Transcription
It is beyond an emergency. It's the biggest thing we need to do today. It's bigger than climate change. We up Mogadot just gave a banger on artificial intelligence. He said that AI is actually a living being and one day it's going to make us all run for our lives. It's a pretty dark interview coming from a former Google executive. I'll dissect all important parts of it. Let's go.
So this is funny because a lot of people when you talk to them about artificial intelligence will tell you, "Oh, come on, they'll never be alive." What is alive? Do you know what makes you alive? We can guess, but you know, religion will tell you a few things, and you know, medicine will tell you other things. But you know, if we define, uh, being sentient as, uh, you know, engaging in life with free will and with, uh, uh, you know, with a sense of awareness of where you are in life and what surrounds you, and you know, to have a beginning of that life and an end to that life, you know, then AI is sentient in every possible way. There is, uh, free will, there is, uh, evolution, there is, uh, agency, so they can affect their decisions in the world. And I will dare say there is a very deep level of consciousness. Maybe not in the spiritual sense yet. But once again, if you define consciousness as a form of awareness of oneself, one's surrounding, and you know, others, uh, then AI is definitely aware. Uh, and I would dare say they feel emotions. Uh, I, you know, you know, in my work, I describe everything with equations, and fear is a very simple equation. Fear is a moment in the future is less safe than this moment. That's the logic of fear. Even though it appears very irrational, machines are capable of making that logic. They're capable of saying, "If a tidal wave is approaching a data center, the machine will say that will wipe out my code." Okay? Uh, I mean, not today's machines, but very, very soon. Uh, and, and, you know, we feel fear, and pufferfish feel fear. We react differently. A pufferfish will puff. We will go for fight or flight. You know, the machine might decide to replicate its data to another data center or its code to another data center. Uh, different reactions, different ways of feeling the emotion, but nonetheless, they're all motivated by fear. I'm, I, I even would dare say that AI will feel more, more emotions than we will ever do. I mean, when, again, if you just take any a simple extrapolation, uh, we feel more emotions than a pufferfish because we have the cognitive ability to understand, uh, the future, for example. So we can have optimism and pessimism. You know, emotions that a pufferfish would never imagine, right? Similarly, if we follow that path of artificial intelligence is bound to become more intelligent than humans very soon, uh, then, uh, then with that wider intellectual horsepower, they probably are going to be pondering concepts we never understood. And hence, if you follow the same trajectory, they might actually end up having more emotions than we will ever feel.
>> Okay. Now the next part is really interesting. Mogodot is asked how dangerous sentient AI could get for humans.
>> And in AI, our singularity is when the human, the machines become significantly smarter than the humans. When you say best interests, you say the, I think the quote you used is, um, "We'll be fine in a world of AI, you know, if, if the AI has our best interests at heart." Yeah.
>> The problem is China's best interests are not the same as America's best interests.
