Bullish on Crypto and AI w/ Raoul Pal (Wolf of All Streets Interview)

0:00Your favorite neighborhood spot grows with Square. Indeed, my favorite neighborhood spot has quickly become Todd Snyder in Williamsburg. Todd Snyder is one of my favorite menswear shops and has supplied me with all the clothes I have needed this quite hot summer. Every business has different goals, but Square is the business platform that supports them all. From opening a new location, selling something new, or just expanding their reach. Indeed, I've seen it with Todd Snyder. In Square, also, you can get real-time insights, so don't wait for end-of-day reports. Go to square.com forward slash go forward slash realvision to learn more about how your business can grow with Square.

0:43That's S-Q-U-A-R-E dot com slash G-O slash R-E-A-L-B-I-S-I-O-N.

0:57Hey, everyone. If you like this podcast, go behind the paywall to get privileged access to the smartest minds in finance. Join the Real Vision community and learn how to become a better investor. Visit realvision.com slash RVpod and use the promo code podcast10, that's podcast10, to get 10 % off our essential membership for the first year. Now, to the top analysis of today's crypto markets. Well, Rel, how are you, buddy? You there? Always good. Can you hear me? Yeah, you sound great. Hey, good to see you, my friend. Good to hear from you, my friend. How are you? You well? I'm doing well. I'm doing well.

1:35Glad I was kind of away for a few days. And, you know, if you leave for four days now, the AI has already taken over by the time you get home. It's ludicrous. It's ludicrous. I can't keep up. It's just impossible. Yeah, so we're going to start there in about two seconds. So listen, just to set the room, everybody, this is our weekly Tuesday spaces. We do this 11 a.m. As I've kind of discussed, I've considered doing this more often, but we're really trying to favor quality over quantity. And the guests have really, truly been incredible. And I know people have valued these conversations. next week.

2:09I have a feeling it'll be a sort of an opposite vibe conversation. I have Jeff Booth, so we'll be talking more heavily, I think, about Bitcoin next week. Everybody, please hit that little arrow button up at the top, share this space, get more people in. And also, of course, this will be available for playback here on Twitter, but also on Spotify, Apple Music, everywhere that we have my podcast. But without further ado, I want to be respectful of everyone's time. So let's get started. Raul, listen, I just sort of already hinted at it with that joke about AI. You've been on a bit of a tweet storm.

2:43I know that you just tried auto GPT. I've seen quite a few things. You and I have had a lot of conversations about exponential technological growth. This is insane. Yes. So, you know, the speed at which it's happening is surprising, even though are the epicenter of AI itself. So people like Ahmed Mosek is a good friend of mine. Even he is like, I asked him to predict where might we be in five years. He goes, I have no idea where we're going to be by the end of the year because things that he thought were going to take a year to develop happened in a month. Don't forget, ChatGPT went from zero to a hundred million users in five weeks.

3:29Literally nothing on earth has ever happened like this before. so we kind of get our heads around that and our heads around mid-journey and it's incredible development in terms of the kind of photorealistic art then we know that text to video is coming and that's going to come at scale as well and that's going to happen in the next month or so and then on top of it we launch all this kind of agent-based auto gpt which basically is creating your own intern but skilled intern to be able to do pretty much anything now it's early stage but it got announced last week and already by the weekend there were about six major projects built using it i don't even know who these people are building this stuff and how they can do it at such speed but it's crazy and that's the first iteration of that so if we think of it as its gbt3 moment where the hell are these agents in a year's time these are basically autonomous AIs that conduct tasks for you and can work together to create complex tasks.

4:35So let's say, I mean, I asked it to write me a newsletter based around the recent developments in AI, but also brand it the exponential age, create a branding for it, and then create it into a PDF. Now, I didn't know how to do the prompts. It was my first time, but I got quite a long way along that with it with only two prompts and i'm like oh my god what does this even mean so it means that the replacement of knowledge workers and the enhancement of knowledge at scale knowledge doesn't become a scarcity expertise doesn't become a scarcity it becomes an abundance thing um nobody's really ready for that but when we're talking about exponential age technologies you know we've already seen um you and i scott been all over this is crypto and how fast that was that was the fastest adoption of any technology and we probably haven't even had the chat gpt moment there yet but then coming hot on the heels is i think the apple announcement that comes of their whatever their big mega show is i think that's going to be their new headset plus this new nerf technologies that i've just been tweeting about which is basically 3d renditionings of spaces which allows you to create photorealistic metaverse experiences okay they've been working on that for a long time so that's coming on top of it we've got the rise of the robots and that's the self-driving cars and within the next couple of years we're going to see a plethora of self-driving um cars around as you know whether it's google whether it's Tesla and a bunch of others start scaling that out.

