Google I/O 2019 – AI Prowess & Hardware Ambition Expands

The annual Google developer conference, Google I/O 2019, just concluded in Mountain View, California. Google CEO Sundar Pichai highlighted that the company is looking to resolve real-life problems through artificial intelligence (AI), machine learning, and other deep learning algorithms. The tech giant is marching ahead of competitors as online search, command and control shifts from the web to voice and users move towards new ways of interacting with their smartphones. Having said that, Google is looking to address the elephant in the room – Privacy and Security. Throughout the keynote, Google executives consistently emphasized Google’s sincere focus and work on the same. Here are some of the key announcements at the event:

HARDWARE

Google’s Pixel Family & Ambitions Expand

  • Google has introduced two cheaper or Lite versions of existing Pixel 3 and Pixel 3XL – the Google Pixel 3a and Google Pixel 3a XL priced at US$399 and US$499 respectively.
  • The launch also marks Google’s ambitions to expand the organic Pixel portfolio with a wider distribution reach starting with the US spreading across multiple carriers such as T-Mobile, Sprint, US Cellular, compared to just Verizon in the past.
  • Also launching in other 12 countries simultaneously – Australia, Canada, France, Germany, India, Ireland, Italy, Japan, Singapore, Spain, Taiwan, and the United Kingdom.
  • In terms of hardware, it is powered by Qualcomm Snapdragon 6 series 670 SoC from last year which features Qualcomm’s Artificial Intelligence Engine (AIE) but unlike the latest Snapdragon 6 and 7 series chipsets which pack in more capabilities.
  • Google Pixel 3a series is positioned to attract the mid- to high- tier segment with Google Pixel portfolio’s USP (or ultimate selling points) being – Latest Pure Android Experience, Faster Security & Software Updates, Google’s flawless AI capabilities out-of-the-box, Augmented Reality (AR) Core ready, eSIM (Google Fi), Titan M Security Module and above all the 12.2 MP dual-pixel Camera experience with features such as NightSight, Motion AutoFocus, HDR+ and Unlimited High Quality photo Cloud storage
  • The Pixel 3a should perform relatively well in the US and some other markets compared to predecessor Pixel portfolio. However, the high-tier segment is the US is a tough sell as the market is quite polarized and dominated by premium and low-to-mid tier portfolios
  • Pixel 3a needs to succeed in the US, which is a critical market for Google to first break into with its hardware ambitions considering the higher advertising ARPU from the American user-base.
  • If Pixel 3a succeeds in the US, it will expand the shrunk high-tier price segment though it will have to compete with older generation flagships from Samsung, Apple and new generation affordable premium phones from likes of OnePlus which are making their mark in the US.

Google’s Smart Home Positioning with Nest

  •  Google Home smart home devices will now sell under the “Nest” brand.
  • As a result, Google’s Home Hub will now be called the Nest Hub. Alongside the rebranding, Google launched a new smart display with a camera called the Nest Hub Max.
  • It supports Google Duo video calls and has all the features of the previous generation Google Home Hub.
  • The Face Match feature in Next Hub Max will show calendar updates for the user in front of the device by recognizing the face through the locally stored encrypted data.
  • The device will act as a control device for other connected smart home products from third-party smart speakers, gaming consoles, smart lighting, smart locks and especially tightly integrating with Nest portfolio such as Thermostats, Bell, Security cameras.
  • It will also support a few gesture control features, such as controlling music/videos using hand movements.
  • Nest Hub Max will cost US$229, and it will be available in the next few months.
  • The Nest branding and higher penetration in the US market could make Google more attractive to consumers compared to Facebook’s Portal or Amazon’s Echo Show, though Amazon is moving in a similar direction with the US$1 billion Ring acquisition from last year.

