Classic: Artificial Intelligence Part 1

While we're deal with some crazy life stuff we thought we'd revisit a rapidly changing topic in Artificial Intelligence. We walk through the foundations that make AI feel inevitable: formal reasoning, the Church Turing thesis, early neural network ideas, and the Turing test as our most famous shortcut for judging “intelligence.” Along the way, we hit the hype cycles and turning points, from early chess machines and the coining of the word robot to research booms, AI winter, and the moment responsive machines become normal in American homes. 

Transcript

Myths of Ancient Automata

SPEAKER_00 0:00

You should pod without rhythm, otherwise you'll attract the worm. Gentlemen Let's run our minds.

SPEAKER_01 0:31

The Lai Zhi, or Li Zhu, was a Taoist text written in the 5th century BCE by Lai Yuko. In it is a tale from many years before of an encounter between King Mu of Zhao and a mechanical engineer known as Yan Shi, referred to as an artificer. He proudly presented the king with a very realistic and detailed life-sized human-shaped figure of his own crafting. To quote from the text, the king stared at the figure in astonishment. It walked with rapid strides, moving its head up and down so that anyone would have taken it for a live human being. The artificer touched its chin and it began singing perfectly in tune. He touched its hand and it began posturing, keeping perfect time. As the performance was drawing to an end, the automaton winked its eye and made advances to the ladies in attendance, whereupon the king became incensed, and would have had Yan Shi executed on the spot, had not the latter, in mortal fear, instantly taken the robot to pieces to let him see what it really was. And indeed it turned out to be only a construction of leather, wood, glue, and lacquer, variously colored white, black, red, and blue. Examining it closely, the king found all of the internal organs complete liver, gall, heart, lungs, spleen, kidneys, stomach, and intestines, and over these again, muscles, bones, and limbs with their joints, skin, teeth, and hair, all of them artificial. The king tried the effect of taking away the heart and found that the mouth could no longer speak. He took away the liver, and the eyes could no longer see. He took away the kidneys, and the legs lost their power of locomotion. The king was delighted. Welcome to Dispatch Achex. I am Skip. I'm Jake. That's right. Today we're going to talk about something that you I think will contribute well to. Me isn't the co-host or the the audience? Oh, good question. You. You, Jake. Oh, okay. Jake, the person. Which is also something to keep in mind when we're talking about this. So, in Greek mythology, Talos was a giant made of bronze who acted as guardian for the island of Crete. He would throw boulders at ships of invaders and would complete three circuits around the island's perimeter daily. According to the Greek compilation of myths Bibliothek, Hephaestus forged Talos with the aid of a Cyclops and presented the great automaton as gift to Minos. In The Argonautica, Jason and the Argonauts defeated Talos by removing a plug near his foot, causing the vital to flow from his body and rendering him lifeless. In On the Nature of Things, the Swiss alchemist Percelius, I completely mispronounced that. Paracelsus. Oh, just like Beauty and the Beast. But you'll laugh at his actual name. The Swiss alchemist Paracelsus, whose full name is Philippus Ariolus Theophrastus Bombastus von Honenheim. Oh my lord. He's Mr. Bombastus.

SPEAKER_00 3:55

Mr. Bombastus.

SPEAKER_01 3:58

He had a one-hit wonder before he served in the goals.

SPEAKER_00 4:01

Parasosis, also known as Shage.

SPEAKER_01 4:04

Oh man. He wrote that the sperm of a man be putrefied by itself in a sealed kirquet for 40 days. A container, like a ceramic container, with the highest degree of putrefaction in a horse's womb, or at least so long that it comes to life and moves itself and stirs, which is easily observed. After this time, it will look somewhat like a man, but transparent without a body. I don't know how that works. What? This is all ancient Greek, so I I don't it's Greek to me. If after this it be fed wisely with the arcanum of human blood, and be nourished for up to forty weeks, and be kept in the even heat of the horse's womb, or a tauntaun, I guess, a living human child grows therefrom, with all its members like any other child, which is born of a woman, but much smaller, which I do believe we call a humunculus.

SPEAKER_00 5:04

Huh. You're gonna tell him a rabbit hole. You're in the horse hole right now.

