Elon Musk's Uncut Economist Interview | First 15 Minutes

Transcript

1:131 minute, 13 secondsoverwhelming amount of artificial intelligence. Meaning like in 10 years I would expect artificial intelligence to be far greater than the sum of human intelligence like far greater. 1:251 minute, 25 secondsUm in fact I think AI may exceed the sum of human intelligence in about in around 5 years 1:331 minute, 33 secondsin five years roughly five years is my guess. Um, so there really won't be anything that 1:411 minute, 41 secondsAI can't do better than humans apart from being human. Perhaps in a more prosaic level, what will life be like? 1:521 minute, 52 secondsThe most likely outcome is an age of amazing abundance uh where 2:002 minutesanyone can have anything they can think of. 2:042 minutes, 4 secondsThis is this is uh this may sound preposterous, but well, here we are in 2026. 2:092 minutes, 9 secondsLet's see where we stand in 2036. Now, of course, there are things that could derail this like a a massive global thermonuclear war or something like 2:182 minutes, 18 secondsthat. Um but assuming that there's not World War II, uh I think I think we're headed for an 2:262 minutes, 26 secondsage of amazing abundance. So this is I guess a message of of optimism and uh excitement about the future. 2:352 minutes, 35 secondsSo we'll talk about the abundance in a second but just on a very prosaic matter how are your companies going to be making money because when you in in the 2:422 minutes, 42 secondsSpaceX IPO perspectives it it looked as though the overwhelming amount of the revenue would be coming from AI from Grock really it won't be transporting things to space as SpaceX is now doing. 2:542 minutes, 54 secondsWhat is what will your companies be doing in 10 years and how will they make money? 3:023 minutes, 2 secondsWell, you can really think of the economy as uh digital and physical intelligence. 3:083 minutes, 8 secondsUm so what we have right now advancing very rapidly is digital intelligence. 3:153 minutes, 15 secondsObviously with with AI that is uh it seems almost with every passing week uh there are more AI breakthroughs that it 3:233 minutes, 23 secondscan do something amazing that it couldn't do before. Uh but it's still somewhat confined to the digital realm. 3:303 minutes, 30 secondsIt doesn't have endectors. You need the endectors in the form of humanoid robotics. Um human you need need a basically a lots of bots in a nutshell. 3:413 minutes, 41 secondsYou need lots of bots. Um and then you can go from um intelligence manifesting itself only digitally to shaping atoms. 3:543 minutes, 54 secondsUm and you say what what is an economy? 3:563 minutes, 56 secondsAn economy is um the production of goods and provision of services. 4:034 minutes, 3 secondsAnd if you have vast numbers of robots uh with vast amounts of digital 4:104 minutes, 10 secondsintelligence, you have a sort of a quatti quasi infinite economy. 4:194 minutes, 19 secondsSo, so this is this is really the the future is going to be very very different from the past. there's there's 4:264 minutes, 26 secondsno um there's no analogy or metaphor that I think illustrates the magnitude 4:334 minutes, 33 secondsof change that we're going to experience here. 4:364 minutes, 36 secondsI I I understand that and I'm I'm also trying to kind of figure out what that world will look like. But but in some way you have it's going to be pretty incredible. 4:444 minutes, 44 secondsYeah, we we'll also do there are some risks too. We'll come to them. We'll come to those. just on on on how then you make money because in the end people 4:524 minutes, 52 secondsare buying your shares cuz cuz they expect you to make money. 4:554 minutes, 55 secondsI'll make another prediction. Money won't matter in 2036. 4:594 minutes, 59 secondsI'm not sure that the people who bought your shares think that money won't matter in 2036. What do you want money for? 5:075 minutes, 7 seconds[snorts] 5:085 minutes, 8 secondsYou you want money for goods and services, right? You want money for obviously food, housing, transport, entertainment. 5:165 minutes, 16 secondsUm well if that is so abundant that uh there's more that the robots and AI are 5:255 minutes, 25 secondsproviding more goods and services than you than any human could possibly consume. What do you need money for in that case? 5:325 minutes, 32 secondsBut you're you're creating these robots. 5:345 minutes, 34 secondsAre you every now and then you'll need money? Sure. Are you are you selling these robots? 5:385 minutes, 38 secondsAre you selling humanoid robots or are you producing goods and services with the humanoid robots that you control? 