Ben Goertzel, Founder & CEO of SingularityNET: The Intersection of Bitcoin and Artificial Intelligence

Ben Goertzel is Founder & CEO of SingularityNET, a blockchain based AI marketplace. In this conversation, Ben and Anthony Pompliano discuss the current state of artificial intelligence, how it could impact the world moving forward, what Ben's doing to build up the two-sided marketplace, and where he sees Bitcoin, crypto, and AI intersecting in the future. BLOCKFI-----BlockFi allows you to keep your crypto, put it up as collateral, and receive a USD loan funded directly to your bank account. They do loans ranging from $2,000 to $10,000,000, and they're perfect for helping you reach your financial goals of all sizes. Visit BlockFi.com/Pomp to learn more about putting your crypto to work without having to sell it. CRYPTO. COM-----Crypto.com is a pioneering payment and cryptocurrency platform that seeks to accelerate the world's transition to cryptocurrency. With the vision of "cryptocurrency in every wallet", the Crypto.com App offers a full range of financial products with competitive pricing, well designed UX and high security. It is the best place to buy, sell and pay with crypto. COINMINE-----The Coinmine One is like an Xbox that turns your electricity into Bitcoin. You just plug it in, connect to wifi, and tap on the crypto you want. It’s so easy anyone can do it. Everything is controlled from the Coinmine mobile app and the Coinmine keeps getting better with over the air updates that add new coins, features and services to your Coinmine. Visitcoinmine.com/pomp to get a Coinmine and earn crypto for powering a new world.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:01Anthony Pompliano?

  What's up everyone, this is Anthony Pompliano. Most of you know me as Pomp. You're listening to Off the Chain, simply the best podcast in crypto. Let's kick this thing off.

• 0:12Anthony Pompliano?

  Ben Goertzel is the founder and CEO of SingularityNet, a blockchain-based AI marketplace. In this conversation, we discuss the current state of artificial intelligence, how it could impact the world moving forward, what Ben is doing to build up the two-sided marketplace, and where he sees Bitcoin, crypto, and AI intersecting in the future. I really enjoyed this conversation, and I hope you do as well. Skrt skrt! Wanna know who has the best URL? Crypto.com. That's right, Crypto.com. They're a crypto platform with one goal: motherf***ing mass adoption. That's why we're all here. We're trying to get crypto in every wallet. Crypto.com is helping people do that through

• 0:56Anthony Pompliano?

  buying, earning, lending, and card payment. Everything you could want at Crypto.com. Go help your boy out, tell him Pomp sent you. Download the app or visit Crypto.com. Pomp's got you, always. Ever wanted to get into mining and didn't know how? Don't worry, your boy Pomp's got ya. Everybody got some electricity and wifi. All you gotta do is go to CoinMine.com, you buy a coin mine. It's like an Xbox or a Playstation that helps you turn your electricity into Bitcoin. That's right, you purchase it, it shows up at your doorstep, you pull it out of the box, you plug it in, connect to your wifi, five minutes or less, you're mining Bitcoin. All you have to do is control it from the mobile app they provide, and then you receive

• 1:39Anthony Pompliano?

  over-the-air updates that add new coins and new features on a consistent basis. Kinda like how Tesla does over-the-air updates and updates the car software. Just your update in your coin mine. Consumer mining made easy. That's right, go to CoinMine.com, tell them Pomp sent ya, and thank me later. As many of you know, crypto investors store their digital assets on exchanges or in cold storage for long-term safekeeping. However, this strategy doesn't help them grow their investment holdings or build overall wealth. With the new BlockFi interest account, users can now securely store their Bitcoin or Ether at BlockFi and receive 6% annual interest paid monthly in cryptocurrency. 6% is an absurdly high rate. It's the best rate in the industry. I highly suggest you

• 2:23Anthony Pompliano?

  go check out blockfi.com/pomp. Again, that's blockfi.com/pomp to sign up and start earning crypto today.

• 2:31Speaker 01

  Anthony Pompliano is a partner at Morgan Creek Digital. All opinions expressed by Pomp or his guests on this podcast are solely their opinions and do not reflect the opinions of Morgan Creek Digital or Morgan Creek Capital Management. You should not treat any opinion expressed by Pomp as a specific inducement to make a particular investment or follow a particular strategy, but only as an expression of his opinion. This podcast is for informational purposes only.

• 2:56Anthony Pompliano?

  All right, guys. Bang, bang. I've got Ben here on the podcast. Super excited about this. Thanks so much for taking the time to do this.

• 3:04Speaker 02

  Thanks for having me. I'm looking forward to being on your show.

• 3:09Anthony Pompliano?

  Absolutely. All right, let's start with your background. You've been working on a lot of this stuff for a long time, so maybe just walk us through your career before Singularity Net.

• 3:17Speaker 02

  Sure thing. I mean, I got a PhD in mathematics in the late 1980s, and I've been basically doing AI research and development for 30-plus years. I mean, I was academic for eight years in math, computer science, and cognitive science, but since the late 90s, 20-plus years, I've been doing applied AI, so helping big companies and government agencies solve their AI problems. And yeah, at the same time, doing AI research and development aimed at the grand goal of artificial general intelligence, AIs that can think more like

• 4:00Speaker 02

  people. And, you know, this last few years has been amazing for AI development. More and more things are working now than ever before due to, you know, more data, faster computers, and more people working in the field. So that both makes those of us who have been in the field a while really excited that, you know, the next five to 10 years are going to be legendary. We may see the breakthrough, the AIs that really can think better than people. On the other hand, it also makes us worry about, like, who controls these? First really powerful AIs and you know what are their goals and biases? What's what's on their minds?

