Presumably you have also seen the IOS keyboard with google (and ads) built right in.
The pattern is pretty clear, one way to drive traffic and ad revenue for the company is to create ways that get you to use Google before you use the platform's chosen search provider, if you do it in Apps you get the "bonus" that you both get the traffic and you don't have to provide the platform provider for sending search traffic your way.
The way this plays out is going to pretty interesting. It gets really interesting if Apple deploys their own web search capability and starts freezing out apps that have a search component and don't use theirs (like the Microsoft phone ecosystem was rumored to do).
Really challenging environment for Google to be in. I'll be expecting lots of similar sorts of things from them in the future.
Google knows that this is a developer conference right? I feel like I'm at CES. I don't care about any of their apps unless there is some really cool integrations I can do. Talk API's or this is just noise.
Currently listening to details about full disk encryption, when/how shaders are compiled for graphics improvements, SELinux improvements, etc. Seems like a developer conference for sure!
I know! Finally! Of course I can't help but see that the message Google took away from Stagefright was that they needed to harden their media frameworks rather than needing OS level security updates for all devices.
All the journalists buzz around, and all the product managers/execs/reps are there to show off for them.
For example, when I went, the wi-fi was horrible, basically unusable - except if you were in media. If you were media, then you got your own area to connect to better internet.
That said, the subsequent days were noticeably quiet in comparison.
Not too many people there to actually go to dev talks, etc. It's very relaxed by then.
Well, this is the keynote, which is generally used to debut things and get people (and especially the press) excited about the new things. I'd expect the subsequent sessions will have more developer specific details.
IIRC, its pretty typical for the front end of the I/O keynote to be loaded with items that are somewhat less immediate and burning interest to developers but greater interest to consumer-oriented press, and the back end to have the big new developer-focused offerings. This keynote seemed to follow that pattern.
Next step in Allo: just automatically pick the best "smart reply" for me, add "smart questions" and we can leave both sides of the conversation to the bots. "Have your AI assistant contact my AI assistant to work out the details" :)
I don't understand why anyone would want to have such a device in their home. How long before they start data mining conversations for better ad targeting? What if a state actor gets access to such devices?
There's enabling state spying and then there's doing so knowingly and then there's doing so provably knowingly. Which do you trust?
Do you think devices like this have no remote vulnerabilities, that nation states don't have the resources to find and exploit them, or that they can but don't?
Yesterday on CNBC Jason Calacanis in response to the question.
Question: "Jason you don't worry about the data Alexa is collecting on you and your conversations that you're having unwittingly around this device?
Answer: "Your privacy is an Illusion, uhh it's been gone for many years, the NSA is listening to any coversation it wants to through your phone already. So the idea that we have any privacy at this point is sort of laughable... https://www.youtube.com/watch?v=B9fnY5rIjb4
In case you are wondering he's bullish on the Echo and Amazon
I've seen the ads my Kindle Paperwhite displays as a screensaver, and... no, I'm not worried about Amazon mining my personal information to start doing better ad targeting.
That might be more the fault of the advertisers than Amazon. Amazon knows what books you will like with a high degree of certainty, but they don't always have ads available that are a close match, and some advertisers set their targeting very broadly because they don't care (they want a wide reach, e.g. for rank boosting) or don't know better.
You're probably not too far off (although nowhere else in Amazon's ecosystem have I seen very good recommendations, and most of their advertising elsewhere seems to be just showing me stuff I browsed on Amazon earlier and didn't buy), but if the ad inventory isn't there to really incentivize that sort of targeting, that's the same thing, isn't it?
No kidding. And my fiance was under the impression for a while that it was the cover of what I was reading at the time... She was probably somewhat concerned.
> "I'm not worried about Amazon mining my personal information to start doing better ad targeting."
I get that better ad targeting may not seem like much of an issue at first thought, and may even seem like a benefit, but I'd suggest that the perceived benefits may start to unravel if you consider it from a different point of view.
