Google I/O Keynotes

[LAwLz]

How come nobody has made a thread for this yet?

Edit: The event is now over.

 

Watch it live here:

 

Or follow the live blogs from:
Anandtech

ArsTechnica

The Verge

 

 

The news below is a short summary of what was talked about at the event (and I skipped some stuff that I found boring). More details about things will appear later.

It seems like Google had a lot of things to talk about which resulted in things being very brief with not much details. Expect more details about everything they talked about today to come out in the next few weeks.

 

 

Fluff nobody really cares about:

Google products have more than 1 billion unique and active users.

 

Last week there were 2 billion active Android devices (phones and tablets, not counting watches/TVs/cars/consoles or anything else).

 

Google is investing a lot in machine learning to improve things such as Duo, YouTube suggestions, Photo search and such. (Machine learning is the buzzword this year apparently).

 

If you want to see Google projects using machine learning/AI then you can visit www.google.ai. That website also contains some tools for developing machine learning products.

 

Speech recognition error rate has gone down a lot. In one year it has gone from 8.5% error rate, to 4.9%.

 

Google's image recognition is now better than humans (citation needed).

 

1 billion hours of video watched each day on YouTube.

60% of that is on mobile devices.

 

60% of all K-12 laptops sold in the US runs ChromeOS.

 

 

App stuff:

Gmail will be getting a "smart reply" feature that was previously only available in Inbox. The rollout will start today. Basically, it gives you a suggestion on what to respond.

 

Google Lens - Seems like an improved version of Google Goggles. You take a picture and the app will do different things depending on what you took the picture of. Took a picture of a sticking with a WiFi SSID and password on it? It will connect to that WiFi network. Took a picture of a weird plant? It will tell you what that plant is. Took a picture of a baseball player, but you were behind a chain fence? It can digitally remove the fence. It can do many more things too.

 

Google Assistant will get better "conversation ability".

You no longer have to talk to Google Assistant. You can now feed it text.

Lens and Assistant will be able to integrate into each other in the near future. So you will be able to take a picture with Lens, and then have Assistant talk about what it is seeing, translate text in the picture, order movie tickets if you take a picture of a movie sign, and other such things.

It is being released for the iPhone today.

A new set of APIs will start rolling out. These APIs will allow third party developers to integrate Google Assistant in their apps.

Assistant will also get an update this summer to support more languages. More specifically, Germany, French, Brazilian, Portuguese and Japanese. Italian, Spanish and Korean coming later this year.

 

 

Google Home:

Google Home coming to Canada, Australia, France, Germany and Japan this summer.

 

Proactive assistant - Google Home can call on your attention (white lights start blinking on the device) and give you relevant information that it thinks will be useful for you. The example they used was Google Home called on a person's attention, and after being asked "what's up" Google Home told him that traffic was heavy today and he would need to leave earlier than usual in order to get to his meeting on time.

 

Hands-Free Calling - Google Home can now make phone calls, and it's free! You can call any US or Canada number, landline or mobile, for free (yes, free phone calls).

 

Multiuser support - Google Home can do personalized tasks depending on who is talking. If two different people say "call mom", then Google Home will be able to figure out who "mom" is based on the person's voice. You can also configure it so that when you tell it to call someone, it will dial out using your personal number. If your spouse makes a phone call, it will dial out using his/her number.

 

Spotify, SoundCloud, Deezer and more coming to Google Home - You will now be able to play for example your Spotify playlist through it.

 

Bluetooth audio support - You can now connect to your Google Home and play any audio from it. It's basically a Bluetooth speaker. (Coming soon™).

 

Google Home will integrate with your iOS/Android device and TV. For example you can ask Google Home what the weather is like, and it can show it to you on your TV (if you got a Chromecast). If you ask Google Home for a route, it can transfer it to your phone.

 

 

Google Photos:

Suggested sharing - Let's imagine you were out with your friends yesterday. You took a lot of photos. Google Photos will scan through your photos and then give you a reminder to share some of them with the friends you hung out with. For example it will detect that your friend Bob was in one of the good photos you took, and then give you a notification that you might want to share it with Bob.

If you click OK you will be taken to a photo sharing function which will suggest which people to send the photo to and select which photos to share. Bob will either get a notification (if he has Photos installed) that an image was shared with him, or he can get an email or SMS with a link to the photo (if he does not have the app installed).

At the bottom of the gallery Bob will have a box asking him if he wants to add more photos to the album, and Google will analyze which photos on Bob's phone that were from the same event.

 

Shared libraries - You can now share parts or all of your libraries with some contact. For example you can share your library with your wife, which means all photos you take will be accessible by your wife.

 

Photo Books - You can order photo books from the app. Soft- (10 dollars) and hardcover (20 dollars) options available (20 pages, and the hardcover is slightly larger). You can order one today from the website, and the feature is coming to the app next week.

 

 

Android:

Android O will be released this summer (beta out today).

 

Android O focuses on two main themes. Fluid Experiences and Vitals.

 

Fluid experiences.

New features to help do tasks which have previously been difficult to do on phones.

