"MIT Professor Max Tegmark: LIVE in Boston"

0:28This is officially a cold open, I guess. Smart. Smart.

0:38Smart.

0:43Smart. Smart. We're so happy to be in Boston. We are so happy. Thank you for being here tonight. And you guys rolled out your nicest weather for us. Yeah. Felt like being home in Canada almost But thank you, thank you, thank you Not only for listening to our garbage But coming out and looking at our garbage There's a better way to say that There's a way, there's a better way to say that So, alright, let's sit down Yeah, we're just excited Sit down So, I, oh wait, somebody left their phone here That's me, I brought it, it's in my pocket I don't know why If I do get a call though, we're just gonna take a quick break Is that your phone?

1:26It is mine Sorry. The camera was on too, Granddad. Yeah, my flash was on. Here, put it over here. And I'm chewing gum. It's a really professional operation here. All right, so... Let me ask you something. Do you guys... How is traveling going for you with all of this? Great interview question. Sean did a little preparation. I know. No, I really want to know that because as of this morning, Will started calling me Katy Perry. because I bring so many outfits with me. Now this is the best you could use. It was just... I know. Well, the true story is the sweet girl who picks out my clothes for me because I don't know how to do that, clearly, she sent me over, God love her, she's a really close friend, but she sent me over two of the exact same outfits today.

2:20Right. So I kind of put this together willy-nilly. It is a new thing for us because we do this thing on our lap, laptops and we wear pajamas. So this whole notion of having to dress and actually have a specific time and all that stuff, it is odd. And that there's people here. And usually one of us is a few minutes late. I am commonly late. Oh, that's you. I am. But here's the, I don't like to be, if you're early, you've wasted time, right? Do you have an issue? No, we've got a lot of milling around going up there. By the way, I go to a doctor's office now late so I don't have to wait the 20 minutes they make me wait.

3:08Right. I just go, right? By the way, you still have to wait. But do they like that? They love me for that. This sounds familiar. I feel like we talked about this, maybe. I don't think we've ever talked about that. I don't think we've ever talked about that. I don't think we've ever talked about that. You're never early or late, are you? No, I'm right on time. You're always right on time. What is that? Right on time. Just because I respect you. Primarily. The implication is I don't respect you. Here's the other thing I discovered the other day. Go, what were you going to say? Well, no, I was just going to say about calling you Katy Perry.

3:37It wasn't just that. I just kept going, I don't know, ask Katy Perry over here. Like that. So it wasn't just bullying. And then we all have, you're going to make me roar in our head. Yeah. And Jason, you've been singing that all day. I can't. I'm so, I can't. Is everybody, when you get a song in your head, you can't get it out? Am I the only one like, but it lasts an abnormal amount of time. Like a week, I'll get something stuck in my head. They say to count down backwards from a certain amount of number and that will make you stop thinking of the, are you doing it right now? Does it just work for songs?

4:10Yeah. Because there's a lot of shit I'd like to forget. Like that kind of. Yeah, so if that works, I want to do that. Well, the other thing that happened to me yesterday was I found my back has been itching, and then we have this wait are you gonna yeah so he's backing into a plug for hypochondriac for his other podcast so Sean we're in conversations in the hotel room and Sean does a lot of this he's like uh-huh what's going on and he gets up against the door and he's going really? like a dog scooches across the rug and scratches his yeah so then our friend Eli is with us who we love and is a very good friend of ours and I'm like what's going on with my back he goes take your shirt off let me see so I took the thing and I'm back and he goes that's shingles.

4:54Truly? Yeah, he said I have shingles. Were you not there? No. No, he went, oh my God, that's shingles. And I'm like, what? I just got the vaccine for shingles. How could I have shingles? Right. Because if there's a vaccine for it, it's in my body. Yeah. Yeah. Right? Yeah. Good health isn't political. No. Say that again. Thank you. No. Thank you. Who's got a pen? Yeah. I used a pen earlier to write my intro, and since we're talking about science, it's a great segue. Fellas. All right. I wanted to tap into the brainpower of this city. Okay? We got a big brain coming out. Uh-oh. This fella has a master's and a Ph.D.

