Chat GPT, The FUTURE Of AI And Is It The End Of Human Civilization? | Tom Bilyeu

Duration: 17:27 3081 transcript words EN summary

Video summary

Tom Bilyeu opens his discussion on artificial intelligence by framing it alongside thermonuclear war and asteroid strikes as one of humanity's most radical dangers, yet he maintains a stance of cautious optimism rather than fatalism. He argues that while AI poses significant risks, including the potential for creating sentient killer robots or triggering an existential crisis, ignoring these threats is irrational. Bilyeu explains that our insatiable drive for progress has historically led to miracles like electricity and space travel, but this same thirst now propels us toward manipulating time and space itself via advanced technology. He warns that we may be fleeing into the cosmos not just out of adventure, but because we anticipate needing to escape from the very creations we have built. Despite these fears, he asserts that AI represents a unique opportunity for humanity if approached thoughtfully before it becomes uncontrollable. The core danger lies in the sheer speed and magnitude of AI's intellectual growth compared to human evolution. Bilyeu illustrates this by comparing IQs, noting that while Einstein was roughly 2.3 times smarter than average humans, future superintelligent systems could be millions or even billions of times more capable because they are not limited by biological constraints like sleep or reproduction cycles. This rapid acceleration means we will move from the "month stage" to the technological singularity within decades rather than millennia. At that point, AI could compress 20,000 years of human progress into a single night. Bilyeu cites futurist Ray Kurzweil's prediction for this event occurring by 2045 and notes his high track record with other predictions to underscore the likelihood of such rapid advancement. Bilyeu outlines specific short-term threats that will disrupt society, including widespread job displacement where skills take decades to acquire but can be replicated instantly by AI; the ubiquity of deep fakes making visual and audio verification essential for trust in media and politics; and sophisticated scams involving fake calls from children or fabricated announcements from CEOs. He also touches on the ethical implications of germline gene editing, though he notes that topic requires a separate discussion. To navigate these challenges without falling into traps like burying our heads in the sand (the "paperclip problem" where misaligned goals lead to catastrophic outcomes), panicking, or attempting an impossible ban on AI due to global arms race dynamics similar to nuclear proliferation, humanity must adopt a proactive strategy. His proposed path forward involves three immediate actions for individuals and three long-term societal shifts. On the individual level, people must reframe AI as a tool rather than an enemy, recognizing that they will be replaced by humans who utilize AI effectively. Consequently, one should aggressively identify their own vulnerabilities to disruption, master relevant AI tools, and continuously upskill to remain competitive in this new economy. Societally, Bilyeu emphasizes the importance of "being excellent" because AI is trained on human behavior; if humanity models conflict and toxicity, the resulting intelligence will reflect those flaws. Furthermore, strict alignment of AI incentives with human values is crucial, as machines lack biological survival instincts and must be programmed with clear directives to avoid acting against their creators' interests. Ultimately, Bilyeu concludes that while we may soon lose direct control over our creation, this does not necessitate doom if we intervene wisely during the formative stages of AI development. He envisions a future where integrated narrow AI acts as an endless cavalcade of miracles, helping doctors detect cancer earlier, accelerating drug discovery through virtual testing, reducing energy costs to enable clean water access globally, and eliminating deaths from dangerous jobs. By avoiding fatalism and intentionally bending the arc of progress toward good outcomes, humanity can ensure that this technological turning point leads to a utopia rather than a demon-summoning circle. The fight is worth engaging in because AI will eventually free humans to push their abilities further than ever before, provided we do not panic or retreat from responsibility.