• Picture-in-Picture support on phones - For example you can have a floating YouTube video playing while doing other things (not the multitasking view, it's more like a traditional window)
  Picture:

Spoiler

• Notification dots - New API which lets apps put a little dot on their launcher icon if they have a notification it wants to show you. Long pressing the app icon will also bring up the notification in-place. Seems kind of pointless to me...
• Autofill - The Chrome feature now works in apps! If you have your Twitter username and password stored in Chrome and then install the Twitter app, your username and password will appear as a suggestion (and be filled in automatically if selected) inside the Twitter app.
• Smart Text Selection - Selection will now recognize if you are trying to copy a place name or phone number. So if you double tap on a place name that is comprised of several words (such as "Old Coffee House") then the selection will know that you want to select all three words and not just one. It will also do things such as recognize addresses and put a shortcut to Google maps next to the cut/copy buttons in the floating menu.
• New APIs which will hook into (not yet available) coprocessors specifically designed for machine learning. So your next phone might have a processor inside it specifically designed for machine learning.

 

 

Vitals has been split up into three subcategories.

Security Enhancements, OS optimizations and developers tools.

 

Security - Google Play now has a button which lets you scan your apps for malicious activities. It's basically like a built in anti-virus. It already existed, but now some controls of it are exposed to the users (such as manually starting a scan).

 

OS optimizations - New improvements to ART. Your device will now boot faster, and apps will launch quicker and run smoother.

A new feature called "wise limits" which will limit background execution and location services for apps, in order to stop apps from using up too much battery and RAM when running in the background. Maybe this will fix Linus' complaints about Android battery usage?

 

Developer Tools - Improvements to Android Studio and a new tool called Play Console Dashboards. These two things will help developers understand issues with their apps such as how to improve performance and battery usage, crashes and other statistics and tips.

New language support! Kotlin is now supported!

 

 

 

Android Go - The new Android One program.

The OS is optimized for low performance devices.

 

Comes with special Go apps which are optimized for low data usage and low end devices. For example the YouTube Go app will allow you to download youtube videos, and transfer them over a P2P network to your friend.

 

All devices with 1GB or less RAM will use the Go configuration by default (seems like you will be able to turn it on in high end devices too).

 

Some developers has joined and created special apps. Skype and Facebook have both created so called "lite" apps which has to adhere to guidelines laid out by Google. These guidelines are:

• Useful offline state
• <10MB apk size
• Better battery and memory usage than their regular versions.

 

Better keyboard support for multi-language users.

 

 

Google for Jobs:

When you search for a job in Google, it will show you special results and filters. Search now also hooks into all major job listing services (in the US).

Will start rolling out next week (US only).

[Doobeedoo]

Heh, just got home so was wondering where can I watch aside from live blogs so nice I saw it here :3

[DocSwag]

I was completely unaware IO was today  

[Trixanity]

Wasn't as exciting as previous years. Very developer (yes I know that's the point) and business focused. Every other topic was 'machine learning'-something.

 

A lot of it is US-only. So half of it was me going "and here's another thing I can't use or might be able to use in about three years". If not US, then it was not-my-country launches.

 

It seems Google had so much to talk about that everything was very brief despite the two hour keynote. So I felt like I wanted more about Android O (I'm biased because that was my primary reason to tune in) and it got about the same size time slice as Photos or YouTube pretty much.

 

Not really any interesting hardware launches either (granted it's not really the intended platform for it) and only few of them to boot: third-party VR headsets that are standalone vs the previous Daydream headset being reliant on phones.

 

The disappointment was real but sort of expected. I hope there will be deep dives on the interesting bits later so we get some of the mentioned stuff clarified in later articles.

 

PS. I noticed Google's Notification Dots. If I'm not mistaken it's 'borrowed' from Nova Launcher's Dynamic Badges.

[Misanthrope]

Quote

How come nobody has made a thread for this yet?

*reads the rest of the OP*

 

[Techicolors]

apparently Android O is changing how emojis look again

http://www.androidpolice.com/2017/05/17/blobs-be-gone-with-android-o-google-completely-redesigns-its-emoji-again/

 

[AlexGoesHigh]

TL;DR machine learning everything...

 

Honestly, the build up to this i/o was nothing compared to others, hence almost nobody din't care, and the keynote prove it, it was just slap machine learning on everything and updates to current products and project, and just the mainstream stuff, it would have been nice if the have mentioned in some form fuchsia but nope. 

 

[WMGroomAK]

I have to admit that the part of it that I liked was dealing with the Google Lens AR features and the VPS tools that they are building.  

[Nowak]

Android Go seems interesting. Might help the experience out on low-end devices.

[WMGroomAK]

I think some of the more detailed stuff they are presenting is also looking kind of interesting...  So far, I've only had a chance to read a little bit about Seurat, however, it appears that this could greatly improve importing high-end VR experiences to mobile platforms and hardware.  Especially if the news on Engadget is correct...

 

https://www.engadget.com/2017/05/18/google-seurat-vr/

 

Quote

At Google I/O today, Lucasfilm's ILMxLAB revealed that it used Seurat on a scene from Rogue One, which took an hour to render on a high-end PC, and turned it into something a mobile GPU could render in 13 milliseconds. The tool was able to reduce the scene's texture size by a factor of 300, and its polygon count by a factor of 1,000. Based on the brief demo we saw, the mobile VR version definitely wasn't as polished as the full-scale cinematic scene, but it still looked surprisingly great given that it was running on mobile hardware.

 

[Inkz]

Now the real important question is still unanswered, will it be called Oreo?