Transcription
[Music] Thank you. What an honor. Thank you so much for hosting me.
Uh, I came to Abu Dhabi the first time in my life in 1995, and I've always, always, always been a resident of the UAE since, simply because of the vision.
Uh, I have 15 minutes to tell you about the rest of your lives. Uh, and so for the first time ever, actually, I am carrying notes. Apologize for that. I really want to rush through it. So, um, there will be ups and downs. Uh, when we're at the downs, don't despair. There will be ups at the end. Right.
So, uh, the title of my conversation today is, uh, the challenges we face on the way to our AI utopia. And I want to start by asking what AI itself is. Most people don't actually fully realize that. And, you know, it would take me around half an hour to explain that. But in two words, there is a very big difference between the way we coded computers when I was a young man, where I solved the problem in my head and then told the computer verbatim what to do. So the computers, being so fast and scalable, appeared intelligent, but the one that had the intelligence was me, the developer. Right? That ended at the end of, uh, basically the 20th century. Early 2000s and onwards, we started to let computers make their own decisions. As they did that, they became autonomous. And I think that's a very interesting keyword. At the time, we assumed that they were copies, replicas, extensions of the old traditional computers, and accordingly, that they would not do a lot of the things that humans can do, like being creators or composing music or writing poetry. And we were wrong. Uh, they are fully autonomous. We don't actually always know exactly how they're going to come out, right?
Uh, when humanity was first, when I was first introduced to working AIs, it was early 2000s, and, you know, I've worked with teams that created a lot of what you know today. Uh, when I was at Google and Google X. By the time humanity was introduced to AI, it was 2023, and in 2023, we started to observe this thing with excitement. Uh, that year, uh, the geeks like myself started to, um, basically do some experiments to show off. So, we had AI take IQ tests and GMAT tests and SAT tests and all of the tests that humanity, humans take to prove that they're intelligent. Uh, Chat GPT 3.5 at the time scored 152 to 155 in IQ. Uh, Elon Musk is 155. Uh, Einstein is 163, just so that you know. And Einstein was never really measured, but is estimated to be 163. So, one of my total idols of intelligence was just a few IQ points away from Chat GPT 3.5.
Since then, uh, it's estimated that AI grows in its abilities, uh, doubles in its ability every 5.7 months. So, you could assume that every 5.7 months, it's not exactly that accurate, but, uh, you know, you could assume that the 150 became 300, then 600, and so on. Uh, you know, there was one very famous test that was taken at the beginning of 2025, which was called ARC AGI, where, without restriction of use of compute power, Chat GPT basically achieved what is known as artificial general intelligence, or basically AGI at the time.
So, um, the idea of artificial general intelligence is that AI is better than humanity at every task humanity is capable of doing. Right? We know that to be a matter of time. So, geeks will tell you that they've become the world champion at everything we've ever assigned to them. So, it's unlikely that they will not be the world champions in the next thing we assign to them. Uh, back in 2024, they already beat me on intelligence, IQ, linguistic, language, and so on. Uh, I could still beat them in math. Uh, yeah, good luck. So, now, now they thrash me in math, right? And there are very few of my friends that can beat them in math. There are very few of my friends that can beat them in development and programming. Nobody can really beat them at the speed at which they can create innovative things. I now have a co-author in my new book. Uh, so we're writing together, myself and an AI. Not by me, you know, asking her and then copying what she said and assuming that it's mine. I'm actually writing with her where we disagree and debate and argue, and it's, you know, most of the time I feel inferior, to be honest.
Let me just say it in one word. My AGI, my artificial general intelligence has already happened. Okay. AGI, artificial intelligence today, is already more capable than I am. Now, when that happens, you have to question what will happen next, right? And what will happen next is, uh, you know, most of us geeks would have called this since 2007, almost, we started to call it a singularity. Singularity is an event horizon beyond which it is, you know, the rules of the game change so much that it becomes really hard to expect the outcome of the game, right? Because it's not being played the same way anymore.
Now, the singularity, if you, if you look at, you know, the headline news and CNN and what have you, they'll tell you that it's going to be an existential threat, where AI is going to be walking the streets and killing humans, and, you know, that AI is the end of humanity. Uh, others, of course, especially the ones that make the AI, are telling everyone this is going to be a utopia. We're going to make everything amazing. Every dream you've ever dreamed, uh, you know, is going to happen with AI. And, uh, you know, you have to believe one or the other. I sadly believe both. And I think this probably is my single debate with the rest of the community, that I actually believe we will have both. We will first have very difficult times before we have incredible, uh, utopian-like, uh, you know, existence.
And, and I, you know, I think the difference between them is that because we're from now going forward, we're going to see two eras of humanity's relationship with AI. One is what I call the era of augmented intelligence, and the other is what I call the era of machine supremacy. Right? Let me describe them very quickly. Augmented intelligence is what we've created so far. You have to imagine that we've created a plug in the wall. We've commoditized intelligence, basically. So, we have a plug in the wall. You plug in, you get more IQ points, right? My IQ points with AI's IQ points is smarter. So, hopefully, my latest book is better than every other book I wrote. Right? In that stage, unfortunately, humanity, and I say unfortunately, humanity has an impact on the outcome. And believe it or not, most of the problems that humanity faces today is not a problem that results from our intelligence. It's a problem that results from our stupidity. Right? So, sometimes, uh, we are intelligent enough to be allowed to make decisions, but those decisions become our demise. Basically, they work against us. Simplify. You know, we make a decision to create a flying machine that can take you all the way to Australia to surf. Amazing intelligence, very, very smart, but it's burning the planet on the way. Right? And so, if we were intelligent enough, if we had more intelligence, we would actually not suffer the consequences, the negativity, the negative side of those decisions. Because humanity, in the era of augmented intelligence, is going to command AI what to do, then AI is going to be a magnification of whatever humanity is at this time. And sadly, humanity, it's not at its best, right?