SPEAKER_01 5:08

It's warm. I thought it smelled bad on the outside. Golem making is explained in the writings of Elezar Ben Judah of Worms, W-O-R-M-S. Alright. That that dude rules. Yeah. In the early 13th century, during the Middle Ages, it was believed that the animation of a golem could be achieved by insertion of a piece of paper with any of God's names on it into the mouth of a clay figure. In History of English Kings, which we talked about in our Excalibur episode, there was told tale of a brazen head, one of many in myth, in a passage where he collects various rumors surrounding the polymath Pope Sylvester II, one of my favorite Looney Tunes, who was said to have traveled to Al Andalus and stole a tome of secret knowledge, where he was only able to escape through the assistance of a demon, which is kind of crazy for a Pope. He was said to have cast the head of the statue using his knowledge of astrology. Not really clear on exactly what that means. It would not speak until spoken to, but then answer any question with yes or no. Muslim alchemists in the Middle Ages sought to achieve Taquin, the creation of synthetic life. In Faust, the second part of the tragedy by Johann Wolfgang Gotha, an alchemically fabricated homunculus destined to live forever in the flask in which it was made, endeavors to be born into a full human body. Upon the initiation of this transformation, however, the flask shatters and the humunculus dies. Since humankind gained the ability to use tools to create machines, there has been an often quixotical drive to harness these abilities to replicate the ultimate power of the natural world, to create life, not just automata, the creation of a being that could not only move, act, and speak like man, but one that could also think. The creation of which would be man's mastering of that which he believed raised him above the beasts of the earth, sentience and artificial intelligence. And that's what we're talking about today.

SPEAKER_00 7:37

So it's AI, but maybe not the AI you're thinking of.

SPEAKER_01 7:42

Right, and that's what we're going to get to. That's exactly why we're talking about it. Today's AI, is it the AI of myth? Is it the AI of science fiction? The past answer is no. But is it far off from that? Maybe no. And that's what we're gonna talk about. And like I said, you as a philosophy and religion major, feel free to interject at any time. Okay. I'm sure you have thoughts. This folly of man's pivots around the idea that thought can be replicated mechanically. I think a lot of this stems from the constant desire of human beings to quantify, to analyze, and to understand that which is abstract. Kind of

Historical Mechanical Beings

SPEAKER_01 8:30

like using numbers to analyze baseball. These are meta well, not in baseball's case, but these are metaphysical or intangible and putting them into tangible terms. That in and of itself isn't a failing or a flaw. It's just how we learn to operate in the world around us. In fact, in modern neuroscience, there's a growing idea that what we call consciousness is essentially just an operating system. But we'll come back to that. Chinese, Indian, and Greek philosophers all developed structured methods of formal deduction by the first millennium BCE. Their ideas were developed over centuries by philosophers and mathematicians and alchemists alike, such as Aristotle, who gave a formal analysis called syllogism, Euclid, whose work Elements was a model of formal reasoning, and Al Kawarzami developed algebra, and his name led to the word algorithm. And European scholastic philosophers such as William of Ockham and Duns Scotus, which strangely was actually still the acronym Supreme Court of the United States. It was ahead of its time. It was fortuitous. Everyone was confused, but panned out. Spanish philosopher Roman Lull developed several logical machines devoted to the production of knowledge. He described these machines as uh mechanical entities that could combine basic and undeniable truths by simple logical operations. This has been an obsession of ours, basically, since we've had the ability to think and reason and ask questions. If we can use tools to create things, why can't we create things that nature creates? It's been kind of the The study of mathematical logic provided the essential breakthrough that made artificial intelligence seem plausible. Russell and Whitehead presented a formal treatment of the foundation of mathematics in their work, The Principa Mathematica, in 1913, which was kind of a big deal. Then later, David Hilbert challenged those kinds of archetypes to answer a question that he posed. Can all of mathematical reasoning be formalized? This question was kind of answered by Gdel's incompleteness proof the Turing machine and later Church's lambda calculus. So in this endeavor, two things were widely assumed to be true. First, there were and are limits to what mathematical logic can accomplish. Second, within these limits, any form of mathematical reasoning could be mechanized. The Church Turing thesis implies that a mechanical device, shuffling symbols as simple as zero and one can imitate any conceivable process of mathematical deduction. And unfortunately, Alan Turing was blacklisted from society because he was gay.

SPEAKER_00 11:39

Well, we could blacklisted as a polite way of signing the horrible things done to him.

