Google I/O is coming and it's time to get excited! It's like Christmas in June! It will be here in just a few short agonizing weeks - and we need to prepare. There is background information you need to know, rumors you should have in mind, and past announcements and acquisitions that need to be remembered. Google always leaves little news breadcrumbs for those that pay attention, and I pay attention.Fanatically.
This post will be part news recap, part rumor roundup, and part speculation. The last time I did this went pretty well, and now it's time for another look at what the little elves at Google HQ are working on. And believe me, they have been busy. I hope you packed a lunch.
Jelly Bean!
Google I/O 2011 was all about Ice Cream Sandwich, so 2012 should be all about Jelly Bean. ICS came out at the end of 2011, so, if Google sticks to their typical 6 month release cycle, Jelly Bean should be fast approaching.
What do we know about Jelly Bean? Well, specifically Jelly Bean? Not much. But this whole article is full of potential Jelly Bean features, so keep reading!
Google Assistant (AKA Project Majel)
Google Assistant is going to be a HUGE launch for Google. If the rumors pan out (which have been frequent, and mostly very consistent), we're talking about the culmination of Search, Google's voice recognition technology, and several acquisitions, along with integration with almost every other Google product. It's going to be a big deal, so I'm going to talk about it, a lot. Starting with some context:
Search Under Siege
Google does not take threats to its core business lightly.
About a year ago, Google had a very public company-wide panic attack over the success of Facebook. In a memo leaked to the public, it was reveled that Larry Page had tied 25% of every employee's annual bonus to the success of Google's social strategy. This was an unprecedented move that sent a message to the entire company: Facebook is scary, it's time to get serious about beating them.
There were primarily two reasons for the social hysteria: 1, Facebook knows a lot about you and could potentially offer better ad targeting than Google (but it turns out Facebook sucks at advertising), and 2, (the relevant one) is the theory that more time on Facebook means less time on Google. You could, potentially, just ask your friends about something instead of doing a search. Fewer searches = less ad revenue.
Just the idea of a potential threat to Google's Search business set the company on a massive, all-hands social coding spree. In about a year they created a social web site from scratch and integrated it into every product in their portfolio. Google can seriously move when it needs to.
The thing is, most of the Facebook threat is arguably theoretical, and largely insignificant for the near future. You could make the case that more time on Facebook doesn't mean less searches, and that Facebook is mostly a useless cesspool.
It's important to keep Google's Facebook freak-out in mind when talking about Google Assistant. Because while Facebook's user stealing abilities are suspect, a very real, very popular threat exists today, and it has shipped on millions of phones: Siri.
Siri is Apple's virtual assistant that ships on the iPhone 4S. You talk to it, and it talks back. Siri could be described as competing with 2 Google products: It's obviously a Voice Actions competitor, in that it responds to voice commands for apps, but it's also, sometimes, something way scarier, it's a search engine.
That's right, Siri is a search engine. It's actually better than a search engine. You can ask it a question and you'll (sometimes) get a direct answer, where Google would traditionally just offer you a page full of links. Not having to dig though links is a big advantage over traditional search, it makes Siri easier and faster than Google.com.
I can guarantee you 90% of these Siri queries (heh) would have been Google searches if Siri didn't exist. Siri's answer service is a direct competitor to Google.com. It's real, it's here, and it's stealing users right now.
Whether Siri is any good or not is irrelevant. What matters it that it's scary. Apple shipped a super popular product (the iPhone) with a headlining feature that competes with Google Search. Think about the reaction to just the hypothetical threat of Facebook and tell me Google isn't in full-on, armageddon meltdown-mode right now over Siri. Siri needs to die.
Now, killing Siri shouldn't be particularly hard for the algorithmic wizards at Google. Once the "Assistant Wars" kick off, Siri won't be able to put up much of a fight, because, once you step out of the "technology is magic" Reality Distortion Field, you'll realize Apple controls very little of what makes Siri work. The voice recognition is powered by Nuance Communications, (who's cofounder, Mike Cohen, now works used to work at Google and helped build their speech tech) and the answer results are mostly queries of Wolfram Alpha. Siri is really just the glue that ties several third party services together.