Companies that are involved in delivering targeted advertising are basing what ads to show you on the profile they're building up about you and what they perceive as similar people to you. The idea being to give you more of what they think you'll like. What this leads to is a filter bubble. When you're in a filter bubble (and I'd argue most of us are), the culture you're exposed to is increasingly limited by your past likes and dislikes, increasingly tailored based on market demographics, far less likely to broaden your range of experience.
We shouldn't welcome the filter bubble, we should act to break out of it, and resisting the mechanisms behind targeted ads is one such way of doing so.
I have an Echo. It's great. To me the Amazon play for this is very straightforward: they want you buying Amazon stuff using it. They don't sell ads, they sell products. When they say that the device only listens for its own name locally, I believe it (sans someone hacking it to spy on people of course, but I don't believe that Amazon deliberately programmed it to do this).
Google is a different story. I imagine that they would try to sell a device that listens 100% of the time and targets ads at you based on what you talk about. Google makes money by selling ads, not by making buying stuff convenient. Their incentive is different.
Ideally, I'd love for Apple to create such a device, but it won't be open, would cost an arm and a leg, and would probably use Siri, which lags behind the Echo in what it can do by quite a bit now.
Amazon does sell ads, actually. A quick search would have shown you this[0], and they show up for Kindle users, on www.amazon.com, and elsewhere. It is highly likely Echo will be involved at some point.
I think the rest of your post is conjecture, but does this change your opinion? Amazon is a diverse business that wants to sell many things, including your eyeballs (or ears in the case of Echo) to others via advertising.
First off, this is all my opinion, not absolute truth.
Second, sure Amazon does ads on the side, but the majority of their money is made by selling their own products. Occam's Razor suggests that they simply want to sell you stuff using one more way. Notice that the Echo didn't do almost anything when it was initially released, except allowing you to buy stuff. That was the part that they worked out before launching it.
Also, don't forget that Amazon also has a number of other devices that are very similar: the Dash button and Amazon Fresh come to mind. The Echo fits nicely in line with it.
I am not saying they couldn't use it for something nefarious: they could definitely abuse it. I just think that they have less incentive to do so than Google.
Sure, some people are worried. But these devices only listen when they hear the wake word ("Ok Google" or "Alexa"), so there's no real concern about the companies listening to anything other than your queries. And 99% of people are not interesting enough for a state actor to care about. If you are a spy, criminal, politician, etc., then I'd agree you should think twice about owning one.
Google and Amazon are huge companies and would face serious liability for falsely advertising that their devices only listened upon hearing the wake word. Moreover, I'm sure someone out there would analyze the network traffic and figure out it did not work as claimed. They have nothing to gain by secretly transmitting data, and everything to lose.
State adversaries are a different story. I don't doubt the NSA might somehow break into Snowden's Echo, but most people don't have such adversaries.
I think it would be entirely possible within US legal framework for Echo/Home to be able to selectively switch from "wake word" to "always listening" mode, on the whim of Amazon/Google or a request of the government. Now, if the government isn't stupid about it and would use it selectively, then good luck tracking it down.
You know what I would want my phone to be doing? To be actually always listening. So that I could just say "Ok Google" at random and interact with Google Now without having to unlock the phone or have it constantly hooked up to a charger. That very simple thing would probably quadruple the usefulness of the whole thing. But Google probably has some "smart" reason for never ever allowing that. sigh
(I mean, what the hell - deep in the options of Google Now I can actually allow the phone to be unlocked by my voice - but if and only if it is plugged to a charger. They had to explicitly code in a special condition to make Google Now less useful.)
(and yes, I know you meant other kind of "always listening")
FYI, some phones actually have this (the Nexus 6 for example). IIRC, not all phones are always listening because of battery concerns, but some phones have chips specifically for listening to "hot words"
Todoist Android app runs a WearSync Service all the time, despite I never went past one. After 6 months and two emails, they still didn't manage to fix that.
Strava Android app runs a GoogleNowAuthentication Service 24/7. I stopped using the app last year, installed in in April 2016, saw it's still happening and hit them an email to support. I never got a non-automated response.
I'm very enthusiastic that apps include various Google additions that are no use to me, but run 24/7. /s Maybe that's how we're going to "need" 6GB RAM phones.
