Could do with this too. Missed everything today due to family stuffs :/
Google I/O 2018
- Thread starter sesevans
- Start date
Associate
- Joined
- 1 Jul 2016
- Posts
- 2,225
Here's, in my view, the most interesting new feature from the event:
Last edited:
Not that I have yet seen but suggest they will start appearing tomorrow. At the moment, most of the articles discuss individual new features, and there are many.
Summary of some of the important points at I/O Day 1:
1. AI will be transformative for Google in healthcare in terms of predicting cardiovascular events. Google sees its AI breakthroughs predicting medical events likely to happen in next 24-48 hours to assist physician.
2. With their AI advances, they will bring "smart compose" to all GMail users later this month.
3. AI will be boosting Google Photos to create "suggested actions" like fix brightness, share photos, fix documents (example given was converting a doc automatically to a pdf file), etc.
Google Assistant was a huge focus on day 1. Google has created 6 new voices to choose from including John Legend (coming later this year). GA is now on more than 500 million devices and in more than 40 automobile brands and in large numbers of appliances. GA is to become naturally conversational with "continued conversations" without having to say OK Google and will stitch voice and visual assistants together. GA coming to Google Maps, as earlier mentioned on this thread. The assistant will make phone calls in the background on your behalf to perform services, as earlier mentioned on the thread.
Google's sister company, Waymo (self driving vehicles) with the use of AI has made huge strides in perception (classifying objects on road) and in behaviour prediction (what will a biker, driver, pedestrian, etc do next). Waymo announced they have suprassed 6 million miles of self driving on public roads for the first time.
As to Android P, which launches in Beta next week, one clever change addresses a concern expressed on this thread often: battery life. Google is teaming up with its UK-based artificial intelligence unit DeepMind on a feature called “Adaptive Battery.” Via machine learning, Android determines which apps you’re likely to use over the next few hours, and which ones are going to remain unused. The operating system devotes battery resources only on the apps and services it expects you to use.
Other Android P features:
1. AI used to help you set phone's brightness.
2. App Actions will help you get to the app you want faster, saving you time.
3. You will be able to scroll through recent apps by sliding the Home button sideways.
4. A Dashboard may help you understand your screen time habits. Might be helpful in managing screen time for children.
And more...
Last edited:
This might help a little:
PhoneArena: "
Qualcomm's Snapdragon optimization and Google's Project Treble should make Android P updates faster
https://www.phonearena.com/news/qualcomm-snapdragon-optimization-for-Andropid-P_id104771
I believe it was more conceptual that an imminent release if we are talking about same demo. Specifically I am referring to camera-assisted walking navigation. The idea combines Google Maps 2D interface with a view from your smartphone camera lens.
https://arstechnica.com/gadgets/201...-ever-augmented-reality-interface/?comments=1
That AI assistant demo was amazing. I am seriously considering a move to android if the next gen pixel is impressive. Their entire ecosystem is quite amazing and they are currently killing apple in innovation. I love my iPhone but google are making it harder to stick to iOS every year.
Soldato
- Joined
- 6 Jan 2013
- Posts
- 22,524
- Location
- Rollergirl
I'm starting to use GA more and more in the car, opening Spotify and playing tracks. Calling people in my address book. The closest I get to a conversion is when the voice responds with "would you like to call home or mobile?"
It's really great when it works, but it really is buggy and a bit hit and miss. If they're developing it further with Android P then that's something that will interest me.
It's really great when it works, but it really is buggy and a bit hit and miss. If they're developing it further with Android P then that's something that will interest me.
@Glaucus @Kainz
You both inquired about a summary of announcements at Google I/O. Here are two summaries I have seen recently:
https://venturebeat.com/2018/05/09/everything-google-announced-at-i-o-2018/
https://www.pcmag.com/feature/361062/the-coolest-things-from-google-i-o-2018
You both inquired about a summary of announcements at Google I/O. Here are two summaries I have seen recently:
https://venturebeat.com/2018/05/09/everything-google-announced-at-i-o-2018/
https://www.pcmag.com/feature/361062/the-coolest-things-from-google-i-o-2018
I have not yet done it but the Beta version of Android P is available for download.