This lack of control should translate to a lack of improvement. So unless Apple goes on a spending spree (and I really don't understand why they don't) they really can't do anything to make the voice recognition better, nor can they do much to improve the information services Siri relies on. In fact, Siri is such a Non-Apple product that Samsung has taken the exact same pieces (Nuance and Wolfram) and built a competitor. So now Google has two major phone manufacturers funneling searches to Wolfram Alpha. It's time to get serious. Once the real competition shows up, Siri should be down for the count.
Google's Answer For Answers
That "real competition," according to the rumors, is called "Google Assistant." I think the other name it goes by is much more enlightening: "Project Majel," named after Majel Barret-Rodenberry, the voice of the Star Trek computer. That should give you a good idea of what they want to build. Google isn't just out to smack-down Siri and Wolfram, they're out to replicate the freaking Star Trek computer. They probably won't come out with a full, working replica at this year's I/O, but Google Assistant is the first step toward that larger vision.
The real problem Siri and Wolfram Alpha have created isn't their current execution (which should be easily surpassed by Google), the problem is that they've changed the search game. It's no longer only about a set of links. Links are great for research, but if the user is asking a question, it's much more helpful to directly answer it.
I think Google is going to take that ball and run with it. They will want to provide you with direct answers for everything, and Google Assistant will handle all that verbally, just like Star Trek. They aren't just building a Siri killer, they're building an entirely new interface for Google. Google.com will start providing you direct answers whenever possible, and Assistant will read those answers to you.
Google's Searchketeers (are they not called that? they should be called that) have been working on direct answers for years. Take a look at this page and you'll see Google has been able to directly handle math equations, unit conversion, currency conversion, the weather, word definitions, movie show times, stocks, time, and flight information for a while. Slap a text to speech engine on top of that and you have something very similar to Siri.
Recently though, Google Search was augmented with a massive database project that Google is billing as "the next frontier in Search." Google calls it "Knowledge Graph," and the project is all about making Search more intelligent. The underlying goal is to understand what a query actually means, and to understand the relationship between queries. Right now though, it pops up a nice little summary box whenever you hit on a database item:
That, right there, is a big part of Google Assistant's brain. Just imagine asking "Who is Majel Roddenberry?" and Assistant rattling off the info in this box with a text to speech engine. Very Star Trek. This is already more info than Wolfram Alpha provides, and this is just their first release.
If the Knowledge Graph/Google Assistant connection wasn't obvious enough, Google left us a little breadcrumb in the Knowledge Graph announcement post. In the closing paragraph, they sign off with:
So there you go, Knowledge Graph is a part of Google Assistant/Project Majel. Now that you fully understand what they're talking about, watch this video while keeping in mind that this is reallyabout Google Assistant.
So now what we know what it will say, what will it sound like? I certainly don't want to listen to something that sounds like my current Android TTS all day...
A Voice That Doesn't Suck
Whenever your phone "speaks," (for instance Navigation directions, or in the case of Google Assistant, everything) it's really using a text to speech engine. A TTS engine saves you from having to record every combination of words in the English language, you can just generate them on the fly with a computer program. Usually, TTS sounds like ass. The robotic voice of most TTS engines is grating, hard to understand, and not something you'd want to interact with all day like you would need to for Google assistant.
In my Ice Cream Sandwich preview article the one thing I completely whiffed on was better text to speech. My reasoning was, 9 months ago (at the time) Google had purchased Phonetic Arts, a voice synthesis company. PA's PR claimed they were at the forefront of speech synthesis research. I figured 9 months was long enough to integrate PA's speech tech, and Google would have it prepped and ready as a new TTS for ICS. Yep, I was claiming they went out and bought an entire company so that navigation would sound better. It all sounds so naïve now, doesn't it? Anyway, that didn't happen. What we got was a different TTS, but not a better one. It certainly wasn't the "completely fluent, perfectly natural” speech tech Phonetic Arts was claiming they had.
So, knowing what we know now, let's take a look at Google's PA acquisition blog post again, shall we?
OH. MY. GOD. They've been dropping Project Majel hints for almost a year and a half. Those diabolical bastards! So that's what the PA acquisition was for, it's the Google Assistant voice. Things are certainly starting to come together.