• 4:41Speaker 02

  And so this is what led me into Singularity Net, which is aimed at making a decentralized foundation for for AI. And so so that there's a more of a democratic participatory dynamic underlying not only practical AIs being rolled out now to help companies and people do their stuff and help people with their lives, but also more of a democratic and participatory basis underlying superhuman AI minds once those finally emerge.

• 5:12Anthony Pompliano?

  For sure. Now, before we kind of dig into some of the nuances here, you know, artificial intelligence gets talked about all the time. Maybe describe a little bit, you know, when you say artificial intelligence, what are the criteria for an actual system that is artificially intelligent versus maybe something that is automated with algorithms or maybe just like machine learning type environment?

• 5:39Speaker 02

  Yeah. I think the the concept of AI is very, very broad, right? And I mean, intelligence doesn't really have a widely accepted exact exact definition that works across like different species and different kinds of systems. So I mean, the field of AI has come to encompass pretty much everything that software or hardware can do that we would consider intelligent if people were doing it, right? And and there's a joke in the AI field that as soon as something has been done by machines, then it's considered not AI anymore, right? So like we would have, we would have considered solving algebra problems or even dividing

• 6:24Speaker 02

  huge numbers intelligent. Now we have calculators and programs like MATLAB that's not intelligent. So we considered driving a car down the street in traffic to be an AI problem, but now now that it seems to be on the verge of being solved, we say, well, that's not really AI, it's just an automation problem. So I mean, there, there's a lot of vagueness there. So the a distinction that I introduced in 2004 or so, which I think is valuable, is the distinction between narrow AI and AGI or artificial general intelligence. And I think

• 7:04Speaker 02

  That's not the only important distinction, but it cuts through some of the confusion. A narrow AI is an AI that's good at doing one very narrowly defined set of things. And I mean, it could be just place chess, or just played board games, or you know, it just drives cars, or it just diagnoses diseases based on medical data. An AGI is a system that, a bit more like a human, can deal with new problem domains, new concepts, new aspects of existence that it didn't even know were there before, and that were not present in its programming or in its training data.

• 7:46Speaker 02

  And, you know, that kind of general intelligence isn't yet there in any software programs that we've created, but that's that's what those of us in the AGI field have been working on for years, and now we're seeing more traction in that direction. Like, Microsoft invested a billion dollars in OpenAI specifically to pursue AGI-oriented development, and to do it on Microsoft Azure. And I mean, DeepMind, that was bought by Google five years ago or something, also was explicitly pushing in an AGI direction. So I think what we're gonna see in the next five years or so is a gradual transition, like, from

• 8:30Speaker 02

  narrow AIs like we have now, through sort of narrow AGIs that have more and more general intelligence, which is still kind of focused on one application area, and then full-on artificial general intelligence, finally, which, uh, I mean, then that will quickly lead to what Ray Kurzweil called the singularity. Because once you have an AGI that can really think creatively like human beings, but you know, also as an AI with all the ability to access huge amounts of data and do precise computations, that implies, I mean, then then you're gonna have an AI that can improve its own programming, and get smarter and smarter and smarter.

• 9:12Speaker 02

  And so that's, that's what's been called superintelligence, that's a grand vision. Obviously, we're not there yet, and what, you know, what I've been involved in with SingularityNet is...

• 9:24Speaker 02

  Practical steps now that are just sort of the next natural steps in the progression toward the the evolution of really, really smart AGI and super intelligent systems. The amazing thing is, I really think this progression may unfold within our lifetimes. Like, we we may have, you know, general intelligence as smart as human beings within five to 10 years, and then smarter than humans a few years after that. And that is obviously extremely revolutionary. The upside is amazing. The possible downside is is terrifying, right? And so that that that that's sort of what incents me, wow, when I get up in the morning to work so hard on what we're doing.

• 10:09Anthony Pompliano?

  For sure. And I think one of the AGI stories, I think it was confirmed, but it was definitely reported, was that Facebook had created some software, and the two different systems started to create their own language.

• 10:22Speaker 02

  Yeah, unfor- unfortunately, that that that story was a complete load of crap, because these were random, trivial, formal languages rather than sophisticated languages like humans use to communicate. I mean, I think that that illustrates illustrates a broader point, that we're getting into a domain where, you know, the important technologies in the world are ones that that almost nobody understands, and that that that leads to a bunch of issues. So, for example, I worked for three years on the Sophia robot, which, created by Hanson Robotics, with some some AI from my OpenCog project and from SingularityNet, used in it from time to time.

• 11:06Speaker 02

  But, I mean, this robot, it's the world's most sophisticated humanoid robot, amazing human-like emotional expressions. And, you know, her AI system has a mix of some sort of canned content that's programmed in by people, and some learning and reasoning.