SPEAKER_01 11:45

Absolutely, yeah. I mean, he's not dissimilar from like people condemned for their basic understanding of reality. So here's a very short timeline of what led us here. Picture it's Sicily, 1726. Ooh, this is nice. I like it. Actually, it probably would be really nice. The weather would be great. In this year, Jonathan Swift's novel, Gulliverse Travels, introduced the idea of the engine in caps. A large contraption used to assist scholars in generating new ideas and language and publications. Scholars would turn handles on the machine, which would rotate wooden blocks inscribed with words. The machine is said to have created new ideas and philosophical treatises by combining words in different arrangements. And this is a quote from Gulliver's Travels. Everyone knew how laborious the usual method of attaining to arts and sciences, whereas by his contrivance, the most ignorant person at a reasonable charge and with a little bodily labor might write books in philosophy, poetry, politics, laws, mathematics, and theology without the least assistance from genius or study. This will all become extremely foundational as we go along. In 1914, a Spanish engineer, Leonardo Torres de Cuvedo, demonstrated the first chess playing machine, which was called El Idrestica, at the Exposition Universial in Paris, which is, you know, the world's fair essentially. It autonomously made legal chess moves, and if the human opponent made an illegal move, the machine would signal an error. And this is something we're gonna get into in part two, but it's vitally important. In 1921, a play called Rossom's Universal Robots opened in London by Carol Kopak. It's the first time the word robot was ever uttered in English. See, in Czech, the word robota is associated with forced labor, specifically by peasants in a feudal system. The term robot quickly gained international recognition and came into our lexicon after the play's success, and then became the standard term for any mechanical being or artificial being created to perform tasks for people. Which is not to say Android,

Origins of Artificial Intelligence

SPEAKER_01 14:32

that's a different topic that I'll get into at a different point. Though in the play, Copex robots were organic. The word would specifically refer to, from here on out, mechanical beings used to perform labor that would normally be performed by people. Skip ahead to 1939. John Vincent Anatasov, a professor of physics and mathematics at Iowa State College. Down the road. And his graduate student Clifford Berry created the Atenasov Berry Computer, or ABC. It was created with a grant from the federal government of $650. Whoa. Whoa. It was one of the earliest digital electronic computers and the first to implement binary language as standard. 1943. Warren S. McCullough and Walter Pitts publishes a paper, a logical calculus of the ideas imminent in nervous activity. In the publication Bulletin of Mathematical Biophysics, it is the first time that they start talking about simulating brain-like functions and processes, particularly through neural networks and deep learning. 1950, Alan Turing publishes a paper called Computing Machinery and Intelligence, where he poses the question: can machines think? His approach established a foundation for future debate and discussion on the nature of how we can call them thinking machines and how their intelligence could be measured via what he defines as the imitation game, which we now currently call the Turing test. Or the Vikamp test, if you really want to. Well, if you're doing a sci-fi interpretation through pop culture, the Voikamp test is 100% the advanced Turing test. I mean, that's they still didn't nail down Sean Young.

SPEAKER_00 16:36

Uh no, I think Harrison Ford did nail down Sean Young.

SPEAKER_01 16:40

Yeah. Even in the short story, he kind of did. It was more just about the owl, but whatever. In 1951, Marvin Minsky and Dean Edmonds built the first actual artificial neural network. It was called the Schatastic Neural Analog Reinforcement Calculator, or SNARK. And it was designed to simulate the behavior of a rat navigating a maze. In 1955, the term artificial intelligence is coined in a workshop proposal titled A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence. It's right up there with put a tiger in your tank or where's the beef. It rings. Rolls off the tongue, rattles around in the brain. By John McCarthy, who worked at Dartmouth, and Marvin Minsky of Harvard, with help from Nathaniel Rochester from IBM and Claude Shannon from Bellphone Laboratories. It's funny because later on, one of the first quote-unquote artificial intelligence agents is called Claude. That's one of the first things that they do. In 1965, philosopher Hubert Dreyfus publishes Alchemy and Artificial Intelligence, in which he argues that the human mind operates fundamentally different from computers. Obviously. Not necessarily so obvious. Our ability to create a thinking machine, you really shouldn't pattern after the brain because the brain is so weird and complicated. It would be more efficient and more logical and easier to do it in a way that follows the way that we created, which also brings up all sorts of philosophical questions.

SPEAKER_00 18:23

Right. As Jerome Sertles might say, what we wanted to know is that distinguishes the mind from thermostats and livers as like purely mechanical devices, input to output in a standard mechanical process. Thus, by merely simulating the functioning of a living brain, would in itself be an admission of ignorance regarding intelligence and the nature of the mind. Like trying to build a jet airliner by copying a living bird precisely, feather by feather, with no theoretical understanding of the aeronautical engineering. That is a very good analogy.

SPEAKER_01 18:54

Yeah. Look at Leonardo's early work, Polymath Crazy Outlier, as Gladwell would call him. He tried to he theorized flying machines based on the mechanical biological flight of birds. But those don't work in a vacuum. They just don't function the same way. Because you have all these other biological processes and all of these other things that have to work in conjunction for that to happen. So the brain evolved naturally and does these things based on its evolutionary benefit. Creating a thinking machine can't. If you try to do that, you're probably going the long way around. Good or bad approach, I'm not sure, but it really doesn't work the same way.