Happy to hear that. I sold my Moto 360 after ~year.
I used it every day and until a software update broke it (it run into infinite loop of crashing Play services). I sent it to warranty and Motorola reinstalled the OS but in meantime I used my old, classic watch. It turned out I didn't miss 360. Other people I know have similar stories.
Hopefully the platform is still alive. Does anyone develop apps for Android Watch? AFAIR there is no special category in Google Play for this.
In the example Google showed us, a graduation photo came
through. The suggested replies were along the lines of
"Congratulations!" and "You look great!" Think for a moment about
what it takes to do that. Google recognized it was a graduation
photo and then went a step beyond just guessing what it was, it
guessed at appropriate responses.
Umm. Its suggestions are no better than a simple like. It's less than greeting card sentiment. And it removes the need to even know what you're looking at. I'm not sure that's really a positive thing. Thread full of messages responding to pictures that the users never looked at. It's bad enough that for many people birthday and anniversary and holiday well-wishes are a simple "Happy Holidays" on someone's FB timeline or a text. Do we really need to automate sending those so that the sentiment is delivered but never actually shared?
Dystopian future headline from December 25th, 2040: Millions of zombie "Happy Holiday" texts, Google to use death certificate database to fight zombies.
I feel like it's only getting worse with this Duo rollout. "We call it 'Knock knock' -- now you can not only see who's calling, but see _why_ they're calling"
Suddenly Gavin Belson curing cancer with compression seems like a better sell.
Guess I am one of the few who got excited about the Duo launch. I always wanted a Facetime like feature for Android phones, a light weight one which just works unlike Skype/Hangout. Though, I feel they should have wrapper both in a same app(may be hangouts), i feel it's great to have a native video calling from Google.
Why isn't this hangouts? Are they going to kill that too? I can't get onboard with any of these Google products because they drop them as soon as they have something "cooler"
Because these days Google releases for iOS first and there are no android apps to build on there, so they don't even try. And then we get the redundant android app which is a port of the iOS app.
Why not just build the same features into Hangouts? It's already got enough momentum, instead they have to start from the ground floor yet again with a completely new app. Is it going to have SMS/MMS support built-in? Do I have to use it alongside Hangouts in the future? What features from Hangouts will I have to give up to use Allo? Are they going to sunset Hangouts, now that I've finally convinced (most) of my friends to at least install it?
I'm pretty tempted to give up on Google and just use Whatsapp. At least I can trust that there's a business behind it in brand and concept. With Google, I haven't the slightest idea how much genuine effort they'll put into Allo. It took them years to fix integration issues with SMS/MMS in Hangouts, and there are still bugs and issues that have been simply ignored.
Google seems to have a lot of internal silos. The same can be observed in (historic?) Microsoft, current IBM, and other corporations. Perhaps the split into Alphabet with Google, et al. as subsidiaries will help? But I doubt it.
Anyways, my point is: Someone has a good idea, they implement it, it reaches the point where it's not easily integrated into another product. Instead of strong corporate leadership directing that the new thing be integrated into the old thing (or vice versa), they allow both to exist. Eventually one or more products have to be boxed up. I guess the main problem will be with discovering when this will happen and which product.
It should also be noted that, like WhatsApp with Facebook, Allo doesn't strictly require you to have an account beyond your phone number. This may be useful in helping it to grow as sign-up will be trivial. And users may eventually attach the app to their Google account (if they have one).
It seems like someone said, "why don't we have a WhatsApp competitor?" and so Bob and Steve made it without thinking about all the different messaging applications already under the Google brand.
Pretty much. Which, on its own, isn't a bad thing. A lot of innovation can happen that way. But see Wave, Talk, Hangouts, Spaces, Allo, G+ [0]: Many communication apps, some mobile, some defunct. They've each added something, they've each lost something.
G+ introduced "Circles". I really liked that concept, it was a smooth way to organize the people I wanted to communicate with. Post a message, only my gaming buddies could see it. Post another, friends and family (and most gaming buddies fell under friends) could see it. And it was under my control and invisible to other users.
