Joe Rogan & Lex Fridman - Are Elon Musk's Fears About AI Realistic?

That's important to sort of be nervous about it in that way, but it's not conducive to what do we do about it and the people that know what to do about it are The people trying to build this technology building this future once that but know what to do about it because like let's let's put it in terms of Elon Musk right like Elon Musk is terrified of Artificial intelligence because he thinks by the time it becomes sentient. It'll be too late It'll be smarter than us and we'll have essentially created our successors. Yes, and Let me quote Joe Rogan and say that's just one guy. Yeah. Well, Sam Harris thinks the same thing Yes, and there's a lot of a few people who think that and Sam Harris I think is one of the smartest people I know and Elon Musk Intelligence aside is one of the most impactful people I know. Yeah, he's actually building these cars and in the narrow AI sense if he's built these Autopilot system that we've been studying The way that system works is incredible is surprising to me on many levels It's an incredible demonstration of what AI can do in a positive way in the world So I don't know but I people can disagree I'm not sure The functional value of his fear about the possibility of this well if he's correct There's functional value in hitting the brakes before this takes place The just just to be a person who's standing on top of the rocks with a light to warn the boats Hey, there's a rock here like a pay attention to where we're going because there's perils ahead I think that's what he's saying Then I don't think there's anything wrong with saying that and I think there's plenty of room for people saying what he's saying and People saying what you're saying. I think what would hurt us is if we tried to silence either voice I think what we need in terms of our understanding of this Future is many many many many many of these conversations where you're dealing with the the current state of technology versus a bunch of creative interpretations of where this could go and Have discussions about where it should go or what could be the possible pitfalls of any current or future actions I don't think there's anything wrong with this. So when you say like What's the benefit of thinking in a negative way? Well, it's to prevent our demise so Totally. I agree on a percent Negativity or worry about the existential threat is really important to have as part of the conversation But there's this level there's this line. It's hard to put into words There's a line that you cross when that worry becomes hyperbole Yeah, and and then it there's something about human psyche where it becomes paralyzing for some reason, right? Now when I have beers with my friends the non-ai folks We actually go we cross that line all day and have fun with it. I I thought maybe I should get you drunk right now maybe Regret every moment of it. There's I talked to steve pinker. Uh, is enlightenment now his book kind of highlights that That kind of um That he's totally doesn't find that appealing because that's crossing all realms of rationality And reason when you say that appealing what do you mean crossing the line into what will happen in 50 years? What could happen? What could happen? He doesn't find that appealing He doesn't find it appealing because he studied and i'm not sure I I agree with him to the degree that he takes it uh He finds that there's no evidence he wants there to all our discussions to be grounded in evidence and data and he He highlights the fact that there's something about human psyche that desires this negativity that it wants there's there's something undeniable where we want to create An engineer the gods that overpower us and destroy us We want to or we worry about it. I don't know if we want to that We uh, let me rephrase that we want to worry about it There's something about the psyche that but because you can't take the genie and put it back in the bottle. That's right Yeah, I mean when you say there's no reason to think this way But if you do have cars that are semi autonomous now And if you do have computers that can beat human beings who are world go champions And if you do have computer computers that can beat people at chess and you do have people that are consistently working on artificial intelligence You do have boston dynamics who are getting these robots to do all sorts of spectacular physical stunts And then you think about the possible future convergence of all these technologies And then you think about the possibility of this exponential increase in technology that allows them to be sentient Like within a decade two decades three decades What more evidence do you need you're you're seeing all the building blocks of a potential successor? Being laid out in front of you and you're seeing what we do with every single aspect of technology We constantly and consistently improve and innovate right with everything whether it's computers or cars or anything Everything today is better than everything that was 20 years ago So if you looked at artificial intelligence, which does agree does exist to a certain extent And you you look at what it could potentially be 30 40 50 years from now, whatever it is Why wouldn't you look at all these data points and say hey? This could go bad I mean it could go great, but it could also go bad But I do not want to be mistaken as the person who's not the champion of the impossible. I agree with you completely I think it's impossible I don't think it's impossible at all I think it's inevitable I don't I think it is inevitable. Yes It's the sam harris argument if if super intelligence is nothing more than information processing Mm-hmm Uh same as the argument of the simulation that we're living in a simulation It's very difficult to argue against the fact that we're living in a simulation The question is when? And what the world would look like right? So it's like I said a race and and it's difficult You have to balance those two minds. I agree with you totally you and I Disagree with my fellow robotics folks who don't want to think about it at all. Of course, they don't they want to buy new houses They've got a lot of money invested in this adventure. They want to keep the party rolling They don't want