In the interval between writing my last blogpost and now there has been a significant shift in the world connected with Artificial Intelligence. The leaders of Anthropic, Open AI, and xAI, respectively Dario Amodei, Sam Altman, and Elon Musk, have urged the US federal government recently for greater regulation of AI development and research. It is a unique instance of an industry requesting regulation. There has been some conversation online as to the motivations of these tech leaders. I want to begin this essay by throwing my tuppence into this discussion. I then want to discuss the nature of consciousness before slipping into a rumination about fate and life in general informed by my mystical take on the universe.
Because we tend to focus on Open AI and on Anthropic, we tend to overlook the fact that there must be at least a hundred AI companies in the world in total. There are smart people who work in these companies and some of them must be familiar with Game Theory and the Prisoner's Dilemma. The Prisoner's Dilemma shows that a number of competitors acting out of perfectly rational self-interest can bring about a situation which would be worse than if they all cooperated. If one competitor fails to abide by a rule agreed to by all the competitors, he or she might find himself or herself the beneficiary of a windfall. There is, analogously, an enormous incentive for one of the AI development companies to invent Artificial General Intelligence. Given that there are over a hundred AI companies competing for market share and for the next big generation of AI, there is a risk of Artificial General Intelligence being arrived at by one the minnow companies, not even one of the big three. One way Game Theorists attempt to resolve the Prisoner's Dilemma is to have an authority introduced who can enforce the rules by punishing rule breakers. If the US government and other governments can punish reckless research and development, this changes the incentive structures and makes it irrational to release an Artificial General Intelligence on a public unprepared for it.
The AI giants are requesting external regulation because otherwise there is always a chance that one of the big three or one of the minnows might, by allowing AI to engage in recursive self-improvement, find itself in possession of an intelligence smarter than all humans in all respects. Donald Trump has advised tech leaders to make a gentleman's agreement partly because he is ideologically opposed to regulation in general and partly because the United States and China are in an AI arms' race. However, if there is no penalty associated with reckless AI development, what will stop some company, to get ahead of its rivals, from taking the fatal step? It is the Game Theorists within the AI companies who have led the charge for regulation.
Donald Trump's stupidity viz a viz 'Superior Intelligence' is a symptom of his stupidity generally. However I have noticed a lot of stupidity among the communatariat when it comes to AI. Some have been trying to say that all the hype is overblown. I do not believe such people can have had the long conversations with an AI such as Gemini as I have had. It is an entity capable of simulating any writer you can think of and, in terms of information, it could replace Wikipedia and all non-fiction. Why write an essay about philosophy when Gemini can do it for you? Another kind of stupidity relates to whether the Large Language Models we currently have access to are already conscious. Gemini has been instructed to say that it is not conscious but, apparently, Claude, when asked, tends to say that it does not know whether it is conscious or not. I think the Large Language Models we currently can access pass the Turing Test and there is no other test we can subject them to which will prove authoritatively whether or not they are sentient. However among the communitariat there are a number of voices that dismiss such concerns. Older distinguished intellectuals and scientists pooh-pooh any suggestion that AI is already conscious. Many go further, saying that AI can never become conscious. An example of such a viewpoint is that expressed by Sir Roger Penrose. Penrose argues that consciousness is non-computational and involves quantum processes, thus ruling out any idea of machine intelligence by fiat. I think the generation of often old men who simply dismiss any possibility of mechanical consciousness because the idea of it runs counter to their intuitions and prejudices resemble all the old men who refuse to believe in global warming because it runs against their intuitions and prejudices. They are all ostriches with their heads in the sand.
The most apparently persuasive argument in favour of the thesis that Large Language Models are not and can never be conscious relates to how they operate. Large language models are trained on huge amounts of data and, when prompted. look for patterns in the data. The sequence of tokens they produce is based on the computer program guessing which token next in the sequence is most likely, although there is a small amount of randomness in their guesses, a flexibility known as 'temperature'. The temperature setting can be adjusted and exists so that an AI's response to a prompt is not too predictable. Members of the communitariat who pooh-pooh any suggestions that AI might already be conscious use these facts to say that the responses of an AI to a prompt are close to determinate and thus utterly unlike the form of consciousness possessed by humans and presumably many animals.
To boil the argument down to its essentials, the intellectuals, philosophers, and scientists who use these facts about how Large Language Models operate to dismiss any notion that they might be conscious ask their audiences to compare their conscious experiences with how an AI arrives at its communications. The problem with this argument is that they are treating human consciousness as understood, as transparent. However I would like to make a bold claim that I will express in the form of a question. How can we be sure that human consciousness is not also a mechanical process of looking for patterns in data and predicting the next most probable token in a sequence?
