Google I/O 2018 Opening Keynote Live-Blog (10am PT)

11:43AM EDT - Hello everybody and welcome to the Shoreline Amphitheater in Mountain View where we'll be live covering Google's IO 2018 opening keynote.

11:46AM EDT - The event will be starting at 10am PT - a little bit over an hour from now so stay tuned.

12:51PM EDT - The keynote should be kicking off in 10 minutes.

01:00PM EDT - The event is starting.

01:01PM EDT - Intro video with cubes - I wonder what they represent.

01:02PM EDT - Lots of pixels!

01:03PM EDT - Google CEO Sundar Pichai is taking the stage.

01:04PM EDT - Over 7000 people attending the event in the theater.

01:04PM EDT - Sundar addresses a major bug in one of the core products - reference to the Cheeseburger Emoji

01:05PM EDT - It's been an extraordinary year on all fronts - an inflexion point in computing.

01:06PM EDT - Google is focused on bringing digital skills to enable creating developers needed for the rising technology needs.

01:07PM EDT - AI enables solving problems in new ways. Last year Google AI has been announced to bring AI to everyone.

01:07PM EDT - Google is opening AI centers around the world.

01:08PM EDT - Healthcare is the most important fields AI is going to transform.

01:08PM EDT - Machine learning assisted diagnosis offers a lot opportunity.

01:08PM EDT - Machine learning assisted diagnosis offers a lot opportunity.

01:09PM EDT - AI can help predict medical events before a patient will become sick.

01:10PM EDT - In general there is a lot of enthusiasm in the medical space with AI/Machine learning diagnosis as it's one of the aspects that can vastly outperform human diagnosis.

01:11PM EDT - Next example for AI - when two people talk over each other, machine learning can separate the audio through audio analysis and image cues.

01:14PM EDT - Now rolling a video on Gboard - Google's Keyboard - and how adding support for Morse Code improved accessibility

01:15PM EDT - GMail's new refreshed design is also bringing AI features.

01:16PM EDT - SmartCompose starts suggesting phrases when starting an email - you just have to press tab to continue to it to autocomplete.

01:16PM EDT - It analyses the context of the email conversation for prediction. It will be rolling out publicly end of this month.

01:17PM EDT - Every day over 5B photos are viewed by users. A new feature called Suggested Actions a new set of actions is presented based on the image contents.

01:17PM EDT - It for example offers already sharing capability to the users who are recognized in the picture.

01:18PM EDT - Say the image is underexposed - the app will offer options to correct this.

01:18PM EDT - If you take a picture of a document, it will convert it into a PDF and clean it up.

01:19PM EDT - An astounding new feature is the colorize feature for black & white pictures. This is a key demonstration of machine learning into a very real use-case.

01:19PM EDT - Sundar is announceing a new TPU generation - TPU3.

01:20PM EDT - The new TPU3 is said to be 8x more performant than TPU2 and will require liquid cooling in the datacenter.

01:20PM EDT - Over 100Petaflops ML capability currently deployed.

01:21PM EDT - Wavenet models the human voice to create a new synthesized voice.

01:21PM EDT - This results in 6 new voices to the Google Assistant - all sounding very natural.

01:23PM EDT - John Legend's voice is coming to the assistant - Wavenet allows to capture his voice and to synthesize it. Coming later this year.

01:23PM EDT - New video demonstrating Google Assistant scenarios.

01:24PM EDT - Scott Huffman taking the stage.

01:25PM EDT - Two years ago the Assistant was announced at IO - today it's available on 500M devices.

01:25PM EDT - It's available in over 40 auto brand and 5000 Assitant enabled devices.

01:25PM EDT - Last year it was launched in India and has gained a 3x uptake in adoption.

01:26PM EDT - We're now covering new enhancements to the Assitant.

01:26PM EDT - Natural conversation is an important aspect.

01:27PM EDT - There's no longer a need to say "Hey Google" everytime to enable it - it will naturally follow the conversation.

01:27PM EDT - The Assistant now understands when you are talking to it and when you're talking to somebody else.

01:28PM EDT - The feature is called Continued Conversation and will be rolling out in the coming future.

01:29PM EDT - "Multiple Actions" follows the conversation and stiches together several requests- it's very hard to parse sentences like these.

01:31PM EDT - "Pretty please" is a feature to make GA require to be polite for requests.

01:32PM EDT - Visually assistive features now coming into focus.

01:32PM EDT - We're looking at the first demo - a Lenovo smart display.

01:33PM EDT - The device is controlled by voice but because it has a screen it's a much more rich assistant.

01:33PM EDT - YouTube & YouTube TV will be added onto Smart Displays.

01:34PM EDT - There's many more of ways that SmartDisplays can make home actions easier.

01:35PM EDT - Inspired by the smart display experiences, this translated into improvements on phone based Assistant.

01:35PM EDT - The assistant takes full advantage of the screen for content response.

01:36PM EDT - Demonstration of food pickup & delivery experience. I'm wondering how far this will actually be adopted outside of small areas.