Like I said in the ICS article, most information about the company has been lost to the ages. Googlecompletely wiped the site when they took over, and the only thing The Wayback Machine can come up with is their "we got bought by Google" park page. I wish we knew more.
Phonetic Arts was, pre-acquisition, focused on using their voice synthesis tech in video games. They claimed they could replicate an actor's voice from speech recordings, so if you wanted a TTS engine of Majel Barret, you could actually do that (personally, I would go with Jarvis). You can check out some of their promotional materials here. They seemed to think their product was so good you could use it in lieu of an actual person.
A "Do" Engine: Voice Actions Upgraded, And Developer Support
I've mostly been talking about the answer service, but Google Assistant is supposed to help you dostuff too, according to this rumor from Techcrunch:
Aside from it just making sense, this is very believable, mostly because #1 is a spot-on prediction of Knowledge Graph. It even uses the same language as the Googlers in the announcement video.
The "Do Engine" makes perfect sense, because that's just an upgrade of Voice Actions. Assistant should be a complete verbal interface to the rest of Google: do a search, send with Gmail, make a Calendar appointment, post a G+ update, ask it where your Latitude friends are, get a stock quote from Finance, find a business with Maps, send a message with Talk or GV or G+. All of this plus the Knowledge Graph answer service will cover just about anything you could ask.
A big, overarching project like this would greatly benefit from Google's recently unified privacy policy. Google use all the Google services to build a profile about you. Take, for instance, the example they gave at the end of the privacy video: "It may even mean we'll be able to tell you when you're late for a meeting, based on your location, your calendar, and local traffic conditions." that definitely sounds like something you would want a smart, helpful assistant to do.
And if the Google stuff doesn't cover everything you could ask, the other really believable part of this rumor has got you covered: developer support. Turning Google Assistant into a platform perfectly meshes with Android's intent system, and it'll be a great feature that no other assistant app can claim. So you'll eventually be able to post on Facebook and Twitter or do anything else you can imagine with your voice.
TechCrunch has Assistant pegged for a Q4 2012 release, but Assistant needs to be at I/O because it is a platform, they need to show this off to developers so they can get their imagination going.
Google Games: A Mega-Merger Of G+, Chrome, And Android Games
The Game Developer's Conference is a massive get-together of major video game developers from all over the world. It features the likes of Nintendo, Sony, Microsoft, EA, Activision, and a million others. If you're a heavy hitter in the game industry, you go to GDC. This year though, there was a relative unknown in the gaming world in attendance: Google!
This year, Punit Soni, the group product manager for Google+, went to GDC, stood in front of the world's top game developers, and declared:
Google Games! A mega-merger of every Google gaming property out there. Google wants to be taken seriously in the gaming space, and combining mobile, social, and the browser into one giant platform would certainly be a big step in the right direction.
Let's examine just what we're talking about here. Google Games would be a massive store housing 4 different development platforms: the Android's SDK, for making java-based apps; the Android NDK, for native apps which will only run on compatible devices; Chrome and Google+ HTML5 apps, which will run on any compatible browser, and Native Client (NaCL), which is a special Chrome plugin that allows you to run native code in the browser (go try Bastion).
If that sounds like a mess, it's because it is. What would they do with that many platforms under one roof? Would everything run on everything else? G+, Native client, and Chrome HTML5 games on Android and Android SDK and NDK games on Chrome? A setup like that would be pretty mind blowing and require a whole mess of emulator magic and horsepower. Not to mention you'd have some games that need a keyboard and mouse, some that need multitouch.
Native Client running everywhere doesn't sound too far fetched. Their product seems custom-built to handle a multi-platform scenario like this. The NaCL team even has "ARM Chrome" support (aka Chrome on Android) on their roadmap.
This video specifically mentions that Star Legends, an Android/iOS native game with a half a million lines of code, was ported to Native Client by one engineer in only two weeks. So is that the solution to the Google Games platform nightmare? Release an Android Native Client implementation, and tell everyone to build games on that? One binary running on Windows, Mac, Linux, Chrome OS and Android, with all sorts of Google+ integration, sounds pretty sweet for developers.
The other solution is to just fragment the hell out of the Play Store. You could have separate sections for Android, HTML 5, and Native Client games but that just seems like it would be making things worse, and it would be confusing for users. It's time for Native Client to shine.