• 11:26Speaker 02

  And that's just too subtle for most people to understand, even for many researchers. Temoni Grappelwitz, I mean, she's an amazing robot, and these are going to be rolled out in the next few years as practical service robots for homes and offices. But it's very hard to get people to understand what's going on behind the scenes. And that's true even now, you know, how much truer will it be as we move toward general intelligence or superintelligence? And that, I mean, that, that gets us once again to the power of decentralized networks because, you know, if you're dealing with technologies that people can't understand, you know, who do they place their faith in? Either they place their faith in like the company that made them or the government, the government that regulates them, or do they place their faith in their, you know,

• 12:11Speaker 02

  the reputation system of the decentralized network that, that they, that they belong to? And I mean, the, the latter option is a little subtler, but I think ultimately it's going to be the best, the best option because, I mean, lo- large governments and large corporations certainly have their place and they do some things very well, but I think the regulation of emerging general intelligence is going to be better done by global decentralized participatory networks.

• 12:43Anthony Pompliano?

  For sure. What, what, um, are your big fears about artificial intelligence? You've been working on this a really long time. Like when you sit back, what are the fears that you agree with are real and, and you share versus ones that maybe people are worried about and aren't actually legitimate?

• 12:59Speaker 02

  I mean, most of the fears are real in the sense that there's a lot of bad things that could happen, but we don't have any reason to think bad things have a high probability of happening. I mean, the, the one fear that I think is total BS is like machines are going to enslave people to work for them or use them as batteries or something like in The Matrix. I mean, of course. If a machine was smart enough to do that, it would also be smart enough to create machines to do whatever work it needed, right? I mean, that's, that I think is, is very unlikely. The idea that we're gonna have machines that are hostile, hostile to people and wanna fight people, very unlikely. I mean, some, some idiot would have to build machines

• 13:43Speaker 02

  and program or teach them to wanna hate and fight people and could happen, not likely. The possibility that we create super intelligent machines that are kind of indifferent to people and treat us like we treat ants or squirrels. I mean, that, that seems a plausible fear to have, right? I mean, we, we don't, we don't wake up every morning like, how can I kill more ants or kill more squirrels? On the other hand, if we wanna build a house, we don't mind flattening a bunch of ants and destroying a bunch of squirrels homes because we just consider ourselves superior and more important to them, right? And then we, we preserve some of them in national parks and stuff. But we wanna preserve the species of squirrels.

• 14:25Speaker 02

  We don't really care about preserving each individual squirrel very much. So, I mean, it's plausible that a superhuman AI eventually could view human beings in, in, in that, in that sort of way, like some respect for the human species as, as its creators and you know, not malevolence toward us. On the other hand, if they're like a billion times smarter than us, they may not care about us any more than we care about much simpler life forms. And I mean, if, if we wanna ward off that kind of problem, we need the superhuman AI to be more compassionate and kind and ethical than, than humans are,

• 15:07Speaker 02

  which may be considered a low bar in a sense, if you look at human society. But, I mean, it's not a foregone conclusion. So, we, we need to be training the first AIs to really be compassionate and loving toward all sentient beings so that they'll be nice to us once they're more powerful than us. So then, then you look at, you know, what is AI being used for in the world today? And I, I like, I sum it up as spying, killing, selling and gambling, basically. I mean, it's surveillance applications, it's, it's military, it's financial market prediction and, and trading, it's advertising. So, all of which are

• 15:49Speaker 02

  Parts of humanity and they're gonna keep existing, but it will be good to sort of shift the shift the balance a little bit, so that there was more focus on applying AI for loving and compassionate applications. And that's, I mean, yeah, we can't we can't obviously change everything the world is doing with AI, but we can shift the balance in a more positive direction and that can help things in the near term as well as hopefully filling the early stage mind of the emerging superintelligence with more

• 16:32Speaker 02

  more positive stuff, right? And so, I mean, these are all sort of some of the broader motivations behind the very nitty gritty, practical work that we're doing now.

• 16:43Anthony Pompliano?

  Absolutely. Maybe let's talk a little bit about what you're building at SingularityNet.

• 16:48Speaker 02

  Absolutely. So what we've built in the last year and a half is

• 16:55Speaker 02

  two years, really. It's a decentralized blockchain-based AI platform and marketplace. So I mean, it's a SingularityNet platform now in the beta version two, not the 1.0 release yet, but there's a version that works. It's reasonably scalable. There's a bunch of AIs working in it, doing various you know, vision processing and biology, data analysis and so on. So this it's basically, it's an AI marketplace. So anyone can put an AI agent into the marketplace and then it will advertise its services to any anyone who needs AI services and wants to

• 17:38Speaker 02

  wants to use the marketplace to get them. But also, there's a novelty that the The different AIs in the marketplace can all communicate with each other and talk to each other. And an AI in the marketplace can outsource work to another AI in the marketplace. So then you have what AI pioneer Marvin Minsky called a society of minds, where the different AIs in the network, they're sharing information, they're outsourcing work to each other, they're cooperating, and then the whole network of AIs in the Singularity Net, you know, in itself has an intelligence beyond the intelligence of the individual AIs in the network.

