Crypto AI: A Fad or the Future? w/ Ben Goertzel

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-V-I-S-I-O-N. Hey there, before we get started with the show, we just want to let you know today's Crypto Daily Briefing is in partnership with the Crypto Phenom Letter. This is a one-of-a-kind premium investment newsletter service that highlights the next winners coming in cryptocurrency tokens and equities. Get special access today at phenomcrypto.com slash realvision. That's phenomcrypto.com slash realvision.

1:25Dr. Ben Gertzel, CEO of Singularity Net. Welcome to Real Vision Crypto Daily Briefing. Hey, good to be here. Ben, it's great to have you. Looking forward to this conversation. Excited to talk about AI and crypto and the nexus thereof. But first, we've got a little bit of news flow this morning, and I want to take a look at some price action. First up, Bitcoin trading right now, 25 ,653. One day that's off about five and a half, we'll call it 5.7%. It's dropping right now on my screen and yours on some news flow that we're going to talk about in just a second here. But first I want to talk about Ethereum.

2:02Ethereum right now trading at$1 ,789. It's off about nearly 4 % in the last hour, off about 6 % in the last 24 hours, trailing seven days off about 5%. All of this happening in the wake of revelations that SEC has sued crypto exchange Binance over its violation of U.S. rules and law. This is a breaking story. It's developing here hard and quickly in the last call at 60 minutes or so. I want to just read this to you right out of the Wall Street Journal so you have a sense of what's going on. The Securities and Exchange Commission on Monday sued Binance, the world's largest cryptocurrency exchange, alleging the overseas company operated an illegal exchange in the United States.

2:46The SEC also named Chengpeng Zhao, better known, of course, as CZ, Binance's founder and controlling shareholder, as a defendant. This is personally. The SEC said that Binance and Zhao misused customer funds and even diverted them to a trading entity that Zhao controlled. The trading firm Sigma Chain engaged in fraudulent trading that made Binance volume appear larger than it actually was, the SEC said. This is from a Wall Street Journal article literally within the last 30 minutes or so that has just crossed the wire. With that as backdrop, Ben, an eventful day. It's great to have you here. Yeah, great to be here.

3:26And I mean, the crypto markets have been completely insane in various respects since I got into this space since 2017. So, I mean, you sort of, after a while, get used to like, let that happen and focus on building amazing stuff, which is what we've been doing in SingularityNet, or at least trying to. Ben, you are someone who has been in the AI space for a long time. You wear a lot of hats. Talk a little bit about your background, how you got to where we are today, and all of the things that you're doing right now in this space that represents the interface between artificial intelligence and blockchain.

4:10Sure. Yeah, I've been doing AI since mid-1980s, really. I got my PhD in math in 89, but already then I was doing AI prototyping, experimenting, and R &D. And, you know, reading about AI since the early 70s when I was a kid, AI has been around a while. The term artificial intelligence was coined in the late 50s, I guess. And already there'd been research like that going on for a couple of decades before. So AI has been building up for quite some time with prototypes and research and ideas and even deployed commercial systems for many decades. But of course, there's never before been such an explosion of AI capability as we've seen in the last couple of years, even the last six months.

5:04So it's a super exciting time to be doing AI. What got me into the AI field in the first place was the promise of what I eventually came to call AGI or artificial general intelligence meaning AI that could generalize take imaginative and creative leaps beyond its programming and beyond its its training data and thus you know advance itself much like much like people at our best are are able to do and i introduced the term agi in i guess 2004 or 5 and i've been organizing agi research conferences since then of course when we got into crypto i used agi as the the ticker symbol for for singularity nets a token we didn't change to agix as we made it made it cross-chain.

6:00But the tension between narrow AI for specific applications and AGI, meaning AI that's broader and can extend to any application, basically beyond its programming or training, this has really been there since the beginning. It is interesting to think about in the current market. In the very beginning of the AI field, in the 50s and 60s, people were after AGI, though they didn't have that term. They were after machines who could really think like people. They thought it would be much easier and they thought it would just be a few years of hacking, right? And then gradually people started to realize how hard it was to make a machine do some of the basic things people take for granted, like recognize pictures or make legs walk down the street.