There is no doubt that GA is being expanded and continually improved---different voices to choose on GA--including John Legend coming later this year, better quality performance and services, more ways to save you time, etc, all highlighted at the I/O conference. One of my favourite concepts was Google Duplex---the demo had Google Assistant making a phone call in the background to book a haircut with the time and date of your appointment simply confirmed for you with no effort by you other than asking.
They now have GA built in to 5,000 devices like appliances, speakers, phones, etc and in 40 automobile branded models with now 500 million units globally. As we move more and more to a voice activated/personal assistant "world", GA seems very well placed (also Amazon) and likely to see massive improvements and capabilities with AI and ML responsiveness.
Cheers for that, will try and check it out later.
Would be interesting to know but the second conversation was the more impressive. Honestly I would have struggled on that phone call. The person the other end was talking all sorts of crap and was very hard to understand.
I am in the same boat having seen the video on Marques YT channel.Call me a horrendous cynic but I just don't buy that they were real phonecalls. I'd imagine they were scripted or something. SorryI'll try to force myself to believe though...
Associate
- Joined
- 1 Jul 2016
- Posts
- 2,225
I'm also somewhat skeptical, but Google's AI blog shows many more scenarios. I think the truth is somewhere in-between.
Fine to be a sceptic of course but the blog seems quite convincing.
Of interest, not all States in the US will allow Google Duplex at launch. 12 of the 50 States require all party consent to recording phone calls, something required by Google Duplex.
I found this statement about whether the UK is an all party or one party consent country. Does anyone know if this is up to date?
"Under the Regulation of Investigatory Powers Act 2000 (RIPA), it is not illegal for individuals to tape conversations provided the recording is for their own use. ... Journalists often record phone conversations but can only use what is said for research purposes if they have not told the person".
Here is current state of play in US:
https://www.voicebot.ai/2018/05/09/...-at-launch-and-will-only-support-3-use-cases/
Call me a horrendous cynic but I just don't buy that they were real phonecalls. I'd imagine they were scripted or something. Sorry
Interestingly, some have found the conversations TOO REAL! There has been social media blow-back to Duplex as some people believe it is not ok to "trick" the recipient of the phone call into thinking the other party was human, with phrases like "uh-hum", etc. The height of political correctness gone crazy? I think so!
https://www.npr.org/sections/thetwo...oogles-new-voice-bot-sounds-um-maybe-too-real
Yes, Duplex is going to identify itself as a robot!
https://www.cnet.com/news/google-says-its-designing-duplex-with-disclosure-built-in/
https://www.cnet.com/news/google-says-its-designing-duplex-with-disclosure-built-in/
On the point of timely updates, it seems that importantly security patches will soon be updated more regularly. I appreciate this is not the same as operating system updates but is an important element nonetheless.
The GMS partner programme and the Android partner programmes are both being reworked, according to XDA:
https://www.xda-developers.com/google-require-oem-regular-security-patches/
Variety believes that the biggest announcement from Google I/O is that Google wants to "augment" everything--AR---all with the help of our phone cameras.
For example, in Google Maps, soon you will be able to open the camera view from within Maps, point it at a Street corner, and get instant directions of where to go. Google is combining Maps, computer vision, Street View and the power of your camera to re-imagine walking directions.
With new features added to Google Lens you will be able to point your camera at the real world and copy and paste say a document onto your phone. Think books and posters. And soon Google wants to overlay info over these identified objects---eg, music videos over copied posters.
Soon your phone will browse the world.
http://variety.com/2018/digital/news/google-io-2018-augmented-reality-1202806541/
For example, in Google Maps, soon you will be able to open the camera view from within Maps, point it at a Street corner, and get instant directions of where to go. Google is combining Maps, computer vision, Street View and the power of your camera to re-imagine walking directions.
With new features added to Google Lens you will be able to point your camera at the real world and copy and paste say a document onto your phone. Think books and posters. And soon Google wants to overlay info over these identified objects---eg, music videos over copied posters.
Soon your phone will browse the world.
http://variety.com/2018/digital/news/google-io-2018-augmented-reality-1202806541/