01:37PM EDT - The new visual assitant will be launched on Android later in the summer and iOS later in the year.

01:38PM EDT - Google Assitant will be integrated into Google Maps this summer.

01:38PM EDT - Sundar taking back the stage.

01:38PM EDT - The mission of Assistant is to help you to get things done.

01:38PM EDT - Making calls is a core aspect of this.

01:39PM EDT - Connecting users to businesses is also a core aspect of this - AI can help with this with connecting you with businesses who do not have an online presence.

01:40PM EDT - For example making an appointment through Assitant to a hair dresser - the Assistant will actually call the business to make a real appointment.

01:41PM EDT - This is very amazing as it demonstrates a full fledged voice "AI" that's able to have conversations.

01:42PM EDT - Another demo demonstrating a restaurant reservation which doesn't go as one would think- the assistant still managed to navigate through the conversation succesfully.

01:43PM EDT - The feature will be rolling out as an experiment in the comings weeks.

01:44PM EDT - Giving users back time is a core goal for Google.

01:44PM EDT - "Digital well-being" is being addressed now.

01:44PM EDT - People are anxious about missing information - "FOMO" Fear of missing out.

01:45PM EDT - JOMO - Joy of missing out.

01:45PM EDT - Google understands your habits and focuses on the content which you are interested in.

01:46PM EDT - I wonder if this creates an echo-chamber in terms of information....

01:46PM EDT - Family Link enables managing kids' screen time.

01:47PM EDT - This is part of the Digital Wellbeing site rolling out soon.

01:47PM EDT - Sundar is addressing the issue of journalism - fake news being brought up?

01:48PM EDT - "Support quality journalism" - We agree!

01:48PM EDT - Google News initiative invests $300M over the next 3 years to collaborate with news outlets to improve the quality of news.

01:48PM EDT - Google has a responsibility to provide news.

01:49PM EDT - AI has been used to improve the relevancy and quality of the news sources.

01:49PM EDT - The new Google News - Trystan Upstill taking the stage to explain more.

01:49PM EDT - With the new GN - 3 things are the key aspects that are improved.

01:50PM EDT - As soon as you open GN - the top five stories are always the stories that you care most about.

01:50PM EDT - Google pulls from various sources, for example it also pulls from local news outlets as it knows your location.

01:51PM EDT - The headlines category contains the large worldwide stories.

01:52PM EDT - The new GN app is builed with the new Material Theme.

01:52PM EDT - There's a new visual format called Newscasts.

01:52PM EDT - It uses natural language processing to bring news content on a topic together in a visually expressive presentation.

01:53PM EDT - Understanding the full story- it's a lot of effort to bring together a lot of sources and to sum it up in its fullest.

01:54PM EDT - Google organizes news stories based on their "temporal co-locality".

01:55PM EDT - It uses its real-time event analysis to create a timeline of an story and its events.

01:56PM EDT - Full coverage is not filtered based on the user's habits, but rather something that will be identical to all users universally.

01:57PM EDT - Newsstand allows you to subscribe with Google to paid publications.

01:57PM EDT - Google build this with over 60 publishers around the world and will be rolled out in the coming weeks.

01:57PM EDT - It's an interesting take on aggregating paid content.

01:58PM EDT - It will be rolled out on Android, iOS in 127 starting today.

01:58PM EDT - We're focusing towards Android now.

01:59PM EDT - Introduction video showcasing Android's strengths over the years.

01:59PM EDT - Android's "openness" is a key aspect that Google's focusing on here.

02:00PM EDT - Dave Burke taking the stage.

02:00PM EDT - 10 years of Android - The T-Mobile G1 was the first phone with Android.

02:01PM EDT - Partners have launched tens of thoustands of phones being used by billions of people.

02:01PM EDT - The growth of Android has helped to fuel the shift from desktop towards mobile.

02:01PM EDT - Android P.

02:01PM EDT - P is an important step towards vision and AI in the mobile space.

02:02PM EDT - On-device machine learning can enable a lot of feature about the OS learning your habits.

02:02PM EDT - Battery life is a key aspect. "Adaptive Battery" is designed to give a more consistent machine learning experience.

02:03PM EDT - It will analyse which apps you will use and depending on the usage patterns it will only enable app wakeups for those relevant apps.

02:03PM EDT - This results into a 30% less wakeups on average.

02:03PM EDT - Moving into screen brightness - most smartphones adapt brightness based on the environmental lightning.

02:04PM EDT - P introduces adaptive brightness which learns how you like to adjust the brightness and adjusts it based on your habits by itself.

02:05PM EDT - Google Launcher - which introduced predictive apps which introduced the predicted apps you'd like to use.

02:05PM EDT - P now introduces predictive actions which does something similar but applies to actions within apps.

02:05PM EDT - This requires developers to introduce these actions onto the app.

02:06PM EDT - This provides deep links into the apps from the launcher.

02:06PM EDT - "Slices" are interactive snippets of app UIs.

02:07PM EDT - Early access for app actions and slices begin next month into the developer previews for P.