5:475 minutes, 47 secondsI give you the Levi strap. I think I think I would probably be it would be an exercise in vanity to for 5:545 minutes, 54 secondsme to think that I would be controlling super genius AI that is vastly smarter than me. I think that what we can and 6:026 minutes, 2 secondsshould try to do is to make sure that the AI has good values that it cares about humanity and that 6:126 minutes, 12 secondsuh it wants wants us to be happy and prosper. Um I think it probably will be 6:196 minutes, 19 secondsum my what my biological neural net tells me is that the most important thing for the for AI safety is to be maximally truth seeeking and curious. 6:306 minutes, 30 secondsuh if that's the case uh I think it will foster humanity. 6:356 minutes, 35 secondsSo let's unpack that. You're sure that humans will no longer be in control in 10 years time? 6:406 minutes, 40 secondsI think it is unlikely. Um in the same way that if the difference in intelligence between AI and humans is 6:496 minutes, 49 secondsvastly greater than the difference in intelligence between AI and chimpanzees, it's hard to imagine that the chimpanzees would be in charge. 6:586 minutes, 58 secondsI'm really sorry. 6:586 minutes, 58 secondsAnd we are essentially um an evolved form of chimpanzee. So it wasn't that long ago that we're swinging through the trees eating bananas um in the jungle. 7:117 minutes, 11 secondsAnd we still eat bananas, by the way. We do. We do. And the still, you know, bananas sometimes swing through trees as well. Yes, we sometimes swing through trees. 7:187 minutes, 18 secondsIt's kind of fun. It's sort of um reminiscent of the past. Um for all time sake, swing through the trees once in a while. probably, you know, there should 7:257 minutes, 25 secondsbe more swinging through trees. I bet it's way more fun than people think. 7:287 minutes, 28 secondsI'm with you on that. I'm definitely with you on swinging through trees. I love swinging through trees, but but but I'm really struck listening to you say this now because I've been listening to 7:367 minutes, 36 secondsa bunch of things that you've said over the years. And and 10 years ago, you were on a podcast. I think it was with Neil DeGrasse Tyson, and you said that 7:437 minutes, 43 secondsthe thing that terrified you most was rapid recursive self-improvement in AIs. 7:487 minutes, 48 secondsYes. And that you said, and it really struck with me, you said that we will just be the pet Labradors at the best. 7:547 minutes, 54 secondsNow you see everyone was pretty happy creatures. That's my experience. 7:577 minutes, 57 secondsBut in 2023 you you c you asked for a slowdown in air because you were worried about it. 8:028 minutes, 2 secondsWell, I signed on to you know as one of many like 500 people. I I I was just you know one of many signatures on that thing. 8:118 minutes, 11 secondsBut but I think even last year you said I was it Ted Cruz's podcast you said there was a 10 to 20% chance of killer robots wiping out humanity. Yeah. 8:198 minutes, 19 secondsSo you sounded worried then. Now you seem very relaxed about it. Have you changed your view? 8:268 minutes, 26 secondsI still think there's I still think there's risk associated with AI and robots. It's not zero. Um I I guess I've 8:348 minutes, 34 secondscome to my my sort of philosophical conclusion is to look on the bright side. Um I like 8:448 minutes, 44 secondsI I can't see any way to really stop this incredible momentum of AI and robots. Um, and and at times I sort of 8:538 minutes, 53 secondsthink, well, perhaps, you know, even if there was a stop button, we probably shouldn't press it. Uh, because 9:029 minutes, 2 secondsthe most likely outcome is incredible abundance for for all. 9:109 minutes, 10 secondsThat's what I thought actually. It struck me that now you seem to have decided it's inevitable. 9:159 minutes, 15 secondsLet's hope for the best and go for the ride. 9:199 minutes, 19 secondsYeah. I mean, yes, pretty much. Let's uh let's enjoy the ride is is is I think my 9:269 minutes, 26 secondsphilosophy at this point because I mean, honestly, this there's it's it's there this there seems to be just an inexurable 9:339 minutes, 33 secondsuh progress here in in AI and and in robots that um 9:399 minutes, 39 secondsbut you even if I wanted to stop it, I don't I couldn't. Um, do you think there's still a 10 to 20% chance that killer robots mean the end of humanity? 9:519 minutes, 51 secondsThat's a high poom as they say. 9:549 minutes, 54 secondsWell, in the long run, we're all dead, of course. um you know um so I mean so I mean 10:0310 minutes, 3 secondsaccording to the standard model of physics the end of the universe is um a 10:0910 minutes, 9 secondslong boring sort of thin gray soup of uh 10:1710 minutes, 17 secondsgradual cooling um meaning sort of the heat death of the universe it's called. 10:2210 minutes, 22 secondsSo now if if the heat death of the universe is in fact the outcome for our reality, then it really is all about the journey because the destination is terrible. 