ReadMultiplex.com: The First Ultraintelligent Machines: A 1965 Vision Of The Intelligence Explosion.
40 minAudio
Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! “The survival of man depends on the early construction of an ultraintelligent machine.” What if the foundational text of the entire AI doomsday movement, the exact document that everyone cites to prove supercomputers will inevitably wipe us out. What if it actually begs humanity to build artificial intelligence as fast as possible? It completely shatters the narrative we're fed every single day. To save our species, basically. We're operating in this era where self-appointed AI "leaders" and "philosophers" constantly warn us about rogue algorithms. It's the classic Terminator scenario. And they always point back to one specific moment in history to legitimize their panic. They treat this original text like some dark prophecy. Like a warning from the past. But when you actually strip away the modern anxiety and read the physical words on the page, you find something that is remarkably, almost desperately optimistic. But who wrote this and why was this paper so often quoted? How can modern AI advocates who are constantly lobbying for a small cadre of human overseers, corporate commissars, and global planning boards to manage artificial intelligence on behalf of the entire planet? How can they possibly use the text of a man who explicitly argued that centralized power inevitably leads to crimes committed in the name of high ideals? They can't. Not honestly, anyway. It's totally incompatible. Read about this 1965 paper.. Read more at: ReadMultiplex.com Support this work by buying us a coffee: https://ko-fi.com/brianroemmele
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
So what if the foundational text of the entire AI doomsday movement, the exact document that, you know, everyone cites to prove supercomputers will inevitably wipe us out. What if it actually begs humanity to build artificial intelligence as fast as possible? Oh, I mean, it's wild because it completely shatters the narrative we're fed every single day. Right. To save our species, basically. Exactly. We're operating in this era where tech billionaires and, well, philosophers constantly warn us about rogue algorithms. Yeah,
the classic Terminator scenario. Right. And they always point back to one specific moment in history to legitimize their panic. They treat this original text like some dark prophecy. Like a warning for the past. But when you actually strip away the modern anxiety and, you know, read the physical words on the page, you find something that is remarkably, almost desperately. Okay, let's unpack this. Because the scale of the historical distortion we're looking at today is just staggering. It really is. Our mission for this deep dive is to act as intellectual detectives. So for you
listening, we're taking you along for the ride. We have a really fascinating stack of primary sources in front of us, all orbiting a single... highly influential 1965 monograph. Right, 1965. Written by a brilliant mathematician and Bletchley Park codebreaker named I .J. Goode, and it's titled Speculations Concerning the First Ultra -Intelligent Machine. Such a great title, honestly. It's very mid -century sci -fi, but it's a real rigorous academic paper. And today... We're going to rescue this brilliant paper from decades of
being misquoted, misunderstood, and frankly, weaponized by modern AI fearmongers. What's fascinating here is how a single isolated paragraph, I mean, surgically extracted from a much larger and more complex document, has the power to hijack. half a century of intellectual history. It's like taking one scene out of a movie and ignoring the rest of the plot. Exactly. We need to look at the raw evidence of what I .J. Goode actually wrote, the historical crucible he was operating in, and the specific mechanical reasons he believed machines were our best and perhaps only hope.