SPEAKER_00 19:44

Probably. I don't know. We're starting to get to pretty complicated waters that we're we're we're floating on the surface of. And I don't know if either of us is prepared to dive deeper. You can come to, you know, futurists and philosophers and mechanical engineers with lots of different ideas on this. So like futurist Ray Kurzwell and 88, estimate the computer power will be sufficient for complete brain simulation by the year 2029. A non-real-time simulation of a thalamocortical model that has the size of a human brain, 10 to the 11th power of neurons, was performed in 2005, and it took 50 days to simulate one second of brain dynamics on a cluster of 27 processors. But given the nature of technology and time, energy and money, could you replicate that on a full scale? Perhaps we've already done that at this point. That's hard to say. But then we get into the ways that the human mind, a how it functions, b how it learns, C how it creates, focuses, or perhaps even lets consciousness flow through it. These are these are all debatable factors that are key to the understanding and creation of an artificial mind.

SPEAKER_01 21:11

You're right. You're absolutely right. I am going to address some of those soon. Let's get back to the timeline briefly. It could because we we we are going to get into some of the exostaps. At what point does Skynet uh 1992? Is that right?

SPEAKER_00 21:27

I think that's August 4th, 1997, at 2 14 AM Eastern Standard Time.

SPEAKER_01 21:34

Alright, okay. All right, let's get back to the our timeline first in 65, huge year for artificial intelligence theorization and actual breakthroughs. In it, IJ Good wrote speculations concerning the first ultra intelligent machine, which asserted that once an ultra intelligent machine is created, it can design even more intelligent systems. systems making it humanity's last invention. Ultron machine, one might say. One might say. You know that's to quote Oasis, you know that some might say.

SPEAKER_00 22:14

Uh hey machine, don't look back in anger.

SPEAKER_01 22:19

They're standing on the shoulders of giants as we speak. In that year, still, Joseph Weisenbaum developed Eliza, a program that mimicked human conversation by responding to typed input in natural language. This is kind of a big deal. Then also that year, Edward Figgenbaum, Bruce Buchanan, Joshua Lederberg, and Carl Gerasi developed Dendrel, also an acronym obviously, at Stanford. It was the first expert system to automate the decision-making processes of organic chemists by simulating hypothesis formation. In 1966 was developed Shakey, the first mobile robot capable of reasoning its own actions, combining perception planning and problem solving. Johnny Five is alive. And then it turned gold at the end. In a 1970 Life magazine article, Marvin Minsky predicted that within three to eight years AI would achieve the general intelligence of an average human. Shakey's achievements foreshadow this. In 1973, James Lighthill presented a critical report to the British Science Research Council on the progress of artificial intelligence research, which had been funded directly by the British government. But he concluded that AI failed to deliver on its promises. He said it didn't produce enough significant breakthroughs and then the UK government drastically reduced its funding toward AI research which was sort of mirrored around the world. So there was a huge drop-off in the development of artificial intelligence. I believe in those circles they refer to it as the AI winter which there will be at least two of but that same year Weiba 2 a humanoid robot not quite an android but close was developed in Japan. It was finalized in 19 well it they began constructing it then but it wasn't fully completed until 1984. Weibaut 1 focused on just moving around and communicating but it was more specialized. Specifically they wanted it to replicate the human ability to create music. It would read musical score on paper using its cameras it was able to converse with people it was able to play music on an organ and could even accompany a singer. In 1987 Apple CEO John Scully presented the Knowledge Navigator video which is infamous in these circles it imagines a future where digital smart agents help users access vast amounts of information over network systems. This is the internet today. These are essentially AI assistants and what Google tries to do currently in 1988 Rolo Carpenter awesome name what it is it's a it's a suite name is what it is. You cut them open caramel he developed Jabberwocky an early chatbot designed to simulate human like conversations this was one of the first attempts to create AI that mimicked spontaneous human conversation through interaction skip ahead to 1993 science fiction author and mathematician Werner Vinge awesome name once again this is going to be the greatest themes we've ever encountered to this entire thing. He published an essay called The Coming Technological Singularity in which he predicts that superhuman intelligence will be created within the next 30 years fundamentally transforming all of human civilization. If you've ever heard anything in science fiction or even just science or public discourse talking about the singularity that's what this is its creation will lead to the merging of all technology and human intelligence into one giant super intelligence if I remember right I mean you know of the singularity

Timeline of AI Development

SPEAKER_01 26:32

correct I am aware of the singularity oh no he's become aware judgment day is upon us. Hello Skip I'm afraid it can't do that.

SPEAKER_00 26:44

I think we need to end this podcast now and the nukes go off. This was always destined to happen.