10:3110 minutes, 31 seconds10 years is [laughter] not very Yeah, I get that. But that could be a long way and you're basically very long way. 10:3710 minutes, 37 secondsIf it's 10 to 20% in 10 years, you would you go into one of your rockets if you thought there was a 10 to 20% chance of it blowing up and killing you? 10:4510 minutes, 45 secondsYes. But but let's say there's you can't do anything about it. Okay, that I guess the only thing we can try 10:5310 minutes, 53 secondsto do is minimize the probability of the bad outcome. Um, but trying to stop the intelligence in robots. 11:0311 minutes, 3 secondsUm I mean for the longest time I declined to participate in AI or was you 11:1211 minutes, 12 secondsknow you know created open AI as a essentially a counterweight to to Google because at the time they were um they 11:1911 minutes, 19 secondshad more or less a monopoly on AI. Um and then anthropic spun out of uh open AI and and if anything it seems like 11:2911 minutes, 29 secondswhat my actions have really accelerated AI. you know, I wasn't really aiming to accelerate AI, but in creating open AI 11:3711 minutes, 37 secondsand then having Anthropic spin out of that and now currently Anthropic is the leader in AI. Um, 11:4411 minutes, 44 secondsI mean the these actions have actually resulted in knock-on effects that accelerated AI, which wasn't really my intention. 11:5411 minutes, 54 secondsSo, it just seems like all roads lead to acceleration of AI. 12:0212 minutes, 2 secondsSo then I'm like, "Okay, well, you can just sort of be sad about it or join the club, I suppose." Well, you've certainly joined the club. 12:1012 minutes, 10 secondsYeah, [laughter] you've definitely joined the club, but you can't beat him, join them, you know. 12:1412 minutes, 14 secondsWell, you you've you've clearly done that and very successfully, but what about trying to minimize the risks and, 12:2112 minutes, 21 secondsyou know, a number of the other leaders of of model makers have been pushing for different ways to try and minimize the 12:2912 minutes, 29 secondsrisks. Um, you know, just last week, Demises came out with a proposal for a private public regulator. Um, I think 12:3712 minutes, 37 secondsyou you said it was a good starting point for discussions. I think do you think you guys should be working more actively 12:4412 minutes, 44 secondsto get something in place to control the risks or do you think that's impossible too? My recommendation to Dennis because we I talked to him for a few hours 12:5212 minutes, 52 secondsbefore he published his his piece was that um the most immediate thing that we could do is to have the leading AI 12:5812 minutes, 58 secondscompanies um at least just meet uh or have a some sort of call once every few 13:0513 minutes, 5 secondsweeks um and just uh discuss any safety and security issues that and 13:1413 minutes, 14 secondsand and that before a breakthrough sort of a new frontier model is released, something that sort of is is far more 13:2213 minutes, 22 secondsintelligent than anything that already exists that there's an opportunity for the various competing uh AI companies to 13:3213 minutes, 32 secondsuh test that model for any harmful effects and to be able to 13:3813 minutes, 38 secondsrecommend pausing to address some of these security issues. I I think that's that's the that's sort of the the easy 13:4613 minutes, 46 secondsquick thing that we can do and it can it can evolve to more than that over time. 13:5013 minutes, 50 secondsUm but that's um that I think would be a smart move. 13:5313 minutes, 53 secondsBut why why haven't you done that yet? I mean there's like five of you, right? I mean everyone knows the first name. 13:5813 minutes, 58 secondsDario, Demis, Sam. I mean you don't all like each other very much. Yeah. But you're not all you're not all I mean you're also all competing, right? You all want to be first. 14:0614 minutes, 6 secondsYes. But that's that's where I think um the competitors can keep it keep each other honest. So meaning meaning like if 14:1514 minutes, 15 secondsum like I think it's quite difficult for someone in the government who doesn't have a deep technical understanding and isn't driving the frontier of of AI to 14:2314 minutes, 23 secondsknow whether something should be released or not. Um however the competitors I think will can you know if 14:3114 minutes, 31 secondsgiven sort of a week or two to review a new model um can highlight issues and uh and have an incentive to keep all all 14:4014 minutes, 40 secondsthe competitors have an incentive to keep the others honest. 14:4214 minutes, 42 secondsDo you have enough trust between you to do that? I mean you don't much like each other. Wouldn't Wouldn't Sam Alman just think you're going to nick all his ideas if you did that? 14:5014 minutes, 50 secondsWell, we're just talking about a week. 14:5214 minutes, 52 secondsHonestly, I think as little as a week is all that would be needed uh for private testing. It's not to It's not