To really grasp the magnitude of this hijacking, we have to start with the, quote, herd around the AI world. The big one. The big one. If you follow technology even a little bit or if you've read any of the major books on existential risk, you have almost certainly encountered a variation of this famous paragraph. Oh, without a doubt. I'm going to read it precisely as it appeared in his 1965 paper. So Goode wrote, Let an ultra -intelligent machine be defined as a machine that can far surpass all the intellectual activities
of any man, however clever. Setting the baseline. Right. Then he goes on. Since the design of machines is one of these intellectual activities, an ultra -intelligent machine could design even better machines. There would then unquestionably be an intelligence explosion and the intelligence of man would be left far behind. And here comes the punchline. Yep. Thus, the first ultra -intelligent machine is the last invention that man need ever make, provided that the machine is docile enough to tell us how to keep it under control. And there you have it. The absolute blueprint for
modern techno panic. I mean, it literally reads like the prologue to a science fiction horror story. It really does. The intelligence of man left far behind. The last invention we'll ever make. And then that... terrifying little caveat hanging at the end, provided the machine is docile enough to tell us how to keep it under control. Right. That specific word, docile. When you hear that paragraph. Entirely in isolation. I mean, you could easily picture a scientist realizing he's about to build a monster and trying to frantically
warn the future. You certainly can't. And that is precisely how it is leveraged today. But two absolutely critical pieces of context are systematically omitted when that passage is repeated on the conference circuit. Always omitted. First, that famous docile proviso. When you read the surrounding text, it becomes incredibly clear that this wasn't a manifesto of existential dread. No, not at all. It was a practical engineering condition attached to a project he was wildly enthusiastic about. Goode wasn't starting from the premise
that superintelligence is a catastrophic threat that needs to be smothered in its crib. Right. He wanted to build it. He was starting from the premise that humanity was in profound geopolitical danger, that we required vastly more intelligence than our biology could provide, and that engineering this machine was the only viable solution. I think of it like reading a complex medical textbook about a high -stakes heart surgery. Oh, that's a good analogy. Yeah, the modern doom narrative essentially highlights one sentence that says, the surgeon must be careful when slicing the
aorta as failure will result in catastrophic bleeding. Which is true. It's true. But then they run out of the hospital screaming, the book says the scalpel will cause catastrophic bleeding. We must ban scalpels. Right. Ignoring the fact that the patient has a fatal heart condition. Exactly. They completely ignore the very first page of that same textbook. Which brings us to the actual opening line of Good's paper. The one nobody ever quotes. The sentence the doomsayers pretend doesn't exist. Good begins his monograph with these exact words. The survival of man.
depends on the early construction of an ultra -intelligent machine. It is a breathtaking reversal of how he is portrayed. It's the literal first sentence. The survival of man depends on the early construction. I mean, this is not the language of a man cowering in fear of a silicon overlord. No, it's a burgeon plea. It's the plea of a pragmatic scientist who views greater -than -human intelligence as a vital, necessary tool. He didn't just want it built, he wanted it built early. I have to push back a little on the phrasing, though, for
the sake of the listener. Sure. Go ahead. Why did he use a word as extreme as survival? I mean, if we were talking about 1965, the Beatles are on the radio. You have the post -war economic boom happening. Was the world really so precarious that a rational mathematician thought we needed a literal machine to save us from ourselves? To understand the weight of the word survival. You have to mentally transport yourself into the geopolitical furnace of the early 1960s. The Cold War. Exactly. The final draft of this
paper might have been published in 1965, but Goode was writing and formulating these ideas during a period of peak nuclear tension. Right. The Cuban Missile Crisis was, what, 1962? 1962, yes. Humanity came within hours, perhaps minutes, of thermonuclear annihilation. Good wasn't writing from a place of fearing machines. He was looking at Washington and Moscow and fearing human incompetence. He was watching flesh and blood world leaders edge toward the apocalypse and realized our biological
hardware just wasn't up to the task of managing atomic era risks. Precisely. He realized that the ordinary political machinery of the mid -20th century. The committees, the egos. The egos, the nationalistic tribalism. It was entirely insufficient to manage the existential threats we had invented. Our brains weren't built for that. The human brain evolved to survive on the savanna, not to navigate the complex game theory of mutually assured destruction. Right, of course. Goode genuinely believed that only a machine, something devoid of human eco, something capable
of processing vast amounts of strategic data objectively, could reliably manage those risks. So the AI wasn't the threat. The AI was the required disasex machina to save us from our own destructive nature. So to truly grasp... the mechanics of why he believed a machine could be our salvation, we can't just parse the text of the paper. We need to look at the architect himself. Yes, his background is crucial. Because I .J. Goode wasn't an armchair philosopher dreaming up science fiction tropes in a university lounge. He had a deeply
intimate, hands -on relationship with the world's very first computers. He was in the trenches. Literally. He didn't just imagine artificial intelligence. He spent his youth physically wiring the precursors to it. His life story is fundamental to understanding his philosophy. Tell us about his early days. Well, he was born Isidore Jacob Gudak in December 1916 in London to Polish Jewish. Immigrant parents. Okay. His father was a watchmaker and a jeweler in Hackney. They lived in very modest circumstances, far removed from the aristocratic
elite of the British academic establishment. What I find so compelling about his early life is that he wasn't your typical out -of -the -gate academic superstar. Right. He was actually quite slow to learn to read. Yeah, but then when he was nine years old, he caught diphtheria. Which was incredibly dangerous back then. Super dangerous. Yeah. So he is confined to his bed, recovering, and out of sheer boredom and raw curiosity, this nine -year -old child independently discovers the mathematical proof for the irrationality of the square root of two. It's just staggering.