5:45from Berkeley. He's a fellow at Princeton. He has tenure at Penn. He arrived here at MIT in 2004, where he still works today. He does it all from physics to cosmology to quantum stuff and computers. He's going to explain what it is. Stephen Hawking. Will Ferrell. Stephen Hawking. Please welcome a guy who can definitely make us all more smart, not less smart. Smartless, you get it? Max Tegmark. What? Max Tegmark. Come on, Max. Oh, my gosh. Get out of here. Here he is. Max, this is Max Tegmar. How are you, man? It's so nice to meet you. Pleasure to meet you, sir. Come on. Please, please. This is so exciting.

6:33Well, see, we have the same stylist. Wow. He wears it a little better than you do. He certainly does. Now, can you do a better job than I just did of explaining what it is that you do? First of all, how do you introduce yourself, call yourself what you do? By the way, you look like a rock star. Well, it depends on what I want. Like if I'm on a long flight and I just want to be left alone, and the person asks me what I do. You say you're a pedophile. Physics. That was my worst subject in high school. Five hours of silence. Right, right. But you are a... If I want to talk, I'd be, oh, astronomy.

7:08I'd be like, oh, I'm a Virgo. Oh, right, right. Or maybe if I say cosmology, they'll be talking about eyeliner and makeup. Okay, so the class you teach is? Oh, it's whatever they want me to torture the students with that year. So it can be either torturing the freshmen who came right out of high school with the basic physics of how stuff moves to torturing the grad students with some advanced topics about our universe. Okay. Or most of my time I spend actually torturing my grad students doing research on artificial intelligence. Okay, okay, good. By the way, this is everything I'm for. Yeah. Will you marry me?

7:51No, I'm kidding. I wanted to go. You probably have a bunch of stuff. You'll have to ask my wife over there. Well, that's... I saw this documentary on artificial intelligence, and what I was surprised to learn is that it's not about, like, robots like the Steven Spielberg movie. It's more about the amount of computing speed that we now can do such that, like I think they said in this documentary, you can put all the books that have ever been written into a computer now, and you're going to tell me whether I'm right or wrong. I bet I'm close to right, but probably wrong. You can put all the books into a computer, and the computer will ingest all that information, separate it out, and be able to give you an answer about anything that you can ask them if the information was in any of those books, from languages to rocks to, Isn't that called Google, though?

8:46Well, I'm sure you could explain that, but that's artificial intelligence. Well, yes and no. So on one hand, yeah, you can take all the books that were printed and put them on a memory card so small that you might have a hard time finding it in your pocket. But that doesn't mean that a computer necessarily understands what's there just because it can store it and kind of regurgitate it back to you, right? And I think the truth is, despite a lot of the hype, that artificial intelligence is still mostly pretty dumb today compared to humans or even cats and dogs. But, you know, that doesn't mean it's going to remain that way.

9:23I think a lot of people make the huge mistake of thinking just because AI is kind of dumb today, it's always going to be that way. Right. Well, shouldn't we keep it dumb? Because if we let it get too smart, et cetera, What is that threshold, the point of no return? Yeah, because remember that thing about Facebook where they started, I don't know if this is true, they started doing AI technology and they started talking to each other and they shut it down? Is that true? Because they were gossiping. Yeah. It's true, but I think Hollywood and media often make us worry about the wrong things. Sure.

10:01What do you mean? First of all, people often ask me if we should fear AI or be excited about it. The answer is obviously both. AI is like any technology. It's not evil or good. If I ask you, what about fire? How do you feel? Are you for it or against it? What are you going to say? Right. It can hurt if you use it incorrectly. Exactly. And the same thing with all tech. The only difference is that AI is going to be the most powerful technology ever. because look, why is it that we humans here are the most powerful species on this planet? Fucking A. Is it because we have bigger biceps, sharper teeth than the tigers?