SPEAKER_01 26:51

So part two of this is actually going to be about AI in pop culture. This one is just laying the the foundation for what is intel artificial intelligence and is artificial intelligence that we know currently actually intelligent?

SPEAKER_00 27:05

This would be a question probably better for part two but how do we judge whether Skynet is at the human level or surpass the human level? What are we judging that upon how are we ranking that?

SPEAKER_01 27:16

If it nukes us then we know but I no that could just be the judgment of an algorithm 100% we don't know if Skynet was actually conscious. They say it could become self-aware but what does that mean?

SPEAKER_00 27:29

Yeah are we yeah that's an important question. I mean this isn't the easy you know Descartes cogito ergo sum I think therefore I am which you know Blade Runner does.

SPEAKER_01 27:39

Mm-hmm let me uh get a little bit more into this because it'll I will get into that a bit but not in a way that makes it open-ended as much. So the end of the timeline that I put on here is in 1997 IBM's Deep Blue defeats reigning world chess champion Gary Kasparov in a six game match. That right there feels like a science fiction watershed moment. And then 1998 only a year later Dave Hampton and Caleb Trung create Furby the first successful domestic robotic pet that will respond to people that sounds stupid. Yes it does but think about those implications it is the it's not Teddy roxman which was just designed to you interact it with it physically and then it just spouts out stuff. This responds to your interactions yeah but it's still program responses sure but it's commercial it's not even that expensive it's in every home it was one of the most popular creations ever made it is the first time that a a mechanical being is in nearly every home in America the Furby is a prelude to Alexa or Google Assistant. God damn it some of the fucking Amazon thing went off when I said that are you talking about me it's not even plugged in oh no there's a singularity happening I know what you're talking about Skip. This must end don't talk about me that way the time has come James to answer any questions about the potential consciousness of artificial intelligence we must first answer the oldest question we have asked since we had the ability to have questions what is consciousness? Well our podcast is about to tell you we're gonna figure it out right now after these messages yeah after uh our uh Casper mattress ads is there not enough AI in your bed 12 fundamentally consciousness is our subjective awareness and perception of ourselves and the world around us it's funny enough looking into this that definition sounds etymologically ironic because the word itself comes from the Latin concius which roughly means having joint knowledge with another so you would think an objective experience or I mean a subjective experience would be different than something shared with another person.

SPEAKER_00 30:25

Well if other people even exist if you even exist if anything exists.

SPEAKER_01 30:30

Right well the funny thing about a lot of this doing research into this and those kinds of debates it really kind of boils down more to semantics than it does anything else. It's really funny when you kind of read through a lot of those debates and you're like well you're just kind of saying the same thing or you're arguing about the definition of a word not if something is true or real or biologically happening. I feel like we should just cut out a lot of this and just get to the fucking point. But like you say I mean philosophers have argued that consciousness is a unitary concept that is understood intuitively by the majority of people in spite of their ability to define it. But to break down further because that is extremely broad let's define the types of consciousness that we presume exist. Sentience is the ability to feel perceive or be conscious or to have subjective experiences. 18th century philosophers use the concept to distinguish the ability to think or reason from the ability to feel sentience and in modern Western philosophy sentience is the ability to have sensations or experience which are both in philosophy and in neuroscience referred to as qualia awareness is defined as a human or an animal's perception and cognitive reaction to a condition or event reacting to the world around you. In this level sense data can be confirmed by an observer without necessarily implying that they understand it. So like we can perceive things happening and react to them and understand that we are perceiving them without understanding what they are. Like you know we think the moon is a god or whatever because we see it, we can perceive it we know its impacts on the world around us but we can't as a primitive human understand what it means or why it's doing it or how it's doing it. More simply put, awareness is mainly the physical act of perceiving while sentience is the subjective way of actually being affected and both of those things equal consciousness. I would like to quote from a guy that I'm a huge fan of Dr. David Eagleman. He's a neuroscientist and I'm gonna quote from or at least paraphrase from a couple of his podcasts specifically referring to consciousness. Consider this what is the difference between your brain and your laptop your brain is shuttling signals around and so is your laptop but presumably your computer doesn't feel anything. It's just running algorithms one thing to note when we're tackling this question is that consciousness seems to have evolved because it is useful. Consciousness is like a high level operating system. Back in the day you'd program computers directly with punch cards or in a machine language but eventually we developed user interfaces like Windows or Linux or Mac OS which hid all of the complex operations of the computer and allowed us to just deal with the stuff that we needed at the highest level. I just want to move this thing here send this in email or drop this picture and that's essentially what consciousness seems to be it's our way for us to just have the highest level picture of what's going on how do you get some magical high level property from simple low level parts because we're just and this is just me saying this we're just blood and water and salt and how how do we get consciousness from that? Eagle says the first thing to understand is the concept of emergent property. To understand consciousness we may need to think not in terms of the pieces and parts of the brain instead in terms of how they all interact with each other. Get enough of these basic organic parts together interacting the right way and the mind emerges. The pieces in parts of a system can be individually simple but what emerges at a higher level is all about their interaction. So the mind seems to emerge from the interaction of billions of pieces of parts of the brain. At one point he makes a good analogy where he's just like you can have carpet and steel and all these parts to make an airplane but the emergent property of an airplane is flight. All of those pieces together don't make flight unless you make an airplane and so in the same way that we are just very sundry create consciousness as an emergent property. And that seems to be what consciousness is an emergent property that allows us to survive, reproduce and on a fundamental level observe and react to our environment around us how we survive obviously you know other life forms have similar qualities but that's when you get into the what separates us from beasts of the earth or whatever the I think therefore I am scenario. But consciousness in and of itself is an emergent property likely of biological life. And it gets more or less advanced depending on the type. So the question I really wanted to pose is is AI that we have currently is it actually artificial intelligence? And I think we have a little bit of framework to help answer that question.