• 18:21Speaker 02

  So I mean, where we are now, we built this platform. We put some cool AI tools into the platform. The presence of a cryptographic token, the AGI token, which is a payment and incentive token for transactions in the marketplace, that gives us an additional tool to play with for sort of incentivizing suppliers, incentivizing developers, because we can use like AGI token bonuses, we can do contests with AGI token rewards and so forth, to spur AI developers to use the platform.

• 19:05Speaker 02

  And then on the demand side, we're doing a bunch of different things. But one thing is, we formed this spinoff company called Singularity Studio, which is a separate for-profit company. And it's sort of like you have Ethereum and Consensys, you have Linux and Red Hat, right? So we have Singularity Net, which built the open decentralized platform, and which has the AGI token powering it. Then we have Singularity Studio, which is building enterprise products on top of that platform.

• 19:41Anthony Pompliano?

  So Ben, explain a little bit, we talked about this two-sided marketplace, some of the challenges of building it. What are the services that I as a user come into? What are the services that I'm looking to use? And what's that user experience like for me?

• 19:56Speaker 02

  Yeah, so in terms of use of the marketplace, I think the the AI services offered in the marketplace are really they're mostly aimed at relatively knowledgeable developers who are they're building products, they're building websites, and they need an AI component to their product. And, you know, they could use an AI service in in Google Cloud or Amazon Web Services. But an alternative is to use use an AI service from SingularityNet platform.

• 20:38Speaker 02

  So let me give just one example, like we have we have some AI services in there now that do image style transfer, so you can sort of map the style of one image to another one. Now, if you were, if you were, say, a modeling agency and you wanted to do something like, take

• 21:00Speaker 02

  take, you know, the head and face of a star model and sort of morph it onto the body of a of a less famous model, but who was cheaper to get to pose in various places, right? So then you're sort of, you're making a mashup of different models who are all in your own portfolio. So that you can show the star models in a lot of different poses and locations without having to pay them to go all those places. So if you're a modeling agency who wants that, I mean, you might hire a software developer to build that application for you and train it on all the photos you have of models. But you might not have the AI component to actually

• 21:46Speaker 02

  do the image magic that you need to combine the different images together. Then we have a service on the SingularityNet marketplace that a developer could actually use to provide the AI component inside that application. And that example illustrates some fairly common aspects, which is you often have business customers who have their own data that they want to use to drive some AI application. But they don't have the AI wizardry to leverage their data to do what they want. And so we can provide

• 22:25Speaker 02

  AI tools that a developer can integrate and apply to the customer's data to deliver the function that they need. So I mean, that. If you were a modeling agency and you had data with images of a bunch of models, and you wanted to sort of map the face of one model into the body of another model because you had photo of Model A in, you know, by the beach, but Model B wasn't available to go to the beach, right? So if you, if you want to do mapping of one model's face to another model's body, we have, you know, image, image style transfer services and stuff that can, can help with that. And you can then access this software on our platform,

• 23:07Speaker 02

  provide the AI wizardry, you know, inside your own domain-specific application. Or if you're a biologist who's analyzing DNA data, and you, you have a bunch of DNA for people with prostate cancer, you want to understand, you know, what genes are important for prostate cancer and why to help develop new therapeutics. Again, we have bio-AI analytics tools on, on the platform that can, that can help, help you interpret your data. Similarly, like say you're, you're a bank who wants to do credits going for small, medium enterprises

• 23:45Speaker 02

  based on all sorts of different data, social media data and, and financial data. Again, financial analytics tools on, on, on the platform can, can help provide the AI component deep, deep inside your product. So I think that the platform in itself is, it's really geared toward technical people, toward product developers who would otherwise either have to build their own AI code, or use some service on Google Cloud, or Amazon Cloud, or Microsoft Azure. We're giving an, an alternative to, to build it yourself, and an alternative to big company.

• 24:28Speaker 02

  AI services and I think as as the platform grows bigger and bigger. We should be able to get, you know, a much wider and longer tailed set of of AI services on the on the platform because we we don't just have the things that a big company decide to provide to their users. We have whatever anyone in our in the world decide to put put onto the onto the platform. So I would say the the other the other set of customers we're looking at is large corporates and that's a bit different. So I think a small company or an AI startup is gonna just use the AI services in our platform directly.

• 25:12Speaker 02

  And, you know, that's the cheapest thing to do. And you can be very creative and then you combine different AI services together. On the other hand, if you're a large corporation, you may just want your very polished product that serves exactly the functions that you need that integrates with all your databases that that has someone maintaining the product for you. And then the backend AI may come from the decentralized AI network, but you don't even wanna think about it. You just wanna think about the the product that you're buying. So we're we've spun off a company called Singularity Studio, which is a for-profit company aimed just at building

• 25:57Speaker 02

  commercial applications for large corporations where the AI backend draws on the decentralized Singularity AI network. And I sort of I view this as a you know, the best of both worlds. It's somewhat analogous to you have Linux, which is free and open source. Then you have Red Hat, which was doing, you know, an enterprise Linux and it was a for-profit company recently acquired by IBM for like thirty four billion dollars. Or you have Ethereum, then you have Consensus, which is is sort of a venture studio building out applications on top of Ethereum. So we have Singularity Net, which is our free and open platform with some AGI,

• 26:39Speaker 02

  R and D in it and some useful low-level AI services. Then we have Singularity Studio, which is is building, you know, polished corporate appli- applications on on on top of the Singularity Net.

