Google I/O 2017 Recap


Hello, my name is Seungwon Lee, I work at LINE as an Android developer. This post is about our visit to Google I/O 2017 held in Mountain View, USA from May 17th to 19th. Google I/O, started in 2008, is an annual developer-oriented conference. Along with Apple’s WWDC, Google I/O is by far one of the largest events for software developers worldwide. The “I“and “O“in the conference name stand for “Input and Output” and also for “Innovation in the Open”. This conference is renowned for Google announcing its future visions. For me, the conference was a chance to learn the changes in the new Android OS that would immediately affect LINE, and to glance at Google’s latest achievements and future plans for web technologies, cloud services and VR (Virtual Reality).

Google I/O Keynote

From mobile-world to AI-world

The hottest topic in the event was AI (Artificial Intelligence); Google shared their various attempts in applying their in-house AI technology onto real services. In the keynote, Google CEO Sundar Pichai proclaimed the ‘Mobile First to AI First’ vision, signifying Google’s aggressive approach towards AI-based business. Correspondingly, presentations at this year’s event mainly dealt with fundamental technologies for an AI-first future and the features of Google services that had embraced AI.

Sundar Pichai introduced, a window prepared to share Google’s research made on machine learning and to communicate with developers. The site will probably be used as a medium for distributing Google’s new AI tools and to make AI easily approachable to the people interested.

new Cloud TPUs

According to Pichai, a new hardware technology is essential to enter the “AI-First” era. To speed up machine learning workloads used on various Google services, Google is providing a new service, Cloud TPU.


Tensorflow Lite for mobile

Tensorflow Lite is a framework based on Tensorflow, optimized for mobile devices to run machine learning on Android O devices that are to be released at the end of this year. Tensorflow Lite enables running deep-learning models on smartphones, enabling deep-learning models run stably in the absence of network.

Seeing the world through Google Lens

Recently, computer vision technology has been exceeding human vision in terms of recognizability, with the use of artificial neural network. Google is providing a new service, converging their existing services such as Google Map, Google Knowledge Graph and Google image search with computer vision technology. The new service, Google Lens, uses a camera to provide meaningful information about objects near the user, in the form of AR (Augmented Reality). For example, users can use Google Lens to get information about restaurants they are walking by, or get menus written in a foreign language translated on the spot. In near future, Google Lens is to be provided in connection with Google Photo and Google Assistant.

Multi-lingual Google Assistant

Google’s speech recognition technology has been refining constantly and the recognition error rate for English has fallen to 4.9% this year. Google’s speech recognition, started out in the past as a simple input interface for Google Now, is now maturing into a true personal assistant system through Google Assistant. Google Assistant is a personal assistant service installed on Google Home and smartphones such as Pixel. The basic features of Google Assistant include calendar, e-mail, web searching and audio service. In addition, Chromecast, IOT connections and user customized services are to be provided. Google Lens will join the clan as a new input interface to Google Assistant. A delightful news was announced that Google Assistant is now available on iOS and is to support more languages; Japanese, French, German will be supported from this summer. Korean and Italian are to be available from the end of this year.

Google Assistant Image:

Source :

Google Photos suggests friends to share photos with

Image sharing features, including Suggested Sharing and Shared Libraries, are now available on Google Photos. Google Photos uses face recognition to offer suggestions with whom to share the photos. Users can now continuously share photos of a particular person, like on iOS’s Photo Stream. Another new service is Photo Books, where Google Photos selects your best photos and turn them into a real photo book.

The keynote addressed not only AI topics, but the following updates on Google’s main services and new projects.

Android O Beta: Elegant & light


Elegant user interface and ‘fluid experiences and vitals’ are the two things Google emphasized to provide to the end users. The most notable changes in the beta version of Android O were the following:

Supporting PIP mode: PIP (Picture in Picture) supports running two apps on one screen like multi-window viewing, but is more advanced. PIP has more focus on actual content. For example, PIP enables playing a video on top of another app, within the same screen.

Battery, Background check: Rapid battery drain is mainly due to apps consuming resources carelessly. On Android O, background apps have restrictions on using resources. The doze mode on Android O is more powerful than ever before, to control background services and unnecessary broadcasts.

Notification Channel, Badges: Notification channel allows developers to customize notification categories. Furthermore, notification badges are provided on the launcher to inform that one or more notifications are yet to be checked.

Supporting Kotlin: For Android developers, the biggest news was that Android officially supports Kotlin from JetBrain. Kotlin is a language that runs on JVM, consists of simple syntax (unlike Java) and is easy to handle runtime exceptions such as NPE (null pointer exception). Bon voyage to Kotlin enthusiasts.