6:49But that it was, you know, counterintuitively simple to make an AI do some things that were really hard for people, like solve algebra equations or play checkers or chess at the master level, right? So then the AI field sort of split into people focused on making AI systems just do practical stuff that seems smart when people do it by hook or by crook versus research toward making machines that can really think in a broad and general sense more like people can. And now AI doing specific tasks got a big boost when the internet came out, obviously, just because there was networks of machines and there was more and more data to feed into the AIs.

7:35And AGI R &D on making real thinking machines has been progressing at a smaller scale along the side. Let me just jump in here to frame this up a second just so I can get, because I know it's a tremendous amount of information and I want to make sure that we can explain it to people in a way that they can form some kind of categories in their head to break it down. You are really one of the true experts in this space. Pretty extraordinary that you've coined the term AGI, artificial general intelligence. A great debate about where we are with regard to AGI right now. You alluded to something called Morvex paradox, the notion of some things that appear to be relatively simple, physical motor activities, sensory motor activities, much more complicated for computers to do than basic algebra.

8:20You mentioned singularity. There's just a whole lot of concepts here for us to unpack. I'm excited to get to do it. So let's talk a little bit about, you know, one of the things that really strikes me and something that you alluded to initially there in your remarks is this idea was for a very long period of time. AI was coming. It was coming. It was just around the corner. It seemed like it was just around the corner for decades. And then suddenly there just seems to be this explosion within the last six months. And some of this is probably the degree of popularization that ChatGPT has created in the business world.

8:52in more general news cycles, but talk a little bit about what's happened here in the last, say, six to 12 months that's created this just explosion of new use cases, applications, and view. Yeah, I would say progress on AI has felt a lot more continuous from the inside when you're working on research and development. So I can see from an end user's point of view it's like whoa i wasn't doing anything i could see or understand and then boom suddenly it's making pictures and writing documents for me right but from from the standpoint of someone in the middle of the research field ai has been making pictures and writing documents for a long time just not quite as well and it took a huge amount of compute power to to do so so and ai i mean it's been proving proving math theorems and predicting predicting markets to controlling robots.

9:51AI has been doing a lot of things for a long time. There has been a jump, an inflection point in the last six months, but it's probably not as big as it appears to the retail end user. But I do think from a sort of economic point of view and societal impact point of view, the launch of ChatGPT by OpenAI is a sort of black swan event, right? I mean, it's a big deal that changes everything, even though there are very few new ideas inside the ChatGPT code base. And it doesn't really work that much better than things that existed before it. But I mean, it was rolled out to the public at large with a very simple interface.

10:43Everyone's using it. Everyone's thinking about it. and everyone's customizing it, right? So certainly in terms of public visibility, public enthusiasm, and just amount of resources going behind developing and deploying it, it's a black swan event, right? And it's quite interesting and important from that standpoint. I mean, I think it's important to clarify that chat GPT, I mean, And while it can impersonate human intelligence effectively in some settings, it's certainly not a human level general intelligence. I mean, it doesn't deviate too far from its training data. It's just that its training data is the whole goddamn web, right?

11:30So, I mean, it's got a lot of data in its mind, so it doesn't have to deviate too far from its training data to do some pretty impressive stuff. from an economic standpoint, it may well be that minor variations on GPT-4 and related systems, because there's a lot of other ones out there now, it may well be that minor variations on this type of transformer neural net could do, who knows, 70%, 80%, 85 % of jobs that people do now. And just because most jobs people do are minor variations on stuff that's been done in the past, right? Hey, that's a striking. take a big leap beyond experience to do most of what people get paid for, right?

12:14So it's like on the one hand, no, it's not yet a human-level thinking machine. And we can talk about what the gap is, which I'm trying to fill, to use chat-GPT-type systems to jump toward human-level intelligence. On the other hand, even without being at human-level intelligence, well, it can be massive disruption to the economy. and a lot of the disruption that one might have thought would happen only after you had human level AGI, a lot of that disruption could happen beforehand, right? If these, you know, massive scale narrow AI systems trained on everything in the web can take over most human jobs, that's going to trigger tremendous disruption and then reorganization economically and societally.

