Google I/O 17 Summary: Mobile, AI, and Innovation
By Ken Dulaney
(Aragon Research) – For anyone looking for new Android or ChomeOS features, Google I/O was much less of a “wow” factor than in previous years. This is not surprising. As the ecosystem for mobile and IoT devices matures, there are less and less new features in the OS. Much of the emphasis in feature enhancement is not specifically in the platform but in the applications and in the integration of cloud enabled features. Google I/O from this standpoint was more impressive than in all the years prior and we expect that future attendees of Google I/O must pay close attention to these improvements rather than the device and operating system news stories of the past.
Mobile First and AI
The conference keynote started with a core message of Mobile First transitioning to AI (Artificial Intelligence) First. For years, we have said that mobile was turning the traditional computing market on its head because much of the software that had been developed prior had never considered varying screen sizes, varying input methods and so on. Software has now begun to mature in these areas by recognizing that adaptability is essential in any software design. The mobile challenge has been largely met and Google has recognized this.
AI is taking the adaptability challenge to a new level under the premise that the entire ecosystem must be interconnected to react to the needs of the user. Such needs are not isolated events but potentially lengthy activity streams that are triggered by intelligent systems that know the user and predict what the user wants when they ask for something specific but expect that all data related to that request be brought forth. The user now expects that if they have input an appointment, that the system will tell them when to leave to that appointment, tell them any details regarding that appointment, and attach future requests to the fact that the user is in a specific situation dictated by that appointment’s context (e.g. location, time, etc.) Should the user ask for a restaurant recommendation, that recommendation should be made with knowledge of the user’s situation in time.
Google and AI – Pushing the Envelope
Google is incredibly well positioned in AI and to inject its capabilities across much of its ecosystem. It can offer not only the intelligence to make comprehensive decisions for the user but it also offers the “glue” that can combine hardware and software systems into meaningful productivity improvements for its clients. Microsoft is talented in AI but doesn’t possess the tremendous consumer information that Google collects every day from its users. That data, combined with Google’s Tensor computing architecture, gives it a big lead. Apple at this juncture may have the data but it’s more narrowly focused than Google and it must join up with either Microsoft or IBM to catch up. Google also offers its products across both the Apple and Microsoft ecosystems ensuring that users of those communities will be continually aware of Google’s prowess relative to those other vendors.
A good example of Google’s AI was its demonstration of Google Lens. Lens uses Google’s AI to recognize many types of objects. Visual recognition is just coming into its own as the next challenge after the tests presented by handwriting and voice recognition have been largely met. Google’s ability to augment reality to provide details on any scene shot by the camera was impressive. A user could take a picture of a restaurant and the system would immediately give the user details about that restaurant. And every one of these events conducted by the user, adds to Google’s knowledge of such places. We believe in the next 10 years that Google will have stored 100s of views if not more of every object on the planet of interest.
Some say that true intelligence is observed when the intelligence is seen to give multiple answers for the same set of conditions. Google’s system would provide the user with the highest probability answer consistently if the data set underneath remained constant. So maybe it’s not the human equivalent of intelligence that is the ultimate challenge for AI. But its usefulness is beyond question. Our point is … who cares?
Google Doubles Down on Android Security
On the enterprise side, Google has proven to us that it takes security very seriously. We believe that enterprise clients should consider Android devices very reliable in this area as long as they are not rooted and that they delivered from mainstream suppliers that are using the native Android build (e.g. Amazon Android should not be considered for enterprise use at this writing). The increased security scanning that the Google Play app will deliver to all Android users that have the Play App installed is a solid move.
Google and the Price of Innovation
Despite its tremendous innovation, Google still faces marketing challenges. It has too many products that stand alone and where users have seen Google pull them as they are usurped by other products or lack of interest. With Microsoft, you have a much more cohesive product plan with both release and end of life dates that have been proven reliable. It could be said that this is the price Google pays for innovation. The company prizes idea creation among many other attributes. But if Google can improve in this area, it would likely rise further among its competitors.
Google I/O 2017 showed that Google is delivering on its promise of improving our lives and for that it must be commended.