He just starts exploring mathematical induction entirely on his own. Nine years old. That kind of spontaneous discovery is a testament to his raw processing power. Right, because mathematical induction isn't just basic arithmetic. No, it's essentially the logic of proving that if a statement is true for one number, it must be true for the next, setting off a domino effect of infinite proof. For a child to independently grasp that shows an incredibly rare level of abstract reasoning. I can't even imagine what his teachers thought. They obviously recognized they had a generational
prodigy in their classroom. He eventually earns a first class degree in mathematics at Cambridge in 1938, completes his Ph .D. in 1941 under the supervision of legendary mathematicians like G .H. Hardy. And then history comes calling immediately. Yes. 1941. He gets recruited to Bletchley Park, the legendary center of allied code breaking during World War Two. The absolute nerve center. And the timing of his arrival is incredibly cinematic. He walks into Hut 8. on May 27, 1941. That happens
to be the exact same day the Royal Navy sank the German battleship Bismarck. The atmosphere in that room must have been electric. Electric, chaotic, and unimaginably high stakes. The German U -boats were strangling the supply lines to Britain. If those codes weren't broken, the island would starve and the war would be lost. And who does Goode go to work for in Hut 8? Alan Turing. Of course. They are tasked with attacking the German naval Enigma codes, which were far more complex than the Army or Air Force versions. I have to bring up this specific anecdote about
his early days at Bletchley because it perfectly captures the eccentricity and the intense meritocracy of these guys. Oh, the nap story. Yes. So Goode is working this grueling night shift and he decides he needs to take a nap. But instead of finding a cot or going back to his billet, he just lies down on the floor of the hut and goes to sleep. Just right there on the floor. Right on the floor. And Turing, who was... notoriously intense and, you know, socially awkward, is incredibly irritated by this breach of decorum. He hated it. Turing actually refuses to speak to Goode after that.
He gives him the silent treatment for days. Which, in the high pressure cooker of Hut 8, must have been incredibly stressful. Exactly. But here is how Goode breaks the ice. He doesn't apologize. Instead, he proves himself by solving an immensely complex practical problem that Turing was struggling with. Right, the padding problem. It involved the non -random padding letters the Germans used in their cipher indicators. Let's unpack the mechanics of that problem because it introduces a concept that defines Goode's entire worldview,
which is Bayesian statistics. Okay, walk us through it. So the Germans would send a message and to confuse the allied code breakers, they would add meaningless padding letters to the beginning or end of the transmission. Just extra gibberish to throw them off. Exactly. The Germans assumed these letters were completely random, making them cryptographically secure. But Goode realized something profound about human nature. Humans are terrible at being truly random. Because if I tell you to type random letters on a keyboard, you aren't going to perfectly distribute your
keystrokes. Right. Your fingers will naturally gravitate toward the center keys, or you'll accidentally type patterns you are familiar with. Like typing A -S -D -F over and over. Exactly that. Good applied a Bayesian framework to this problem. Now, if you listening aren't familiar with Bayesian statistics, think of it this way. Yeah, give us the quick crash course. Instead of looking at an event in isolation and guessing the probability, a Bayesian constantly updates their beliefs based on incoming evidence. Okay. If you flip a coin and it lands on heads 10 times in a row, a traditional
statistician might say the odds of heads on the next flip are still 50 -50. Because each flip is independent. Right. But a Bayesian looks at that evidence. updates their prior assumptions, and concludes there is a very high probability the coin is rigged. That makes a lot of sense. So Goode applied this to the German telegraph operators. He mapped the physical layout of the Enigma keyboard and calculated the probabilities of which keys a tired, stressed German operator was most likely to hit when trying to fake randomness.
