To understand why the current boom in the field of AI is happening, we must first understand how we got here.
This quest we are on all began with Alan Turing and his bold claim that there is a test that we can do to determine if a machine is intelligent. It’s post World War II, 1950s and Alan Turing, widely regarded as one of the pioneers of what we know today as modern computing published his paper titled “Computer Machinery and Intelligence”. In it, he puts forth a proposition – that we will know when machine is intelligent a text transcript of a conversation between a machine and a human is given to another human and that person cannot determine which participant is human and which is machine.
This gave way to the first wave of academia and corporations putting their intellectual and financial resources to make such a system. Let’s take a moment and contextualize this – this is the 1950s, 20 years from the personal computer and 30 years before the internet.
Computing power was expensive and no way near the mammoth that we have today. Thus came the first AI winter. Despite standing on sound statistical and mathematical theories, the practicality of applying these theories was limited. One interesting system that came out of this period was an early chatbot called ELIZA which has a fascinating similarity in User Experience to some of the systems we see today
The next important moment to mention to understand why now came in the 1970s. The personal computer is now upon us and unleashes a wave of innovation and access in our society never before seen as the price of computing compared to its capabilities was lowered.
The 1970s and 1980s produced some of the groundbreaking research which lay the foundation for the landscape we have today such as Geofrey Clinton and backpropagation as a technique to have a self learning system.
However, the right ingredients were not in place for the explosion we see today and we hit another period of slow progress in the space due to limited funding because the usefulness of the systems being built was not meeting expectations.
Fast forward to 1990 the internet is being quickly adopted and is truly revolutionizing the world around us. We started digitizing our surrondings – representing people, processes, communities and the physical objects around us digitally. A side-effect of doing this is that we started and continue to this day producing TONs of data. Today, everything from your car, smart devices, your interactions with applications, applications themselves – almost everything is producing data at all times.
Now we are warming up, the theory of the 1950s now is meeting increased computing resources and the foundation of data which will become essential is being laid. Great!
Enter the 2000s, the internet is now a staple, PCs are staples, now they even fit in your pocket and compared to decades before that device in your jeans is thousands of times more powerful than computing systems that use to be so large they fill a large room.
Now comes the Cloud – another revolution in our society. Pay-as-you-go computing, storage and networking offerings. For the next 15 years we also had improvements to the hardware as well as the approaches and algorithms used for machine to learn such as the landmark paper “Attention Is All You Need” by eight scientists working at Google
This was the mixture that led to the breakthrough innovations we see today. Machines trained on massive sets of data on extremely powerful and sophisticated hardware able to generate new data based on the training they have received but with no rigid rule in place on what that output should be – thus came the ChatGPTs, Geminis, Claudes etc we see today.
If Alan Turing were alive today he would likely be deeply fascinated at the systems we have today – the scale and usefulness of these systems (both for good and not so good causes) is incredible and deeply intriguing. Spend just a few minutes with your favourite AI chatbot, print the transcript and if you could have handed it over to Alan Turing and ask him which is the machine and which is the human, he might not be able to tell. A shining moment to the testament of human ingenuity and our ability to stay with and solve a new problem without a precedence being in place on what the solution should be. Now here we are with systems that can seemingly pass the Turing Test, now we must ask “is this intelligence?”.
Thank you for reading! If there is something I can help you with, feel free to reach out via the contact form above or via my services