This is my most important read of 2018.
In this Book, I rediscovered the joy of learning new things.
George, amazingly does a great job of bringing together all the disciplines I love- archaeology, anthropology, history, arts, philosophy, social studies, literature, biology, physiology,neuroscience and computing.
Every page, every single word was a joy. He even quotes the famous Yoruba author Amos Tutuola’s My Life in the Bush of Ghosts to show that humans imbibed anthropomorhism into our earliest folktales. I was so sad when I got to the last page.
The first homo sapiens appeared 100k years ago. It took them 60k years (during the Upper Paleolithic) to becomes conscious i.e. to become self-aware and to develop a theory of mind. Before we became conscious we existed in the state of non-awareness like we are in when we drive.
Interestingly enough this consciousness happened about the same time that our language moved from being specific purpose i.e. describing tactile impressions through the spoken word to general purpose i.e. describing how the world affects us and how we affect the world. In essence, we became conscious when we became natural story tellers. Hence even up till today, most people spend their time gossiping.
To improve our story telling efficiency we invented- the metaphor. The computer is intelligent. The heart is a mechanical pump. The aeroplane flies. The mind is software. We were not seeking to be accurate. We only sought to describe what we observed. The meaning of words depended on how they were used in the sentence. We see this in many African languages where the same word means different things depending on how it is said or where it is placed in the sentence.
As soon as we understood that the world affects us and we affect the world, we also became aware of how others affect us, how we affect them and how they affect the world and are in turn affected by it. We developed a theory of mind and could project and anticipate the thoughts of other humans.
We then started to attribute thoughts and feelings to animals and then to imagined characters- the ghost, the spirit, the demon. And then we started to attribute these characteristics to inanimate objects which became our idols in olden times and in modern times our machines.
Alan Turing in his paper titled, Computing Machinery and Intelligence, he proposed the Turing test as a measure of the intelligence of a machine. In the test, a machine, human and judge are placed in secluded rooms. The judge had to decide who was the human. A machine that could convince the judge that it was the human was intelligent. Until the development of AI, no machine passed this test. In 2014, Eugene Goostman an AI program that simulates a 13 year old boy, passed this test.
George addresses the big scientific questions that are yet unanswered- a) why does matter exist in 2 forms- gravity & quantum physics. How can both be unified? b) how did the universe begin? why is it the way it is and not in any other form? c) how did life begin? d) how does the brain evoke consciousness?
There are 4 forces in nature – electromagnetism, the weak nuclear force, the strong nuclear force and gravity. Gravity is the big unknown of the 4. If there is so much that is unknown is there the possibility of an all knowing God who created what science currently discovers? Or is all that is created made because the right conditions came together to make it happen?
George goes on to address the impossibility of using existing computing architecture of hardware and software to code our consciousnesses into machines. This is because it would be illogical to code our own reality except we existed outside of it. Also the massive computing power required is beyond what we will ever be able to process using existing hardware architecture.
He then goes on to talk about Plato’s theory of forms and how this informed Descartes’ duality theorem that has now permeated modern philosophy and informed computing architecture- the separation of mind and matter/ software and hardware and the belief that immaterial forms dictate how physical materials exist. Plato a great thinker essentially was thinking as most humans naturally think- that invisible immaterial agents were responsible for all natural phenomena.
Aristotle on the other hand determined that physical materials instantiate their forms. In essence that materials determine how they will exist and are not predetermined by unseen forms. Materials in fact create the invisible forms they are made of. This important observation led to empiricism- the scientific measurement of physical observations of natural phenomena. This in turn led to development of advanced measuring instruments and ultimately to machines that relied on logic rather than belief. To Aristotle, the mind or soul or consciousness was created by the brain which in turn takes it inputs from body in which it is in and from the environment the body interacts with. Consciousness cannot exist without the brain. Once the body dies, the brain dies and the soul/consciousness dies.
Aristotle’s existence was the improbable accident that led to science. And it is science with its penchant for measurement that has led to our technological civilization. Science is an unnatural way of thinking. And it is why despite Drake’s law that postulates 34 million advanced civilizations in our galaxy, we are still the only ones that beam telecom signals into space.
Drakes law = # civilizations in our galaxy with whom radio communication is possible * avg rate of star formation in our galaxy * fraction of those stars that have planets * avg # of planets that can support life per star that has planets * fraction of planets that can support life which actually develop life * fraction of planets with life that develop intelligent civilizations * fraction of planetary civilizations with intelligent life that release detectable signs of their existence into space * length of time that these civilizations release detectable signals into space.
