Many people are writing about what Google *will*
discuss at tomorrow’s Google I/O conference, but unless they have clairvoyance, leaking, or guessing based on the schedule, it’s difficult to be definitive. I will be attending tomorrow and wanted to discuss what I think Google *should*
discuss during the event.
Last year, Google rolled out the TPU, a machine learning inference accelerator, and they also showed how the company used machine learning-based AI in Google assistant, bots, Google Home web and Android experiences.
Google should talk about updates on all those fronts. Bots so far have been a market and user failure and think like Facebook and Microsoft did in their conferences (F8 and BUILD which I attended), Google should talk how they see bots working in the future. I think we should here more examples of how AI is everywhere in all the Google experiences of Chrome, Android, G Suite and possibly Pixel. Google, like Facebook and Microsoft will go out of their way to show how much farther ahead than anyone in AI with frameworks, applications and hardware. They will talk about the thousands of ways they are using AI and reinforce that Tensorflow, created and then open-sourced by Google, is the #1 distribution on GitHub.
Google was closed, if not cagey, last year on details of the TPU aside from their blogs and it was a challenge to get at last year’s show, determination on what a TPU actually was (ASIC, DSP, FPGA) and what is actually did (machine learning training, inference, or both). Google also compared the TPU to much older CPUs and GPUs so I’d like to see more contemporary comparisons.
: Google made most of their enterprise cloud announcements at Google Next 17, so I think on volume, we shouldn’t hear as much about the cloud as last year. I do think we should see some fancy updates on cloud/web apps like Photos, Music, Mail, Search, Maps, Music, YouTube, etc., likely leveraging AI above.
Virtual/augmented reality: Google VR hasn’t appeared to take off as strongly as I think the company expected, and Samsung and Oculus is making more traction in mobility while Oculus, HTC Vive and Microsoft are defining more immersive experiences at the higher-end. Google should put much more of a focus on VR this year than they did last year, maybe even launch their own self-contained HMD?. Few VR developers I talk to at developer events in Austin are excited about Daydream VR so I think they will need to do a good job reinvigorating the developers.
Regarding AR, Microsoft has sucked a lot of wind out of the Google ecosystem with a successful HoloLens and Windows Holographic OS, even though Google was first with Glass. Google should to trot out much better headsets and tools. Who knows, maybe it’s time to trot out Google Glass 3? Google Glass 1 is the consumer-focused device sitting in one of my boxes in my garage, Glass 2 is the less talked about one in developer’s hands right now.
Related to all three (AI, AR/VR, cloud), we very well could see some real details on autonomous cars via Waymo. This pulls together AI, AR and the cloud, so as a show of technical bragging rights, we could very well see details. Companies like NVIDIA and Intel are starting to get more credit than Google for powering autonomous cars, so Google should trot out Waymo details.
Google already had their enterprise show at Next 17, so while Google I/O isn’t enterprise focused, I think Google should reiterate some of the key messages and products they sent at Next 17.
Android apps on Chrome OS: Last year, Google announced Android apps on Chrome. Google must address the “less-than-optimal” experience of Android apps on Chrome OS. I spent months with Samsung’s latest Chromebook Plus and the experience felt slow and confusing.
As for Hangouts Chat for enterprise, only a few customers are able to test it and therefore must not be fully baked. Slack and Cisco Systems Spark is fully released, Microsoft Teams is fully available to all customers, but Hangouts Chat isn't broadly available or complete. I don't quite know what the delay is, but it would be nice to be able to test. Google should provide an update on Chat and Jamboard.
Obviously we should hear more details on Android O. I’d like to hear Google address how it will attack the next billion smartphones, many of which today are flip phones in emerging regions. Ubuntu phone OS dropped out of the race, Microsoft isn’t interested in it, Apple either, so this space should be target-rich for Google.
Google last year how it decoupled security updates from base OS updates and the company should provide updates on it and how it is working. Compared to iOS and even Windows 10, Android is way behind on user upgrades driven by handset makers unwilling to make the investment into newer operating systems and carriers unwillingness to update their payloads.
AI is great, but we can’t forget that like Facebook, Google’s primary business is advertising. Therefore, it shouldn’t be a surprise that everything Google does, with the exception of some of its advanced development projects, is designed to build better personal profiles to increase ad CTR (click-through rate) or create platforms to deliver advertising.
AI exists at Google to make better user experiences so you will use Chrome, Android, Pixel, Google apps and services more often and disclose, share and upload more information. Newer, machine learning-based AI, are exceptional at parsing important profile information from all pictures and videos.
I’d like to see Google discuss how it is clearly disclosing how it is using AI, what it is doing with those very personal details, how it is secured and disposed of and how this consumer data is kept separate from enterprise data.
After having attended both Microsoft’s and Facebook’s developer events, Google I/O should round out much of the field until Apple’s WWDC event in a few weeks. I’m looking forward to being dazzled and if not that informed. I’ll write about the show here in a few days.