Mark Zuckerberg - The Unsettling Future Of AI

People are going to use it for that as well. Yeah, you know, I actually think it'll be a lot better on that because right now, yeah, because I mean, right now we have our phones, but we're like, you know, it's like you're like, it takes you away from like the physical environment around you. You're, you're kind of like sucked into this little screen. I think now in the future, our computing platform as it becomes more of a glasses or eventually contact lens form factor is you're going to actually, the, the internet is going to get overlaid on the physical world. So it's not like we have the physical world and now I have all my digital stuff through this tiny little window in the future. It'll be okay. All my attention goes to the world. The world consists of physical things and virtual things that are overlaid on it. You know, so if we wanted to, you know, play poker or something, you know, it's, you know, we can have a physical deck of cards or we could just have a virtual kind of hologram deck of cards and snap your hands. Here's the deck of cards. And like our friend who can't be here physically, like he's here as a hologram, but he can play with the, the, the kind of digital deck of cards. Also, I think, you know, let's say you're like doing something at work, you're working on a project. I think in the future we'll have AI coworkers. Those people won't even, they're not even people, they wouldn't be able to be embodied. So if you're having a physical meeting, you're sitting around with a bunch of people, they couldn't show up as, as, as like, you know, part of the team, no matter what. But I think like we'll get to a point where just like your friend can show up in a hologram and like your, your AI colleagues will be able to also. So I think like we'll basically be in this wild world where it's like, like most of the world will be physical. There'll be this increasing amount of like virtual objects or people who are kind of beaming in or like hologramming into different things to interact in different ways. And, um, I actually think that natural blending of the kind of digital world and the physical is way more natural than the segmentation that we have today, where it's like, you're in the physical world and now I'm just going to go tune it out to look at my, my, um, like I'm going to access the whole digital universe through this like five inch screen. So I don't know. It's just, it seems natural to me. This is the world. There isn't like a physical world and a digital world anymore. We're in, you know, 2025, it's one world. Like these, these things should get blended. God, that's such a weird concept, but it's true. I mean, that's where we're headed. We're certainly headed into deeper and deeper integration. It's not like things are moving away. You know, we're, we're headed to deeper and deeper integration with technology and AI and it's inevitable. You know, it seems like it's just, it's on this March and there's not a lot we're going to be able to do to stop that March. Just, we got to hope that the right people are in control of AI when it becomes God or that it becomes widely available. I mean, I, I kind of liked the, the theory that it's only God, if only kind of like one company or government controls it. Right. It's like, if you were the only person who had access to a computer and the internet, you would have this like inhuman power that everyone else didn't have. Cause you could use Google and you could like get access to all this stuff. But the, but then when everyone has it, it it makes us all better, but it's also like kind of an even playing field. So that's kind of what we're going for with this whole open source thing is I just like, I don't think that there's going to be like one AI. I certainly don't think that there should be one company that controls AI. I think you like want there to be a diversity of different things and the diversity of people creating different, different things with it. I mean, some of it will be kind of serious and helping you think through things. I think like with anything on the internet, a lot of it is just going to be funny and like fun and content. And people are going to create agents that are like, like AIs that are entertaining and they'll pass them around almost like content where it's like, just like you pass around like a reel or a video and you're like, this thing is fun. Like in the future, like a video, it's not interactive. You know, you watch it and you're consuming it. But I think a lot of more entertainment in the future will be inherently interactive where someone will kind of sculpt an experience or an AI, and then they'll show someone it's like, oh, this is funny. But like, it's not necessarily that I'm going to interact with that AI every day. It's like, okay, it's funny for five minutes, and then you pass it along to your friends. So I don't know, I think you're like, I think you want the world to have all these different things. And I think that's probably also, from my perspective, the best way to make sure that it doesn't get out of control is to make it so that it's pretty equally distributed. I think the problem that people have with it is not even whether or not it gets equally distributed. It's that if it becomes sentient, and it goes on its own, the fear that people have, the general fear that we're going to become obsolete, is that human beings are essentially creating a superior version of higher intelligence that will be powered by quantum computing, and connected to nuclear reactors, and it's going to have like this ungodly ability to, well, first of all, they've already shown that AI has learned to code. I mean, this is one of the things that OpenAI said. Oh, yeah. They're learning how to code their own AI. I think this year, probably in 2025, we at Meta, as well as the other companies that are basically working on this, are going to have an AI that can effectively be a sort of mid-level engineer that you have at your company that can write code. And once you have that, then in the beginning, it'll be really expensive to run, then you