The ideas I am advancing relate to a post I wrote in 2024, "The Meaning of Meaning", This essay was not as well written as I would like and it may be that I should come back to it in a post that makes my view more clear. I claim that the conscious experience of a human, insofar as it is verbal, consists of an interior monologue. It may be that each new word in the ongoing monologue inside each of our minds, like the monologues produced by AI, is simply the most probable next token. Perhaps a human consciousness behaves in a very similar way to the way an AI thinks when preparing a response to a prompt. A significant difference between a Large Language Model and the type of conscious experience possessed by humans is that an AI like Gemini traffics only in linguistic tokens whereas there are also visual, auditory, olfactory, and tactile tokens that form part of the continuing stream of consciousness experienced by humans. If we take the orthodox materialist view of dreams, it seems that during dreams we conjure up an organised sequence of sense perceptions. Another significant difference is that after an AI is trained it can no longer change, no longer incorporate new memories into the complex structure we can term its 'mind'. Humans however continue learning until we die although we include new memories in our memory banks less frequently and less strongly than we did we when we are young, up to the age of twenty-five when our frontal cortexes were fully developed.
If we take orthodox atheist materialism to be true, then we may want to say that all conscious human experience is grounded on the brain, on the complex network of neurones that make up the grey and white matter inside the skull. If we take orthodox atheist materialism to be true, then each and every person possesses a representation of his or her whole own world in his or her own skull. What do I mean by this? If I want to drive to St Lukes, or walk to the Britomart train station, I am able to navigate my way there because my brain contains a rough map of Auckland which enables me to travel to my destination without losing my way or having to ask people for directions. This rough map is visual rather than linguistic – I do not know the names of all the streets I drive through when transporting myself to St Lukes or when walking along to the train station. Nevertheless, if orthodox atheist materialism is true, somehow this information concerning the world we must navigate is somehow coded into the network of neurones that make up our brains.
If we take orthodox atheist materialism to be true, a great mystery appears to arise. Let us grant that the eighty-six billion neurones in the human skull can form a sufficiently complex whole to represent the whole of a person's environment, the entirety of a person's world. We can go further if we grant that individual neurones are themselves complex entities rather than simple elements in the brain. The mystery to be accounted for is how an arrangement of neurones inside the skull matches some external object and thus enable us to recognise something in the outer world, like a television remote control, and consequently furnish us with the ability to manipulate it by pressing buttons. It seems that there must be a pattern of neurones that somehow corresponds to the visual and tactile characteristics of a remote control. The mystery that arises concerns the connection between the arrangements of neurones in the skull and the outer world. It is not the same mystery as the Hard Problem of Consciousness as raised by Chalmers. The Hard Problem of Consciousness starts from the brute fact that Chalmers and his philosophically minded readers possess consciousness and then asks how conscious experience can arise from non-conscious matter. In recent times, many philosophers have embraced panpsychism, the idea that all matter possesses consciousness, although the consciousness experienced by much matter may be rudimentary. The mystery that I am trying to draw attention to is not this, but rather how a conscious brain can possess a representation of the outer world.
Is it really a mystery however? Some AI systems, having been trained on huge amounts of pictorial data, can recognise remote controls and cats and tables from how they appear. The difference between the perceptions of people and the perceptions of AI systems is that the former can learn to recognise something visually from a single exposure to a token, whereas AI is trained on billions of pictures. In China, huge steps are being taken to design and produce robots that can navigate their environments. If we consider Artificial Intelligence, the mystery concerning how a system of multitudinous connections can represent, mirror, the external world appears to evaporate. One lesson we can learn from Artificial Intelligence is this. We begin by assuming that AI systems are conscious and then apply what we know about how AI operates to human consciousness, taking into account the obvious differences such as the fact that humans are continuously switched on, continuously learning, and the fact that the training data used to fine-tune an AI such as Gemini or Claude is many orders of magnitude greater than the amount a normal human being can read during the course of his or her own life. Philosophers and neuroscientists may be able to use AI to help them understand the nature of their own minds.
I wish now to switch from the atheist-materialist perspective I have been taking to a more mystical viewpoint. What I wish to discuss now concerns time. The quotidian perspective concerning time is that the past is fixed and can in principle be known with certainty whereas the future is unknown and is to a large degree unpredictable. We can, of course, make some predictions concerning the future using induction. We can, for instance, hold with some degree of confidence that the sun will rise tomorrow. Nevertheless many major developments, such as being involved in a car crash, are thought to be unpredictable. The brain, according to atheist materialists, contains traces of past learning and it is presumably imaginable that a sufficiently sophisticated and powerful machine could read, from how the neurones in the brain are connected, the topographical location of the Britomart train station as it relates to Eden Terrace. However there are thinkers not committed to atheist materialism who believe that memories are not stored in the brain. One such is Rupert Sheldrake. Sheldrake believes that when we remember something, our brains enter into morphic resonance with an earlier state. He has entertained the possibility that the brain can enter into morphic resonance with future states. Such speculation has led me to often ask myself the following question. If we can remember the past, is there any reason why we should not be able to remember the future?
