The Google I/O developer conference, which was canceled last year due to the outbreak, is back with a live webcast at 1am today.

More than two hours of live can be said to be full of content, the whole highlight. The theme of the conference, “Building a More Usable Google for All,” aims to provide every user with critical help at the moment, and it was a real reminder of what technology is about to change lives.

This conference is no longer a pure technological show-off, but a demonstration of the future world after integrating intelligent technology into life.

Android 12 is no longer boring

Google has unveiled a new design language for Android12, called Material You.

The design takes the look level up quite a bit. It can match color combinations based on your choice of color or a custom wallpaper, choosing a more beautiful and harmonious color combination for you, which becomes the theme color of Android 12. Not only that, Google has also used the rounded corners of the new ICONS and widgets to make them look smoother and smoother.

During the conference, Google confirmed that the design would first appear on the new Pixel, which will be released this fall, meaning that the official version of Android won’t be available until the fall. Google also said that in the future, “everything from web applications to hardware” will adhere to this design language, which means wearables, smartphones, tablets and IoT projects will all be compatible with Material You’s new palette of colors.

Android 12, on the other hand, not only focuses on the visual design, but also simplifies a lot of the interaction features, and simplifies the animation to make users feel more fluent.

After simplifying the system, Android 12 is more responsive and has a higher power ratio. Google says it has reduced CPU time required for core system services by up to 22%, and reduced the large core CPU usage of system servers by up to 15%, which should be very battery friendly.

In addition, Google has further protected users’ privacy.

The first is the new privacy Kanban, where users can see all their permissions, as well as what data is being accessed, how often, and the applications that have used it. Users can also remove permissions from applications at will.

The second point is that when the app uses the microphone or camera, there will be a privacy indicator in the upper right corner, which is similar to iOS and more intuitive than Android 11. It is presented in the form of a small dot with color.

The third point is to ensure privacy security through the design of Android’s private computer core and other technologies. It can be thought of as a privileged space within the operating system, where data is kept locally for the device to learn and use, and not from outside prying eyes.

After talking about Android, Google said it wanted to build a mobile-centric network that would allow computers, cars and other devices to collaborate.

Finally, there’s the Easter Egg: Android 12 will also feature a digital car key. Google says it’s working with BMW to introduce digital car keys to more brands in the future.

Wear OS is back online

Google is the oldest player in wearables, but for a variety of reasons it hasn’t been the first to be seen as a leader in wearables.

At the conference, Google announced a major change: it will reinvent WearOS in conjunction with Samsung and Fitbit, which Google has already acquired.

Samsung has been using its own Tizen OS for wearable devices for the past few years, but when Google relaunch Wear OS, Samsung decided to team up with Google and integrate Wear OS with Tizen OS.

Battery life in the integrated Wear OS will be significantly improved, APP loading times will be shorter and animations will be smoother.

AI is the big deal

Based on powerful AI algorithms, Google showed off two AI-based models — LAMDA and MUM — during the keynote.

Natural Language Platform — Lamda. “We have revolutionized the ability of computers to understand natural language,” says Sundarpichai.

If you had a conversation with Pluto or a paper airplane, what would you get?

For this reason, Google has built a complete artificial intelligence dialogue model, LAMDA. Through machine learning algorithm, it can not only select the most accurate answer trigger from the pre-configured response list, but also adjust the response result through active learning.

This means that Lamda will allow for more intelligent question-and-answer conversations, and will be able to provide suggestions and answers based on context.

A new unified model for multitasking – MUM.

The capacity of MUM is 1000 times that of the existing Bert, and it can simultaneously process text, pictures, videos and other information, and finally obtain unified information with strong reference value.

At the conference, Google demonstrated a case of Mum. The user told Mum that he had successfully climbed Mount Adams and wanted to climb Mount Fuji next year. What should he do to prepare for it?

MUM provides the most reasonable climbing route, time and equipment requirements based on photos, videos, road maps and other information provided by users, as well as the seasonal and geological differences between Fuji and Adams.

According to Google, MUM will combine a deep understanding of the intent of your query with the world’s information to solve the most complex searches. Mum, like Lamda, is currently in the development stage, and no specific date has been given for when Google will be available.

In addition, Google officially unveiled the TPU IV at the conference.

With 4,096 TPU V4 chips in each device, a POD can provide more than one Exoflops (10^18 floating point power) of AI computing power.

Google also revealed that it is working on a new tool that will make it easy to delete entire photos of a person from a photo library, for example, an ex that you want to forget but frequently appears in albums.