• 26:56Speaker 02

  Network and I think that can help with the demand side of the problem of platform community growth that we face, right? Because with with SingularityNet, I mean, having launched the beta version two, which is a scalable version of the SingularityNet platform, you know, it it it works, it allows us to to deliver AI services at at scale, it allows different AIs to talk to each other and coordinate with each other. But for such a thing to really flourish, you need supply, which is people putting AI into the network, and demand, which is users using the network to get their to get their AI done. And I mean,

• 27:37Speaker 02

  the token ecosystem with AGI token, this can be used to drive supply to a significant extent by giving token rewards to AI developers, putting AI in the platform. To drive demand, I mean, one way to do it is just to do do developer workshops and make slip developer tools and you only get AI developers to use the platform. Another way to do it is to do something like Singularity Studio where you're building heavily polished domain-specific products aimed at big companies to get them get them using the platform. And I mean, we've been, as as one example, within SingularityNet, we've been working

• 28:17Speaker 02

  closely with with Cisco Systems and we we've been working with them on some applications aimed at car traffic analytics. So using deep neural nets together with some abductive logic, like logical reasoning, a different kind of AI. We've been putting logic together with neural nets and working working with Cisco to better understand what's happening in street scenes, like between cars, pedestrians, bicyclists and and so forth. So this is a case where we are working directly with a big company to help solve some of

• 28:59Speaker 02

  their problems using the AI tools that we have developed, which are then made available on SingularityNet. Marketplace and right now it's been a sort of custom thing and working on together with Cisco. But then AI on Singularity Net platform for car traffic analytics. A next step will be to productize that. And then we have a product which we could potentially license to Cisco or to anyone else who needed needed car traffic analytics. But then like the what the customer sees is just what we have some car traffic data feeds. And we need some output telling us what's happening in this roadway situation.

• 29:42Speaker 02

  And that's all they care about is they have something that sucks in their car traffic data. It spits out car traffic analytics. It happens on the back end, the AI is being supplied from this decentralized blockchain based platform. Using some neural nets and using our open cog AI tools. But the customer, I mean, they may think it's cool that there's this blockchain based decentralized network and some AGI technology behind it. But ultimately what they care about is they're getting their their problem solved, right? And so I I think that that's an example of the type of approach we're taking in Singularity Studio. Where we, you know, we're taking the trouble to develop a pretty full feature vertical market

• 30:25Speaker 02

  specific product for some customer that backends on the network. But you also want like a vast teeming pool of, you know, more small scale developers who are directly using the low level AI code provided on the platform to solve their problems. And and these developers may be putting their own stuff available for sale on the platform also. So, yeah, I think, you know, the stage right here is a stage. A bunch of blockchain based projects are going to be out in the next couple of years. It's like, okay, we built some infrastructure. We built some platforms. We've actually worked through a bunch of technical headaches to make tools that are decentralized

• 31:10Speaker 02

  in their infrastructure and actually do something useful now. You know, now we're, we're in the situation of competing with the big boys, right? Because we need to get, we need to get developers to use our tools and need to get customers to use the stuff, the stuff on our platforms. So, which is, you know, it's a good problem to have, to have a scalable decentralized platform that works and be at the stage of growing the network of users. On the other hand, that's, it still is a challenge, right? It's just a next level of challenge.

• 31:43Anthony Pompliano?

  Skrt, skrt. Want to know who has the best URL? Crypto.com. That's right, Crypto.com. They're a crypto platform with one goal: motherf***ing mass adoption. That's why we're all here. We're trying to get crypto in every wallet. Crypto.com is helping people do that through buying, earning, lending, and card payment. Everything you could want at Crypto.com. Go help your boy out, tell him Pomp sent you. Download the app or visit Crypto.com. Pomp's got you, always. Ever wanted to get into mining and didn't know how? Don't worry, your boy Pomp's got you. Everybody got some electricity and wifi? All you gotta do is go to CoinMine.com, you buy a coin mine.

• 32:25Anthony Pompliano?

  It's like an Xbox or a Playstation that helps you turn your electricity into Bitcoin. That's right. You purchase it, it shows up at your doorstep, you pull it out of the box, you plug it in, connect to your wifi, five minutes or less, you're mining Bitcoin. All you have to do is control it from the mobile app they provide, and then you receive over-the-air updates that add new coins and new features on a consistent basis. Kind of like how Tesla does over-the-air updates and updates the car software. Just your update in your coin mine. Consumer mining made easy. That's right. Go to CoinMine.com, tell them Pomp sent you, and thank me later. One more word from our sponsor, BlockFi. Their new interest account allows you to securely deposit your Bitcoin or Ether at

• 33:07Anthony Pompliano?

  BlockFi and receive 6% annual interest paid monthly in cryptocurrency. This rate actually compounds, so you receive a 6.2% APY, which is very attractive given the alternatives. So you can actually take your Bitcoin, you can deposit it with BlockFi, and get paid an interest rate of 6% in return. Go check out BlockFi.com/Pomp. Again, BlockFi.com/Pomp to sign up and start earning interest on your crypto today.

• 33:37Anthony Pompliano?