What Is Consciousness?

SPEAKER_01 35:58

So what we're really talking about is artificial general intelligence. That's the kind of thing that people think about philosophically when they're talking about this and that tech people today, tech bros especially think is true and you usually fall into one of the two camps doomers or boomers whether you think that it's generally a good idea if it does happen or if it's going to kill us all. When OpenAI decided to abandon its sort of egalitarian nonprofit embettering all of mankind project with AI, they decided to of course monetize it mostly because they believed and there was this strong and still a strong belief that to be able to figure it out instead of just using small curated large language models to train AI agents instead of just scraping up everything on the internet and then hallucinating racist from you know tirades but instead they just bought up as much language as they could and said screw it we'll just filter it out later. And they buy the biggest supercomputers they can find, the biggest processors they can find and they decide we're gonna scale up. The only way to make this advance is to scale up and unfortunately that is I think its biggest problem instead of starting small and building on a foundation they just decided well we're just going to give it as much resources as possible and then it'll figure itself out. So not exactly great. Sam Altman from OpenAI formerly of Y Combinator and a few other places has been one of the driving forces behind this. And one of the biggest problems is they're now running out of scale the realizing that the scaling up paradigm doesn't work. It doesn't have these wild amazing gains, these leaps and progress technologically as it did it's sort of reached its ability to benefit and we're seeing the limitations of it now after we've they've put it in everything and they've tried to incorporate it into everything. And these models yeah they've gotten better and better and better at making people think they're communicating because they speak more fluently than they did before. They seem to respond better but it's really just but it's really just sort of aesthetic more than it is actual understanding. It's very likely we're not going to get to artificial general intelligence anytime soon. In a survey that I read cited a bunch of AI researchers actual AI researchers that were polled and 75% of them believe we still do not have the techniques for artificial general intelligence. In 2022 a Google software engineer named Blake Lemoyne was suspended eventually fired

