In this article we are going to discuss the evolution of AI, its origins and development to here and where it may be heading. Many books have been written on the subject, so this is going to be a short but hopefully informative and abbreviated history to here and a peak into the beyond. We will continue to write about the evolution of AI in our blog and bi-monthly newsletter and build upon this article.
We are entering a period of technological accelerated change over the next 3-20 years, the likes of which has never been experienced by humankind. Some technologists have identified or labelled the culminating point to which this accelerated change will lead as "The Singularity" which can be viewed as the "end of the beginning" after which the singularity epoch of human history begins. What exactly is "The Singularity"? The “Singularity” is a projected future point in time when exponential developments in computer (artificial) intelligence and computational power result in the most dramatic and irreversible change in human history. The origins of the term lie in physics where "The Singularity" is described “as a point at which a function takes an infinite value, especially in space-time when matter is infinitely dense, as at the center of a black hole. “The emphasis in this sentence is "takes an infinite value".
To put this in some perspective, scientists are now estimating that AI (artificial intelligence) capacity will "double" every "3 months". You may be familiar with Moore's law who forecasted that computational power and capacity would double every 2 years. AI's capacity is however forecasted to double every 3 months. That represents an approximate multiple of 128x growth every two years. That number is harder to grasp. Now, add in the projected leap to "Quantum Computing" which is estimated to be 158,000x faster than the fastest supercomputer on the planet today. What is Quantum Computing? Up until now all digital evolution has been built on 1's and 0's. Quantum computing is a "paradigm shift" away from ones and zeros into the world of "electrons". It is not a huge leap to grasp managing one's and zero's in spite of the complexity of how computers are configured to do what they do, but managing electrons which are moving at quantum speeds - in random motions - to organize vast amounts of information is a shift into the beyond or science fiction territory.
When you merge AI with Quantum Computing, which is estimated to be 5-10 years away from industrial implementation, the possibilities become harder to grasp.
AI - A Short Summary of How We Got to Here or Chat GPT
Chat GPT (Generative Pretraining Transformer) took the world by storm as the fastest adopted software in history. Its origins started in the early 2000's with the advent of a project that was funded by an instituion of the Canadian government called "Deep Mind" which was led by Geoffrey Hinton one of the pioneers of AI. Prior to 2006 it was challenging to get neural computational networks to learn complicated things but with advances in technology and computational power, they found better ways of initializing networks. These methods were referred to as "pre-training" which in turn led to deeper "more intelligent" networks. All this work led to a breakthrough in speech recognition technology in 2009 which then got disseminated to Google, Microsoft, Apple, IBM and to the evolution of "Siri" for example.
In 2012, students of Geoffrey Hinton made yet another breakthrough in "object pattern recognition" systems, the ability to determine what the content of a picture, for example, contains. A combination of human labeling of "objects" to begin with and then mapping thousands of the same images down to the 1's. zeros and pixel data created complex pattern recognition systems that begain to recognize objects with astounding accuracy, up to 97% of the time.
It does this through a process referred to as "back propagation". A computer algorithm and network is able to consume and organize vast amounts of data and organize it into what appears as highly intelligent or "artificial intelligent response" or AI. This is the methodology that is used by Chat GPT. While this method is highly sophisticated and fast in its ability to retrieve and organize data into apparent intelligent responsiveness, it's weakness - at this point in the development - is that it is not designed to "reason" with an human level of sophistication.
AGI - Artificial General Intelligence: Next Level
The ability to reason in a way that mimics how a human being functions which is through a highly complex inter-connected neural, physical and biological network is not within it's current configuration. However, that could be described as the goal of AI or coming as close to that as possible.
AGI is still years away from being realized but it is coming! Ray Kurzweil, who is another prominent figure in the AI world made a prediction in 1999 that computers would pass the "Turing Test" in 2029. What is the "Turing Test", you may be asking? The Turing test was defined by Alan Turing - the man who invented the modern day computer and broke the Nazi's encryption codes - as being the point in time when you could have a substantial conversation with a computer - that had the capacity for "multi-chain" reasoning - for many hours, days even and that could convince you, it could pass for being human. Kurzweil made this prediction in 1999 at the time many of his peers thought it may take another one hundred years to pass this test.
Advances in computing science have been greater than the majority anticipated. Today, he stands by his forecast that computers will pass the "Turing Test" by 2029.
It is anticipated that quantum computing could become commercially viable within the next 5-10 years.The combination of Quantum Computing and AI will accelerate modeling advances in complex neural intelligence models that will make AGI a reality and kick start a LEAP in humanity that is in many respects unquantifiable. Now - whenever that point arrives - and it is going to be sooner than later, "The Singularity" that we described at the beginning of this article will take shape. We have not discussed the merger of "Robotics" with AI and Quantum Computing, but this combination will lead to increasingly agile and intelligent robots that look, talk, and feel like human beings populate the world all over. Tesla for example is working on exactly this kind of project with Optimus.
In addition - as if this were not enough to comprehend - scientists see the merger of AI and human beings as inevitable. The day is coming when integrating complex AI smart platforms and chips into human beings will beocme a reality. The intermediary technology to do this is "Nanotechnology" which can develop nano-bots to deliver “payloads” whether that is information, medicine, or something else, that can communicate with parts of our bodies and in this example our neo-cortex which will connect our brain with AI.
In a recent article by Goldman Sachs the three professions they projected would be most impacted were Office & Administrative Support, Legal and Engineering. We think it is far too early to say. The Quantum Leap ahead is impossible to quantify. Take a pencil and draw a straight line up. That is exponential growth on a scale never seen before.
If this article has left your head spinning, we fully understand because ours has not stopped. How will these sophisticated computing intelligence networks impact businesses, stocks, trading, portfolio management and every aspect of our lives is subject matter for future articles. In this article we wanted to first provide a brief introduction and perspective of how AI got here and where it is going!