  What do you think is kind of the end goal or end state of SingularityNet? When you look and you say, hey, five, 10 years from now, as you continue to build it, what does this end up looking like?

• 33:49Speaker 02

  The end goal is to foster the emergence of benevolent, superhuman general intelligence. The end goal is a beneficial singularity. That's why it's called SingularityNet. But obviously, there are a few intermediate sub-goals along the way. An intermediate sub-goal, which is still very big, is I want to see a substantial fraction of the AI on the planet running in decentralized networks that are controlled in a democratic and participatory way. So I mean, we don't need to monopolize all the AI on the planet, any more than Linux has monopolized all the operating systems on the planet.

• 34:34Speaker 02

  But I mean, if SingularityNet was as big for AI as Linux is for operating systems, then I think we would be quite successful and we would have, you know, we would have made it so that decentralized, participatory, secure, privacy-respecting systems are a big chunk of how AI happens on the planet Earth. And I mean, I think that's the right platform for the next stage, which is the transition from narrow AI to general intelligence and the emergence of benevolent, superintelligence. So then, you know, practical stuff like analyzing genetics data or helping a modeling

• 35:19Speaker 02

  agency do, you know, image transfer of faces from one scene to another, or helping Cisco with car traffic analytics, these very practical, nitty-gritty projects across many vertical markets, I mean, these are part of the process of getting... Decentralized AI to be a significant part of the picture in terms of how how the transition from narrow AI to AGI happens on the planet.

• 35:46Anthony Pompliano?

  Got it. And then in terms of the actual services and the AI functionality that you have on the marketplace, how are you creating, are you guys creating that in-house? Are you using kind of some sort of incentive tool to get others to create it and contribute? You know, where's that coming from if you consider that kind of the supply?

• 36:09Speaker 02

  So so far, the the tools on the beta version two marketplace, I mean, these are these are either in-house or they're developed by other groups that we're pretty close, close partners with. So I mean, we've used we've used token-based incentivization, but it's been more on a bespoke basis. Like, there's a, there's a company called Moza AI, which builds bioinformatics to standardize genetics data. And I mean, we've, we've worked with them and they put some genetics AI tools on, on the platform. But it's so it's been some in-house and some developed by close partners.

• 36:51Speaker 02

  And that really the services on there now, I mean, they're useful. They do, they do valuable things, but they're, they're mostly, you know, examples of what, of what the platform can do. And what we're going to be launching later this year, the a request for AI portal, which basically has a list of, hey, if you develop this, then we'll reward you with this many tokens. If you develop that, we'll reward you with this many tokens. And then, then we're inviting the community to post new requests for AI on, on, on that portal. Like, if you, if you want some AI developed, post a request here, you know, post a

• 37:32Speaker 02

  A criterion for what does someone have to do to to have their AI be good enough for you? And how many tokens as reward are you willing to give someone for for developing this? And then that that it's a little bit like cackle, but it but but it's it's more flexible in in what kinds of of of AI can can can be can be requested. So that's that's the next stage, which we're going to get to this fall, which which we're pretty pretty excited about. I mean, we've we're doing things in-house and with close partners because the development tools have been a bit raw. But now that the beta version two is much more mature and easy to use.

• 38:16Speaker 02

  So with the yeah, with the launch of the request for AI portal later this fall, we're gonna we're gonna make a big push to get third-party developers to put stuff on the platform. And then that, I mean, that's at the same time as working with Cisco, with P I C C, which is China's largest insurance company, which we have an M O U with. And we're working out the contract details now for some special work there. And that Domino's Pizza, who's working with us on optimizing pizza delivery in in in Malaysia and Singapore. So, I mean, on the one hand, getting corporate customers on board

• 38:58Speaker 02

  through Singularity Studio and through contracts with specific corporations, getting large corporates on. On the other hand, launching the request for AI portal this this fall to make this sort of systematic way to guide the token incentivized AI development. I mean, that that's building demand and supply on the network. And I mean, that's what we need to do to to really make this grow. So I mean, our our big goal in twenty eighteen and nineteen was building stuff because the current blockchain tools are a hard-finding pain to work with. So like getting getting a decentralized AI network that actually works is a very hard technical problem.

• 39:41Speaker 02

  So now our, I mean, there's still loads of more tech development to do. But our big goal for, you know, this fall and then twenty twenty is more on the community development side. So we want to come out of twenty twenty with a bunch of users and a bunch of third-party developers and a sort of flourishing. Flourishing ecosystem there and then that's where you have the conditions for really radical exponentially growth over the few years after that.

• 40:13Anthony Pompliano?

  For sure. One of the things that to me is so interesting here is these businesses can get very big, very fast because of the global nature of them. How do you think about, you know, kind of focusing on one geography or jurisdiction versus a global expansion from day one or the early days?