The Google Game Center, Powered By G+
Google is also supposed to be developing a gaming social network, according to Business Insider (or, in their typical Apple-centric parlance: "An iOS Game Center clone." I guess they've never heard of Xbox live).
BI says the service will provide "social" achievements and leaderboards, but isn't sure if Google+ will be involved. Really though, these days, when isn't Google+ involved? Imagine Google bringing OpenFeint or iOS Game Center functionality in-house with all the identity and social stuff handled by Google+. BI doesn't say this, but "Game Center clone" implies the service will handle multiplayer matchmaking too, which would certainly make sense. This one must give the OpenFeint guys nightmares, and good riddance.
But wait, there's more! Venture Beat, in the article reporting on Google's GDC appearance, also had some interesting statements from the G+ product manager that seems to line up nicely with this:
So now we're even more into Xbox Live territory, one-upping their gaming voice chat with gaming video chat. Even simple games like online poker get more interesting when you can see your opponents. They even mention Native Client! I think we're on to something here.
It's almost hard to imagine Google hitting the gaming scene this hard. A single binary that runs across all its products, with achievements, leaderboards, identity and matchmaking all powered by a sizeable social network. This provides a boost for Chrome, Google+, and Android. If they can pull it off it will be amazing.
Project Glass
Project Glass is Google's wearable, heads-up display tech. It's being cooked up by Google's "crazy idea" division, known as "Google X." The video above is the target for what Project Glass hopes to look like some day. The reality of the matter is that, today, Glass is barely functional. Because it's in the very, very early stages, it might not even make it to I/O, but I'll include it in this article for completeness sake.
Sebastian Thrun, the head of Google X recently went on Charlie Rose to talk about Project Glass and and some of his other products.
Because the user interface is only visible to the wearer, you can safely use the glasses in public without revealing anything about them - which, apparently, makes these the latest, must-have fashion accessory at Google HQ. So while they may not be officially talked about, I expect lots of Googlers to be running around sporting the futuristic headgear. They just love to tease us.
The (Cheap) Nexus Tablet
Pop quiz! What's the beat selling Android tablet out there? The Galaxy Tab? Nope. Asus Transformer Prime? Wrong. The Xoom?! Hahahaha, no. Give up?
It's the 7-inch Amazon Kindle Fire. It's actually the Kindle Fire by a lot. According to Comscore,Amazon owns over 50% of the Android tablet market. This is a big problem for Google, because theKindle Fire is missing the Play Store, Gmail, Maps, Music, Books, Movies and all the other stuff that makes Google money. It's basically the nightmare scenario for AOSP, a popular, forked version of Android with all the Google bits cut out.
The Fire's success isn't a big secret, there's really nothing special about it: The browser isn't very good, the Amazon App Store is a mere shadow when compared to the Play Store, and the software is based on Gingerbread, so it can't even run tablet apps. The Fire does one thing right: it's cheap. At $200, it's less than half the normal going rate for a tablet.
So Google needs to fight this, because, 1, they are losing money, and 2, they need to send a message to all the other OEMs: "Forking Android is a bad idea, and we will crush you if you do it." You can bet the OEMs are paying close attention to the Kindle, to see if mutiny really is a viable option.
So, expect a 7-inch Nexus tablet built as cheaply as possible. Eric Schmidt even gave us a timeline:
That was in December, which means 6 months should put the release right around I/O, barring any delays. Eric Schmidt occasionally says some crazy things, but it's hard to imagine him messing up something like this. We aren't totally relying on Eric Schmidt though, we had the old CEO of Google giving us the timeframe, now we've got the new CEO dropping hints as well:
Translation: Yes, the Kindle Fire is selling well, we're working on a competitor.
Asus is the rumor mill's favorite candidate to build the Nexus tablet, mostly because they already showed off the MeMo 370T - a 7 inch, quad core tablet for $250 - which seems to hit all the sweet spots needed for a Nexus tablet. After the showing at CES, 370T suspiciously disappeared. The rumor mill claims it was scrapped in favor of a Google collaboration. Sounds plausible.