6:13So that's basically robots combined with AI and so on and so forth. I mean, it keeps going and keeps going. I mean, I even looked at this when I, when Elon bought Twitter, I said, he's no interest in, in a social media platform. That's all a front. What he's actually buying is one of the best sources of human interaction for training ai and he's basically announced that x.ai is is that it's the ability to get natural human conversation both in long form because that's the new long form tweets plus the kind of new sub stack ideas that are coming plus short form plus video plus audio um that's a very powerful data set and i'm sure he'll close off the apis to twitter so then he gets the data set now why does he want that well x.ai is just part of a bigger plan which is he also obviously owns dojo which is this massive supercomputer to create ai processing for for the self-driving cars for tesla but in addition he's building the optimus robot so you've got robots and you've got all the humans in the world interacting on twitter um and you've got the cars it's like okay this is moving very very fast yeah there's never been anything like it and now it seems like even the fast things that took years take days i read yeah gbt i read that auto gbt was only conceived two and a half weeks ago yeah idea so what what we're seeing here and people are going to have to get used to this it's going to be very very unnerving and we're all feeling it it's like i cannot keep up with this is we kind of began to get to grips with metcalfe's law so that is these kind of exponential adoptions yeah we've seen it in in crypto we've seen it in the internet we've seen it in mobile phones all great this however what's going on in ai is actually ai and the all of these technologies together is creating something called reed's law which is metcalf's law squared and we've i've gone through and checked and i don't think there's any example of this in humanity um and we're gonna have to deal with this at scale globally in a ridiculous period of time.

8:27Yeah, I guess the next question becomes what's left for us, right? And I'm not a fearmonger who believes that AI is coming for our jobs. I think that this just increases the individual human productivity by many, many multiples. But this is sort of, I joke about travel agents being replaced, obviously, by the internet and Kayak and Expedia and such, but it feels like we're just going to have thousands of jobs that are travel agents. And it's going to be next month. Six months. Yeah, or next week at this rate. I mean, yes. Yes, it's very disruptive to the labor force. As I've said to people who are arguing about sticky inflation, this is like a nuclear bomb of deflation that's just hit the well.

9:10I think it's the biggest advancement of any technology since probably the splitting of the atom in its massiveness of importance. it is it is fundamentally a change in society how we work how economies run and that is all underway i mean i've been talking about this and writing about this for the last two or three years saying this is all coming and that i didn't realize how fast this was going to happen and what it was actually going to feel like but it's like you know we're all in the media business as well as investing in a whole bunch of other things i mean scott you've been in the music business well that's completely changed you write newsletters that's completely changed you create content And that's completely changed.

9:49The investing world, well, that's going to completely change. I mean, what's left? I mean, everything. Everything is going to change. I don't have the answer for that. And I feel like I'm so far behind it. Even now, I'm having meetings in the coming weeks with people who are master prompters and understand this stuff better than me. Because I do feel like it's either you learn how to prompt these and get the most out of it, or you will literally be left behind. I'm not saying that there won't be things for us all to do, but even from the first iteration of ChatGPT, I was using it to help edit my newsletter.

10:23I immediately saw at least the value there, but like you just said, you could literally have it conceive of, name, brand, and write your newsletter. I'm not going to do that. But if you don't, somebody's going to do it. And then Scott's going to be walking around LA looking for another job. It's just the issue we're facing. I mean, it's very real because there's some 24-year-old who's figuring out how to do this who's coming off the old work.