10:42No, it's because we're smarter, right? So obviously if we make machines that are way smarter than us, which is obviously possible, and most researchers in the field think it's going to happen in our lifetime, then it's clearly either going to be the best thing ever or the worst thing ever. My question is, when it's the worst thing ever, By the time it becomes the worst thing ever, then we're fucked. Then it's too late. So let's not let it be the worst thing ever. So that's the catch, though. We humans have had to play this game over and over again where technology got more powerful. We're trying to win this race, making sure the wisdom with which you manage the tech keeps pace with the power of the tech.

11:24We always use the same strategy, learn from mistakes. But it seems like the big safeguard that we have as humans that we don't yet have with machines is that we have ethics, we have empathy, we have emotion. And what is the computer program that you would need to put together to inject that into this new machine with all of this information? Can we put some snuggles in it? That's a fantastic question. What's the snuggle recipe? So you're hitting exactly my wish list. If you want to have ever more powerful AI that's actually beneficial for humanity, right? So you can be excited rather than horrified about the future.

12:00there are three things you're going to need. First, you're going to need the AI to understand our human goals and then get it to actually adopt the goals and then to actually keep those goals as it gets smarter. And if you think about it for a little longer, each of these are, like, really hard. Suppose you have a tell-your-future self-driving car to take you to Logan Airport as fast as possible, and then you get there covered in vomit and chased by helicopters, and you're like, no, no, no, no, that's not what I meant. and the car goes like, that's exactly what you asked for. You know, it clearly lacks something.

12:31See, they do talk like that. Literally, he's the Terminator. He sounds like the Terminator and he's talking about Terminator stuff. So we humans have so much more background knowledge, right, that a machine doesn't take part because it's like a very alien species of a sort. So that's hard for starters. And then suppose you can get that right. But let me stop you there. Is there any chance of getting that right? In other words, the formula, the equation that equals emotion, responsibility, ethics, can you even create a computer equation for that? I think right now we don't know how to do it. It's probably possible.

13:11We're not working enough on it yet. The catch is, though, computers are just like, you know, if you think of a baby that's six months old, you're not going to explain the fine details of ethics to them because you can't quite get it yet. by the time they're a teenager, they're not going to want to listen to you anymore, those of you who have kids out there, right? But you have a window with human children when they're smart enough to get it and maybe still malleable enough to hopefully pay attention. That's fascinating. With computers, though, they might blow through that so quickly. That window.

13:44It makes it harder. Did you see Ex Machina? Did anybody see Ex Machina? I did. That's amazing. I did. What did you think of that? Yeah, let's give it up for Ex Machina. No, no. What is the most accurate film to science? Like, is it HAL in 2001 or Ex Machina? Those are my top two, actually, because HAL emphasizes this key thing that the thing you should really fear in an advanced AI is not that it's going to turn evil, but that it's going to just turn really competent and not have its goals aligned with yours. That's what happens in HAL. Like, no spoilers. Right. And like that taxi I mentioned.

14:15I think 60 years in, we're good, but... But then the other thing you should also worry about is even if you can solve all these things, and I think it might take 30 years to figure this out, which is where we should start now, not the night before some folks on too much Red Bull switching on. I've got two cans in me right now. Not in the can. Not in my can, but in here. Right the superintelligence away from him. But the other thing is, even if you manage to solve those technical problems, which we should have more research on, you also have to worry about human society. because just think for a moment about your least favorite leader on the planet.

14:54Don't tell me who it is so we don't offend anyone in the audience. Thank you for thinking he's a leader. But just imagine their face for a moment here, okay? Now imagine they are the one who controls the first superintelligence and take control over the whole planet and impose their will. How does that make you feel? Not good. None of it makes me feel great. Listen, after a lifetime of doing all this stuff with the robot, how does it feel to talk to an actual robot? Like, that must feel... And we will be right back. And now, back to the show. So is there, do people have proprietary right over certain stuff?