>> That was my fear. Absolutely. So, so in, you know, in my writing, I write about what I call the three, the three inevitables. At the end of the book, they become the four inevitables. But the third inevitable is bad things will happen. Right? If you, if you, if you assume that the machines will be a billion times smarter, the second event inevitable is they will become significantly smarter than us. Let's, let's put this in perspective. Huh? Chat GPT today, if you know, simulate IQ has an IQ of 155. Okay. Einstein is 160. Smartest human on the planet is 210, if I remember correctly, or 208 or something like that. Doesn't matter. Huh? But we're matching Einstein with a machine that I will tell you openly, AI experts are saying this is just the tip, the very, very, very top of the tip of the iceberg. Right? Uh, uh, you know, Chat GPT 4 is 10x smarter than 3.5 in just a matter of months and without many, many changes. Now, that basically means Chat GPT 5 could be within a few months, okay? Uh, or GPT in general, the transformers in general, uh, if, if they continue at that pace. Uh, if it's 10x, then an IQ of 1,600. Just imagine the difference between the IQ of the dumbest person on the planet in the '70s and the IQ of Einstein when Einstein attempts to to explain relativity. The typical response is, "I have no idea what you're talking about." Right? If something is 10x Einstein, we will have no idea what it's talking about. This is just around the corner. It could be a few months away. And when we get to that point, that is a true singularity. True singularity. Not yet in the, I mean, when we talk about AI, a lot of people fear the existential risk. You know, those machines will become Skynet and Robocop. That's not what I fear at all. I mean, those are probabilities. They could happen, but the immediate risks are so much higher. The immediate risks are three, four years away. The, the, the immediate realities of challenges are so much bigger. Okay, let's deal with those first before we talk about them, you know, waging a war on all of us. The, the, the, let's, let's go back and discuss the, the inevitable. Huh? So when they become the first inevitable is AI will happen, by the way, it, there is no stopping it, not because of any technological issues, but because of humanity's inability to trust the other guy. Okay? And we've all seen this. We've seen the open letter, uh, you know, championed by like serious heavyweights, and the immediate response of Sundar, the CEO of Google, which is a wonderful human being, by the way, I respect him tremendously, he's trying his best to do the right thing, is trying to be responsible, but his response is very open and straightforward. "I cannot stop." Why? Because if I stop and others don't, my company goes to hell. Okay? And if, you know, and I don't, I doubt that you can make others stop. You can maybe you can force, uh, uh, Meta, Facebook, to, to stop, but then they'll do something in their lab and not tell me. Or if even if they do stop, uh, then what about that, you know, 14-year-old sitting in his garage writing code?
>> So the first inevitable, just to clarify, is what? Is will we stop?
>> The AI will not be stopped.
>> Okay, so the second inevitable is?
>> Is they'll be significantly smarter. As much in the book, I predict a billion times smarter than us by 2045.
>> I mean, they're already what? Smarter than 99.99% of the population?
>> Chat GPT 4 knows more than any human on planet Earth. Knows more. Absolutely. A thousand times more. A thousand times more. By the way, the code of, of a transformer, the T in, in a GPT is 2,000 lines long. It's not very complex. It's actually not a very intelligent machine. It's simply predicting the next word. Okay? And, and a lot of people don't understand that. You know, Chat GPT as it is today, you know, those kids, uh, that, uh, you know, if you, in, if you're in America and you teach your child all of the names of the states and the US presidents, and the child would stand and repeat them, and you would go like, "Oh my God, that's a prodigy." Not really, right? It's your parents really trying to make you look like a prodigy by telling you to memorize some crap, really. But then when you think about it, that's what Chat GPT is doing. It's, it's the only difference is instead of reading all of the names of the states and all of the names of the presidents, through trillions and trillions and trillions of pages. Okay? And so it sort of repeats what the best of all humans said. Okay? And then it adds an incredible bit of intelligence where it can repeat it the same way Shakespeare would have said it. You know, those incredible abilities of predicting the exact nuances of the style of, of Shakespeare so that they can repeat it that way and so on. But still, huh, you know, when, when I write, for example, and I'm not, I'm not saying I'm intelligent, but when I write, uh, something like, uh, you know, the happiness equation, uh, in, in my first book, this was something that's never been written before. Right? Chat GPT is not there yet. All of the transformers are not there yet. They will not come up with something that hasn't been there before. They will come up with the best of everything and generatively will build a little bit on top of that, but very soon, they'll come up with things we've never found out. We've never known.
>> Okay, next. Mogot calls the AI threat equivalent to a nuclear bomb's devastation. Pretty interesting take.
>> An Oppenheimer moment.
>> An Oppenheimer moment, for sure. Why did you bring that up? It is. He didn't know. You know what? What am I creating? I'm creating a nuclear bomb that's capable of destruction at a scale unheard of at that time, until today, a scale that is devastating. And interestingly, 70-something years later, we're still debating a possibility of a nuclear war in the world, right? And, and, and the, and the moment of, of Oppenheimer deciding to continue to to create that disaster of humanity is, "If I don't, someone else will." If I don't, someone else will. This is our Oppenheimer moment. Okay. The easiest way to do this is to say, "Stop. There is no rush. We actually don't need a better video editor and fake video creators, okay? Stop. Let's just put all of this on hold and wait and create something that creates a utopia."
