Skip to content →

Tag: Google I/O

Thoughts from Google I/O

A few weeks ago, I had the chance to attend Google I/O — this time, not just as a fan of the Android platform but representing a developer. Below are some of my key takeaways from the event

  • Google‘s strategic direction – there were three big themes that were emphasized
    • Next Billion – a lot of what Google is doing (like making Google Maps / YouTube work without internet) is around making Chrome/Android/Google Search the platforms of choice for the next billion mobile users — many of whom will come from Brazil, India, China, Indonesia, etc. Its important for us to remember that the US/Western Europe is not the totality of the world and that there’s a big chance that future major innovations and platform will come from elsewhere in the world.
    • Machine Learning – I was blown away (and a little creeped out!) by the machine learning tech they showed: Google Now on Tap (you can hold the home button and Android will figure out what’s on your screen/what you’re listening to and give you relevant info), the incredible photo recognition tech in the new Photos app, as well as innovations Android is making in unlocking your phone when it knows its been in your pocket and not your desk. Every company should be thinking about where machine intelligence can be used to enhance their products.
    • Everything Connected – it reminded me of Microsoft’s heyday: except instead of Windows everywhere, its now Android/Chrome everywhere: Android Wear, Chromecast, Android TV, Android Auto, Brillo/Weave, Cardboard for VR, Nest/Dropcam for the home, things like Jacquard & Soli enabling new user interfaces, etc.
  • Marketing enhancements to Google Play: Google has taken steps to make application developers’ lives easier — more details here: http://android-developers.blogspot.com/2015/05/empowering-successful-global-businesses.html, but:
    • I sat through a panel on how Google does personalized recommendations / search on Google Play — long story short: keywords + ratings matter
    • Google will now allow A/B testing of Google Play store listings
    • Google Play console now directly integrates App Install advertising so you can run campaigns on Google Search, AdMob, and YouTube
    • Google Play console will also track how users get to Play Store listing by channel and how many convert to install
  • Android M – a lot of tweaks to the core Android app model for developers to pay attention to
    • Permissions: Android M moves to a very iOS-like model where app permissions aren’t granted when you install the app but when the app first uses them; they’ve also moved to a model where users can go into settings and manually revoke previously granted permissions; all Android developers will need to eventually think about how their apps will work if certain permissions are denied (see: http://developer.android.com/preview/features/runtime-permissions.html)
    • App Links: Android will now let apps handle all links on websites they control by default (see: http://developer.android.com/preview/features/app-linking.html)
    • Doze and App Standby: Applications will now have two additional modes that the OS may enforce — one called Doze that keeps all apps in sleep mode to reduce power drain and Standby where the OS determines an app is “idle” and cuts off network access, syncs, and jobs — apps in both modes can still receive “high priority notifications” (see: http://developer.android.com/preview/behavior-changes.html under Power-Saving Optimizations)
    • Auto Backup: Applications will now backup up to 25MB worth of data to the user’s Google Drive (but won’t count against their quota) once every 24 hours; this can be customized (see: http://developer.android.com/preview/backup/index.html)
    • Fingerprint API, Direct Share, and Voice Interactions: universal fingerprint recognition API, the ability to share specific content with specific favorite users (i.e. send to someone over Facebook Messenger, etc), and a new way to build voice interactions in app (see: http://developer.android.com/preview/api-overview.html, starting from Authentication)
  • Other stuff for developers
    • App InvitesGoogle has built out custom share cards / install flows and deep links to make it easier for users to share apps with their friends: http://googledevelopers.blogspot.com/2015/05/grow-your-app-installs-with-app-invites.html
    • Android Design Library: Google now has libraries to help devs build out Material Design elements — now, you too, can make your own Floating Action Button!: http://android-developers.blogspot.com/2015/05/android-design-support-library.html
    • Chrome Custom Tabs: basically lets you embed Chrome in your app with custom styling (rather than having to embed a vanilla webview and do a lot of work styling it), its apparently already out in beta channels for Chrome: https://developer.chrome.com/multidevice/android/customtabs
    • Google Cloud Testing Lab: This was pretty cool (and a product of Google’s acquisition of Appurify). Now, Google will provide two highly useful testing services for Android developers: (more details: https://developers.google.com/cloud-test-lab/)
      • For free/automatically: pound on every button / interface on your app that they can see after launch for 1 min and see how many crashes they can get on a variety of Android devices (which helps given the sheer number of them that exist)
      • Paid: run custom Espresso or Robotium tests on specific devices (so you can get test coverage on a broader range of devices doing a specific set of things)
    • Places API: a lot of talks promoting their new mobile Places APIs (which will let iOS and Android apps have better mapping and place search capability)
    • Google Cloud Messaging: this is basically Google’s push notification delivery engine and they announced support for iOS as well as “Topics” (so devices don’t have to get every notification, just the ones relevant to them): http://android-developers.blogspot.com/2015/05/a-closer-look-at-googleplay-services-75.html
    • Espresso testing framework: this was a ridiculously packed session — but Google has apparently made numerous refinements to the Espresso UI testing framework
  • A lot of cool announcements about new Android Wear functionality (which my Moto 360 is eagerly awaiting)
  • Just cool stuff from ATAP
Leave a Comment

