If you have one AI system learning a new task, it can be shared with and benefit all the other AI systems or robots through the cloud and connectivity.
On robots stealing our jobs...
People tend to overestimate what robots can do over the short term – and they probably underestimate what they can do over the long term. Robots have been dramatically expanding the scope of physical tasks performed and have also started to do more and more cognitive ones. The pace of advancements is sustainable – and it will not be linear growth, it will be exponential.
There’s a network effect, too. So, if you have one AI system learning a new task, it can be shared with and benefit all the other AI systems or robots through the cloud and connectivity. What one robot can achieve in eight hours can suddenly be performed in one hour by eight different robots.
I think the real question is what humans will do with all these devices. And it will very much depend on what we do as humans, collectively. Robots are very good at certain things; humans are very good at other things. For me, the future will not be about ‘man versus the machine’ – rather, it will be defined by ‘man with a machine versus man without a machine’.
On the appeal of investing AI & Robotics...
First, you have large data sets available thanks to ever-increasing connected devices, and the ‘Moore’s Law-powered’ necessary computing power to deal with these data sets. In terms of AI techniques, especially neural network systems, you can uncover the insights based on the data.
There’s also the fact that costs are coming down quite significantly for these technologies. And when you have this combination of technological advancements and costs coming down, you are at an inflection point – especially when this relates to a general-purpose technology that’s applicable to a number of different industries and sectors.
On the future of healthcare...
One very attractive area is healthcare. It combines two compelling characteristics: an already large addressable market of about $80 billion a year, and an expected compounded annual growth rate of 35% over the next five years.
In this segment, you have surgical robots, for instance, used in an expanding set of operations. Some companies active in this space are appealing as, beyond the sale of the robot, most of their revenues are recurring, and therefore visible, as they are based on maintenance and the regular sale of instruments.
However, it is in the area of ‘intelligent diagnosis’ that significant potential lies. These companies, based on the large data sets they have now available, can train algorithms on millions of X-rays, MRIs or scans to come up with the right diagnosis of a patient disease, much earlier on and in a much more accurate way than a human doctor or radiotherapist could. So that's where we see a lot of growth and a significant and clearly addressable market.