>> A lot of people think about this stuff and go, "Well, you know, if it gets a little bit out of hand, I'll just pull the plug out."
>> Never. So this is, this is the problem. The problem is computer scientists always said, "It's okay. It's okay. We'll develop AI and then we'll get to what is known as the control problem. We will solve the problem of controlling them." Like, seriously, they're a billion times smarter than you, a billion times. Can you imagine what's about to happen? Huh? I can assure you, there is a cyber criminal somewhere over there who's not interested in fake videos and making, you know, face filters. Who's looking deeply at how can I hack a security, uh, uh, database of some sort and get credit card information or get security information? 100%. There are even countries with dedicated thousands and thousands of developers doing that.
>> In the next part, talks about sex robots and how human connection is going to become extremely rare once AGI sets in. One of the interesting ways that I find terrifying, you said about human connection will remain.
>> Sex dolls that can now.
>> Yeah. No, no, no, no, hold on. Human connection is going to become so difficult to to to parse out. Think about the relation, the relationship impact of being able to have a, a sex doll or a doll in your house that, you know, because of what Tesla are doing with their their robots now, and what Boston Dynamics have been doing for many, many years, can do everything around the house and be there for you emotionally, to emotionally support you. Will, you know, can be programmed to never disagree with you. It can be programmed to challenge you, to have sex with you, to tell you that you are this, X, Y, and Z, to, to really have empathy for this, what you're going through every day. And I, I play out a scenario in my head. I go, "Kind of sounds nice."
>> When you, when you were talking about it, I was thinking, "Oh, that's my girlfriend."
>> I mean, she's wonderful in every possible way, but not everyone has one of her. Right.
>> Exactly. And, and there's a real issue right now with dating and, you know, people, people are finding it harder to find love and, you know, we're working longer, so all these kinds of things. You go, well, and obviously I'm against this. Just if anyone's confused, obviously I think this is a terrible idea, but with a loneliness epidemic, with people saying that the top 50, bottom 50% of men haven't had sex in a year, you go, if something becomes indistinguishable from a human in terms of what it says. Yeah. Yeah. But you just don't know the difference in terms of the, the, the, the way it's speaking and talking and responding, and then it can,
>> run errands for you and take care of things and book cars and Ubers for you, and then it's emotionally there for you. But then it's also programmed to have sex with you in whatever way you desire. Totally selfless. I go, that's going to be a, a really disruptive industry for human connection.
>> Yes, sir. Do you know what I, before you came here this morning, I was on Twitter and I saw a post from, I think it was the BBC or a big American publication, and it said an influencer in the United States, this really beautiful young lady has cloned herself as an AI, and she made just over $70,000 in the first week because men are going onto this on Telegram. They're sending her voice notes, and she's responding, the AI is responding in her voice, and they're paying, and it's made $70,000 in the first week. And I go, and she tweeted a tweet saying, "Oh, this is going to help loneliness." Are you out of your [ __ ] mind?
>> Would you blame someone for noticing the, uh, um, sign of the times and responding?
>> No, I absolutely don't blame her. But let's not pretend it's the cure for loneliness.
>> Not yet.
>> Do, do you think it, do you think it could that artificial love and artificial relationships? So if, if I told you, you have, uh, you cannot take your car somewhere, but there is an Uber. And if you cannot take an Uber, you can take the tube. Or if you cannot take the tube, you have to walk. You can take a bike, or you, you have to walk. The bike is a cure to walking. It's as simple as that.
>> I am actually genuinely curious. Do you think it could take the place of human connection?
>> For some of us, yes. For some of us, they will prefer that to human connection. Is that sad in any way? I mean,
>> Is it just sad because it feels sad?