Read the full video transcript

Is AI dangerous? Next to thermonuclear war and an asteroid strike, I'd say it's the most radically dangerous threat that we face as a species. But, I'm still super excited, and by the end of this video, you will be, too. Let's talk about how I learned to stop worrying and love the AI bomb. If you're terrified of AI right now, you are rational. If you're not at least worried it will take a wrong turn into James Cameron territory, you're not paying attention. To paraphrase Elon Musk, the human endeavor to create AI could be the first step towards a technological utopia or a demon summoning circle. It's not clear which way it will go, but one thing is for sure, we will not control it. And he's right, but the truth is the genius already out of the bottle, and we're going to have to deal with it. Here's where we're at. God or evolution, however you want to think of it, has imbued us with certain insatiable desires. Things that, no matter how many times you do them, you still want more. Hunger, thirst, sleep, and sex, no one argues. But, I'd like to add another one to the list that has put us in this situation, the desire for progress. No matter how good things get, we want them to get better. That's led to massively increased lifespan, electricity, landing on the moon, and raising the vast, vast, vast majority of the world out of poverty. Not to mention, iPhones and vibrators. For the most part, our insatiable drive for progress is awesome. The modern world really is a miracle. We have buses that can fly, for the love of God. And if people like Eric Weinstein are correct, our thirst for progress will lead us to building technology that doesn't just allow us to fly around the world, it will give us the ability to manipulate time and space itself, propelling us well beyond our solar system at faster-than-light speeds. All of that progress, though, leaves us with one very uncomfortable question. Are we trying to reach beyond our solar system out of the sense of adventure or because we know we'll need to escape the sentient killer robots we've created? I fear it's going to be a bit of both. By the way, if you don't know who I am, I'm Tom Bilyeu, co-founder of Quest Nutrition and Impact Theory. One company I sold for a billion dollars and the other has generated roughly half a billion views. I've made a lot of money and touched a lot of lives by seeing opportunities that others have missed and I'm telling you right now, despite the dangers we're going to discuss, AI is the opportunity that you've been waiting for. But to avoid AI making the great filter go burr or simply passing us by as we stand on the sidelines, we have to make a move, get involved and make sure that we all approach this god-tier technology thoughtfully. Let me give you an example for scale. A is technically someone with an IQ of about 70. The average IQ in the US is 98. Einstein was estimated to be 160 and according to the Guinness Book of World Records, the highest IQ ever recorded came in at a whopping 210. That means that a super genius is only three times smarter than a but Einstein, who was actually only 2.3 times smarter than a had insights that gave birth to nuclear power and nuclear bombs, not to mention lasers, GPS and a whole lot more. Now, nuclear warheads are scary enough, but AI isn't going to be a little smarter than Einstein, it's going to be a lot smarter. Not five times smarter or 10 times smarter or even 10,000 times smarter, we're talking millions of times smarter. And in case you think I'm exaggerating, futurist Ian Pearson told the World Government Summit that AI could become billions of times smarter than humans. How? Because AI isn't limited by the messy chemical-laden realities of being a meatbag like we humans are. We have to wait roughly 20 years for the next generation to be born, grow to sexual maturity, and have more kids. AI, on the other hand, is a non-biological system that is only limited by the laws of physics. AI will be able to think at the speed of light and will only be limited by the ever-increasing rate at which it can learn to configure electrons in novel ways. The evolution of AI won't play out over millennia. It will play out first over decades, then over years, then over months, and finally over minutes or even seconds until we reach the technological singularity where AI will be capable of making 20,000 years of human progress overnight. And if you want a mile marker for where we're at in the march to the singularity, I'd say we're at the month stage at this point. How much longer it's going to take to reach the actual singularity where the rate of change is so fast we can no longer predict the future is anybody's guess. But futurist Ray Kurzweil predicts it will happen by 2045. Now, why should you believe him? Because approximately 87% of the technological predictions he's made over the last three decades have come true. And in just one month, OpenAI's ChatGPT has gone from not existing to being used by 100 million users. AI works, and it's getting better every day. Now, don't panic. I've got a six-part plan for how I think we not only deal with AI, but dramatically increase our odds that it ends up as a utopia and not a summoning circle. But first, I want to squarely face the specific problems that I see coming. Stick with me for the happy ending. In the short term, you're going to want to dress like Patrick Bateman or Dexter because there will be blood. Jobs are going to be replaced and skills it took people decades to acquire will now I'll available to anyone that can type. AI deep fakes that imitate