15:31Or does one country control a lot? Like, who's leading? China's leading the AI. Are they not? The U.S. and China are both very strong. I mean, most research suggests that U.S. is still kind of ahead. But there's a lot of hype around. Both countries, of course, try to... If anybody's serving USA, I'm going to lose it. And I'll bet when they say the USA, they're talking about MIT, and they're probably talking about him. Well, you know how that goes. Both countries are trying to claim that the... Both countries' researchers are trying to claim that the other one is headsets should get more funding.

16:04That's how we researchers always do it. Oh, I got it. But seriously, the interesting key here, I think, is ultimately, it's not really going to matter which country gets it first. Is it going to matter most? Is it going to be us who control the machines? Or are they who control us? I mean, but it really is. All joking aside. I'm obsessed with the Terminator movies and anything sci-fi. Sure, all joking aside. Yeah, joking aside. Let's put the jokes aside and let's talk about it. No, but that's kind of the idea behind a lot of Hollywood movies and stuff. Is that what if the AIs become more intelligent than the humans?

16:38But here's the thing. I saw something on 60 Minutes years ago, which fascinates me to this day. And I'm not going to get this right. Andy Rooney's eyebrows That's it You ever wonder why? I opened my own desk drawer And I got a tie from the 1968 Democratic convention It's got soup on it I don't like soup Nobody knows who Andy Rooney is Yeah

17:06It's also His Dax Shepard impression But anyway There was this guy the interviewer was interviewing the scientist who claimed to have come up with this idea. This thing was like wrapped around his ear and was like tied to the side of his head. And the interviewer was asking him a question like, what's the population of Utah? And all he had to do was think of the answer and it popped up on the screen. Do you know what I'm talking about? This kind of, yeah, well, this sort of stuff you can already do if you have a connection to Google. Oh, yeah. Ergo, you saw it 15 fucking years ago. What are you talking about?

17:43No, no, but that you... Oh, you mean Peter Jemey? Is it Jeopardy? No, that you can, that if you think of a response... He got it. Oh, okay. He heard you. Answered it. Let me move on to another question, something current. So, you know, we were talking earlier about... We're going to kiss it out later. You mean when you read out people's brainwaves? Well, I found it fascinating. I saw the segment where the guy was thinking an answer, and it popped up on the screen. The third time is really clear. Hey, Sean, your best robot voice, go, quick. He's going to sing Katy Perry, watch. Do you want to play a game of war?

18:26No. Do you want to play tic-tac-toe? Another 15-year-old. Jason, robot voice. Do you want to play a game? That's it.

18:39He's teeing himself up. This is the voiceover artist here. Let's have it. Do you want to play a game? Okay, no matter how I say it, it's just the gayest computer ever. I'm sorry, Dave. I can't do that. Do you want to play a game? Like, yeah. Okay, good. Sorry. And what's your best computer voice? I'm sorry, Dave. I can't do that. I just love that hell scene again. because it points out what you really should worry about. But coming back, there are two things. One, again, to summarize, we need to make sure the machines can actually align their goals with ours because if you have something much smarter than us that has other goals, we're screwed.

19:25It's like playing chess today against the computer that has a goal to win when you have... It's no fun. What's the next... Because when somebody says AI, all I picture are those mechanical dogs that walk around and they don't really do anything. They're just like, oh, look, we invented a robot dog, and it doesn't really do anything. So I want to know, like, what's the next thing that we can use it like that's... Is it mechanical cats? Could it be? Yeah, no, you know what I mean? Like, what's the... Oh, imagine the mechanical cats doing cats. Oh, my God, that would be amazing. Sorry, we'll get back.