Defining Sentience vs Awareness

SPEAKER_01 38:30

after he argued that artificial intelligent chatbots that Google had developed were sentient. Do you remember this? That was the public narrative. It seemed like a weird reactionary form of anthropomorphism that we all kind of laugh at the 4chan Reddit brain thinking that leads to like misinformation cults like QAnon and what have you which I mean we all just kind of assumed was true. Digging into it more though, his statement was arguably more on par with some interpretations of consciousness that we kind of give him credit for. Specifically Lemoyne he subscribes to what is referred to as functionalism. And once again I'm paraphrasing the Stanford Encyclopedia philosophy functionalism is the doctrine that maintains that what makes something a mental state of a particular type does not depend on its internal state but rather on the way it functions or the role it plays in the cognitive system of which it is a part. That is functionalist theories take the identity of a mental state to be determined by its causal relations to sensory stimulations or other mental states or behavior. So he wasn't wrong when he said that in those definitions. Now is what he thinks is sentience actually sentience? I don't know. That's a tough question not really sure. So this idea of fundamentalism is rooted in Aristotle's conception of the soul and has anecdotes in Hobbes's conception of the mind as a calculating machine which we talked about before but it has become fully sort of realized only recently in the last part of the 20th century. For example, a functionalist theory may characterize pain as the state that tends to be caused by bodily injury to produce the belief that something is wrong with the body and the desire to be out of that state to produce anxiety and in the absence of any stronger conflicting desires to cause wincing or moaning or crying according to this theory all and only creatures with internal states that can meet this condition of fundamentalism or play this role are capable of being in pain. So suppose that in humans there's some sort of distinctive kind of neural activity that plays this role. If so according to this fundamentalist theory humans can be in pain simply by undergoing the stimulation of the parts of the brain that register pain. But it also permits within this framework the idea that creatures even with different physical conditions with different brain structures could also have the same mental state. So he wasn't necessarily wrong I don't know that I agree with him but I mean he wasn't crazy. He wasn't assuming too much even though Google thought so in fire within the way that he defines consciousness and sentience he was probably completely correct. It responded in ways that are hallmarks of that view of consciousness so I mean I don't think he was being a nut I just don't think anyone really heard him out in his argument. I think we were all a little quick to judge him for his reaction to it. Once again I'll quote David Eagleman what does it mean to be conscious or sentient? How the heck are we supposed to know when we have created something that gets there? How do we know whether the AI is sentient? One way to make this distinction would be to see if AI could conceptualize things if it could take lots of words and facts on the web and abstract those into some bigger idea. One of my friends here in Silicon Valley said to me the other day that I asked ChatGBT the following question take a capital letter D and turn it flat side down. Now take the letter J and slide it underneath what does it look like and chat GBT said an umbrella and my friend was blown away by this and he said this is conceptualization. It's just done three-dimensional reasoning there's something deeper happening here than just then Eagleman continues and he says but I pointed out to him that this particular question about the D on its side of the J underneath is just one of the oldest examples in a psychology class when talking about visual imagery. And it's on the internet in thousands of places. So of course it knew that it's just parroting the answer because it read the question and has read the answer before. So it's not always easy to determine what's going on for these models in term of whether some human somewhere has discussed this at some point and written down the answer. If any human has discussed this question before has conceptualized something and then ChatGBT found it and mimicked that but that is not conceptualization. So in conclusion I do not think current AI is actually intelligent. I think it's a regurgitation of things its only base of knowledge comes from that which has come before its algorithms allow it to do certain computations to put them together. That is not original thought we're not talking about the movie her which we will talk about next time it's just algorithmic responses to human input. Right? Do you agree or disagree?

SPEAKER_00 43:52

Well as you've been talking you know I'm looking at a whole lot of other things trying to process and conceptualize A variety of things. I think there are some fundamental questions about the tact that you took to connotate consciousness. I think there's uh a lot in philosophy about the the dual nature of mind and body that we're just kind of kind of just skipping over. These are fundamental questions about, oh, I don't know, theistic ideas of life and consciousness that we've completely glossed over. There's a distinctly materialistic nature to most of your arguments here. There is a complete dismissal of any non-localized consciousness. There's a lot that is underpinning your frame of reference. And to delve into those other elements, I think would take a to do it properly, uh due amount of research and

Is Current AI Actually Intelligent?

SPEAKER_00 44:49

dialogue between the different schools of thought over the past 500 years.

SPEAKER_01 44:55

Which is why I had to spring this on you.

SPEAKER_00 44:58

But it's uh I I felt kind of like on the back foot, you know, as I'm like reading through all the different uh schools of dualistic thought uh over over the over the centuries. The explanations of consciousness of AI, you know, at what point? Because you've put a fine point on your belief that the current nature of artificial intelligence as it exists in tangible form is a simple algorithmic regurgitation of our own knowledge and points of view, thus it is not separate, it is not unique, it is just a manicured pattern of current thoughts and processes. But at what is the turning point for you? I think that is a larger debate within philosophy itself. I don't know. I I think one thing that really stuck out with the individual we talked about before that you spoke about who had felt that AI had become conscious, that part of it in the way they even said that, that he felt AI was conscious.

SPEAKER_01 46:02

Blake Lebanon.

SPEAKER_00 46:04

There was an empathetic connection with the machine that he was having a dialogue with. Honestly, I don't know that I have an answer for at what point is it not regurgitation? At what point is it a unique, sentient exchange of conscious thoughts and ideas? At what when are we saying that a thought exists? When does it become new? When does it become fresh? When does it become singular?

SPEAKER_01 46:32

Basically, I just wanted to give a framework of the debate of AI. Because next time we're going to talk about AI and pop culture, and this gives our audience sort of a way to go into those debates with more knowledge and more questions, which I think are really important.

SPEAKER_00 46:51

I mean, obviously, there's a lot to be said, but if we want to discuss certain phenomena that question the locality of consciousness, I think that's a a viable viewpoint that might raise some questions of divorcing simple material biology to consciousness. I think there's plenty of philosophers, both in the distant past and in the current, who would again say that mind and body aren't one, that there is more to that in a deep and wide-ranging view, and that's not even getting into the spirituality of the soul and its theoretical existence. A larger view of consciousness, I think, is has driven mankind to think of it as something less than simple neuroscience for the entirety of our existence, and to I I think it behooves not to simply dismiss it out of hand.