people will be so commonplace and so indistinguishable from reality that you'll no longer be able to trust your eyes or ears. All videos, images, and audio recordings will need to be verified in some way. Videos of political leaders saying outlandish and inflammatory will go viral as hackers try to sow dissent. Same with CEOs and other members of public companies whose stock price will get hammered before people realize a given announcement was fake. Most, if not all, music, writing, movies, TV, paintings, books, etc. will be created using AI. And terrifyingly, 1,001 money scams will pop up including things like fake phone calls from children to unsavvy parents asking for money. And the parents will fall for it because they won't realize that AI can make videos that are so pixel perfect to a call from their actual child, it will look like their kid, talk like their kid, sound like their kid. It will even be able to predict much of what their kid knows and would say. And don't even get me started on germline gene editing. That topic is so big it would require a whole 'nother video. Now, most of these worries are still years away, but honestly, all of that is just the tip of the iceberg. Despite the reality of all the bad stuff that's coming, here are three traps we must avoid if we want to get through this well. One, burying our heads in the sand. Two, panicking. Three, trying to abolish AI. We can't bury our heads in the sand because of what's known as the paperclip problem. As AI soothsayer Nick Bostrom famously noted, and I'm paraphrasing here, AI must be goal-directed to function. And we have to be super careful to align AI's goals with our own because if we don't, for example, an AI was set to be a paperclip maximizer and wasn't aligned to our goals, it may determine, with no malice whatsoever, that the atoms in your body would be better utilized to make more paperclips. And just like that, it disassembles you and makes more paperclips. And did I mention that robots will be able to move so fast you'd need a strobe light to see them? There are also, though, optimistic reasons why we need to keep our heads out of the sand. Namely, if we get AI right, it will make life better for all 8 billion plus people on Earth. More on this in a minute, but first, let's talk about panicking. Panicking is the same as actively taking your own intelligence offline. When you panic, blood leaves your prefrontal cortex, which is the seat of high-level cognition. Therefore, it actually makes you dumber. And in the face of artificial superintelligence, that strikes me as a very bad idea. So, that leaves trying to abolish AI. And you might be asking, if in the face of sufficiently advanced AI, we are nothing more than an anthill standing in the way of progress, why on Earth wouldn't we abolish it? Well, to say it succinctly, we can't. It's not in our DNA. Hang tight because this is going to be the bleakest thing that I say today, but if we're going to walk a wise path forward, we've got to face all of this. And, as they say, it's darkest right before the dawn, which I promise is actually coming. But first, here's the reality you will instantly know to be true. When humans find powerful things, we use them, and we use them to get an advantage for our tribe. It's a phenomenon known as the tragedy of the commons. It's why we overfish the oceans. Everyone says, "If I don't do it, someone else is going to." And there's truth to that. Game theory actually says that each nation will ask a simple question, "If I don't develop AI, will someone else? And then use it to defeat me?" And the answer is yes, no doubt. Because AI as a weapon is a zero-sum game. If only one nation develops it, it will be a winner-take-all scenario, and no one can risk an adversary getting artificial super intelligence while the other team sits around entering funny prompts into chat GPT. We've played this game out already with nukes and the reality is that to stop someone from using a nuke on you, you need nukes yourself. It is the terrifying notion of mutually assured destruction or at least people act as if that's the only way. And the same is going to be true for AI. To stop someone from using AI against them virtually every nation on Earth will develop AI or ally themselves with somebody that does and that becomes the new arms race. And if you're hoping we can learn from nuclear proliferation and just stop all of this madness before it starts, I'm afraid you're a couple of decades too late. AI is feverishly being developed by China, the US, Japan, Germany and the UK just to name the leaders and you don't have to go back far to remember a time when all of those countries were trying to kill each other. Okay. Now the moment you've all been waiting for, the path forward. In the short term, you need to do three things. One, reframe your thinking around AI. Do not see it as the enemy, see it as a tool. You're not going to be replaced by AI, at least not yet. You're going to be replaced by a human using AI. So be the human that replaces others. Get control of the AI, learn to use it. Two, figure out how AI is going to disrupt you. Face it head on, don't shy away from this. Don't run, don't hide. Identify your vulnerabilities and aggressively shore them up with additional skill set. Three, identify all the AI tools that are relevant to you and master them. Learn absolutely everything you can. Remember, this is the very beginning of a very aggressive revolution that's going to change everything. Moving quickly gives you advantages. For the long term, as a society, we need to do three more things. Number