19:58You're saying what's the next thing we can look forward to enjoying out of science? Yeah, like in the pop sense. Okay, in the pop sense. So first of all, so just to finish off what we've been talking about, But, you know, Hollywood, it makes us associate AI so much with robots and the Boston Dynamics dogs. You should check them dancing, by the way, if you haven't. Dancing dogs, robot dogs. The dancing Boston Dynamics robots. Super cool. But the biggest impact right now AI is having is actually not robots at all. It's just software, right? I mentioned this improvethenews.org project we're doing, which is just a little academic thing.

20:35But if you think about social media, that's all about AI. One of the reasons people hate each other so much more in America now is the effect of AI. Not AI that had the goal of making people evil, but just had the goal of making people watch as many ads as possible. But the AI was so smart at figuring out how to manipulate people into watching ads that it realized that the best way to do it is to make them really angry and keep showing them more and more extreme stuff until they were so angry they wouldn't fragment. And if you get really, really pissed off, you then research that thing even more, and then you get more ads and all that stuff.

21:08Yeah. So that's one. All of social media. I mean, yeah. Another one is, let's talk about some positive things, because AI, intelligence, right? It's human intelligence that's given us everything we like about civilization. So clearly, if we can amplify it with artificial intelligence, we can use it to solve all sorts of problems we're stumped on now, like cancer and lifting everybody out of poverty and so on. Will there ever be a... Go ahead. Sorry. No. So I was just going to say, like, another pure software thing that has nothing to do with bots is use AI for doing better science, better medical research, for example.

21:43I was just going to ask about that. So is there any, like, I read a long time ago about, like, you know, like you put a locator chip in your dog or your cat, whatever. Can you, I heard that they might be making, like, a chip that has all your medical files, and you put it under your skin so you can just scan it. because filling out all those fucking forms over and over, it's like, I just filled out the form, and now you're asking me the questions all over again. Read the thing I just spent 20 minutes on. Stop calling the doctor once a week. I'm dying. What? Right, I'm like, what? I was like, what's your name?

22:12I just filled three forms out that says what my name is. I think I'm personally going to pass on that chip implant and just ask the hospital to have a less Stone Age computer system. But seriously, of course. Did you hear about Sean showing up to doctor's appointments late so that he doesn't have to wait. It's riveting. I don't know if that's AI or whatever. Something huge that happened this year, for example, is biologists have spent 50 years trying to figure out just from the DNA how the shape is going to turn out to be of different proteins that our bodies make. It's called a protein folding problem.

22:44And then Google DeepMind solved it. No way. Yeah, with an AI. And now you can develop medicines faster. So this is a fantastic example, I think, of AI for good. But another one... But then the robots that are probably going to have the biggest effect, I think, on us and the job market the next five years are probably cars, actually, just autonomous vehicles. That's pretty cool. I'm worried about everybody with their cars and the automatic driving or whatever, and then they show up, and then people are just going to show up at the valet dead in their car. It's going to be really... You know what I mean?

23:20People are going to get in their car, and then they're going to be like, Like the pizza guy shows up and he's like slumped over. You open that door. Yeah, just out. You know what I mean? Like that's what I'm worried about. Now, with the combination of what you know about computers, what you know about space, what you know about intelligence, I know, I know what you want to know, honey. Guess what Sean wants to know. He wants to know if there's aliens, but we're not going to ask him that. No, not that. What I want to know from you is, based on your knowledge of all those areas, does it seem possible to you that there is the requisite amount of intelligence and technology at a place other than Earth?

24:07Ooh. See, he asks a different way than I would. Well, of course it's possible. Although, you know, my guess, based on spending a lot of years dealing with telescopes and thinking about these things is that when we, for the first time, get way more advanced technological intelligence showing up on this planet, it's probably going to be in our lifetime, and it's probably not going to come from outer space. It's going to be something we built. What do you mean? What do you mean? So you're saying AI. What do you mean that we're building something to bring them here? What are you talking about? No, I mean, we're basically...