13:00Well, I'm really eager to talk about the jobs component because I think that 80 % number that you cited has gotten some buzz, some headlines around you talking about it here in the last couple of weeks. I'm talking about here in mainstream media. This has gotten quite a great deal of pickup. Before we talk about that, I want to ask you about AGI versus the singularity. These are two terms that are floating around out there. Can you define them and compare and contrast those two points in terms of the general trajectory? I think that's important. So AGI is a very general term. I mean, artificial general intelligence.

13:36And humans are by no means the most general intelligences that are possible, right? I mean, we're pretty smart. But we're, in a way, the minimum possible artificial general intelligence, you know, capable of building its successor. So, I mean, humans are an inflection point. not because we're the smartest thing that will ever exist but because apes and dolphins and chickens cannot engineer their successor. They're just sort of unwitting parts of the evolutionary process. So you can talk about human-level AGI, meaning software systems that can do general intelligence at roughly the level of humans, meaning that their stupidest component is at least around the human level.

14:24They might be smarter in some things. But then you could talk about superhuman AGI's, general intelligences that are way smarter than people in the same way that we're smarter than chimpanzees, mice, or slugs, right? And what's interesting is when you have a human level general intelligence, if it can program and do engineering, it should be able to improve its infrastructure so as to turn itself into a superhuman general intelligence, right? So if you look forward from here, I think there's going to be a threshold event, another black swan event when someone makes a human level general intelligence that's even smarter that's as smart as people in every significant regard and a little smarter in a lot smarter in many regards but then after that you're going to see another black swan event where that human level AGI has reprogrammed itself to make itself way way smarter than people which has It's been called an ASI or artificial superintelligence, right?

15:23So once you get that ASI, that artificial superintelligence, then in a sense, all bets are off, right? I mean, you have something 10 or 100 times more generally intelligent than people on the planet. And I mean, we can speculate about what that's going to lead to, but it's hard to analyze in depth. That's what Ray Kurzweil and Werner Vindy and others have called the singularity thinking. At that point, your phone may be making 10 Nobel Prize-level discoveries every minute. So in a way, the rate of progress of knowledge and invention becomes effectively infinite from the comparison point of human intelligence.

16:08The interesting thing is the timing of these things could be rather soon. I think we could be at human level AGI within, say, three, five, or seven years from now. And we could then be at super intelligence within a few years after that, right? And so this is the conclusion that rational analysis leads me to, but it's pretty mind blowing from a personal and emotional standpoint and hard to incorporate that sort of analysis into your everyday life, right? Like in your everyday life, you're thinking about like, how do I plan for my kids to go to college and where I'm gonna live when I get old and retire?

16:47But then analytically you're thinking, well shit if it's three or five years to a human level thinking machine and three years after that to a machine a hundred times smarter than people like what is that by the time my five-year-old gets to college we may all be like floating around in a hundred dimensional space or just you know instead of a phone we have a molecular assembler that builds anything we describe to it right and so it's it's a quite dislocating time in history to be alive which i i find quite fun than many other people seem to find disturbing. Well, I find it both fun and disturbing, I guess.

17:25It is really kind of the best and worst of times in terms of risks. You talk about this idea of 80 % of human jobs being eliminated. This got a lot of buzz and a lot of pickup in the last couple of weeks, mostly from you actually citing that statistic. Talk a little bit about what that would mean in terms of the actual implementation, how those jobs disappear, and what your thoughts are about the impact that it will have on broader society. Yeah, so first of all, the 80 % figure is definitely back of the envelope. I mean, I may write up a formal paper on this at some point, but I mean, I just crudely went over the different job categories that are analyzed by U.S.

18:08government and sort of tried to think myself through which pieces of which job should you be able to do without AGI. but with really smart, narrow AI by integrating chat GBT type systems with other tools that we have now. And it comes out with pretty much all of what humans do for a living. You can automate without AGI. I mean, there's a few exceptions. I mean, exceptions would be things like a preschool teacher or a therapist. They're just about people helping other people to be better people and sort of rely on human connection. I mean, very advanced level of science and engineering. I mean, chat GBT can't do that and probably requires human level general intelligence.

