- Google will make the Android Runtime (ART) a Mainline module in Android 12. Google plans to include the Android Runtime (ART) as an updatable Mainline module in Android 12. By making ART a Mainline module, Google will be able to exercise much more control over it, in line with the overarching theme of mainline modules. ART as a Mainline module would allow Google to update it without needing a system OTA update. If Google makes it mandatory for OEMs to preload a Google-signed ART module (likely delivered in the form of an APEX), then Google will be able to retain control over pushing updates to ART on all Android devices. Google would be able to push updates to ART through the Google Play Store, and OEMs wouldn’t be able to make modifications to ART. This in turn inches towards the behavioural consistency that third-party app developers would expect across the ecosystem.
- MY. GAMES and Google launched a program to support game developers. MY. GAMES together with Google announced the launch of the Game Drive program, which is aimed at supporting mobile game developers. Companies will select the most promising participants, give them an assessment and recommendations for project development, and some studios will be able to apply for investments from MY. GAMES Venture Capital.
To participate, you must submit an application by January 8, 2021. The most promising authors will be invited to a special event-Demo Day, which will be held on January 28 in an online format. There, developers will be able to get advice from experts from Google and MY. GAMES Venture Capital, and in a few months expect to get a second consultation on the project.
The game must pass the soft launch stage to provide basic product and purchasing (UA) metrics for review.
The first season of Game Drive starts on November 18 and runs until March 1, 2021. Currently, developers from Russia, CIS countries, Western Europe, Turkey, Israel, Africa and the Middle East can participate in the program.
- Apple released a version of TensorFlow for macOS Big Sur. Apple has released a fork of TensorFlow, Google's development environment for artificial intelligence and machine learning, optimized for Intel and Mac computers running on the new Apple M1 chip. Apple says that by taking advantage of the ML computing platform in macOS Big Sur, the Mac-optimized version of TensorFlow 2.4 allows developers to use the new 8-core M1 processors and their GPUs in training.
The new TensorFlow package, optimized for Mac, promises to lower the entry threshold, which will allow companies to train and deploy models easier and cheaper than before.
According to Apple, the New tensorflow 2.4 fork is learning on a 13-inch MacBook Pro with M1 7 times faster.