Societal change is inevitable. It seems hack to analogise with reference to the printing press. Look what happened, an explosion of communication. Dominance of the book for centuries. Expanding literacy. Progressive shaping of society resulting in this era.
We are only where we are because we stand on the shoulders of the giants who went before[1]. Not just the giants. There is massive amount of human contribution that is never accounted. The unseen heroes and the occasionally rediscovered thinkers and doers.
Along the way of history those who battle the battle of glass half full or glass half empty chatter away. We are either in a glorious age or a minute away from Armageddon. Polar ends of our future, both stories have merit. Who has a crystal ball that works?
I’ve been aware of neural-networks and joked about Bayesian Belief Networks for at least two decades. Having been involved in the business of data analysis that’s no surprise. Even so the rapid advance of a multitude of different forms of artificial intelligence (AI) is a surprise.
Talking generally, we have this foolish mental picture of the world that everything is linear. Progression from one state to another takes proportionate steps forward. It’s a hangover from the analogue world which is where we were until the 1960s/70s.
This fetish for straight lines and normal curves is deeply embedded. It’s odd. Although a lot of rules in nature do have a linear form, one that Sir Isaac Newton would recognise, there’s far more that follows other rules.
In the last few weeks this fetish played out at a global scale. We are all treating climate change as if it’s a water clock. Drip, drip by drip the climate changes. A reaction to a progressive degradation. Yet, environmental reality might have a step change in degradation ahead.
In my view it’s right to try to vision ahead about the path AI technology might take. It’s right to consider more than progressive development and evolutionary change. Information systems have a habit of either falling into disuse or marching on at the pace of Moore’s law[2].
Another example. The math of Fourier transforms has been around a long time. Doing Fast Fourier Transform (FFT) in the 1970s required a couple of chunky cabinet full of power-hungry electronics. For the few, not the many. Today, every smart phone[3] in the world can crunch FFT algorithms. For the many, not the few.
Can we use a simple graphical representation to say where AI is going[4]? Will “intelligence” double every year or two? Well, I suspect that developments will go faster than a doubling. Like Moore’s law these conditions tend to become self-fulfilling. It’s a technological race.
[Why? To a machine there’s no sleep. To a machine there’s 86,400 seconds in a day. Everyone is meaningful and useful. To a complete and successful electronic machine only a tiny fraction of its operating time needs to be spent fixing itself. Or that might be one job left to us.]
POST: The impact of this high speed race makes interesting study U.S. Should Build Capacity to Rapidly Detect and Respond to AI Developments – New Report Identifies Workforce Challenges and Opportunities | National Academies
[1] Sir Isaac Newton, English scientist, “If I have seen further, it is by standing on the shoulders of giants.”
[2] https://www.asml.com/en/technology/all-about-microchips/moores-law
[3] https://www.bbc.co.uk/news/business-38320198
[4] https://www.nature.com/articles/d41586-024-03679-6
John Vincent 