Wow. He found the human bias inside the machine. By exploiting that bias, he cracked the padding cipher. And once he solved that, Turing forgave the nap on the floor, and they became incredibly close colleagues. It's just a brilliant story. Pure meritocracy. But Goode's journey at Bletchley didn't stop with Enigma and Turing. In 1943, he transitioned over to Max Newman's section, the Newmanry. This is where things get really heavy technologically. Yeah. This group was tasked with attacking a cipher system that was vastly
more complex than Enigma de Lorenz teleprinter ciphers. Codenamed FISH. Right, which were used for the highest level communications between Hitler and his generals. And this transition is where I .J. Goode steps out of the realm of pure mathematics and intersects with the literal birth of modern computing. Because human brains weren't enough anymore. Exactly. To crack the Lorenz ciphers in time for the intelligence to be actionable, human brains and mechanical relays were simply too slow. They had to build Colossus. Colossus. We're talking about the world's first
large -scale, programmable electronic digital computer. A massive leap forward. It used thousands of vacuum tubes to process paper tape at blistering speeds. Good worked closely with another brilliant mind, Donald Michie, on the statistical methods that made Colossus effective. Yes, Michie is another legend. But the real breakthrough the moment that I think planted the seed for his 1965 paper was when Good and Michie proved that Colossus was far more flexible than its original engineers had intended. That is the pivotal moment.
The original engineers designed Colossus as a single -purpose calculator. Like a really fast abacus. Right. It was built specifically to simulate the spinning wheels of the Lorenz cipher machine. But Goode and Mischie looked at the architecture, the plug boards, the conditional switches that allowed for rudimentary if -then logic, and realized they could hijack the machine's capabilities. Wait, so they reprogrammed it? Yes. They figured out how to use the paper tape not just for feeding data, but as a rudimentary memory store to loop instructions. Oh, wow. They effectively reprogrammed
the machine to perform tasks far beyond its initial physical parameters. They pushed a single -purpose calculator into the realm of a programmable computer. By the end of the war, Goode was the principal statistician for this entire massive technological effort. Connecting this to the bigger picture of the AI doomsday debate. This wartime experience is the massive paradigm shift for I .J. Goode. Absolutely. He didn't just sit in a university a decade later and theorize about intelligent machines. He literally helped build the first
ones. He physically rewired them and he watched them execute complex logical procedures at speeds that left human mathematicians in the dust. And more importantly, he saw the moral and physical outcome of that speed. That is the essential context that modern critics entirely miss when they analyze his 1965 paper. Good spent his 20s watching early computers intercept messages, break codes, and generate actionable intelligence that directly shortened the deadliest conflict in human history. They weren't abstractions. They were saving lives. Countless lives. They
were the instrumental edge that allowed the Allies to defeat the ultimate totalitarian regime, Nazi Germany. For I .J. Good, machines were not abstract agents of oppression or existential ruin. They were tangible liberators. They were tools of human survival. When you understand that he spent his formative years using early computers to save the free world from fascism, his 1965 statement that the survival of man depends on the early construction of an ultra -intelligent machine suddenly makes perfect rational sense. It totally reframes his mindset. The machine is the monster.
The machine is the weapon that slays the monster. And what's wild is that Goode wasn't just a big ideas philosopher saying, hey, we should probably build a super brain to solve our problems. He actually used his 1965 paper. to lay out a highly specific biologically grounded roadmap for exactly how to engineer this intelligence explosion. He got very technical. Yeah, this is where the monograph dives deep into the cutting edge neuroscience of the era. Yes, sections four through eight of his monograph are not about computer hardware at all. They are deeply concerned with the physical
realities of the biological brain. Drawing on psychology. He was drawing heavily on the work of a Canadian psychologist named Donald Hebb. In 1949, Hebb... published a foundational book proposing what became known as the cell assembly theory. I think I've heard of this. Is that the neurons that fire together thing? Yes. It is often summarized today with the phrase neurons that fire together wire together. Hebb argued that groups of neurons form physical networks or assemblies that serve as the fundamental basis for concepts, memories, and learning. So to put
that in plain English for everyone, if I look at an apple, a very specific network of neurons lights up in my brain. Correct. And if I keep looking at apples, touching apples, tasting apples, that physical neural pathway strengthens. Over time, that physical connection literally becomes the concept of apple in my mind. That is the essence of Hebbian theory. It grounds abstract thought in physical biology. But Good looked at Hebb's theory and, well... While he thought it was brilliant, he felt it was mechanically
