes an input image and generates an output mask or bitmap for the foreground. It also provides a mask and bitmap for each one of the subjects detected.
It enables cool use cases such as background swap, or add unique effects to people, pets, or objects with just a few lines of code.
To get started go to the ML Kit documentation.
Android Studio Hedgehog is now stable, and it’s packed full of upgrades, such as improvements to Android vitals in App Quality Insights and the new Power Profiler.
The Android SDK Upgrade Assistant now makes it easier to upgrade applications to target Android 14.
Finally, there are a number of tools to make it faster to develop using Jetpack Compose including Preview’s Gallery Mode, Compose State information in the Debugger, and Compose Multipreview templates.
We have several new blog posts for you to check out!
Read more on Virtual Machine as a core Android Primitive — the Android Virtualization Framework (AVF) that will be available on upcoming select Android 14 devices, and how it will offer platform developers and privileged app developers new capabilities.
Also, read more on the launch of KSP2 Preview, which introduces support for the new Kotlin compiler (code-named K2) and a new standalone source generator that provides more flexibility and features than the current Kotlin compiler plugin.
Finally, learn more about how NordVPN boosted the speed of its login user flow by 60% using Baseline Profiles.
Check out the video on Enhancing Jetpack Compose performance, where Tomáš demonstrates techniques for performance through measurement, debugging, and strategic optimization.
The easiest optimization technique is keeping Jetpack Compose updated to benefit from the performance improvements we implement under the hood. Another one is to enable R8 and generate a baseline profile, which can significantly optimize code execution. And thirdly, if you change state rapidly in some cases, explore the option to defer the compose phase to prevent unnecessary recomposition. Check the video for a more in-depth explanation.
We had several exciting new features launched in Android Jetpack:
- In Activity 1.9-Alpha ComponentActivity now implements OnUserLeaveHintProvider to allow components to callbacks for onUserLeaveHint events.
- In Webkit 1.10-Alpha we added a new API to WebSettingsCompat to control the experimental Android WebView Media Integrity API behavior.
You can see all the AndroidX release notes here.
That’s it for this week with Gemini, AICore, ML Kit, Android Studio, AVF, KSP2, Performance and more!
Check back next year for our next update from the Android developer universe and happy holidays! 💫