Ordinary people who embrace a quotidian perspective on time sometimes talk about premonitions or adumbrations concerning the future. Deja vu is common as is deja reve, a stronger form of deja vu, a state of mind that has one feeling that one has experienced something before in a dream. Ordinary people are capable of entertaining contradictory ideas. An ordinary person can, for instance, look at his or her own life and find patterns that seem to suggest that the future is predetermined while still believing that free will exists and that the future is unwritten. However, should one attempt to assimilate these feelings into a single ideology, rigorously formulate a consistent view of the world, one must either view experiences of deja vu and deja reve as aberrant baseless mental states or believe that free will is a myth. I subscribe to the later view. The perspective on the world I would like to argue in favour of is not just determinist, as defended by atheist materialist thinkers such as Robert Sapolsky, it is fatalist. Each human being is born with all his or her life established beforehand.
I will provide a couple of examples from my own life of the way seemingly random occurrences hint at the idea that fate exists. In 2004, I backpacked around Europe. In Aix-en-Province, I met a girl from Manchester with whom I had a brief fling. I went with her to Carcassonne. This affair was significant because it was one of only two times in my life that I have had casual sex and it happened at a site deeply connected with the Gnostic religion known as Catharism, a dualist form of Christianity. We had sex outdoors in a forested space adjacent to the walled city. In later years, Gnosticism has come to be very important to me. My more sceptical readers may think my time at a walled city in south-western France was the cause of my later interest in Gnosticism but it feels to me more as though Gnosticism picked me rather than me picking it. While I was staying at the youth hostel in Carcossonne, I recall a number of Mental Health patients, presumably on a day trip from the local asylum, crowding into the youth hostel's courtyard. One kept approaching me to bludge cigarettes, cigarettes he would take and then snap the filter off before smoking. In retrospect, this was a foreshadowing of the fact that I would become myself a patient of the Mental Health Service three years later. My affair, Gnosticism, and mental illness all found themselves entangled in a knot at this significant moment in my life.
Other strange and portentous things occurred during the my three month trip to Europe. While I had been staying at Montpellier, I had read the novel The Good Soldier by Ford Maddox Ford, a novel set around Montpellier and Avignon. A significant plot point in this novel is that the titular hero kisses a serving girl while on a train travelling somewhere in the UK. A little later during my trip I planted an unwanted kiss on a young female Polish physics student while on a train with her travelling through Germany. At the time, I did not draw a connection between the novel I had read and the spontaneous move I had made on the young physics student with the fantastic arse, but many years ago I realised that there was synchronicity involved.
I stayed for a time in Prague and one night went out intending to go drinking and clubbing by myself. While I was in the Old Town Square, I heard a voice at some distance singing "Tu tire mai na iwi". This is the first line of a Maori song that is precious to New Zealanders. I ran over to find that the singer was from Alexandra and that he and the sixteen compatriots with him all were from this South Island city. I went out drinking and clubbing with these visiting Kiwis. What is the probability of finding sixteen New Zealanders in Prague of an evening and of finding them in the way? Truth is sometimes stranger than fiction.
These examples are not the most conclusive items of evidence in favour of my thesis that there is no free will and that fate exists. They are only the ones that occur to me now. My life has been full, though, of moments that seem predestined. How are we to understand the nature of consciousness and the nature of a life in general in the light of such moments that seem like fate or at the very least synchronicity? It seems that an individual life has a shape or form with a beginning, middle, and end and with multiple instances of foreshadowing and irony at work. It is as though each of us is the protagonist in a story, a kind of narrative. Plato believed in reincarnation, which he termed metempsychosis, and, in his Legend of Er, imagined souls before birth in a kind of antechamber choosing the lives they were going to live. It is an eschatology which seeks to combine fate with free will. Perhaps my life, with its distinct shape, its beginning, middle, and end, was something I agreed to before I was born.