Android Go & Stand-alone VR headsets

Android Go: Android Go is the lite version of Android O. Android Go is designed to consume less memory and less battery and is intended for budget smartphones with less than 1GB of RAM. YouTube Go, optimized for Android Go, can be downloaded to view videos offline for free.

Stand-alone VR headsets: Daydream is a VR platform developed by Google. To use Daydream when it was first released, you needed a VR headset and an Android device that supported Daydream. Unfortunately, not many smartphones supported Daydream. This year, Google announced the launching of stand-alone VR headset that requires no smartphones or PCs. Furthermore, these stand-alone headsets will support WorldSense, a new technology that tracks headset locations independently without relying on extra sensor devices. Google is cooperating closely with QUALCOMM to create the industry standard for these headsets. The actual release is expected to be at the end of this year, through HTC VIVE and Lenovo.


A general review of the Google I/O event

Here are notes on my overall impressions on this event.

Nothing like last year

My first impression was that Google put a lot of thoughts on resolving the problems that surfaced in the last year’s Google I/O. Having been the first ever outdoor Google I/O, many attendees ended up getting sunburned from getting exposed to the strong San Francisco sun in long queues. This year, every attendee was given sunglasses, sunscreen and water bottles in their welcome kit. Also, to prevent long waits, each session took reservations in advance. Furthermore, there were water stations running on bicycles that would come to the most crowded spots, unlimited beverages and snacks, and raffle events. In the evening, blankets were given out so the attendees could stay warm in the sharp temperature drop of the San Francisco sunset.


At Sanbox Dome, booths had been set up for you to see the demonstrations of the Google services introduced through out the sessions and meet up with the session speakers.

Beverage Mixer using Google Assistant SDK

Attendees tried out services such as Android Pay with NFC and Google Daydream VR and even do some coding using the latests APIs at Codelabs.

Visits to tech company offices in San Francisco

We had a chance to visit Airbnb and Google, which happen to be located in San Francisco. We spent time with some employees there sharing about their corporate culture, working environment and current business goals. And of course, we didn’t miss out their office cafeterias (known as top IT firms must-haves).

Google Campus: Freedom and discipline

Google is a company that never fails to fascinate the world with every new technology it releases. Having the last 10 years of dominance in the IT market, Google’s employee size and office size were both enormous. Google’s office is called ‘campus’, and campuses were divided by service sectors; platform, application, cloud and so on. The site was so spacious, people used ‘Google bikes’ to move within the site.

The office we visited was a new campus that has recently opened for cloud and server sectors. The office was still under construction—who could imagine how humongous it will be when completed. Coincidentally, it was lunchtime when we visited, giving us an opportunity to see many Googlers enjoying lunch at the cafeteria. Menu items ranged widely from standard ‘western’ dishes to Japanese and Spanish cuisines. There was even a sushi train—definitely not something you would expect to see in a company cafeteria.

Despite such flexible and free environment, everyone seemed to have very heavy workloads. Each employee works voluntarily on their own projects, and gets evaluated on their performance accordingly. Moreover, their evaluation result is open to everyone in the office. The result is used as the indicator of their position in the company. Promotion is not determined by supervisors or co-workers—employees wishing to be promoted actively request for recommendations from colleagues, and get assessed by a group of people who has no relation to the promotion candidate.

Airbnb headquaters: Air-bed and breakfasts

Airbnb is one of must-haves for travelers. I wondered what their office would be like.

The philosophy implied in the company name (a combination of comfortable Air-bed and breakfast) was crafted into their office; there were plenty of comfortable spaces and the atmosphere was great. A fascinating cafeteria was located on the first floor, and other floors all had a resting area, fully stacked with various types of beverages and fruits for the company members.

The office had countless workstations where you could be on your own. In fact, many employees were working at these stations instead of their desks. It was impressive how the office itself resembled the multi-cultural and relaxed vibe many worldwide Airbnb hosts exude.

You can bring dogs to work at Airbnb. The respect for individual freedom was so evident through Airbnb’s effort for embodying it into real work environment.

The company used to have not so many members a few years back. To expand their services, they have been doubling their size every year lately. Projects are executed in groups, consisting of planners, designers and developers. As much personal freedom is guaranteed, the company expects corresponding commitment from their employees.


New services tend to excite developers, and this excitement turns into inspiration for developers. As seen in the keynote, many of Google’s main services are stable now, and the company is putting a superb amount of efforts into AI. Not only did they research but they have been applying the research outcomes onto many of their services to benefit users. Likewise, LINE is striving to serve its users by discovering, developing and applying new technologies on many of its services.

As I close this post, I’d like to encourage developers to stay tuned to the worldwide trend, for example, follow up on Google I/O, like you are reading this post now. Keeping up with the latest trend and news will surely keep you inspired and energetic.