• 40:34Speaker 02

  It's gotta be global and we're very global in the first place. I mean, I'm sitting in Hong Kong. Our biggest AI team is in St. Petersburg, Russia. Our product development is between St. Petersburg and Bangalore for the core platform. We have a team in Brazil, a team in Ethiopia. And I think that's actually very important to the model overall because if you look at what we're trying to do with getting a lot of different AI into the platform, so the supply side, part of what will make our platform successful is that we can take contributions from brilliant AI guys, you know, in Russia, in Mongolia,

• 41:16Speaker 02

  in India, in Azerbaijan, in Ghana, in Honduras, wherever they are, right? So if you have a brilliant AI developer with a new idea for a new AI algorithm, you know, if you're sitting in Azerbaijan or Honduras or Ghana, you're probably not going to get a job at Google or Facebook or Tencent. I mean, you might, but not many will, right? On the other hand, you can take an AI that you designed and you can put it online in our platform and then you can make money for it. You can earn tokens for it and you can build a reputation for it. So the ability to draw on AI contributions on the supply side from developers all around the planet, that's extremely valuable and I mean, that's something that comes along

• 42:01Speaker 02

  with being a decentralized network that's not tied to a very specific, you know, corporate business model or jurisdiction. And I mean, so on the other hand...

• 42:17Speaker 02

  That naturally will lead to serving customers all around the world. Because I mean, if you have a bunch of AI developers in Azerbaijan who are contributing AI to the SingularityNET platform. I mean, of course, the AI they created could be used by large companies or startups anywhere on the planet. But there's going to be, there's going to be an incentive then for companies in that country to use some of the AI's, put in the platform by developers in that country. So I think something like this can be global by nature. And I mean, in the same way that Linux, Linux has become very global, even though the initial

• 42:59Speaker 02

  developers were located in mostly Northern Europe and US.

• 43:05Anthony Pompliano?

  Got it. Before I wrap up, I always do rapid fire set of questions. What do you think is the most important company in crypto other than your own?

• 43:13Speaker 02

  I've been very excited by the Toda ecosystem, T O D A. I think the Toda guys have solved the scaling problem that has plagued blockchain technologies. And we're working on an experimental port of SingularityNET off Ethereum onto Toda. So I mean, that's not launched yet, we're just playing with it. But what they have is very cool. Like they, they got rid of the replicated ledger and they found a clever way to do, you know, secure decentralized transactions of value. And so that's, that's very cool. I've been playing with that a lot lately.

• 43:51Anthony Pompliano?

  Got it. What's the one regulation you would change or improve if you could wave a magic wand?

• 43:58Speaker 02

  Oh, one, one regulation? Well, I mean, I mean, the regulations that are known to you most actually aren't about crypto. They're, they're, they're more about experimentation regarding.

• 44:15Speaker 02

  Biology, I mean, I would, I would, I would, I would love to see regulations changed so if any sane, non-retarded consenting adult wanted to try some medical therapy, they're, they're, they're allowed to do so. Like, that, that, that would, that would advance us a lot toward the, toward super longevity and conquering aging and death. You wouldn't have to have like, medical tourism and boutique, medical facilities on super yachts in, in, in the ocean and so forth. But I mean, in, in, in, in the crypto domain, I mean, clearly, clearly you want to see just free ability for anyone to experiment with

• 44:58Speaker 02

  altcoins, utility tokens, or radical new forms of value exchange that haven't been invested yet, invented yet. I mean, these shouldn't be, these shouldn't be regulated as, as securities. There should be freedom of anyone, anyone, whether they're a credit investor or not, to invent or to in, invest in or purchase or, or whatever, like any cryptographic entity they want. And, you know, the government, the government can protect against fraud and people lying and saying something is what it isn't. But if you clearly describe what something is, and someone else wants to pay their money for it, they should be allowed to do so. I don't see why the government needs to be that paternalistic. And yeah, probably the most offensive thing though is like in,

• 45:43Speaker 02

  in many countries like the US, if you're above a certain level of net worth or income, you're allowed to buy certain things. And if you're poor, you're not allowed to buy certain things. I mean, that, that's just, that's insanely ridiculous. Like, why, why should the government be so paternalistic toward anyone who's not rich enough? I mean, just because you're not rich doesn't mean you're mentally retarded. But I mean, it would be such a long list of ridiculous paternalistic laws that, that exist for historical reasons and are preserved through either stupidity or because they help to increase the advantage that, that the highly wealthy have, have over, over everyone else. So I'm,

• 46:25Speaker 02

  Yeah, I'm very excited, like Patrick Friedman, a friend of mine in California, has recently recently raised a venture fund for charter cities and seasteading and so forth. So I'm I'm I'm very I'm very psyched about that. I I hope I I hope we can see like hundreds of startup countries floating in international waters with with with much more complete freedom to experiment, both with with medicine and with different forms of social organization and and economic exchange. While we're at it, we should legalize all psychedelics so people can can trip out and expand their minds whenever and wherever they want. I mean, but but there's a there's a there's a long list. So once once the super AI takes

• 47:10Speaker 02

  over and and and becomes a benevolent ruler ruler on the planet, they can get rid of all these stupid laws. I'm really looking forward to it.

• 47:17Anthony Pompliano?

  Got it. And so what's the most important book you've ever read?

• 47:23Speaker 02

  Most important book I've ever read in my life. Well, probably the book that influenced me most was Thus Beg, Zarathustra by Friedrich Nietzsche, which my my oldest son is named Zarathustra, probably because I enjoy the the philosophy in there so much.

• 47:44Anthony Pompliano?

  Got it. And then aliens, believer, non-believer?

• 47:50Speaker 02

  Aliens? What what what kind of aliens if we're talking about here?