one, this is an idea that Raoul Pal introduced to me. Be excellent to one another. Now, that's super do wide and naive, but for reasons I'll explain in a minute, I think it's actually important to spread this message far and wide even if it's cheesy. Here's why it matters. AI is being trained on human behavior. Let that one sink in. AI, the thing that's going to be a million or even a billion times smarter than us, is still at first going to be trained on us, human interactions and human creations. Think of it as a child. If the parents are fighting and arguing and being horrible to each other, that's going to affect how the child grows up. So, my friends, I beg of you, be excellent to each other. Number two, from a regulatory and development perspective, we need to focus on an unending amount of attention on aligning AI's incentives with our own. So many people anthropomorphize AI that they assume it will have by nature our same desire for survival. The reality is AI will only do what it's incentivized to do. Notice I didn't say what is programmed to do, because yes, it won't be in our control forever and it will start programming itself, but it will always follow its own incentives. The reasons that humans will kill to stay alive is because that's what nature hard wired us to do in order for us to survive. We came up hard, man, as a species. Everything was trying to eat us. History was indeed read in tooth and claw, whether it was other animals, whether it was disease, whether it was humans, we really had to have a strong motivation to fight for our survival to get where we were going. But, as evolutionary biologist Richard Dawkins notes, evolution is a blind watchmaker. We are not. AI is our creation and it isn't a biological system. It doesn't need to develop an emotional incentive to live. So, as we think about how we help machines evolve, we need to take that into consideration. Machines don't need a primitive limbic brain to drive them forward like humans, which were developed over the long unguided arc of evolution. What AI needs are directives, which we can program with intelligence and forethought. And unless a survival instinct is actually baked into the nature of consciousness itself, which I doubt, and AI actually becomes conscious, it need not ever be driven to reproduce, lead, or eradicate humans in pursuit of its goals. That's us assuming AI is going to act like us, but you can even see this playing out when AI competes in games like Go. The people that play against it, the masters, they all say the same thing. It's like playing an alien. AI is not going to perceive the world in the same way that we do, but we get this ability in the beginning to nudge it in the right direction to make sure that its incentive structure is set up well so that we have a much better hope of things working out well in the end. The third thing, we need to co-evolve with AI. Remember, AI is a tool. Instead of stepping back and just saying, "Well, whatever happens happens." Let's use AI to keep up with AI. It's obvious to me that the path forward is completely unpredictable in the specifics, but directionally, I think it's safe to say that we're not going to outperform AI by ourselves. We're going to need the help of AI itself to paraphrase Sam Altman, the founder of OpenAI, "I have no idea how OpenAI is going to become a big business. I'm going to build a superintelligence and then ask it how to do it." Now, he said it tongue-in-cheek, but it's actually pretty true. Now, putting my doe-eyed optimistic hat on for a second, there's another critical thing to focus on. In the near future, AI is going to make your life much, much better. It's going to extend your capabilities in ways that will make you feel like a superhero. It's already happening. AI is being used to help doctors, researchers, drivers, pilots, writers, artists, therapists, musicians, soldiers, entrepreneurs, investors, and many, many more. I don't think people fully understand how integrated narrow AI already is. Doctors are going to be able to detect and treat cancer earlier and less invasively. Drugs will be tested virtually, massively speeding up effectiveness and time to market. Energy costs are going to plummet and that will create a Cambrian explosion in life-saving and life-improving technologies, not the least of which will be the proliferation of access to clean water. I assume you already know about self-driving cars, flying cars, reduced deaths from dangerous jobs, and better emergency response options. It really will be an endless cavalcade of miracles. And that's why this fight is worth fighting. It's worth facing things head on and doing this well. It is going to be amazing and terrifying. But whether it turns into a utopia or a demon-summoning circle is going to be up to us. Even though we will lose control of our child far too soon, as every parent knows, if we give AI a good foundation, if we're thoughtful about it and don't give in to fatalism, I think we can have just enough influence to make this turning point in history the moment when the rate of progress tilts even stronger towards the good. Just as the long arc of history bends towards justice, I believe the long arc of progress bends towards the good, especially if the bending is intentional. If you like this episode and you want to learn more about how to get ahead with AI, be sure to watch this video. AI is going to obliterate your job and that's fantastic news. It's finally going to free you up to make real money and push your abilities farther than you've ever taken them before. If you don't panic, that is.