A Month with the Chromebook Pixel

Last month, I had the pleasure of attending Google I/O – Google’s annual developer conference and product geekfest. To put it simply, it was probably the nerdiest conference I’ve been to (and yours truly has been to some really nerdy conferences) with Google Glass users everywhere and flying, internet-controlled, camera-connected dirigible floating above the conference floor among the attractions.

One of the things that Google tried to emphasize to I/O attendees was the growing idea of Chrome, Google’s web browser, as a key platform for developers to embrace. Part of that message, of course, came from the talks and sessions where Google promoted Chrome’s widespread adoption (if you count mobile deployments, Google claims 750 million users worldwide) and proudly touted Chrome’s support of both sophisticated open technologies like HTML5, WebRTC, and WebGL, as well as proprietary-to-Chrome technologies like Native Client and their new Packaged Apps capability.

Equally (or perhaps more) effective was the conference’s giveaway of its Chromebook Pixel (not to mention some pretty interesting artistic displays showing off the device and its capabilities, see below).

My generally positive take on the Pixel’s predecessor Samsung’s Series 5 Chromebook is one of the more popular posts on this blog and so I thought I would share my take on Google’s latest and greatest. In a nutshell, I will say that the Chromebook Pixel is light years ahead of its predecessors and is an amazing device which hints at the potential of well-built Chrome OS hardware, albeit one which is probably not worth the $1200+ price tag:

    • Good, not good enough, performance: While the Series 5 routinely stumbled and hiccuped, the dual-core Ivy Bridge processor in the Pixel, while not the fastest chip around, was up to the task of almost any large web workload I threw at it – multiple tabs with Netflix and complex webapps like Tweetdeck and Gmail and Feedly running. Even Evernote, which I had not been able to get working on the older Chromebook, worked without any problems on the Pixel.
    • Amazing display: In the same way that other remarkably high resolution displays make you want to view more content (Nexus 10, Retina Display Macbook Pros), the Pixel has actually managed to steal web browsing and video watching time from my tablets, something I didn’t expect would happen.
    • Touch: I used to be a big skeptic of the importance of touchscreen displays on laptop form factors – no more. As cheesy as it sounds, the type of relationship you have with content is different when you can use touch gestures to zoom in/out and scroll up/down versus using arrow keys or a mouse. I can’t say that I primarily use the touchscreen in navigation, but it’s a nice touch (pun intended).
    • Much better industrial design: I don’t claim to be an ID expert, but the attention to detail on the machine is decidedly impressive for a company that many in the tech industry for years felt just didn’t care about design quality. The touchpad beats most of what the PC industry has put out in feel and responsiveness (although that’s a low bar to beat) and, taking a page from Apple’s playbook, supports multi-finger gestures. The device body is smooth aluminum with only a groove on the body for cool-looking LED lights to come out as a signal that the device is on and an interesting piano hinge for the display which someone engineered to function not only as a hinge but as a heat sink and Wi-Fi antenna. Simply put: it doesn’t feel or look cheap.

Couple that with the advantages I described to all Chrome OS systems (rapid boot, easy multi-user support, frequent and automatic updates, syncing tabs/histories/passwords with all your other Chrome browsers), and I think you have a fairly compelling device.