10:52Really crazy. Mark, I see you have your hand up. By the way, for the other guests who we do have up here, we will be adding more. Just feel free to jump in whenever you want. We don't need to be polite, Mark. No, I just want to ask a question. So the question is quality. I don't see it. um you guys have thoughts on or ral do you have thoughts on on the quality yeah this tool is fast and it does things fast but at least my experience is the quality is mediocre at best hey everyone we're gonna take a quick pause and hear a word from our partners we'll be right back

11:33yeah yeah i'd argue differently mark um hello by the way haven't spoken to you in a while thank you well um i'd argue opposite i mean we've used it a lot i've spent a lot of time playing around with it we use it a lot at real vision of a whole different things and it's the quality of the trumps that gives the quality of the output uh the output is literally extraordinary i've used it whether it's for personal issues um i've used it for business planning business strategy. I've used it for guide and assistance. It's more powerful than I think we yet understand. I think there was a, you know, when you see the interviews with, let's say, Sam Altman and Lex Fridman, or you see the one that the CEO of Microsoft did on 60 Minutes, all of these guys are like, we don't even know what we invented here.

12:26And we're seeing it learn itself in ways we didn't understand, which is what we saw with Go. So this is not a fixed state thing because it's an exponential in how it learns as well that whatever we think of it today is not where it is tomorrow. So I think it's very dangerous to underestimate the power of this and the quality statement today because the quality statement today was different than it was yesterday and it was two months ago. So I would just keep a very open mind to what you think quality is versus the quality of the inputs and how to leverage this more powerfully. I mean, don't forget this AutoGPT was built on top of ChatGPT, and that's leveraging the same toolkit, but also the internet as well.

13:15So that's become even more powerful than we possibly imagined. Yeah, Raul, I was going to ask you specifically about that. Obviously, one of the criticisms of ChatGPT is that it's using a data set that's 2021 and before. Does AutoGPT now leverage real-time data? I haven't even had a chance to look into that. And also just really quickly to Mark, I can tell you that the quality argument is so two weeks ago. I know that sounds nuts, but it improves literally with your ability to prompt, like you said, and just with the evolution of ChatGPT4, it's really crazy. Go ahead. But how much of your newsletter has been generated by the tool?

13:54None of it is generated by the tool, but I'm a simpleton. I generally use it just to make sure that I don't have grammatical errors to edit and make myself sound a bit smarter, but I'm still not comfortable yet releasing that as well. Yes, I agree with you on that regard. Raoul, how much of your content? You can talk about how great the tool is, but I don't have much evidence of quality output that I've received i've received some crappy stuff but so we use ai for video editing um so uh descript allows us to put in a creator also traits transcript instantaneously transcripts not 100 perfect with that transcript we can then just go through the transcript look at chunks of the video we want to take out create a shortcut it creates that edits the video and does it or we can take bits of speech out or you can put its own speech in we've also used it for creating content we've been experimenting internally with creating content in my voice and other people's voices um original content and it can do that pretty well um we're already seeing it for newsletters so we have newsletters that um the editing first pass or the structure of a newsletter is written by AI and we've only been using it for a month.

15:20So the marketing team have been using it. The content team have been using it. The product and engineering team are building on it and using it. So it's kind of coming at us from every angle, but everybody's new to it. We're all trying to learn how to get the most out of it. I'm seeing other people writing entire newsletters written by AI. And you're seeing a lot of posts on Twitter now that are completely AI written, these kind of long-form posts. And they're pretty decent. And again, these are people who are not experts in doing this either. So, yeah. All right. Well, you went from great downgraded to pretty decent.

15:56So if I keep asking you questions, maybe we'll get to what I think, which is not very good. I don't know. I've been testing it, Mark. And I will say some of it is really, really excellent. It's just that I haven't figured out how to work it into my flow yet. Right. And like, I think that I'm assuming that Raul is probably similar. The thing I struggle with is that I still want to read it and check it first, right? And have the original idea. But there are plenty of people who have just put it out into the wild to do its own thing. I think the craziest part is how fast it's iterating on itself.

16:29And so, like, what we say today might be completely different in a week. I mean, AutoGPT, literally, it's been like three weeks. And I have a friend who just booked an entire vacation using AutoGPT to test it with a full schedule, with child care for his kids. It downloaded things to his computer, like apps that it needed to be able to continue the process without him doing any of it. I mean, it's pretty absurd. Yeah, I certainly agree. Bill, I saw you had your hand up there. I don't know if you had a comment. Yeah. Good morning, everyone. I've just been thinking about this for, I don't know, like decades.

17:10I mean, I had the first machine learning classes I had 30 years ago at Stanford. I mean, things have evolved significantly, but a lot of the basic ideas of deep learning, generative models, reinforcement learning, all the concepts that we've been building on for 30 years. I mean, it's largely a function of Moore's law playing itself out, right? And so I think, partially to both Mark's point and Raul's point, I think that we're in a exponential learning phase right now, and you're basically dealing with an infant that is learning at an exponential rate. and that infant is basically the sum total of everything happening in AI right now.