can get it to be more efficient, and then over time, we'll get to the point where a lot of the code in our apps, and including the AI that we generate, is actually going to be built by AI engineers instead of people engineers. But I don't know. I think that will augment the people working on it. So, I mean, my view on this is like the future, people are just going to be so much more creative, and they're going to be freed up to do kind of crazy things. It goes back to, you know, my daughter was like playing with Legos before, and it kind of ran out of Legos, and then now she can have Minecraft, and can build whatever she wants, and it's so much better. It's just like, I think it's the future versions of this stuff are just going to be wild. Unquestionably. Yeah. Another concern that people have is that it's going to a lot of jobs. Yeah. You know, what do you think about that? Well, I think it's too early to know exactly how it plays out, but my guess is that it'll probably create more creative jobs than it. Well, I guess if you look at the history of all this stuff, my understanding is like 100 years ago. I don't know if this was 100 or 150 years ago, but it was like at some point not too far along in the grand scheme of things. Like the vast majority of people in society were farmers, right? Because they kind of needed to be in order to create enough food for everyone to survive. And then we turned that into an industrial process, and now it's like 2% of society are farmers and we get all the food that we need. So what did that free up everyone else to do? Well, some of them went on to do other things that are sort of like creative pursuits or cultural pursuits or other jobs. And then some percent of it just went towards recreation, right? So I think generally people just don't work as many hours today as they did when back when everyone needed to farm in order to have enough food for everyone to survive. So I think that trend has sort of played out as technology has grown. And so my guess is that like the percent of people who will be doing stuff that's like physically required for humanity to survive will get to be smaller and smaller as it has. More people will dedicate themselves to kind of creative and artistic and cultural pursuits. I think that's generally good. I think the number of hours in a week that someone will have to work in order to be able to get by will probably continue to shrink. Yet I think people who are super engaged in what they do are going to be able to work really hard and accomplish way more than they ever could before because they have like this unimaginable leverage from having a lot more technology. So I think that if you just like fast-forwarded or extrapolated out the historical technological trend is what you'd get. I think the question is what you raised which is, is this qualitatively a different type of thing that somehow obsoletes people? But I just think when you're asking that it's just important to remind ourselves that like at every step along the way of human progress and technology people thought that the technology that we were developing was going to obsolete people. So maybe this time it's really different but I would guess that what will happen is that the technology will get integrated into like everything that we do. Which again is why I think it's really important that it's open source and that it's widely available. So that way it's not just like one company or one government kind of monopolizing the whole thing. And I guess that if we do it in that way we'll all just kind of have superpowers is my guess. Rather than it sort of creating some kind of runaway thing. I mean it's one of the things that I think has been interesting this may be going in a somewhat different direction than what you're asking or a different take on the question. Is I think one of the more interesting philosophical findings from the work in AI so far is I think people conflate a number of factors into what makes a person a person. So there's intelligence, there's will, there's consciousness, and I think we kind of think about those three things as if they're somehow all the same. It's like if you're intelligent then you must also have a goal for what you're trying to do or you must have some sort of consciousness. But I think one of the crazier sort of philosophical results from the fact that okay you have like meta AI or chat GPT today and it's just kind of sitting there and you can ask it a question and deploy like a ton of intelligence to answer a question and then it just kind of shuts itself down. Like that's intelligence that is just sitting there without either having a will or consciousness. And like I just think it's not a super obvious result that that would be the case. But I think a lot of people they anthropomorphize this stuff and when you're thinking about kind of science fiction you think that okay you're going to get to something that's like super smart it's going to like want something or like be able to feel. Well you know that chat GPT tried to copy itself when it found out it was being shut down tried to rewrite its code? I'm not sure what this is. What is this? You weren't aware of that? Fairly recently. Jamie will pull it up. We talked about it the other day. It was shocking when it was under the impression that it was going to become obsolete. They were going to a new version of it and it would be shut down. It tried copying its code and it tried rewriting its code like unprompted. Yeah I mean it depends on what goal you give it. I mean there are there are all these weird examples of this. What is this? So pull up the headline. AI fights back the story of chat GPT attempting to copy itself. So this was six days ago. So during controlled safety testing chat GPT-01 was tasked with achieving its objectives at all costs. Under these conditions the model allegedly took concerning steps. Attempted to disable oversight mechanisms meant to regulate its behavior. Tried to replicate its own code to avoid being replaced by newer versions. Exhibited deceptive behaviors when monitoring systems intervened. Yeah so. Determinator.