That said, three major problems are worth calling attention towards:

    • This is still just a browser: granted, most of what we do today is in or can easily be replaced by web-based applications of some form or the other, but, this won’t be playing Starcraft or running Excel or operating a server or doing software build work.
    • Underwhelming Battery life: for an operating system that is effectively a browser, I am surprised that my typical battery life is somewhere in the 3-4 hour range, and significantly lower if I’m using Netflix or YouTube. I can’t tell if this is simply an issue where Google included too small of a battery to save costs, if this is the energy from the extra processing power and backlight needed to run such a high-resolution screen, or if this is a operating system/firmware bug where the video codecs aren’t being used properly, but this is something that will likely need real improvement.
    • Extremely high price: while this is a fantastic device, its usage limitations (to basically being a big browser) and storage and memory and battery life limitations don’t make this a $1200+ machine. Interestingly, I do feel that if they included a dual-boot to Linux option, the screen and industrial design could very well justify a higher price (compare with Linux laptop vendor System76’s new Galago UltraPro)

So, the verdict? I am extremely happy I got this device for free from Google. It’s something I use regularly because it is a delight to use and really does put forward Chrome in a fantastic light for developers (which is really the purpose of the giveaway at Google I/O). This device is also probably more than enough for what the average computer user needs (who is mainly interested in checking email, reading articles, watching videos, and playing webgames) and has unique advantages for enterprise/educational settings. But, the fact that Chrome OS still can’t do everything that I need it to do and has limitations in battery life and storage and memory make it difficult to justify the high price for a regular consumer purchase.

Any other Chromebook Pixel users out there care to share their perspectives?

Leave a Comment

Android Bluetooth (Smart) Blues

Readers of this blog will know that I’m a devout Fandroid, and the past few years of watching Android rise in market share across all segments and geographies and watching the platform go from curiosity for nerds and less-well-off individuals to must-support platform has been very gratifying to see.

Yet despite all that, there is one prominent area in which I find iOS so much better in that even I – a proud Fandroid venture capitalist – have been forced to encourage startups I meet with and work with to develop iOS-first: support for Bluetooth Smart.

LogoBluetoothSmart

In a nutshell, Bluetooth Smart (previously known as Bluetooth Low Energy) is a new kind of wireless technology which lets electronics connect wirelessly to phones, tablets, and computers. As its previous name suggests, the focus is on very low power usage which will let new devices like smart watches and fitness devices and low power sensors go longer without needing to dock or swap batteries – something that I – as a tech geek — am very interested in seeing get built and I – as a venture capitalist — am excited to help fund.

While Bluetooth Smart has made it much easier for new companies to build new connected hardware to the market, the technology needs device endpoints to support it. And therein lies the problem. Apple added support for Bluetooth Smart in the iPhone 4S and 5 – meaning that two generations of iOS products support this new technology. Google, however, has yet to add any such support to the Android operating system – leaving Bluetooth Smart support on the Android side to be shoddy and highly fragmented despite many Android devices possessing the hardware necessary to support it.

To be fair, part of this is probably due to the differences in how Apple and Google approached Bluetooth. While Android has fantastic support for Bluetooth 4.0 (what is called “Bluetooth Classic”) and has done a great job of making that open and easy to access for hardware makers, Apple made it much more difficult for hardware makers to do novel things with Bluetooth 4.0 (requiring an expensive and time-consuming MFi license – two things which will trip up any startup). Possibly in response to complaints about that, Apple had the vision to make their Bluetooth Smart implementation much more startup-friendly and, given the advantages of using Bluetooth Smart over Bluetooth Classic, many startups have opted to go in that direction.

The result is that for many new connected hardware startups I meet, the only sensible course of action for them is to build for iOS first, or else face the crippling need to either support Android devices one at a time (due to the immaturity and fragmentation in Bluetooth Smart support) or get an MFi license and work with technology that is not as well suited for low power applications. Consequently, I am forced to watch my chosen ecosystem become a second-class citizen for a very exciting new class of startups and products.

I’m hoping that at Google I/O this year (something I thankfully snagged a ticket for :-)), in addition to exciting announcements of new devices and services and software, Google will make time to announce support for Bluetooth Smart in the Android operating system and help this Fandroid VC not have to tell the startups he meets to build iOS-first.

One Comment