18:04And I think Raoul made the point that the expert in AI is saying he doesn't know what's going to happen. And I think the reason is that we intuitively understand, and a lot of people don't, that...

18:23I don't know. Did everyone lose Bill there? I did. If you guys can hear Bill, give a thumbs up if you can. We, we, we, oh, I'm sorry. Did you lose me? For a second, but it seems you're back. I just say, you know, we tend to think about exponential technologies at decade scale. Now, because every developer on the planet is going to start using these tools, we're going to see them at sub-month scale, which we've never experienced as a human species before. And so I think Mark's point is fair, but at the same time, we've never seen an infant grow up in days as opposed to years. And I think that's what we're about to experience here.

19:03And we don't necessarily know the implications of that, but I can tell you we're not going to be having the discussion in a year of, is the quality here great? It's going to be, how is your organization becoming exponential because you're able to take advantage of these tools and what they can do for you? Who are you replacing? How are you measuring productivity? the KPIs that we're going to be contemplating around this are going to be emerging over the next few weeks and months. And I'm telling you, every company on the planet is thinking about it. I just came back from A360, which is Peter Diamandis' event.

19:42And half the event was reworked to basically help organizations become exponential thinking, AI-centered organizations. And everyone is completely focused on how to integrate these tools into their organization right now. Now, we may find that - Yeah, but Bill - Go ahead. Let me just ask one other quick question. So, I think your analogy is so spot on that it's an infant and we're treating it like an adult, right, which is the new thing. We all want to be our kids' friends. But I think that the challenge I see is this tool, one, it uses all data, two, and that will improve. It doesn't produce any insight.

20:33It actually imagines facts, right? It makes stuff up when it gets stumped. and it doesn't sound natural because it's processing language using rules and that's not what humans do. So I just struggle with whether it is going to actually become intelligent. I actually don't think it's intelligent. I think it is artificial, but I don't know. I'm a skeptic at this point of this particular tool being so overwhelming I mean, because I hear people talk about, oh, I used it. But then when I ask people, okay, show me something you produced with it, there ain't a lot. Yeah. I have a slightly different experience.

21:22Maybe get back to it. But my experience is right now is every developer that I know is generating code at a significantly accelerated rate, quality. 42 % of all GitHub code is now on a daily, weekly basis is now generated by... That's right. That's right. And so the point is that you have every developer in the world not only developing code, but also integrating now via these APIs that do give you access to real-time decisioning. They give you access to live financial data, live travel data.

22:05I lost Bill again there. I don't know if that's how... Sorry. Twitter doesn't shut the phone off. That's the problem. Twitter doesn't shut the phone off. That's the problem. Anyway, you're getting calls. Yeah. So farmers are going to be making real-time decisioning now. So I get that writing a newsletter may sound a little combination between a Harvard graduate student and a second grader. I don't mind that. But what I see is productivity gains exploding globally over the next few years as a result of this. Go ahead, Tom. Yeah, thanks, guys. Thanks for calling me up. So just a bit of background.

22:49So I wrote for a long time for Masari and I do my own independent research. A lot of that is macro-focused and crypto-focused. And I could tell you every day, you know, I sit down and with a blank page, try to think about a new and interesting topic to write about. And there's always, you know, three or four that I have going. Leveraging chat GPT, as long as you can build the scaffolding for what you want to write, I've been able to, you know, three to five X my output, you know, from simple things like here's the world's reserve currency. And then, you know, the projections of where it could be going forward.

23:20I try to build that part myself. The background chat GPT gives me all of it. as long as I'm asking the right questions and plugging in the right data, it can substantially improve what you're looking for. So I think on the research, the writing side, all of these integrations are going to 10x the researchers who can ask the right questions and know what they're doing and build the right frameworks and really leave a substantial piece and portion of all of those folks who can't do this thing behind. To Raoul's earlier point about social media, I think we're probably talking to a substantial amount of bots and you don't actually already know it.

