Posted by Jeevan Jayaprakash


In this issue, we look at the recent breakthrough made by Google owned DeepMind, the rise of a new challenger to Snapchat as well as the tech that Parisian retailers are using to learn about their customers.

We love user feedback, so if our digest is too long, short, amazing or boring — just let us know at hello@himumsaiddad.com.

Hi Mum! Said Dad


Google owned DeepMind makes another breakthrough

DeepMind, the AI startup behind AlphaGo (the robot that beat 18-time World Go champion, Lee Seedol, 4–1) has taken another huge stride towards building computers that are able to think for themselves. Scientists at DeepMind have managed to build a “differentiable neural computer” (DNC) that is able to solve problems without having any prior knowledge. Examples of such problems include planning the best route between London Underground stations or working out relationships in a family tree.

This is a significant achievement because in the previous approach that utilised neural networks (a computer system modelled on the brain and the nervous system), the neural network could only access data contained within its own network. However, the DNC allows neural networks to access previously inaccessible data (such as text stored in a digital format) from external sources.

Computer scientists say that a general purpose DNC could have a number of useful applications. For example, computers will be able to generate video commentaries, extract meaning from text and reason with humans in natural conversations.

The next step for the DeepMind team will be to test and prove the DNC in use cases that are directly relevant for the real world. Any significant advancements over the next few months and years from the DeepMind team is likely to be good news for the recently launched Google Assistant.

An artist’s impression of a Differentiable Neural Computer (DNC). Image source: DeepMind

Tribe to go after Snapchat with ‘augmented messaging’

Tribe is a ‘one-touch video walkie-talkie’ app that lists a user’s friends and groups as clickable square tiles. Users are able to send short, impromptu videos to friends by holding down a tile.

How is that any different from Snapchat I hear you ask? Well, yesterday Tribe rolled out 2.0 of their product to iOS and Android. The update includes a new feature called ‘Magic Words’, a speech recognition technology which overlays video messages with interactive subtitles. For example, if a friend mentions meeting up for a coffee, a button to put this into your calendar is overlaid on your friends message. Similarly, if they mention a particular consumer good, a link to purchase via Amazon is surfaced. This demo from Tribe shows off the amazing potential of what the founder and CEO, Cyril Paglino, calls ‘augmented messaging’.

For Tribe, the beauty of Magic Words is that it is the standout feature that is likely to drive users to try out Tribe at the very least. However, there is also the added bonus that it is clearly a great route for monetisation. In other words, an unobtrusive means of monetisation that users are more likely to be forgiving of. If they can gain enough traction, Tribe may just be onto a winner.

Tribe has attracted $3 million of seed funding in a round led by Sequoia Capital (who have previously invested in the likes of Apple, PayPal, Instagram and WhatsApp).

Watch this space.

A message that references the Sony PS4 produces a ‘Buy on Amazon’ button. Image source: TechCrunch

Luxury stores in Paris are using tech to learn more about consumers

Luxury brands are looking to recreate the concept of the web ‘cookie’ in the physical world. The likes of Armani and Dior have been utilising motion-activated screens in Parisian stores in order to tell consumers more about the products they are picking up (not too different from Estimote’s video enabled beacon that we mentioned a few digests ago). These screens serve a dual purpose — firstly, they provide a consumer with relevant, contextual information at what is an important juncture in a shopper’s journey and secondly, it provides retailers with invaluable data about the effectiveness of their displays.

A startup called Retency has also been equipping retailers with antennas that detect the unique frequencies emitted by smartphones. This allows retailers to track a customer’s movements through a retail store as well as identify returning customers. Retency go even further, claiming that they are able to identify the smartphones of shoppers who go on to make a purchase by cross-referencing sales records with the data from a retailer’s antennas.

The implications of this tech is very exciting for retailers but obviously concerning for privacy conscious consumers. Being able to link data points about an individual customer such as what products they picked up, what they ultimately bought and the next time they return to a store allows for programmatic targeting via digital screens.

Seeing targeted adverts online on your own personal computer is one thing but on a massive display in a retail store that is in view of strangers? Some may feel that might be taking things a little too far.

Personalised digital screens are polarising. Image source: Flat Vision

Originally written as part of Hi Mum! Said Dad’s Weekly Digest.

If you would like to receive our Weekly Digest straight to your inbox, please drop us an email at hello@himumsaiddad.com.