From the time I was a child, there were three things that I had decided would fundamentally change the world. The first was low-cost space travel, the second was nuclear fusion, and the third – the most important – artificial intelligence (AI). I spent summers reading up on neuroscience, information theory and mathematics, and I learned to code in C. Looking back on my path to becoming co-founder and CEO of an artificial intelligence company, it appears the die was cast at an early age.
Marvin Minsky was one of the first AI researchers that caught my attention. I fell in love with the simplicity of his book on perceptrons (cunningly called Perceptrons). I’m a mathematician and physicist by training and one of the lessons you learn is to build the simplest possible representation of the thing you are trying to model – think of cows as perfectly spherical elastic bodies. Minsky caught my attention because he was ultimately trying to build a mathematical theory of the brain.
That’s when the light turned on and I realized that rather than trying to build a copy of the brain, we should understand the axiomatic foundations of intelligence instead. I realized that this approach held great opportunity for AI.
I believe in a general-purpose AI – sometimes called horizontal AI or Artificial General Intelligence (AGI). This approach brings the ‘build once, deploy everywhere’ mindset to AI instead of building isolated silos that are capable of only a tiny subset of human intellectual power. Building highly specialized AI solutions makes sense in many contexts but the problem is that each solution is a one-off. It’s a little like custom-built cars as opposed to an assembly line.
At DimensionalMechanics, we use an axiomatic approach to constructing an AGI. We also look at AI not as something esoteric that might have an application in the future, but as something that can benefit us right now. We want to see AI incorporated into every aspect of our lives.
We believe that by teaching a platform to think like us, we can give it ‘expert intuition’ – giving our AI the power to form human-like conclusions from a corpus of knowledge that it has built up over time – helping us solve problems from media to medicine. The most important expert intuition is developing a perception of what you as a human and consumer want or need, ranging from ultra-accurate recommendations for which movies to watch next, to how to spend money more effectively, invest better, improve one’s health, and so on.
To some extent AI already touches all of us, from those annoying NLP-based support lines (‘I’m sorry, I didn’t get that – could you please repeat or say something like ‘pay my bill’ or ‘close my account’) to Siri, Cortana and now Alexa. The advantages are pretty obvious. Already, CAD systems (Computer Aided Diagnostics) are helping doctors and there are AI platforms that make targeting recommendations for the military. AI – propelled by the overwhelming zettabyte (1 billion trillion bytes) volumes of data per year by 2017, an increase in compute power and new approaches to the craft – is predicted to grow an estimated Alvin Toffler, there are three waves of human development – the first wave being agrarian, the second being industrial, and the third technetronic. We are perhaps in the early stages of the fourth wave – a movement away from the omnipotent technocrat, to knowledge and understanding becoming universal commodities. Artificial intelligence will form the nucleus of this age and it’s thrilling to help build that future.