Talk: Adds nothing, it's just another messenger.
Hangouts: ... video and audio chat. Screen sharing.
Spaces: Seems like a more socially oriented Wave?
Wave: Neat, got stuck in performance hell and was hard to sell. The invite system didn't help. Collaborative documents. Integrated chat bots with the document much like this @google thing in Allo seems to be.
Allo: Talk - gmail account + (end-to-end encryption OR @google chatbot). It's WhatsApp meets part of Wave meets Google Talk.
Take G+, integrate Spaces into that. Make Hangouts part of that. Make Allo a part of Hangouts. Move Hangouts to end-to-end encryption model (broken with @google chatbot present, but not if you leave it out).
[0] I'm excluding GMail because it, at least, has a fairly well established place on its own and is, fundamentally, an email platform. The accounts are tied to the other services, but GMail is easily its own standalone product with a very large user base.
I just wished that they hadn't killed Talk on mobile. I had converted all of my friends from AIM/MSN to Talk, but that was the last time everyone actually used the same messaging system. Now people use Skype/Facebook/Talk/SMS/Snapchat/??? whatever the new fad is, a few breakaway and use it.
In my humble opinion Google isn't solving new problems anymore. They created great company thanks to solving very important problem: people couldn't find right things on the Internet.
Now it is all about ads and reducing number of taps required to order a pizza. The keynote is just boring.
Because I always wanted a chat app that I could talk to AI with, not my actual real life friends, as they're too busy not giving a fuck at another $chatapp from Google and are still mad at me at the last two I was trying to make them use.
Nah, I would very much appreciate an AI I could talk to. But none is on the horizon yet - I'm still waiting until one will be able to hold the conversation and have something interesting/relevant to say.
Mm, yes, hoover up that Google marketing hype and spew it back out here to create the echo chamber necessary to convince people these technologies are just around the corner.
Duo -- video calling, competitor to Skype? A "knock-knock" feature -- you see a preview of the caller before picking. Uses quic protocol, built by webrtc team, claims seamless transition when connections switch from wifi to cellular. Graceful degradation if network quality goes down.
Allo -- video chat, e2e encryption and expiring messages. That looks good but unless someone else audits, I'll be suspicious.
Android N -- Vulkan for games, that looks promising. New jit compiler, 75% faster app install speed, 50% reduction in size. File-based encryption. Seamless updates -- phone downloads system image in background. Then on next boot it switches to new image. Also split screen to see multiple windows at the same time. Those are some nice new features.
The coolest part of the whole keynote (till now) has been developer tools.
Android Studio 2.2 . Improvements to the compilers, many X improvements to instant run, automatic test generation, C++ debugging and constrained layout editors.
AKA the return of plugins! to be fair it's the most interesting announce in the whole conference so far.
By the way, this is clearly a first step for the merging of ChromeOS and Android. It is also possible that these "apps" will have limited access to Android APIs. In fact, I'd like to see exactly that in my Chrome browser on my desktop just to give it a try.
I wonder if they are going to hijack every URL? I hope they enable some sort of opt-out. I want URLs to keep sending people to the right place, the web.
There's already dozens of pages that detects you're on a phone and send you directly to the play store and that you need to bypass using the "request desktop page" just to get access. Or event worst, sites that say let's make the mobile version less accessible while telling people it's better to install the app (Linkedin I'm looking at you)
This seems like a new step to give more power to publishers than to empower users.
Even during the presentation they said "We want you [developers] to be in control of the experience"
If I don't have app X installed it can mean I don't know the app. But also it can mean I find it a resource hogging, data gathering piece of turd and that's why I'm using a website.
I see them aligned in terms of what they are trying to achieve UX-wise, one serves the community that is already building web apps, one serves the community that's already building mobile apps, both get you low-friction, native (or native-like) UX.
I must be missing something; all of this feels very uncoordinated.
The primary entry point into Android Instant Apps are normal URLs which get shared over Email, SMS, or Instant Messaging, but for those URLs to ever get shared, they must already exist and work on the current web. So you still have to put in the effort of building a full Web experience.