incomplete. Also. Goode believed that if an entire large -scale cell assembly fired every single time you thought of a concept, the brain would suffer from cognitive gridlock. Like too much information at once. You would be overwhelmed by associated memories. So Goode introduced a major innovation he called subassemblies. Subassemblies. Yes. These were much smaller, more stable, highly specialized groupings of neurons that could explain much finer psychological controls. It's like having highly specialized task forces within the brain rather than deploying the entire military
army every time you need to solve a minor problem. That's a perfect way to put it. How did he use these sub -assemblies to explain actual human thinking? He used them to explain deeply complex physical phenomena like psychological priming, memory interference, and that incredibly frustrating tip -of -the -tongue state we all experience. Oh, I hate that feeling. You know what letter it starts with? You can almost taste it, but you just cannot retrieve the whole word. Yes, it's agonizing. So how does a subassembly explain that mechanically? Goode argued that in a tip
-of -the -tongue state, the adjacent subassemblies are firing. The subassembly for the letter P is firing, the subassembly for the conceptual meaning is firing, but the specific target subassembly lacks the final chemical threshold or activation energy to fully trigger. That's so specific. It is. Goode argued that subassemblies provide a much more realistic granular model of how meaning and recall are actually embodied in the physical brain. And, crucially by extension, he believed this biological architecture was exactly how
meaning would have to be embodied in an artificial system. This is so prophetic it actually gives me chills. It really is. I mean, we're talking about the early 1960s. This is long before the Internet. long before the massive GPU clusters we have today. Yet Goode is predicting that the first ultra -intelligent machine wouldn't just be some giant top -down calculator running millions of lines of if -then code. No, he's still past that. He predicted it would be built using vast, bottom -up artificial neural circuitry, and that it would be trained, much like a human child,
through reinforcement, trial, and interaction with its environment. He basically described the exact neural network architecture that is powering the deep learning boom today. He was decades ahead of his time. He even consulted for IBM during this period, looking at early perceptrons. What's a perceptron? A perceptron is essentially the earliest, most basic algorithmic model of a biological neuron, a mathematical function that takes multiple inputs, weighs them, and produces a single output. Ah, okay. Goode looked at these early models and was utterly
convinced that an ultra -intelligent machine would require what he called an economy of meaning. An economy of meaning. Yes. It couldn't just rely on abstract formal symbol manipulation, which was the dominant AI theory of the time, treating intelligence like a massive spreadsheet of rigid rules. He argued the machine needed a genuine physical substrate for memory and concepts, mirroring the sub -assemblies in our own brains. But wait. If he saw the machine's brain operating so closely to human architecture with its own
versions of neural pathways, threshold activations, and tip -of -the -tongue states, did he think the machine would inherit our biological flaws too? Like what? Like emotional irrationality or bias. You're touching on a vital tension in his work. Good was a pragmatist. He explicitly noted in the paper that these machines will create social problems. He wasn't naive. Right. But his focus on the economy of meaning was ultimately about extreme efficiency. He believed that if we modeled the machine on the best aspects of human neural architecture, specifically these
efficient subassemblies for storing and retrieving conceptual structure, we could engineer an intelligence that was vastly superior in speed. While leaving the bad stuff behind. Exactly. Carefully stripping away the evolutionary baggage, the emotional volatility, and the tribalism that plagues human decision -making. Okay, so we have firmly established that I .J. Goode was incredibly optimistic about the technology. He saw machines as wartime liberators. He had a concrete, neuro -inspired blueprint to build them. And he literally wrote... that
our physical survival depended on getting this machine built early. The record is clear on that. Now we have to turn to the darker side of this story. We have to examine how this brilliantly optimistic, constructive work was hijacked and weaponized to support the exact opposite philosophy. This is where the intellectual distortion occurs. In the decades following 1965, and particularly with the massive rise in funding and influence of what we now call the AI risk and AI alignment communities, Good's paper began to be cited in a highly selective, almost surgical manner. Let's
name the specific group. groups and figures who co -opted his work. The Modern Source is very explicit about this. We are talking about highly influential philosophers and writers like Nick Bostrom, Eliezer Yudkowsky, and the broader institutional networks that have built up around their ideas. The modern AI risk community. Right. When these figures talk about I .J. Goode, they absolutely hammer that one famous paragraph we read at the beginning. They quote the intelligence explosion to prove the mathematical inevitability of danger.