02:08PM EDT - Google introduces "ML Kit" which is a new set of APIs for machine learning.

02:08PM EDT - This enables both on-device ML as well as Google's cloud based services.

02:08PM EDT - It's based on TensorFlow Lite and uses the NN API - but it will be also available on iOS.

02:09PM EDT - One of the key goals over the last few years it to simplify the UI of Android for expand it to more users.

02:09PM EDT - Google is demonstrating new feature's on Dave's phone.

02:10PM EDT - The demonstration here is about the transition from the navigation bar towards a simplified home button with contextual actions and gestures.

02:11PM EDT - This is similar to the navigation of the iPhone X's bar - at least Google's take on it.

02:11PM EDT - The new navigation system is said to be a lot faster.

02:12PM EDT - The volume slider now changes the media volume by default now in P. Seems to be another feature that Samsung had first introduced on Android O.

02:13PM EDT - Sameer Samat taking the stage to demonstrate "Digital Wellbeing"

02:14PM EDT - "Helping people with their digital wellbeing is more important to us than ever"

02:15PM EDT - Android P will show a dashboard based on the amount of time you spent on applications and a usage history thereof.

02:16PM EDT - YouTube will be showing total watchtime across mobile and desktop.

02:17PM EDT - P will introduce time limits on apps - it will give you a notification and grey out the app after you exceed the limit. The goal here is to limit your usage.

02:18PM EDT - "Shush" will disable all notifications on the phone when you turn the screen downwards. It will still enable notifications and calls from important poeple.

02:19PM EDT - "Wind down" mode puts the phone into silent mode and greys out the screen after a set time to enable you to help you to go to bed.

02:19PM EDT - The Android P beta is announced today and will be available on the latest Google Pixels.

02:20PM EDT - Alongside the Pixels there's 7 OEMs that will provide beta firmware builds.

02:21PM EDT - Oneplus, Nokia, Vivo, Sony, Xiaomi, Oppo and Essential are part of the beta program.

02:21PM EDT - Jen Fitzpatrick taking the stage to cover computer vision.

02:21PM EDT - Google Maps has covered over 220 countries.

02:22PM EDT - Advancements in AI now will be able to automatically add new streets and buildings.

02:24PM EDT - A new updated Google Maps will show content and events in your area.

02:24PM EDT - This new tab in maps is called "For you", with new places to visit and personal recommendations such as trending restaurants.

02:24PM EDT - It encourages visiting new places.

02:25PM EDT - The locations are scored based on a system "Your match" based on your history and why it recommends it for you.

02:26PM EDT - Group matching is a feature to organize events with your friends. You can share a shortlist of places, and your friends will vote on locations.

02:29PM EDT - Google now explaining location determination based on the phone's camera.

02:30PM EDT - This is an AR type feature which uses the camera to analyse the location and correlate that with the location on the map.

02:30PM EDT - It will overlay interesting information such as business locations over the camera viewfinder.

02:30PM EDT - Basically this really is proper augmented reality for Google Maps.

02:32PM EDT - Google Lens is also introducing these capabilities - Starting next week Lens will be natively integrated into OEM camera applications. A list of vendors take part - Samsung seems to be notably missing here.

02:32PM EDT - Google Lens is also introducing these capabilities - Starting next week Lens will be natively integrated into OEM camera applications. A list of vendors take part - Samsung seems to be notably missing here.

02:33PM EDT - Lens allows OCR and copy and paste functionality.

02:33PM EDT - It will automatically looks up the text that you're selecting and give you contextualized content.

02:33PM EDT - It will automatically looks up the text that you're selecting and give you contextualized content.

02:35PM EDT - .

02:35PM EDT - Last thing: making lens work in real-time

02:36PM EDT - All the Lens features will be coming the next few weeks.

02:37PM EDT - CEO of Waymo (Alphabets's self driving car business) John Krafcik

02:38PM EDT - Waymo is the only company in the world with a fleet of self-driving cars.

02:38PM EDT - Video demonstrating the experience in a trial in Phoenix, AZ;

02:40PM EDT - Phoenix will be first public program where it will enable people to call Waymo calls to get around.

02:41PM EDT - Waymo uses deep learning and machine learning for things such as pedestrian detection.

02:42PM EDT - This allowed for a 100x reduction in the error rate.

02:43PM EDT - Explaining "Perception" and "Prediction" AI uses in autonomous driving.

02:44PM EDT - Dimitri is explaining the use of ML on the combination of sensor data to enable high precision perception and classificaiton of objects.

02:45PM EDT - Waymo cars have driven 6 million miles and have gained a lot of data in terms of real scenarios.

02:46PM EDT - The neural networks are trained on the TPUs and Google's datacenters.

02:46PM EDT - In simulation, Waymo has "driven" over 5B miles.

02:47PM EDT - Google's demonstrating how ML allows for filtering sensor noise - such as snow picked up by the LIDAR system.

02:49PM EDT - The keynote wrapped up - Thank you everybody for being with us. We'll be covering more news and sessions over the following days.

02:49PM EDT - Thanks!