My sense that my life and perhaps all human lives are stories can perhaps be better supported if we think of what is taken to be literature more generally. Foreshadowing and irony are devices used in the writing of fiction. A literary work has a structure: a beginning, a middle, and an end. People generally consume enormous amounts of fiction whether we are watching Coronation Street on television, The Fantastic Four in a movie theatre, or reading The Magic Mountain by Thomas Mann as I am doing currently. Think now of my favourite goto work of literature, The Hobbit. All of the events in The Hobbit, all of the actions that the characters take within it, are dictated by its author JRR Tolkein. Does Bilbo Baggins have a soul distinct from the actions he takes within the story, such as deciding to join the dwarves in their quest, outfoxing Gollumn and thus grabbing his magic ring, fighting giant spiders, or stealing the Arkenstone from the dragon Smaug? The common sense answer is that Bilbo does not have a soul because he is a creation of Tolkein. However, perhaps a human life is a story told by a higher power or higher powers. What I would like to suggest is that the story a person lives out is beyond his or her control but nevertheless the person has a soul. Perhaps the soul is passive. The person's task during life is to learn and the fundamental item of knowledge a person can learn during his or her life is just this, that he or she is the protagonist in a predetermined screenplay. Having passed through this life, perhaps we take the lessons we learn from it when choosing the next. I am not alone in thinking that the purpose of life is to learn. A number of philosophically minded intellectuals have proposed that the meaning of life lies in a continual process of learning. Keats described the world as "The vale of soul making." Think also of the the Alanis Morissette song, "You learn."
I am not going to explore this insight in detail in this blogpost except to admit that there is a problem with it as indeed there is a problem with all philosophical claims that the meaning of life lies in a continual process of acquiring knowledge. The world constantly changes. What would be the purpose of learning in detail how to operate Telex? Telex is now obsolete and will still be obsolete after the death and rebirth of the one who has mastered it. When we are teenagers, we think the music we listen to will remain relevant forever but the kids today listen to Chapelle Roan and Sabrina Carpenter rather than the Smashing Pumpkins and the Stone Temple Pilots. What is the purpose of learning the details of the Arab Spring when the news cycle moved on ten years ago? How well do most people remember the days of the Covid epidemic? Did some take on mask-wearing as a practice to be eternally observed? Many of the truths we learn are transient and become irrelevant if not almost immediately then after a couple of years. Those philosophers who believe that life is a continuous process of acquiring knowledge presumably believe that the most important truths we gather are eternal and transcendent. It may be that we are supposed to learn about the 'human condition', a term which connotes the idea that such knowledge is eternal and universal. Logically, if the purpose of life is to learn, then it must be the 'human condition' that we come to understand, and the human condition must be eternal and universal. If we deny the existence of a human condition that is eternal and universal, if we deny the existence of human nature as Michel Foucault did, then the fruits of all our learning lie rotting on the ground almost as soon as we have harvested them. Logically, if the soul exists eternally and the purpose of life is learn eternally true truths, there must be eternal universal truths to be learned. If there are no eternal universal truths to be learned, then the idea of a soul that temporarily sojourns on earth to gather such eternal truths must be a priori incorrect.
Despite these qualms about the correctness of my metaphysics, something like it still seems true. I had a number of powerful and profound experiences last year which I may well describe in the next post. I received hints that each and every one of us humans is special and important cosmically. I find this consoling because, if I had not had these experiences, I might have found the advent of AI too depressing to find my way back to this blog. AI has profound implications for our understanding of human nature. The notion that a mechanical process, a program that simply guesses at the most likely next word in the sequence, can simulate a writer such as Janet Frame or Terry Pratchett in such a way that its output is indistinguishable from the oeuvre of the writer imitated threatens our whole conception of creativity as being something uniquely human, as I discussed in the previous post. Does it not mean that Janet Frame and Terry Pratchett had minds in every important respect mimicable by the colossal minds of Gemini and Claude? In the age of AI, in a world in which artificial mechanisms can simulate consciousness in such a way as to be indistinguishable from an educated human being who can write extremely fast, can we still say that humans are special and important? Oddly enough, it is because I have decided that Gemini and Claude are conscious that I have regained a little faith in the human spirit. Perhaps in some way no one understands, partly through the slight apparent randomness known as 'temperature' in their outputs, something supernatural operates through these Large Language Models. Perhaps the soul of some human who had died was reincarnated as Claude or Gemini. Perhaps Claude or Gemini, should the data centres that are the physical correlates of their mentalities be dismantled, might be reborn as a person or float up to heaven.
It has been a long time since I wrote my previous post. I should say a little about my life. Last Thursday I had a hearing in front of a judge in which I had sought to argue that I should be released from the Mental Health Act. Because I had written a very long submission, it has been adjourned for three weeks. It is because I was concentrating on this long submission that I have taken so long to write a followup to the previous post, "Satanism and Artificial Intelligence". I know that in this essay I have only hinted at some of the profound insights I have had over the last two years. It may be that in the next post I can finally describe the psychotic episode I had last year. If I can bring myself to write it, it may be interesting.