23:57I was working with a company the other day that essentially trains itself on all of your tweets, all of the content that you want. Say you want tweets about cryptocurrency, DeFi, emerging trends. You plug that in, it scrapes Google News, it scrapes your Twitter feed, writes it in your voice. And every four hours, we'll send out a tweet based on that. and it keeps training itself until you get the content right. And then you basically don't have to be out of the way anymore. And this stuff's already happening for Twitter, LinkedIn, and others. So I think we're talking to bots. We don't even know it.

24:31And I think the folks who can train these things and ask intelligent questions are really going to be the ones who jump ahead. Hey, everyone. We're going to take another quick break and hear a word from our partners. We'll be right back to the Real Vision Crypto Daily Briefing.

24:49Go ahead, Mike. And after your comments, I want to move into how this actually impacts the crypto market. Go ahead, Mike. Oh, hello. Thanks for having me. I think this to me fits into the macro that I say, okay, well, there's a lot I don't know about the specific trees of the artificial intelligence, but it's definitely the stage of suddenly from gradually. Because I remember hearing about this when I was an undergrad, playing with that on a 486 computer, just wasn't ready. But to me, this is what corporate leaders are getting that are typical, like Federal Reserve people don't understand about significant deflationary trends that you mentioned, Jeff Booth nearly nailed in his book, The Price of Tomorrow.

25:27I'm glad you're going to have him on soon. To me, that's what people really missed last year and are missing now when I see things like producer price indexes dropping at the fastest pace in history since 1948. And we have the Fed still taking, I just look at this, You just have to assume that the technology is moving so fast, it's so deflationary. If you don't, you're falling behind. And that's a key thing I get at all conferences I go to, metals, to agriculture, to cryptos, to energy. It's a deer in the headlights look at what's happening with technology. A guy out on the farm said recently, Mike, I went to this event and John Deere showed me their new technology.

26:03It's shocking what they can do now with planting technology. That to me is what's accelerating. And so I look from a macro standpoint with cryptos is this is deflationary trends that people thought were kind of over are re-accelerating. And in the macro, we still have the Fed still tight. So that's why I rope it in. And that's why I like to listen to more of the specifics. Because to me in the macro, this is the big picture of we're going to see by the time we talk next year is all this is going to tilt towards people. It's going to replace jobs. But it's a transition towards higher productivity, lower prices for everything.

26:36Yeah, I'm sure everybody on this panel probably agrees that inflation is pretty much done and we'll be talking about deflation more in the future. But Bill, you said something really interesting. And you were talking about basically the fact that everybody's talking about this in every meeting and that you have to get ahead of it or you're going to be left behind. I said the same thing. And I have to be honest, I remember all of us saying the same thing about Bitcoin and crypto for the last two or three years. And so I don't know. I'm not saying that that's a point against certainly against AI, but I can certainly remember when MicroStrategy bought Bitcoin and Michael Saylor was hosting 2000 CFOs and every company was going to have to add Bitcoin to their balance sheet or be left behind.

27:21So are we getting ahead of ourselves here or is this different or is that all happening for crypto? Raul, I would love your take on that because we've had these conversations so many times. Yeah, so look, the crypto market, as we know, is driven by two things, the network adoption model, which is what makes it an exponential. And in addition, it's liquidity. So we've gone through the liquidity down cycle. So guess what? Less people are interested in crypto, less people are interested in VC, because there's not enough money around all of that stuff. So let's zoom out. I spend a lot of time speaking to the world's biggest institutions.

27:54They are all over this space still. Everybody's done their homework, everybody's set up, but they tend to be price followers. I also, by the business that I set up, Suns Magic Studios, we talk to the world's biggest brands. All of the fashion brands, all of the sports companies, a lot of the media business and the music business is all hyper-focused on this space. They understand that Web3 equals customer loyalty in a new way and that community is their new business model. That is not going away. so as price recovers and liquidity recovers we will see yet again another phase of acceleration if we think of the previous acceleration phase we saw the rise of nfts the rise of defi those were the two big things and the rise of layer twos so three big things happened the last cycle what is this cycle going to be well the most likely outcome is going to be the applications cycle.

28:53So I think we will see a lot of this. And then we've got some super cycles on top, which we know a lot of people have talked about. And I've been trying to drive the discussion on stuff like digital ID and how important it's going to be in an AI world. So I think that we've seen this with almost all technologies that come in ebbs and flows. We saw an enormous amount of money in VC going into the Web3 space. There's a lot of people building. Sure, some people will run out of runway and that's the normal course of events. But out of the other side, there's going to be a whole load of developments.