Google i/o 2017 – Brain game

Reply to this post

Superior brains being used to make its services the best.

• Google held the first day of its annual developer conference and in its keynote, it highlighted the features and improvements that it is making to its ecosystem to keep users engaged while gathering and categorising as much data as it can.
• Artificial Intelligence headlined the event with Google’s leading expertise now being implemented in everything that it does.
• These included:
  • First, Google Lens. This is machine vision similar to what many others have also announced but in Google’s case I suspect it will work properly.
  • This can be used to identify items which combined with search to bring up relevant information about it.
  • This stretches from the history and background of a place to the ratings users have given to restaurants and shops.
  • Others fall short in the ability to identify items as well as in the digging up of relevant information about the item.
  • This is because the AI they are using to power the service is not nearly as advanced as Google’s.
  • This functionality is being rolled across all of Google’s properties to enhance everything Google does such as the Photos app, Maps, Daydream and so on.
  • Second, AutoML. This is a research project within the Google.ai initiative.
  • It is neural network that is capable selecting the best from a large group neural networks that are all being trained for a specific task
  • While few details were disclosed, Google said that the results achieved to date were encouraging.
  • This is a hugely important development as it marks a step forward in the quest to enable the machines to build their own AI models.
  • Building models today is still a massively time and processor intensive task which is mostly done manually and is very expensive.
  • If machines can build and train their own models, a whole new range of possibilities is opened-up in terms of speed of development as well as the scope tasks that AI can be asked to perform.
  • RFM has highlighted automated model building as one of the major challenges (see here) of AI and if Google is starting to make progress here, it represents a further distancing of Google from its competitors when it comes to AI.
• Google also gave updates on all the current products and services including the next version of Android: Android O.
• Most relevant updates included:
  • First, Android. There are now over 2bn active Android devices in the market but I suspect that there is meaningful multiple device ownership.
  • For example in Brazil there are more mobile phone connections than there are people, highlighting that multiple devices are owned by a large number of people.
  • This is a trend that is mirrored in many other emerging markets.
  • Every Google Android device has a Google sign-in and for the other Google services, the figures are closer to 1bn which also includes those that have iOS devices.
  • Hence, in terms of real unique users rather than devices, I think the numbers are much lower.
  • This is important because it is unique users that generate the revenue for Google and hence they are a better measure of the real penetration of Android across the globe.
  • Second, Android Go. This is the relaunch of the failed Android One project which aimed to put smartphones in the hands of more users which obviously, requires much lower cost.
  • Android Go is like a mini-mode of Android O which runs in an optimised way on devices with memory down to 512MB of RAM.
  • Google’s apps have also been optimised to run in this highly constrained environment.
  • Importantly, functionality has been added that focuses on saving data usage as well as offering complete control of data usage from the device.
  • For the lower income users, data has become almost like a currency and this gives them much better control of their “spending”.
  • This looks like a much better proposition than Android One which was highly restrictive to the handset makers.
  • However if they start tinkering with Android Go (as they always do), there is a good chance that all of these good improvement will vanish into thin air.
• While this is not the most exciting i/o event in terms of new announcements, it is what is going on with AI that has the most implications for Google’s outlook.
• AI is now embedded in everything and because Google is clearly the global leader it has the scope to make its services richer and more intuitive than anyone else’s.
• This is critical because this is how Google will win over more users to its services, generate more traffic and therefore more revenue.
• However, I think that much of this is already embedded in the share price and I continue to prefer Baidu, Tencent and Microsoft.