Why should you care about AI chips on Apple and Huawei phones?

Earlier this year, both Apple and Huawei talked a lot about the new processors on their phones, that will have game-changing, neurotic processing engines, specifically designed to handle computing. AI math. But why should we care, do you really need a phone that supports AI?

What none of these manufacturers do is bring game-changing apps to go with these powerful chips. Apple's A11 Bionic supports Face ID and Animoji, while Huawei's NPU Kirin 970 processor offers automatic photo settings to make images look better, translate with Offline Translator on Mate 10 phones. But that is all the features are good, but not yet called the revolution.

The Next Web editor commented that all next year's flagship products will all use AI chips, and challenging phone manufacturers is proving worthy of rumors. Ideally, developers take advantage of new hardware and bring applications to exploit new computing capabilities. But what can we expect?

According to Baofeng Zhang, Huawei's vice president of software, it is important to remember that most applications using AI in the cloud can be easily modified to take advantage of neuron chips, using Use Kirin API or open source frameworks like TensorFlow and Caffe2.

Doing so will help with faster processing, less dependence on connectivity, increased security because data does not have to leave the device to do the job.

Zhang explained that the 3rd party applications will first work with voice recognition and image technology, then tools that use natural language processing such as searching for what is stored on phone.

This is followed by more advanced tools such as video editing, for example when you shoot a 1-hour video game and the app picks out the highlights when you issue a voice command. Hannah plays the ball '.

The application must also be smarter when the phone has a smarter chip

That will replace the generation of today's applications, which only feature Smart-Cuts, essentially using sound, and video transfer data to find 'interesting' scenes. AI chips will make much smarter decisions.

What are the big guys doing with AI chip applications? Microsoft works with Huawei to open offline translation capabilities for its Translator application with NPU chips. The feature is now available but does not yet support voice translation when there is no network connection.

Google plans to bring more services and APIs by the end of the year na. Hope they can help developers exploit Huawei's NPU and similar processors.

What about the next generation of AI chips? Zhang said that they might have a higher computing power, but the development of frameworks like TensorFlow Lite that developers still use to build applications, could affect the development path of microprocessors. this.

Now do you need a phone with an AI chip? Honestly, there is still no remarkable application to take advantage of AI chips, so this is still just an early reception phase. But if Huawei is right, many phones will use AI chips and it will be time for developers to take advantage of new chips to create truly game-changing applications.