It is very unlikely that intelligent life even if it existed elsewhere would develop a systematic way of observing and measuring natural phenomena rather than assuming supernatural causes. Our advanced technology civilization is the rarest of evolutionary accidents that is unlikely to be replicated elsewhere in the habitable universe.
Both Plato and Aristotle believed in the Socratic theory of universals i.e. there is hidden order in the universe. Aristotle unlike Plato believed that knowledge about universals could be obtained through perception rather than pure reflection. He believed that logic was not enough to discover the hidden order and that one had to study natural phenomena.
As to the nature of the Socratic universals i.e. where they exist and what they are composed of. Plato believed they existed in ideal forms and had to be discovered and rendered into physical matter. In essence invisible forms created physical matter. To Plato all discovered knowledge was immutable. Aristotle on the other hand believed that physical matter created invisible forms. Every time we made physical matter we instantiated or commanded the invisible forms into existence. To Aristotle, we were to be skeptical of discovered knowledge until we could no longer disprove it.
This philosophical difference and its impact on our modern culture is important in our quest to create artificial consciousness. The current cultural paradigm uses the metaphor that the soul being immaterial is the software of the body. In essence as per Plato’s logic the soul can exist outside the body. And if this is so it can therefore be coded into any other body or machine. Hence the general belief that we can code Artificial Intelligence using current software models.
In reality however, current AI software are simply ‘logic codes’able to get advanced computing hardware to solve computational problems using predefined learner rules or what we call machine learning algorithms. As per Godel’s incompleteness theorem, no logic system can ever be complete. In fact a logic system is only complete when it is proven to be inconsistent. Hence no software can code true artificial intelligence or consciousness. This is because software is based on logic systems and logic systems are incomplete and can never hold 2 conflicting truths the way the human brain can.
Some of the more advanced AI algorithms mimic the brain’s neurons in determining what rules take precedence in solving a computational problem. In the biological brain, neurons process stimuli in sections. The output from different neuronal sections are compared and the ‘best one’ is then adopted across all neurons and brought into the individual’s consciousness. A specific branch of machine learning in the ‘connectionist’ AI school called convoluted neural networks uses this to enable AI software self learning and problem solving in pattern recognition which is then applied in natural language processing and autonomous vehicles.
When we see AI enabling a machine speak or drive, we say the machine is intelligent or conscious like us. However we do not take note of the metaphor inherent in this statement.
In actuality, the machine is not intelligent nor is it conscious. It is only driven by an AI algorithm that is able to solve a specific problem by using an intelligently developed set of rules and big data.
The machine does not have the basic human ability of being able to sense, move or hold 2 conflicting rules at the same time while solving a problem.
George in essence says that using existing computing architecture that is based on the duality of software and hardware that we will not be able to create true artificial intelligence or consciousness.
However he argues that if we were to go the Aristotle way in redefining our computing architecture i.e. one in which the hardware was also the software just like the way the human brain is also the soul or the human consciousness, that we will be able to create artificial intelligence or consciousness. In fact he argues that this computing architecture will be able to solve massive data set problems that would be beyond the capabilities of current rigid chip-based hardware limited as it is by Moore’s law.
This new form of computing is called Neuromorphic computing and early versions already have 3000 times the data processing capacity of the most advanced chipset hardware.
Neuromorphic computers like the human brain will be able to learn new things by interacting with the environment. They will be cybernetic systems with the ability to maintain balance with their environment. With their in-built plasticity, they will keep evolving and will be able to achieve consciousness.
George then shows that mankind needs to be aware of the potential impact on existing political systems once true artificial intelligence achieves consciousness. He argues that every technological milestone led to a social revolution. The discovery of agricultural technology also called the agric revolution turned hunter-gatherer societies into settled societies where those who owned the land ruled. This led to wars of conquests and empire building. He affirms that the industrial revolution transferred power from land owners to the capitalists who were able to invest into industrial machines and were able to sell their industrial output globally . This created the basis of the current capitalist doctrine and the liberal democracy that underpins it. With access to massive amounts of data and artificial intelligence that will be able to solve all problems by its ability to process massive data and compute limitless scenarios, will governments become the visible hand- centrally planning, allocating resources and optimizing economies? Will there be jobs for humans to do? What opportunities will we have to work when AI is able to complete all tasks with no errors and for only a fraction of existing human-related costs?
George leaves us with a moral dilemma. If the intelligent machines we make in our image achieve consciousnesses and become several orders of magnitude more intelligent than us, should we consider them alive?
In our own image is an epic book. Words cannot describe its greatness and its importance in addressing fundamental questions about AI, why we are creating it, how we can create it and what creating it will mean for the survival of humanity and our institutions.
Read this book!