We've also recently gotten word of Nvidia's "Kai" Platform, which is all about cheap, quad core tablets. Base the Nexus tablet on that and you're down to $200. Sounds about right to me.
Speaking of Nexus Devices:
A Whole Family Of Unlocked, Google-Sold Nexus Devices
The Wall Street Journal has a pretty impeccable record when it comes to Android rumors. By my count, they successfully called the launch of Drive, the Music Store, and device sales in the Play Store. So when they float a rumor, I listen.
The latest rumor from the WSJ brings news of a whole family of Nexus devices:
It's like a brilliant phoenix rising from the ashes of the disbanded Android Alliance. A portfolio of Jelly Bean packing Nexus devices, with at least one from every major manufacturer (can you even name 5 OEMs?), that will be sold unlocked, directly to consumers in time for Christmas. Wow.
Anyway, two thoughts: 1, It's going to be awesome to be able to pick from a large group of devices. I'm not all that satisfied with the Galaxy Nexus hardware, but I have to have stock Android. It's really a shame I have to stick with it because it's the only stock Android device available. More phones with the latest updates and more options for getting the One True Version of Android will be fantastic. Once you go Nexus you never go back.
One last thing, if you aren't giddy enough already: I've seen this reported in some places as "5 Nexus devices" but the report actually says "as many as five manufacturers." There's nothing that says a manufacturer can't build a tablet and a phone, so you could end up seeing more than 5 Nexus devices. Either way the device section of the Play Store is about to get a whole lot more crowded.
And it looks like Google is gearing up for a bigger device section on the Play Store, as I pointed out on my Google+ page (you should totally circle me), Google recently added a link to the "Devices" section on the front page. It's now a first-class section with a color scheme and icon. It definitely looks like a focus on selling devices is coming. Here's hoping the second time's a charm.
Google's Home-Entertainment System
Remember the WSJ's aforementioned impeccable rumor record? Well here's another one to throw into the pile. This time they claim Google is going to release a hardware component to Google Music.
Also interesting is that Google, for the first time ever, is dipping their toe into the commercial hardware waters with this device. They're designing it themselves and it will be a Google branded device.
The WSJ specifically called out Sonos Inc. as a target, and one of the products they make is basically a Wi-Fi radio. Plug it in, hook it up to Wi-Fi, and you can stream Pandora, iTunes, and Spotify, and it's all controlled from a phone or tablet. Sonos' pricing definitely leaves some room for improvement. They're asking a whopping $300 for the smallest, speaker-equipped device (the one on the right). Google thinks that's a little high, too: the WSJ says "Google was hoping to offer the entertainment system... at more-affordable prices than a similar device made by Sonos Inc."
At Google I/O 2011, Google demoed Project Tungsten. A home music player, which is sort of along the same lines as this rumor.
Tungsten was a self-contained Wi-Fi Google music player, and "a bridge to the Android@Home Network" (whatever that means). It was a cool little tech demo, but did anyone expect this to be an actual product? I suspect Tungsten will be the "father" of whatever this home-entertainment device is. The I/O 2011 design is a little raw; the amp+speaker configuration will probably be cut for an all in one device or a box-only Google Music bridge, where you use your existing sound system.
The WSJ also says "In the future, such a device could potentially stream other forms of digital media such as video, one of these people said." Which makes it sound like a device you would add to your home theater stack. Does that seem a little redundant to anyone else? They already have a set top box, Google TV, why not built Google Music streaming into that and call it a day? The hardware guts are going to have to be very similar, there's very little difference between streaming music and streaming video. I guess if you were really going for a low price point, music-only would let you use cheaper chips, at the end of the day you still need a small computer, though. The home entertainment market (like everything else) is trending towards integration, so a separate device doesn't seem to make much sense. So if they really release this, what does this mean for Google TV?
Android@Home
Oh Android@Home, what ever happened to you? @Home was supposed to be the Android team's home automation initiative. They promised light blubs and switches with embedded wireless tech that could be controlled from an Android app, with much more goodies to come. They specifically called out building their wireless protocol into lamps, alarm clocks, thermostats, and dishwashers. The world was supposed to be controllable from an app.