SOFTWARE

Android Q

  • Google says there are 2.5 billion active Android devices with the majority being smartphones which translate to roughly 60% share of the total active smartphones globally.
  • These users can now expect their smartphone activity to become even more secure.
  • Google announced that innovation, security & privacy, and digital well-being will be the key themes of the new Android release, which is called the “Android Q”.
  • Google Maps and Google Search will finally feature an incognito mode. On the security aspect, the two-step verification feature will now be available for all Android smartphones.
  • The new version of Android will also have more parental control options.
  • Google claims to have added more than 50 new features focusing on privacy and security in the Android Q update.
  • Further, Android Q is ready for the next generation of smartphones, which will feature foldable screens and 5G connectivity.
  • Google has closely worked with Qualcomm to dynamically optimize the apps running a Qualcomm powered 5G smartphone in terms of performance (e.g., switch to faster frame-rates, higher resolution graphics streaming, optimized battery life, etc.) whenever the Android Q based 5G smartphone connects to a 5G network
  • Google says 20 carriers are expected to go live with 5G networks this year and over a dozen 5G-ready smartphones will be launched by the end of 2019. Counterpoint also forecasts the 5G ecosystem to blossom faster than previous generation rollouts.
  • For devices with foldable screens, Android Q will have a screen continuity feature that will allow a seamless app and gaming experience while switching from a folded screen to the unfolded screen, and vice-versa.
  • Android Q will also feature a Focus Mode, which disables distracting apps until you want to use them again. The feature will be available in Android Pie and Android Q later this year.
  • The new OS will also have a Smart Reply feature which will provide intelligent suggestions to minimize navigation through different screens. For instance, if someone sends a meeting request with an address, the smart reply feature will provide options to agree to the meeting as well as the option of opening the address in maps.
  • It will also have the option of choosing a dark mode, which will help minimize battery consumption.
  • Android Q beta 3 is now available on 21 devices, including the Pixel devices.

AUGMENTED REALITY (AR)

  • New features & integrations in Google Search: A feature called “full coverage” will be available in Google Search, starting later this year. It will provide in-depth details of the searched event from multiple sources. Podcasts will also be available. Google will also add AR features such as bringing 3D models searched on Google into the real world, using the camera.
  • Integrating Assistant, Photos & Camera on Google Lens: Google is indexing the physical world using Lens. New features will combine the power of Google Assistant, Camera and Photos to solve real-life use cases. For instance, Google Lens will read a signboard in the preferred language and even make recommendations by pointing the camera on a food menu. User can choose the mode of output, such as hearing the recommendations or reading the translated signboard in the preferred language. These features will be available in Google Go, a search app for entry-level smartphones.

ARTIFICIAL INTELLIGENCE (AI)

  • Google Duplex to Minimize Transaction Complexities: With Duplex, Google is trying to remove transaction complexities for the actions we perform on our smartphone. In a nutshell, Duplex performs basic tasks for the user. Initial testing for Duplex was for making restaurant reservations, but now Google is extending Duplex capabilities to a web search. The feature will intelligently navigate through required actions pages for finishing the desired task. For instance, it can make bookings by intelligently choosing and auto-filling user details (including payment), and finally making a booking, while the user remains in control using only voice. More detail on this will be out later in 2019.
  • Advancements in Speech Recognition: Speech recognition is a combination of models, that translate sound bytes into phonetic units, assemble phonetic units into words, and predict the likelihood of their sequence. Such models are complex and require more than 100GB of storage and network connection. Google announced that it has been able to shrink such complex speech recognition models from 100GB to 512MB, allowing these models to be available in a smartphone locally. This eliminates network latency and makes assistance super-fast, which finally eliminates the need for tap-and-touch interface. Google demonstrated multiple voice requests made on the new Assistant, which was much faster than normal tap-and-touch control system. The new assistant is coming to Pixel phones later this year. It will also be available in the Waze navigation app, with driving-friendly features.
  • Live Captions and Live Relay: A new on-device feature “Live Captions”, will allow users to turn on captions for any podcast or any video including even self-recorded videos. Meanwhile, another feature “Live Relay” will allow hundreds of millions of hearing-impaired people to have voice calls by converting the written text to speech and vice-versa. These features again were achieved by shrinking the storage requirements of such complex models to a fraction.

MACHINE LEARNING-AS-A-SERVICE (MLaaS)

  • Google announced its third generation Cloud Tensor Processing Units (TPU) Pods; each Pod made up scalable 1000 TPUs which can deliver upto 100 petaFLOPS.
  • Google will let developers rent “slices” of the machines formed with these powerful TPU Pods for quick training of large ML models (e.g., image classification) in just a few minutes.
  • Basically, Google has transformed this into an on-demand (MLaaS) Machine Learning-as-a-Service.
  • Google is competing fiercely in AI, cloud computing and server space with more integrated self-developed solutions to differentiate with competitors.

Overall, this year’s I/O focused more on Google’s growing stance on Security & Privacy, democratizing AI for the masses and showcasing its hardware ambitions to scale the Google services faster and better than its partners. This messaging is very similar to Satya Nadella’s new Microsoft has been sharing for the last few years. For Google, with its data-driven and more so consumer-centric advertising business model its more difficult to convince the various stakeholders in the tech space. But this is a step in the right direction to position itself better than the rival Facebook, which continues to struggle with its missteps on this front.