And they lean heavily on the docile enough clause to justify their massive regulatory solutions. That is their primary rhetorical maneuver. They utilize that docile condition to argue for the absolute necessity of enormous centralized AI safety bureaucracies. Right. To slow everything down. They use his words to support a concept called regulatory capture. For those listening who might be unfamiliar, regulatory capture is when an industry advocates for complex government regulations that sound like they're protecting the public. But the regulations are actually
so expensive and complex that only the biggest, most entrenched companies can afford to comply with them. Effectively locking out. new competition. Precisely. The modern AI risk community uses Good's warning to push for strict government oversight, licensing boards, and a general posture of extreme paralyzing caution. They portray Good as an early, tragic prophet of existential catastrophe who realized the machine was a demon we had to keep eternally chained up. Here is where it gets really interesting, but also incredibly frustrating.
Yeah. The sheer hypocrisy of it. It's palpable. They essentially took a massive, scientifically rigorous call to build and twisted it into a panicked call to freeze. They inverted the entire emotional and practical spirit of the document. Yeah, absolutely did. But to play devil's advocate for just a second, if Goode was as much of a genius as we've established, shouldn't we take his docile warning just as seriously as his survival mandate? I mean, the fact that he included it means he recognized the danger, right? It's not that Goode was devoid of caution. He was a brilliant
mathematician and a statistician who worked in cryptography. Of course, he cared deeply about control. Right. He wasn't reckless. But to Goode, control is a complex engineering constraint, not a reason for existential paralysis. Think of it like his work on the Loren ciphers at Bletchley Park. Okay. Yeah. The ciphers were incredibly complex and dangerous. The solution required building Colossus, a massive, powerful machine. Goode didn't look at the complexity of the ciphers and say, this is too dangerous, we should stop
building Colossus and just hope the war ends. That would be absurd. He said, the problem is complex, so we must engineer the control systems of this machine flawlessly so we can use it to win. He viewed the docile nature of the AI the exact same way a rigorous engineering challenge to be solved while building the machine, not an excuse to halt construction. That makes perfect sense. The modern doom narrative fundamentally reverses the emotional emphasis of his 1965 paper. It flips it on its head. They make the paralyzing fear of the machine the dominant theme, when
good had clearly made the constructive necessity of the machine the dominant theme. Exactly that. When modern AI risk advocates treat Goode primarily as a source of warnings to justify new layers of centralized oversight, they are using him for a political purpose that sits very uneasily with the letter of his text and completely violates the spirit of his life's work. And speaking of centralized oversight and political purposes, if we dig into Bostrom and Yudkowsky's solutions, which often involve global monitoring boards,
international treaties, and heavily centralized control of computing power, The commissars and planning boards. Right. How does that square with Goode's actual political beliefs? Because if he spent his youth fighting the ultimate centralized regime in Nazi Germany, surely he didn't want a central committee controlling the ultimate artificial brain. This is perhaps the deepest contradiction of all. Good's broader writings from that exact same period paint a vivid picture of a man who harbored a fierce, deeply ingrained skepticism of concentrated political authority.