Something must have happened, because due dates came and went and we didn't hear a peep out of Google about the project. The original plan was to have the first devices out by the end of 2011, which obviously never happened. The last time we officially heard about it was last I/O, so here's hoping they bring it back a year later and show us some serious plans.
There is a glimmer of hope, the Google entertainment device mentioned above was originally cooked up by the @Home team, so they are still around, and cooking up something.
Linux and Android Working Together
The one definite Jelly Bean bit of news is an upgrade to a fresh Linux Kernel. Android is built on Linux, but it has traditionally been a fork of Linux. Linux 3.3 marks the start of the Android Mainlining Project: the merging of the world's most popular Linux-based OS with the proper Linux Kernel. 3.3 should let you boot directly into an Android user space with a stock kernel, and 3.4 will bring even more Android stuff (like power management) into the fold. This should increase Android's hardware compatibility, make life easier for developers looking to port Android, and help out Linux distributions that want to support Android programs.
If you are at all interesting in why forking the Linux Kernel is a Very Bad idea, watch the first few minutes of this awesome talk from GKH, a Linux Kernel developer. Basically, Linux development is scary fast, like 7600 line changes per day fast, and you want to be on that train, not chasing it.
So life for Googlers should get a lot easier. You know how OEM modifications slow the upgrade process when Google releases a new version? How they have to re-port all of their changes to the new OS, and it usually involves lots of duplicated effort? Google has basically been doing that, this whole time, with the Linux Kernel. Having Android be officially supported should allow our beloved Android engineers to spend less time merging all those kernel changes and more time developing awesome features.
UI Scaling For External Displays
If you remember this, good for you! You're paying attention! This is a holdover from my 2012 preview article. Basically, with Ice Cream Sandwich came the idea that you could support tabletsand phones with the same software. In fact, a quick settings tweak and you can turn your ICS phone into a tablet! Same code, same apps, different layout.
The above example requires a reboot, but imagine if you didn't need a reboot, and when you jack your phone into a bigger screen (like say a TV, or a Lapdock) Android could automatically switch to tablet mode. That's the idea that was floated to the Android team at the last I/O:
Turns out, they've been thinking about it! They're worried about the impact on developers, but with a full year to think it over, I bet they've come up with something.
Asus, it turns out, is really impatient, and has whipped up their own version of this for the Padfone.
Bringing this in-house would be a great boost toward a "one device to rule them all" future. Combine this with some kind of wireless screen sharing, and it would be really nice if Android could automatically take advantage of a bigger screen when it is available. We never really got a timeframe for this, but, like they said, they're working on it.
Local Recommendations And Discovery
Late last year, Google bought a company called "Clever Sense". Their sole product was Alfred, a restaurant recommendation app. They billed themselves as a "Pandora for the real world." You tell it what restaurants you like, and it will recommend you other places based on the food type and other people's ratings. It will even do group recommendations, so if you are with a group of Alfred users, you can even tell it who is in your group and it will mash everyone's profiles together and try to pick something everyone will like.
If this sounds interesting to you, you can still give it a try! In stark contrast to the "Shut down everything" mentality of the Phonetic Arts Acquisition, Clever Sense's web site still works and the Alfred app is still up and running in the Play Store. You can visit Alfred in his acquisition purgatoryright here.
So, Pandora for restaurants. What do you do with this? The most obvious move is integration with Google Maps. In Maps, you already rate stuff and leave comments, so it being able to recommend restaurants seems like a no-brainer. Plus that restaurant list on the right looks like a great spot for an ad! (If you worked at Google that would excite you.)
All The Other Stuff
Amazingly, this isn't even the half of it. I haven't even mentioned all the other things due for an I/O appearance: Google TV, Chrome, Chrome OS, YouTube, Google Drive, Commerce, or Maps. All of this stuff is in the I/O Schedule and we have no idea what Google will talk about.
So a new version of Android, a huge step forward in voice control, a massive gaming initiative across all Google platforms, some Project Glass sightings, 5 (or more) Nexus devices that cut out the carriers, a home entertainment device, Android@Home, a merge with Linux, UI scaling, and local recommendations. If even half of these rumors pan out it's going to be an absolutely crazy show!
I/O starts June 27th and lasts until the 29th. We'll be counting down the days.
0 comments:
Post a Comment