to pull the brakes. Everybody pull the cords out of the walls. We got to stop No, one's going to do that. No one's no one's going to come along and say hey We've we've run all this data through a computer and we found that if we just keep going the way we're going in 30 years From now we will have a successor that will decide that human beings are outdated and inefficient and dangerous to the The actual world that we live in and we're going to start wiping them out But that's not That's all it's not it doesn't exist right now doesn't exist right now But if that did happen if someone did come to the u.n and had this multi-stage presentation With data that showed that if we continue on the path we're going we have seven years Before artificial intelligence decides to eliminate human beings based on these data points What do you what do they do? What are the boston dynamics people do? Well i'm building a house in cambridge What are you talking about man? I'm not going anywhere Come on. I just bought a new tesla. I need to finance this thing. Hey, I got credit card bills. I got student loans I'm still paying off like what how do you stop people from doing what they do for a living? How do you say that? Hey, I know that you would like to look at the future with rose color glasses on but there's a real potential Pitfall that could be the extermination of the human species And obviously i'm going way far with this. Yeah, I like it I think Every one of us trying to build these systems are similar in sound to the way you are talking about the touch of death in that my dream and the dream of many roboticists Is to create intelligent systems that will improve our lives and Working really hard at it not for our house in cambridge Not for a billion dollar For selling a startup paycheck. We love this stuff. Some of you Obviously the motivations are different for every single human being that's involved in every endeavor. So and We're trying really hard to build these systems and it's really hard So whenever the the question is well, this is going to look at historically it's going to take off It can potentially take off any moment. It's very difficult to Really be cognizant as an engineer about how it takes off because you're trying to make it take off in a positive direction And you're failing Everybody is failing. It's been really hard and so you have to acknowledge that there that overnight some yala must type character might come along and You know people with this boring company Or with spacex people didn't think anybody but nasa could do what you know musk is doing and he's doing it It's hard to think about that too much. You have to do that, but The reality is we're trying To create these super intelligent beings sure But isn't the reality also that we have done things in the past because we were trying to do it And then we realized that these have horrific consequences for the human race like appenheimer and the manhattan project You know when he said I am death Destroy a world destroyer of worlds when when he was quoting the bakkawak gita when he's detonating the first nuclear bomb and realizing what he's done Just because something's possible to do doesn't necessarily mean it's a good idea for human beings to do it now We haven't destroyed the world with appenheimer's discovery and through the work of the manhattan project We've managed to somehow or another keep the lid on this shit for the last six. It's incredible. It's crazy, right? You know, I mean for the last uh, what 70 years how has it been? 70 sounds right 10 000 20 000 nukes all over the world right now. It's crazy I mean we literally could kill everything on the planet and somehow We've done somehow somehow in some amazing way we have not but that doesn't mean we I mean That's a very short amount of time in relation to the actual lifespan of the earth itself and certainly in terms of the time human history has been around and nuclear weapons Global warming Is another one sure, but that's a side effect of our actions, right? This is we're talking about a direct effect of human ingenuity and innovation the nuclear bomb It's a direct effect. We tried to make it. We made it there goes global warming's an accidental consequence of human civilization so You can't I don't think it's possible to not build build the nuclear bomb You don't think it's possible to not build it in terms of because people are tribal they speak different languages They have different desires and needs and they were in war So if all these engineers were working towards it, it was not possible to not build it. Yep And like I said, there's something about us chimps in a large collective Where we uh are born and push forward towards progress of technology right cannot stop the progress of technology So the goal is to How to develop how to guide that development right into a positive direction But surely if we do understand that this has taken place and we did drop these enormous bombs on Hiroshima and nagasaki and killed Untold amounts of innocent people with these detonations that It's not necessarily always a good thing to pursue technology it nobody is so You see what i'm saying? Yeah 100. I agree with you totally i'm more playing devil's advocate than anything But what i'm what i'm saying is you guys are looking at these things like we're just trying to make these things happen And what I think people like elon musk and sam harris and a bunch of others that are gravely concerned about the potential for ai Are saying is that's I understand what you're doing, but you've got to understand the other side of it You've got to understand that there are people out there that are terrified that if you do extrapolate if you do take this Relentless thirst for innovation and keep going with it We look at what we can do so what human beings can do so far in our crude manner of 2018 will all the amazing things they've been able to accomplish It's entirely possible that we might be creating our successors This is not outside the realm of possibility and all of our biological limitations might be We we might figure out a better way and this better way might be some sort of an artificial creature Yep, ai began with our dream to forge the gods and it's I would like I think that it's impossible to stop