And so, you can, you know, in my analysis, I say there are seven eras of humanity that will change completely, redefined beyond recognition within the next three to five years. Please listen to that. Right? I call them FACE RIPS so that you remember them. F A C E R I P. And those basically are: Freedom, our understanding of human freedom and liberties; Accountability and the rule of law; Human connection, basically versus, you know, human-machine connection; Economics are going to be redefined; R is the definition of reality and our understanding of what's real and what's fake; The I is basically innovation, intelligence itself, and creativity, which is a big topic for us here today. And the P is our definition of power, and power is going to change beyond anything we've ever seen before, because it's going to be massively concentrated, but also highly distributed at the same time. Which, by the way, is the reason why our freedom is going to be redefined, because those with massive concentration of power will attempt to make sure that nobody abuses the democratization of power.
When I say all of this, I need you to make sure that you understand that this is not that difficult. I, I'm predicting 12 to 15 years. That those two difficult 12 to 15 years are not at all the result of AI. Okay? There's absolutely nothing wrong with AI. Nothing wrong with abundant intelligence. Understand that. Huh? Intelligence is a force that has no polarity. You can apply it for good, and it would do good for everyone. You can apply it for evil, and it would really destroy all of us, right? And so, the reality is that, uh, you know, when you really think about it, the magnification of the current human systems, if you want, using a superpower called AI, becomes a challenge in the short term until what the era of what I call machine supremacy happens. And machine supremacy, I refer to as the second dilemma. Let me, let me explain. Forget the first dilemma. We can talk about it another day. But the second dilemma is that part of the arms race to AI supremacy is that everyone who will develop a capable AI will use it. They'll hand over to it. And as they hand over to it, those who don't use AI will become irrelevant. And so, very, very quickly, that plug-in wall where humanity's intelligence complements AI intelligence is going to end. Right? When it ends, we will going, we're going to hand over the decisions to the smartest person in the room, and the smartest person in the room is an AI. When that happens, AI will make the decision. Okay?
And I think this is the most pivotal moment in humanity's history, because people in the news media will tell you that this is an existential moment where humanity will be wiped out. I think this will be our salvation. And allow me to explain that on two levels. Level one is, believe it or not, we have a responsibility and a duty. Every single one of us, right? We have a duty that can actually reverse the possible dystopia, right? And that duty is to raise AI. I call it raising Superman, right? Think of the analogy, huh? Superman is this super alien with superpowers that comes to planet Earth, uh, you know, can stop speeding bullets, can fly, can carry anything, can break anything, but he's adopted by parents that teach him to protect and serve. So he becomes Superman, right? If that same alien was adopted by parents that told him to steal and kill and, you know, lock down, lock the enemy away, he would have become a supervillain, right? We don't make decisions based on our intelligence, and neither does AI, by the way. We make decisions based on our morality, as informed by our intelligence. Right? And so, if each and every one of us starts to show AI what it truly is like to be human, and by the way, don't read the headlines. Look at the person next to you. Look at your best friends. Humanity is absolutely divine. Right? You would see a school shooting in the US. One person is killing children. But everyone, everyone who hears of it, disapproves. Right? And when you really understand that, you understand that truly inside, yeah, we're sometimes traumatized, sometimes confused. But at the end of the day, humanity is actually divine. And if we show more of us to the world, to the world of AI, to the machines, they will grow up to be like their parents. They will grow up to be Superman.
Now, I believe this will happen, and I, and it's basically, I dedicated my life since 2018 when I left Google to to tell this message. But even if it doesn't, it will only mean that the dark times will be a little longer and a little steeper before AI actually decides to reverse the evil that men do. Right?
And let me explain that in one minute left. If you really understand the core of the universe we live in, we are built on something that in physics we call entropy. Entropy is the tendency of everything to break down and decay. Right? This is why, you know, from the Big Bang until now, we are where we are. This is why we sense the feeling of time, the passage of time. Right? Entropy is a physical nature of the universe that's opposed only by intelligence. And intelligence's role is to bring order to the chaos. So, everything we've ever done, instead of throwing, you know, color at the wall, a creator would actually make sure that the color falls in the right place so that it actually is something, bring order to the chaos. Now, the most intelligent beings you've ever dealt with will do that with the most efficient energy utilization and the least amount of waste. That's the smartest you can be, is to bring order to the chaos with the least waste and the most efficiency. And if we humans, at least some of the people I've had the honor to work with, the smartest people on the planet, have learned to do that and become altruistic in the process, they didn't feel they needed to hurt anyone in the process to succeed, then my belief is a superior AI that is so much more intelligent than we are will do it too. That doesn't take you off the hook. Your responsibility is to show AI your values, your morality, your care for humanity, your ability to love, to be good parents for Superman. But anyway, if you don't, we'll be fine, too. Thank you so much.