The excellent Consumer Electronics Show (CES) has always been my personal mirage. Having a first-hand insight into upcoming consumer trends is potentially very advantageous for my strategy role here at Verne Global. However, as was the case this year my good intentions, and even my event registration, were overhauled by the final touches to the year’s corporate planning.
Nevertheless, I stayed connected to CES watching many of the video reports and articles from the event and I was far from disappointed with my efforts. Automotive manufacturers are some of our largest customers, so I was fascinated to find 588 exhibitors listed under "autonomous vehicle vendors". This industry is primed for significant advancement this year and as demonstrated by the amazing steering wheel and pedal-less car (shown above - a modified Chevy Bolt) from Chevrolet (Image courtesy of General Motors).
Halfway through the show I was wondering how pervasive the new AI/machine Learning technologies were in the bleeding-edge CES product announcements. For amusement, I took the “Best of CES 2018” products and started researching which ones were just great ideas and which were driven by the new AI product differentiation techniques. To my surprise seven out of the thirty-two products had documented AI technology integration. This is a huge percentage for such a broad category as consumer electronics. My initial ranking of them is:
- ElliQ (Shown below - an Alexa+ type device aimed at our ageing population)
- Aipoly (Next generation shopping, wander in and gather stuff then leave – no check-outs)
- AR4 (A security camera that recognises you versus your adversaries)
- Intel Movidius (DNN inference power and tuning a USB away)
- Nissan Leaf (AI power safety and utility in an electric car)
- LooxidVR (VR headset that monitors your eye movements and brainwaves)
- Buddy (A family companion robot)
ElliQ - an Alexa+ type device aimed at our ageing population.
At first glance there appears to be little correlation between these diverse products, however as you imagine the AI components used to achieve their utility have one or both of two common technologies, except the Movidius VPU – natural language and machine vision.
Machine learning based natural language engines like Alexa, Siri and DeepL are so much more effective than their predecessors that anything less now appears hopelessly crude. Over the holidays I was challenged to perform some traditional customer support tasks on one of my holiday gifts. The Interactive Voice Response (IVR) system was late-life and I immediately started to lose patience as soon as it repeated questions. It’s amazing how quickly we embrace better technology. Machine vision is traditionally coupled with autonomous vehicles monitoring the road, but it has far more applications like monitoring your eye’s focus/movements and the groceries that you’ve selected.
I recently analysed most VC-funded AI start-ups and categorised them into technology buckets. Machine vision and natural language were the two biggest clusters followed by drug discovery. We already have a solid representation of these performing deep neural network training in our Icelandic data center and expect the trend to accelerate in 2018.
Two other products caught my attention at CES the first being a trail blazer in the application of robotics. “My Special Aflac Duck”, is a stroke of marketing genius driven by real compassion and I'm glad to say it deservedly won the "2018 CES Tech For A Better World Award". The robotic duck is a social robot designed to offer lots of emotional comfort to childhood cancer sufferers. The duck becomes a constant companion and friend to the children who confide to it their mood through applying a mood card to its chest. It then reacts in the most appropriate way. It can also partake in medical procedures like a chemotherapy IV drip to provide company when family members are at work or unavailable.
Imagine evolving this concept for folks with drug and/or mental health issues. A friendly robot who, with top-notch natural language support and mood recognition, you could confide in that relapse was possible and get some pointed friendly advice and speedy human assistance. As with the early internet, we have yet to scratch the surface of the full potential of the new AI-based technologies.
Secondly, Hyundai’s hydrogen-fueled Nexo (shown below) finally caught my attention as verging on the superbly practical. I live out in the country about 65 miles from central Boston, so electric cars with sub 300-mile ranges and the need to recharge for hours are about as practical as matchbox cars for me. The Nexo being fuelled by hydrogen and having a 370-mile range has a good chance of making it through my day with a reserve.
The Hyundai Nexo shown at CES 2018.
I’m now hopeful that I will make CES next year and see over half the award-winning products powered by AI technology. Perhaps we will have a couple of them training their neural networks in our HPC and intensive compute optimised data center, up here in Iceland.