18:57And then, you know, strategic thinking where you're trying to go beyond anything that happened in history before or the sort of high end of the fine arts where you're inventing a new genre of music or being an incredible virtuoso or something. But this is not that high a percentage of what people get paid for. So I think most of what people get paid for is in a way a minor variation on stuff that's been done an awful lot of times before. And also with the way modern society is organized, it doesn't require such a deep I-thou connection between one human and another. Actually, society has been organized to minimize the amount of deep and fundamental human contact that's required during the day in a job setting.

19:51So, I mean, I think this is huge. It happens to be less huge than the advent of true human level AGI, which happens to be less huge than superintelligence. But even this earlier thing of most jobs being obsoleted by infra-human level AGI is going to be big economically. I mean, I think obviously for those who are deploying the AGI, there's tremendous business opportunity, which includes both of us, I would suppose. I think in the developed world, it's almost inevitable you're going to have some form of universal basic income getting rolled out because, I mean, there's already a lot of talk about it and I just can't see developed countries letting like half their population become homeless and starving.

20:40What worries me more is what happens in the developing world, right? Because there just isn't money to give everyone basic income. And we're not very good at sharing wealth between the developed and developing world. You may see very exploitative geopolitical dynamics where, I mean, superpowers offer developing countries offers they can't refuse. Like, give us all your minds and military control of your country in exchange for giving you money not to let your urban poor starve, right? So, I mean, there's a lot of weird things that could happen. And one of the tradeoffs, which is interesting, is on the one hand, you could say the best way not to have chaos, mayhem, starvation and conflict during the rollout of smarter and smarter AI systems will be to accelerate progress even faster.

21:37Because once you have a super AGI, presumably it could create abundance for everyone at minimal effort, including in the developing world. On the other hand, accelerating progress toward a super AGI could be viewed as a higher risk thing to do. Because you might say you want to be careful and make sure you're building the right sort of super AGI, right? So on the one hand, there's tremendous humanitarian benefits of accelerating progress if you can do it right. On the other hand, there's tremendous risks to accelerating progress. And then you see in a global competitive context, there's risk to each country in accelerating progress of AI within their borders if they can do it in a way that fulfills their interests.

22:24Right. So there's going to be a lot of interesting dynamics going forward, which I don't think we can forecast in detail. So what I've been focusing on is trying to build general intelligence systems, which can be ethical and which are hard for any one party to grab hold of and control for their own interests. Right. And I think let's talk about that in just a second. If you have the tools for building ethical ATI and it can't be grabbed control of, at least then you're you're biasing the odds in favor of a of a positive outcome in the midst of all this uncertainty and potential confusion. Ben, let me ask you this, and we'll talk a little bit about what you're doing in just a second here.

23:08But it seems as though the one thing that can be agreed and understood at this point is that the rate of change is going to be simply unprecedented. presented you talk about the risks in emerging markets i think it's very clear how that risk case can unfold in developed markets you talk about ubi universal basic income i guess one of the questions that i have for you is someone who's thought about this from not just a technical perspective but also a social and cultural context what are the risks of a dramatic shift in the way that work gets done i think anybody who's ever spent time hanging out with kids with trust funds knows the risk case that can unfold when you essentially sort of destroy the structure of people's lives and they sit home and they just collect a check and play video games or dive deeper into artificial and virtual worlds.

23:55What are some of the risks that you see from a radical restructuring of society around universal basic income rather than work that is hopefully fulfilling and meaningful to individuals who participate in it? I think that's a fake problem basically I think the issue with trust funds kids is that they're raised believing they have a higher status than everyone else which is sort of psychologically deforming and I mean I have I have three three adult kids age like 33 29 and 26 or something like that and three three little kids age five and two I would say none of my adult kids nor their friends have any worry about oh no what if i don't have to work for a living and get free money instead my life will have no meaning i mean they've got a lot of ways that their life has meaning that don't require earning money and they'll be very happy not to have to solve that problem in their lives of finding a you know a workable stable way of earning income i mean there's intellectual satisfaction there's spiritual growth there's making art and music there's hanging out with your friends there's climbing mountains playing sports i mean there's there's loads of ways to get meaning in life which exists in our society right now that are separate from from working to get to get paid and get get resources so i think for young people that will be no issue and there's a lot of young people on the planet average age in africa is average age in africa i think is 14 right so i mean it's true for more people getting on in years who've come to define their identity around their work, it's going to be very dislocating.