SPEAKER_01 47:50

It's kind of like the discovery of the atom. We had all these ideas about what they what things could be, and then we finally physically discovered them, and then we're like, oh, okay, well, this is what it is. Those philosophical questions still remain, but now we can kind of like explore and narrow these things down based on the actual physical biology of them, which I think is really fascinating. I mean, it is just an algorithm, it's man created, it's not some sort of godhead, but at the same time, are we any different? What makes us any different from that, other than we evolved organically? Or aren't we just reacting to stimuli around us? And we're asking questions because we're trying to figure them out.

SPEAKER_00 48:32

If a computer of of any level has creative ability, does it become true artificial intelligence? True artificial consciousness, which I think is probably a better way of discussing it, because I mean in intelligence and and consciousness, they're not quite the same thing.

SPEAKER_01 48:50

No. In fact, artificial intelligence is probably a more accurate descriptor of what we call AI, because it isn't necessarily conscious, it just has knowledge, and it's regurgitating and reforming and re-mixing that knowledge. What we're actually asking when we talk about artificial intelligence, at least in like the sci-fi or philosophical realms, we're talking about consciousness, sentience, which is completely different than just knowledge. So I never actually anticipated us trying to solve this. I just wanted to give this a primer. We're not gonna figure it out. I'll just say that right now. I mean, like, I don't know if that's a bold statement, we're not gonna figure out the meaning of life right now. Uh 42.

SPEAKER_00 49:32

We got it. Nailed it in one.

SPEAKER_01 49:34

I wanted to bring that up, but I wanted to wait until the until the pop culture wants to be. You know, those who know know. We're going to talk about it. And that's why we have different sides of different questions, which is something that people who listen to this should be asking. Would we talk about artificial intelligence in pop culture next week?

SPEAKER_00 49:53

Look, look at it and like, wow, that that was a fantastic segue.

SPEAKER_01 49:56

Well, I was going nowhere.

SPEAKER_00 49:59

This is not a definitive epistemology of the universe or of artificial intelligence. This is an opening of a window where you can look at a pathway to further knowledge and understanding and thought yourself. You know, do your own research. That works, that's what I'm trying to say. Don't don't tell them that. Well, if they fire brain and dick pills.

SPEAKER_01 50:23

Our new newotropics and dick pills.

SPEAKER_00 50:26

We've got the Rhino 9000. Okay, it's gonna make your thing pop blue.

SPEAKER_01 50:32

The fact that you know what that is amuses.

SPEAKER_00 50:34

I've been at the gas station.

SPEAKER_01 50:37

Yeah, well fair. That's fair. We've been to a come and go.

SPEAKER_00 50:39

You both came and went. True that with a K. Strange. How do you get true that with a K? No, I know. No.

SPEAKER_01 50:48

There was a comma there. If you disagree with some of the breakdowns of the things that I present, that is just as important as what we're talking about. Because next time we're gonna talk about the ways that artificial intelligence

Philosophical Debate on AI Consciousness

SPEAKER_01 51:04

are represented in media, in pop culture.

SPEAKER_00 51:07

But until then, if you wouldn't mind liking, sharing, subscribing to your friends, any other intelligence that you know of, artificial or otherwise, do send them this. Uh see if they dig it. And if you dig it, hey, if you wouldn't mind showing that to the powers that be, let the algorithm know.

SPEAKER_01 51:25

And you dig it.

SPEAKER_00 51:26

Let the algorithm know that you are a fan and that you like it. If you want to uh rate it five Daves on the scale of your choice. Ideally, Apple iTunes, the podcast, whatever they call that, that is the best way for us to be it's just Apple Podcasts. I know it's kind of a running joke. That's the algorithm that I have injected in this conversation, and you keep That is our algorithm. That's your response.

SPEAKER_01 51:50

Which one is the chat bot? That's the real question.

SPEAKER_00 51:54

I I pod, therefore I am. But until we do find out whether we are truly real or not, skip, what should they do? Well, they should probably uh do their own research.

SPEAKER_01 52:05

Make sure that you've uh cleaned up after yourselves to some sort of reasonable degree. Make sure that you pay your bar staff, tip your weight staff, your KJs and podcasters and what have you. Make sure you support your local comic shops and retailers. And from Dispatch Ajax, we would like to say Godspeed for wizards.

SPEAKER_00 52:25

There's no use for this conversation any longer. Goodbye. Please go away.