...and Android can automatically enhance progressive web apps to provide experiences that are capable of being indistinguishable from native. Plus, it will work just fine everywhere else. Because it's the Web.
And you pretty much have to get close to that point anyway, since Instant Apps are useless without preexisting, functioning links for the app to hijack... why would you then duplicate the effort to have a version written in Java? Access to payment APIs is compelling for the B&H demo, but for Buzzfeed?
...and wouldn't you already be doing your damnedest to provide a similar experience for iOS or desktop users, on the Web itself?
Developers differ on preference for whether they want to do "desktop web + mobile app" or "web for all platforms".
Google, rather than taking a firm stand on which is "right", is providing tools so that developers can take either approach. Google, in this area, is more opinionated on the kind of UX that should be achieved than the route.
Now, you make some good arguments for why Progressive Web Apps might be the better route.
Constraint layouts with a visual editor. While a significant part of the iOS developer community doesn't like to work that way apparently there's not better solution out there?
I've not watched the keynote yet but I'm hopeful the announcement of Home will also come with opening up of Google Now's voice search similar to Alexa. I really want to like Google Now but it's just so limited at the moment - let me control Spotify, Hue bulbs, and my TV with it please!
Oh, and let me address it with something other than "OK, Google". Let me name it, or set the trigger phrase, or do an interpretive dance. Anything but that awful phrase.
I might be totally wrong about this, but I think the reason "OK Google" or "Hey Alexa" are immutable phrases is because the listening for them is implemented via an ASIC chip at the hardware level, in order to save on battery life. That is, instead of a software based `while(listening) {...}`, an actual hardware component looks for the correct wave forms from the microphone output.
I looked into this recently. Voice triggering is typically done on DSP and the digital processing power is typically already small compared to the microphone (which is not much itself). For power optimized applications the DSP implementation is power optimized (low speed, low leakage) and processing is done in two steps: a coarse, basic recognition with some false alarm probability, with an accurate second step. For a plugged device there's no need to super optimize, the DSP part should be in the single digit mW.
A fixed trigger makes the system simpler however: no configuration to manage, no risk of having the kids randomly changing the trigger for something funny, etc.
What happened to Google? I remember back in the days when Apple said that they will reviewed all the apps before they appear on the App Stores. Now it looks reviewing apps before release is not cool anymore.
It totally failed to keep crap out of the store. The population of apps still fell to Sturgeon's Law. There was more opportunity for reviewers to do arbitrary crap to look like they were doing a job. Perhaps they did keep some of the more obvious violations out.
New Communication apps: New Messaging app Allo. It can add emojis, whisper or shout and has smart replies like Inbox. It can also reply to images. And has the Google smart assistant.
This is so DOA, bots can do most if this already in existing messaging app. Not putting this in hangout is a death blow. I'm not installing another chat app.
i think you're really wrong about this one. people are tired of tying their google accounts to additional services. they really did need a fresh start in messaging
I didn't know about Google voice search, but I am betting it was just a voice interface to a normal search. Nothing like answering questions, keeping contexts, etc. like they were advertising today. Siri, however, started out like that from the beginning.
But for so much resources and combined talent, I would expect more than new sets of emojis. They way they were dishing it out was as if they brought peace on earth. Honestly, I would have felt embarrassed to mention it even in passing. (And what was with all those jeans and dress code in general?)
Also, if you are copying Apple and Amazon, which are known for their attention to detail, how about improving things a bit? At the end of the voice demo, the app purchases the tickets (without showing prices, locations, times, etc. (and how do I set up those in the first place?)). Now, the kicker: the phone displays a QR code and asks the owner to show it at the ticket counter at the movies. Himmm.... Really? How about skipping that too and just buzzing/vibrating/texting the user with that info when actually at the movie theater? After all, they spent a whole lot of time talking about context awareness, location awareness, NLP processing, etc. Didn't Siri do this much years ago?
Anyway, too many other things like that. I don't have time to list them all. I am just saying that a lot of people expected something more, and more cohesive.
Can you really not think of anything better? Let's put it like this, if you had billions of dollars to create something new, what would you want to create?