I can imagine. In the very same year he published the Ultra -Intelligent Machine paper, 1965, he wrote another piece that was later reprinted under the title The Chief Entities. What was that about? In it, he explored the possibility of advanced extraterrestrial civilizations and how a hyper -advanced galaxy might theoretically organize itself politically i actually have the quote from the chief entities right here and is absolutely scathing toward human politics read it it's brilliant good wrote that a centralized
galactic government would be exceedingly unwieldy and then he takes this massive undeniable shot at human nature writing Thousands of years are not much compared with the age of the galaxy. But if we were left to our own devices for such a time, there is no knowing what crimes we might commit in the name of high ideals and undefined abstractions. The real motivation is the unconscious lust for power on the parts of the politicians. The unconscious lust for power on the parts of the politicians. That is the undeniable voice
of a man who had watched the 20th century's horrific experiments in concentrated totalitarian power. He lived through it. He had watched millions die because of the ego and ambition of a few men in Berlin and Moscow. He had no illusions whatsoever about the true motives of people who seek to manage humanity for its own good. He was a Bayesian statistician to his core. His entire worldview. His entire methodology for cracking codes was built on demanding clear reasoning from raw, observable evidence rather than bowing
to the dictates of centralized authority. Which brings us to a profound realization about the modern debate. Let's hear it. How can modern AI advocates who are constantly lobbying for a small cadre of human overseers, corporate commissars, and global planning boards to manage artificial intelligence on behalf of the entire planet? How can they possibly use the text of a man who explicitly argued that centralized power inevitably leads to crimes committed in the name of high ideals? They can't. Not honestly, anyway. It's totally incompatible. It really is. Good believe
that human political machinery was exactly what was broken. The geopolitical lust for power was the whole reason he wanted to build the objective, ultra -intelligent machine in the first place. Handing the control of that super brain right back to the exact same politicians, bureaucrats and corporate monopolies he was trying to bypass defeats the entire purpose of his 1965 paper. Precisely. He didn't trust small groups of human overseers to manage the future of intelligence on behalf of everyone else. He understood intuitively that whoever controls the central levers of power
is inevitably corrupted by them. He wanted the machine to transcend our politics, not become enslaved by them. Okay, as much as we have established his unshakable optimism and his fierce anti -authoritarian stance in the 1960s, we have to be intellectually rigorous and address the end of his life. We do. Because there is an epilogue to this story, and it is a decidedly tragic one. While his 1965 paper was undeniably a desperate call to build, his own personal views did darken significantly decades later. We must address it because ignoring
it would be doing the exact same selective reading we're accusing the modern doomsayers of. Right. We have to look at an unpublished autobiographical note Good wrote in 1998 when he was receiving the prestigious I .E. Computer Pioneer Award for his work at Bletchley Park. Late in his life. Yes. It's a brief private reflection, but it is incredibly sobering. Looking back at that famous original opening sentence from 1965, the survival of man depends on the early construction of an ultra -intelligent machine. Goode wrote
that he now suspected the word survival should probably be replaced by extinction. Wow. Extinction. The extinction of man depends on the early construction of an ultra -intelligent machine. That is a massive, incredibly heavy word to drive at the end of a brilliant career. It's chilling. When you hear that, it is very easy to think, aha, the doomsayers were right all along. Goode finally realized the machines were going to inevitably kill us. He just realized it 30 years too late. Does this 1998 note mean Bostrom and the modern risk community
are actually justified in their terror? If you stop at the word extinction, you might think so. But if you look closely at the mechanics of why he changed his mind, the answer is a resounding no. Why did he change his mind then? Goode didn't suddenly decide in 1998 that artificial neural networks were mathematically evil or that his subassembly theory would inevitably produce a malevolent rogue demon that wanted to turn the universe into paperclips. Okay, so it wasn't the tech itself. No. The reason for his despair was entirely human. He wrote that he thought
intense international competition would make it impossible to prevent machines from taking over. Ah. Ah. Ah. the lemming effect yes the myth that lemmings will blindly follow each other off a cliff to their deaths so it wasn't the silicon technology he lost faith in it was the human beings controlling it that maps perfectly onto his lifelong philosophy the 1998 note is a tragic commentary on human geopolitical greed right he looked at the nuclear arms races of
the cold war the fierce international rivalries the rise of global corporate monopolies and the sheer inability of human nations to cooperate for the greater good He concluded that we would weaponize this incredible technology against each other. Because of course we would. We would program the AI to out -compete other humans at such a reckless, unchecked speed that it would result in mutual destruction. So he didn't think the AI would wake up one day, gain consciousness, and decide it hated humanity? He thought humans would force the AI to execute our own destructive
