[Shohei Yamada’s Re:config.sys]Generative AI and Computing at the Pixel Edge – Computer Hour

The Google Pixel smartphone series has been updated. The latest version, the Pixel 8 Pro, was released on October 12. By performing various processing operations on the device, many AI-based functions can be used without relying on the cloud. This has been the trend in the past, but the new Tensor G3 processor will further accelerate it. How is Google trying to transform the smartphone experience?

Pixel 8 is out of control

From a look and feel perspective, I think the unbranded Pixel 8 is much nicer than the Pixel 8 Pro, which doesn’t really look like the Pixel 7 Pro, even though it is a slate device with a simpler form factor.

Pixel series as shown in the table below.MigratesI’ve been doing that.

The Pixel 8 has a size that can’t compare to the Pixel 5, but it’s the smallest and lightest since the Pixel 6.

I personally liked the Pixel 5 because it was small. Instead of storing it, I still use it almost every day while training at the gym. The width of its body is 70.4 mm. Pixel 8 is 70.8mm, so the difference is 0.4mm. This can be called pioneering.

These days, smartphones with large screens are very popular, but this comes at the expense of ease of use. Many users cover the body with a case, which makes it difficult to fit in the palm of your hand. You probably pick up a smartphone and have a feeling of giving up. There are many users who believe that it is simply out of their control. I can see why smartphones that can be worn diagonally have become popular.

However, the decision was made to make the screen size of the new Pixel 8 phone smaller than the screen size of its predecessor, the Pixel 7, and the width of the body was reduced. The image initially shows the Pixel 8 sitting on top of the Pixel 7. It appears to be a good size, but when you pick it up, you realize how small it is. It’s out of my hands. Many users seem to think that Google has heard their voices.

As a result, we followed the transformation shown in the table below.

The width is 70.8 mm compared to 73.2 mm on the Pixel 7, a difference of 2.4 mm. As the screen size has become smaller, the height has also become much lower, making the body size much easier to deal with. With this, you won’t feel like you’re out of control.

It seems that they listened to the voices of Japanese users who wanted a compact smartphone and decided to follow this trend. The Pixel series seems to have a fairly high number of units sold in Japan, so it could be said that they’re being considerate of Japanese users, but I don’t feel bad about taking care of them.

A combination of cloud and on-device processing

With Pixel, you can perform magical editing on the photos you take. For example, Eraser Magic can erase extraneous objects in photos with great accuracy, and this feature can be used even when the Pixel is offline without relying on the cloud.

In addition, the recorder application can record audio and at the same time, if you download data for the supported languages, you can record text in each language, but this can also be done during a disconnection.

One similar feature is real-time translation, which translates audio into the target language and displays it as text when the audio is detected on the device. Although it can be used offline, it cannot be used with recorders, so you have to choose an app. .

Additionally, the ability to use the Pixel on device, such as adopting Translate and Google Lens on the cloud, shows that there are differences in each use case, and trial and error are iterated to achieve the perfect implementation.

What’s amazing is that even with a two-year-old Pixel 6, if the software is up to date, the magic of the eraser and copy with recorder that the latest Pixel phone offers can be used offline.

Of course, processing speed, etc. depends on the processor’s performance. That’s why Google is working hard to improve the processing performance of its Tensor processor. Performance differs between the 2nd generation Tensor equipped with the Pixel 7 and the 3rd generation Tensor equipped with the Pixel 8.

Performance affects the processing time. For the end user, the less time it takes to remove extraneous objects from the image, the better. Slow computers are a source of confusion for non-technical end users.

It’s a simple story, but the feeling of wishing smartphones had higher processing performance is quite nostalgic. Non-gaming users are unhappy with the processing power of their smartphones during typical smartphone use cases such as enjoying social media, taking photos, listening to music, watching videos, and browsing the web. This is because there are fewer users and fewer instances where people feel this way.

However, it will be a different story in the future when edge computing, i.e. on-device AI processing, becomes the new norm.

Nowadays, we live in an era where we can connect to the Internet via Wi-Fi even on airplanes, so we may not think twice about turning the Internet connection on and off. If you are not satisfied with the in-flight entertainment, download the movie you want to watch on your smartphone and enjoy it before boarding.

For those who cannot or do not want to do this, they should board the plane and enjoy the last bit of connection, then quickly switch their smartphone to airplane mode, give up enjoying the content and focus on charging.

Try putting your phone on airplane mode or turning off Wi-Fi and mobile networks to see what you can do.

The same goes for laptops running in places where Wi-Fi isn’t available, and you’ll find yourself almost desperate and unable to do anything. I’m disappointed that I can’t do anything like this. However, in the future, smartphones will no longer be like this, and the same will be true for personal computers. From now on, more and more things can be done on the device. The era of smartphones and AI-powered computers has arrived.

Of course, this does not mean that communication is no longer necessary for intelligent life. Communication is necessary to import the latest content from the outside world, and communicating with others is impossible without communication.

However, it is desirable and expected that we will be able to handle computing locally without relying entirely on the cloud. In the future, as applications like generative AI become more and more popular, the concurrent use of simply dumping data into the cloud and having a terminal receive and display the processing results will likely continue forever.

It is said that the extent to which Google can improve the performance of Tensor as its processor will depend on the size of the investment, but I would like to see the end result of its development using different approaches. If Google doesn’t do it, who will?

It’s not too late to redefine the role of smartphones. Because experience goes further. The Pixel 8 series is arguably a stepping stone before that. So you can use it with confidence. There seems to be no reason to hesitate in obtaining it.