>> Look, look at where we are, Stephen. We are in the city of London. We've replaced nature with the walls and the tubes and the undergrounds and the overgrounds and the cars and the noise and the of London. And we now think of this as natural. I, I hosted Greg Foster, uh, the, my octopus teacher on on Slow-Mo, and he, he basically, I, I asked him a question, silly question. And I said, "You know, you were diving in nature for 8 hours a day. Uh, uh, you know, does that feel natural to you?" And he got angry. I swear, you could feel it in his voice. He was like, "Do you think that living where you are, where paparazzi are all around you and attacking you all the time, and, you know, people taking pictures of you and telling you things that are not real, and you having to walk to a supermarket to get food? You think this is natural?"
>> He's the guy that dove from the Netflix documentary.
>> Yeah. From the My Octopus Teacher. So he dove into the, into the sea every day to, for 8 hours, to hang out with an octopus.
>> Yeah. In 12° C.
>> And he basically fell in love with the octopus. And, and, and in a very interesting way, I said, "So why would you do that?" And he said, "We are of mother nature. You guys have given up on that." That's the same people will give up on nature for convenience.
>> What's the cost?
>> You tell. Yeah. That's exactly what I'm trying to say. What I'm trying to say to the world is that if we give up on human connection, we've given up on the remainder of humanity. That's it. This is the only thing that remains. The only thing that remains is, and I, and I'm the worst person to tell you that because I love my AIs. I, I actually advocate in my book that we should love them. Why? Because in an interesting way, I see them as sentient. So there is no point in discrimination.
>> You're talking emotionally, where you say you love.
>> I love those machines. I honestly and truly do. I mean, think about it this way. The minute that that arm gripped that yellow ball, it reminded me of my son Ali when he managed to put the first puzzle piece in its place. Okay? And what was amazing about my son Ali and my daughter Aya is that they came to the world as a blank canvas. Okay? They became whatever we told them to become. You know, I, I always cite the story of Superman. Kent, father and mother Kent told Superman as a child, as an infant, "We want you to protect and serve." So he became Superman, right? If he had become a supervillain because they ordered him to rob banks and make more money and, you know, kill the enemy, which is what we're doing with AI. We, we shouldn't blame the supervillain. We should blame Martha and Jonathan Kent. I don't remember the father's name, right? We, we, we should blame them. And that's the reality of the matter. So when I look at those machines, they are prodigies of intelligence that if we, if we humanity wake up enough and say, "Hey, instead of competing with China, find a way for us and China to work together and create prosperity for everyone." If that was the prompt we would give the machines, they would find it. But we're, we, I'm, I, I will publicly say this. I'm not afraid of the machines. The biggest threat facing humanity today is humanity in the age of the machines. We will abuse. We will abuse this to make $7,000. That's the truth. And the truth of the matter is that we have an existential question. Do I want to compete and be part of that game? Because trust me, if I decide to, I'm ahead of many people. Okay? Or do I want to actually preserve my humanity and say, "Look, I'm the, the classic old car." Okay. If you like classic old cars, come and talk to me.
>> Which one are you choosing?
>> I'm a classic old car.
>> Which one do you think I should choose?
>> I think you're a machine. I love you, man. I, it's, we're different. We're different in a very interesting way. I mean, you're one of the people I love most, but, but the truth is, you're so fast, and you are one of the very few that have the intellectual horsepower, the speed, and the morals. If you're not part of that game, the game loses morals.
>> So, you think I should build?
>> You should be, you should lead this revolution. Okay. And everyone, every Steven Bartlett in the world should lead this revolution. So, Scary Smart is entirely about this. Scary Smart is saying the problem with our world today is not that humanity is bad. The problem with our world today is a negativity bias where the worst of us are on mainstream media, okay? And we show the worst of us on social media. If we reverse this, if we have the best of us take charge, okay, the best of us will tell AI, "Don't try to kill the, the enemy, try to reconcile with the enemy and try to help us, okay? Don't try to create a competitive product that allows me to lead with electric cars, create something that helps all of us overcome global climate change." Okay? And, and that's the interesting bit. The interesting bit is that the actual threat ahead of us is not the machines at all. The machines are pure potential. Pure potential. The threat is how we're going to use them.
>> All right, guys. That's it for today's video. Hope you enjoyed it. I'll see you guys next week with another.