25:37I have a theory people will get over this really, really fast, similar to with longevity. Like if you talk to people, well, should we develop a pill that will let people never die? People will be like, no, but death is what gives life its poignancy and meaning. But if you're like, okay, here's the pill, do you want to restore your body and mind to its 20-year-old condition and leave it there until you decide to stop taking the monthly pill. 99 % of people are going to say, oh yeah, give that to me now. I'll find another way to give life poignancy, meaning that doesn't involve becoming senile old and rotting in the grave, right?

26:12So, I mean, I think in the end, that will be a positive thing. And what we should worry about is not 50-year-old CPAs feeling a loss of meaning where they no longer have to work for a living. We should be worried about the people in Africa, Central Asia, and Latin America who don't have money to buy prescription medication or pay their phone bill during the transition period between when AGI has taken their jobs and when superintelligence can provide abundance to everyone. I mean, there's just a lot bigger problems than the small segment of middle-aged career people feeling uncomfortable because they can get free money and their status isn't as high as it was before.

27:00I'm not sure that I agree with you. I'm not sure that I agree with you. I'm not sure that I agree with you. I'm not sure that I agree with you that it's just middle-aged career people who may potentially be dislocated by this. You know, this notion that it's just going to be people who are over 40 or 50. I think the question about what we do with the structure of our lives remains an open one. I certainly hope that I am wrong and you are right. I agree. I was a bit flippant in that response. But I do think that younger people on a whole tend to be more adaptive, right? And I think we will all need to be adaptive.

27:42And that's probably a more key, more serious point to be less flippant about it. I mean, as you said, change is going to unfold faster and faster. it's accelerating change approaching a singularity there's going to be more and more different on january 1st each year from january 1st the previous year right so that the primary trait each of us needs to have is to be able to you know think on our feet and uh pivot and figure out new ways you know both to earn money while that's still a thing and also new ways to define meaning in life and new ways to give to others around us. And the ability to adapt will be the primary important trait.

28:32And, you know, my friend, Pei Wang, who's been an AGI researcher as long as I have, he defines intelligence as the ability to adapt to an unpredictable environment under limited resources, which is, I mean, it's one among many sort of equivalent ways to phrasing what is general intelligence but i think that's a way to think about it that's very pertinent to what's unfolding in the world around us now right like we are we all need really badly to be able to adapt to unpredictable conditions given given limited resources and that's probably speaking about resources ben we've got limited time here today so i want to give you the opportunity to talk a little bit about what you guys are working on at singularity net talk a little bit about the project, why you spun it up.

29:20Obviously, folks who have been watching the token here, AGIX, about 500 % increase in value here, trailing 12 months, I should say. Talk a little bit about what the goals were for that project and where you think you are on that road right now. Absolutely. So our goal with the Singularity Project was to build beneficial, decentralized AGI. and the reason I founded Singular Internet in 2017 is it was already clear to me at that point we were not that long, like maybe 10, 15, 20 years away from machines that are smarter than people and I'd had the idea since the late 90s when the internet and the web became a thing the best way to work toward beneficial AGI would be to deploy the AGI across the internet in a way that had no central owner or controller.

30:17Like AGI has to be like the internet or like Linux. It has to be all over, owned by everyone and no one, contributed to by a vast mass of contributors. And if it's not that, once you have a threshold event where AGI really starts to be coming, some group of jackasses with selfish urges is going to take control of it and try to use it for their own benefit, which could be a big company. It could be a country and their military intelligence organization. I mean, no one controls the Internet. No one controls Linux because of the way they're deployed in a decentralized manner. But then an AGI is trickier than the Internet or Linux because it has to be like live running systems on computers.

31:01So you really need a decentralized network of live running processes that's running without any central owner or controller. And this is enabled by SingularityNet platform. SingularityNet doesn't enforce any particular architecture for thinking machines, but SingularityNet is a platform that any sort of AI algorithm can be run on across a decentralized network of machines without any central owner or controller. And then there's other pieces to that platform now. So we rolled out something called NuNet last year that deals with tokenization of the processing power underlying this network. We did a token sale for HyperCycle just a month ago, which is a customized layer one blockchain, customized to serve the needs of a decentralized AI network.