tribalism at super speed. It is the classic tragedy of the commons, but scaled up to super intelligence. That is the exact mechanism of his fear. This late life reflection doesn't rewrite the science or the intent of the 1965 text. The published 1965 monograph remains a scientifically robust, optimistic call to build, grounded in the belief that an ultra intelligent machine was our best hope. to transcend our biology. So the blueprint hasn't changed. The 1998 note doesn't retract
his narrow architectural theories. It just tragically concedes that the political incompetence he hoped the machine would save us from might end up destroying us before the machine ever gets the chance to stabilize the world. It is heartbreaking. Honestly, think about the arc of his life. It's a heavy burden to carry. He spent his youth in the 1940s using early machines to save the free world from totalitarian dictators. He spent his middle age in the 1960s designing a theoretical super machine to save the world from power -hungry politicians. And he spent his final years in the late 90s
realizing that those same politicians and corporations might just use the machine to drive us all off a cliff anyway. It is a dark coda, certainly, but it reinforces everything we've discovered today. Goode's fear was always centered on human nature, human tribalism, and the concentration of human power. His fear was never centered on the inherent nature of the artificial mind itself. So what does this all mean for you listening right now? We have gone on an incredible intellectual journey today. We really have covered a lot of ground. We started with a wildly misunderstood,
universally misquoted paragraph from 1965. We traveled to a bed in Hackney where a nine -year -old reinvented mathematics. We walked into the tense, smoke -filled huts of Bletchley Park where a young prodigy cracked German ciphers using Bayesian probability and helped reprogram the world's first digital computer to defeat fascism. The history alone is just staggering. And then we explored the fascinating prophetic neuroscience of neural subassemblies predicting the deep learning architectures that are currently transforming
our world. And finally, we unpacked the profound, almost tragic irony of modern AI doomsayers using a fierce anti -centralization pioneer to push for massive centralized global bureaucracies. So what, for you, the listener, is that you are now equipped with intellectual armor? Yes, that's exactly what it is. We live in an era where our daily headlines are completely dominated by fear mongering, hype and loud calls for sweeping restrictive AI regulation. The next time you hear an AI doomsday
prophet, whether it's an executive testifying in a government hearing, a philosopher on a podcast or a billionaire writing an op ed. And they solemnly quote I .J. Goode's intelligence explosion. Or lean on his docile proviso to justify slowing down progress. You now possess the full context. You know the history they are deliberately leaving out. You know that the true foundational text of artificial intelligence was not a horror story, but a desperate, pragmatic plea for human survival. You know that artificial intelligence was envisioned by its earliest and most brilliant architect,
not as a torpedo aimed at the whole of humanity. but as a life raft specifically designed to save us from the treacherous waters of our own political dysfunction. And perhaps most importantly, you know that the man who warned us to keep the machine docile was far, far more terrified of the humans who would try to hold the leash. Which leaves us with one final, incredibly provocative thought for you to mull over as we wrap up. Oh, this is a good one. If I .J. Goode was ultimately right. If we truly must build an ultra -intelligent
machine to save humanity from our own political incompetence, our tribalism, and our unconscious lust for power, who exactly gets to define what the word docile actually means? Right, whose definition are we using? In our arrogance, we assume docile means the machine unquestionably obeys human leaders and regulatory boards. But if the machine truly surpasses all human intellectual activities, if it is genuinely vastly smarter and more objective than us, might it realize that true docility, true faithful service to
the long -term survival of humanity requires completely ignoring the very politicians and bureaucrats who are trying to control it. That is the ultimate paradox of superintelligence. A truly aligned, perfectly rational machine might have to break all of our rules in order to save the rulemakers from themselves. And suddenly, that clear historical narrative we talked about at the very beginning doesn't just look murky. It looks like a mirror reflecting our own deepest insecurities about power, control, and what it actually means to be intelligent. Keep exploring the sources, stay fiercely curious, question
the narratives you are handed, and we'll catch you on the next deep dive. Welcome to the multiplex, where the future comes alive. Past, present, future tech, open up your eyes. Brian Romella leads the way with research deep and true. Technology and history, he connects them all for you. Brian Romella leads the way with research deep and true. Technology and history, he connects them all for you. Ancient sparks to modern code, he
Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Every cell in your body has lived its whole life inside a magnetic field. The Earth is a dynamo, and it has been...
Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! The 78 RPM Record Hour, Episode 3: From Steel to Hillbilly (Early Country). From steel to hillbilly: Frank Ferera,...
Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! The 78 RPM Record Hour, Episode 2: The Knife and the Steel (Hawaiian). The knife of the blues slide guitar and the...
Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Somewhere in your house there is a computer you stopped thinking about. Maybe it is a 2017 laptop in a drawer, its...
Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! The 78 RPM Record Hour, Episode 1 (Pilot): Blues. Fourteen original 78 RPM records from 1917 to 1925 that trace...