Google IO 2018 Keynote / IO Extended

Earlier this week, I attended a Google IO Extended event in Stockholm. It was really nice event organized by Google to extended the reach of Google IO across the globe and those of us who would not be able to attend live. The main event was watching the IO keynote live, but before it started we had some good team games (my team did so-so I must admit) and talks on WebRTC and real-time communications using standard web technology rather than custom plugins. Watching a keynote like this live in a theatre setting gives a different experience from just reading about it post-facto or watching it on your own where it is easy to be distracted (which is something that was addressed in the keynote…).

Product Announcements

The press has covered the IO keynote and announcements at length and breadth. There is no point for me in trying to do the same here… so I just list what I found the most interesting themes.

Distraction reduction. Several new features were related to reducing the amount of distraction offered by mobile devices. I really like the idea of making the screen turn grey scale when it is bed time – but I sure hope you can enable it without using voice commands.

Maps and navigation. The new navigation features for pedestrians that were presented are just great! Being lost in a big city on foot and looking for the right direction to walk as you exit a subway station is a common experience across the globe. Having the phone recognize buildings and streets and showing you the way is just brilliant and solves a real-world problem that I have experienced many times.

Lens. The Google Lens does a fantastic job with text recognition in photos. Having it in the camera app is potentially really useful. Especially when the underlying recognition features are offered to other app developers via ML kit.

Voice control. Many of the new features were related to improving the abilities of voice control. It is very impressive what is being done here, but I still do not see that voice is a good way to interact with a device. I can see it being useful in the case of driving on your own and needing to fiddle with things in the car, but usually we have the whole family and voice commands are very unnatural. Same thing in the home, it is much easier to just grab the closest computer and do a quick search to get things done than to talk to a speaker. During my regular commute I either bike – in which case there is a lot of noise – or take the train, where talking to a device seems totally silly when you can rather use touch or type.  I don’t want other people to hear what I am doing?

Artificial Intelligence. Almost all the new features are powered by AI in some form or the other. Overall, it kind of boils down to pattern recognition and spotting recurring events. Having a learning system run on your phone that learns your habits is a very powerful way to improve the user experience, and a good complement to the traditional ‘big’ learning that goes on in the cloud. Actually, making AI personal like this is kind of a new thing – in the past, it was mostly about making systems understand the world better by working on lots of data in aggregate. Now, we are using AI to look at a single person.  Sounds scary, but can be an enabler for lots of good features.

The keynote started out with showing how AI can be used to make the world better, with advances in medical diagnosis and assistance to disable people. This no doubt true, but it is just as much used to help companies earn money 🙂

Real News. Google are making changes to Google News that aim to make it easier to get subscription revenue to serious news sources. Basically, gathering up news from a lot of sources, and making it easy to buy access immediately, via Google’s existing store system. This is definitely a good move that hopefully can help the struggling real news business. The question is how well it will actually work in the end, and whether this can make people more willing to pay for real journalism. Good move. The daily briefing feature seems intriguing and maybe I am willing to try it. Rather than using my current routine of reading the local paper (on paper) and checking trusted news sources like European public broadcasting companies and the CNN. Would be nice to have it on a computer though and not just on a phone.


The final part of the keynote was all about autonomous driving, and being a simulation fan I perked up when Waymo described how they train their systems. In addition to driving quite a bit with real cars on real roads:

They have done 1000x more training in simulation, using simulated worlds to provide inputs to their driving algorithms.  This makes it much easier to test rare events and dangerous events.

Some of the simulation work was targeting the hard problem of doing driving in bad weather and snow.  That is one of the things that have always made me very skeptical about fully autonomous vehicles – sure, they work on a sunny California highway… but what about a snowy winter night driving on small road through a Swedish forest? Maybe we can get there.

The simulations sure look cool!

Some Skepticism

For all the cool new stuff being presented at the keynote, I came away feeling a bit like they did not really apply to me… and there was one big thing missing: privacy.

Not for me. Most of the things presented at the keynote seemed to be made to work well in the USA and in English. Historically, many Google features have never made it to other parts of the world, or just very selectively. Which means that the default mode on my part is to assume that whatever is being announced will not work here i Sweden unless explicitly stated. Cases in point:

  • Google Auto is still not available in Sweden, despite my car having shipped with support for it several years ago.
  • The Google Family feature that I would really find useful is not available here.
  • Services like ordering food are clearly only going to work in certain geographies as they rely on specific partnerships with companies.
  • Google Assistant might finally be coming to Sweden, but it has been missing for quite a few years now. Which also means all the talk about voice commands are actually Not Applicable as they will not work here.
  • Voice control features are a potential future global inequality problem – small languages like we have a ton of in Europe risk being left behind as it is much more profitable to just target a few large languages. The work needed to support a new language would seem to be approximately constant, regardless of the size of the speaker population.

I was also rather surprised to see nobody mention privacy. GDPR is coming up May 25, and still Google made announcements that covered services that would be based on making more information available to Google’s systems in order to make sense of the world. Where the social trend would seem to be to want to reduce the amount of personal data being uploaded. Also, why not launch local storage in the EU that guarantees that data from EU citizens stay inside the EU rather than flowing globally?  That would make it a lot easier to endorse using Google Apps in schools for example…


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.