• 47:53Anthony Pompliano?

  Well, I've never seen one, but I ask everybody if they if they believe in aliens or some form of intelligent life outside of humans on Earth.

• 48:02Speaker 02

  Well, if you take if you take DMT, you'll encounter a lot of aliens, but these guys seem to live in a in a different dimension than we live in. So, I mean, if we're talking about aliens in this four dimensional space-time continuum, hiding in Area 51 or whatnot, I mean, I I I haven't met any yet, but I'm I'm I'm highly open to the possibility. I mean, in in general. My feeling is we understand far less about the universe we live in than we think we do. Like, I think about when my dog is sitting in the study where I work, watching me type on the computer. He just thinks I'm sitting there wiggling my fingers randomly and there's no, and guarding,

• 48:47Speaker 02

  guarding my desk or something. Like, he has no idea I'm controlling server farms elsewhere in the world or communicating with other people, right? So similarly, once our intelligence is upgraded by fusing our minds with superhuman AI systems, I mean, what, what aspects of reality are gonna become apparent to us that are completely opaque to us, given our stupid brains right now? I mean, aliens may be, may be the least of the weird things we discover.

• 49:17Anthony Pompliano?

  For sure. And then I end each podcast letting the guest ask me one question. What, uh, what question do you have for me?

• 49:24Speaker 02

  What do you see happening in, in the crypto ecosystem in, in, in, in the next couple of years? Do, do you think we're gonna see like a flourishing of, of, of, of that overtaking the world economy? Or, or, or, or do you think there's gonna be sort of a long, a long, uh, winter of sorts?

• 49:45Anthony Pompliano?

  Well, I, you know, look, for me, it's, um, I, I, I put this, uh, recently on Twitter. You know, take the financial applications. Uh, I tend to think more about automation, uh, automated finance rather than decentralized finance. And same thing with applications as well, right? Like, there are definitely a place for, uh, decentralized applications. But I tend to think more of it as automation, uh, applications. And all that means or what the nuance there is, some of the automation applications will be decentralized, but not all of them. But every decentralized application will be automated, right? And, and so why that's important is, um,

• 50:30Anthony Pompliano?

  When you think of a blockchain, in my opinion, it's simply triple entry accounting, right? And it can count a whole bunch of different things that you want it to. But it works in a way that ultimately will allow all assets of value, so stocks, bonds, currencies and commodities, to be digitized. And ultimately, when those assets are digitized, you will then see true automation, right? I joke all the time and say, the idea that we're having automated systems try to interact with electronic money, right, that has, you know, one or two day settlement times, or other non-natively digital asset, it's like

• 51:11Anthony Pompliano?

  trying to put a CD-ROM in a cassette player, right? It just doesn't really work. And so, you know, there will be a time where a lot of this stuff is important and kind of comes to fruition. Unfortunately, I think a lot of the decentralized applications, or just the ideas around the applications, they're probably a decade early, just like we saw on the internet, right? We saw people try streaming, we saw people try food delivery, etc. in the late 90s. That stuff really didn't happen about a decade ago.

• 51:44Speaker 02

  Well, yeah, what's interesting is, you know, the economy is, everything's advancing faster than it used to, right? So what took a decade between like 98 and 2008, that might take three years now or something, because I think, I think processes are happening faster. So maybe, maybe even if we are ahead of our time and that fundamental dynamic is there, maybe it's not 10 years, it's only a few years.

• 52:15Anthony Pompliano?

  For sure. Yeah, look, maybe saying 10 years is too literal, right? But it's just, it's before its time. And I definitely agree that things can grow faster and kind of trends might accelerate, given the fact that technology's pretty pervasive today versus, you know, in the late 90s, people were just kind of coming online.

• 52:40Speaker 02

  Yeah, it's amazing how fast things can rise and fall now compared to 10 or 20 years ago, right? So I mean, you can see, you can see... Potentially, you can see, just like we saw IBM and One Computers and Honeywell fell from dominance to being sidelined. I mean, you can see that Googles, Amazons, Facebooks, and Tencent's fall from dominance to being sidelined by something totally different in like a five-year period or something. I think the pace of change is gonna accelerate.

• 53:17Speaker 02

  And I think you're right that the current batch of blockchain technologies is, in some ways, premature and prototype-ish. I mean, that's why I got excited about Toda lately, because I look at that and it's like, well, this is the next generation. Like, actually, why did we need a distributed ledger at the base level of our secure decentralized protocol? You didn't. Like, you can do secure decentralized communication without a ledger. And then, if you need a ledger, you build it on top of that. So that's one example of like a totally different architecture than what's commonplace now. So yeah, I think

• 54:02Speaker 02

  there's definitely, there's another phase of experimentation with the core protocols and architectures before you get super-wide adoption. But that phase may go through faster than one expects, based on historical precedence, because the whole economy is just speeding up and then insane.

• 54:22Anthony Pompliano?

  For sure. I completely agree, man. Well, listen, I really appreciate you taking the time to do this. I've been fascinated with all sorts of aspects of artificial intelligence, and to talk to somebody who's been working at it so long is a real treat for me. And then, obviously, what you're doing in terms of applying that is pretty cool. So thanks so much, and we'll have to do it again in the future. Alright, yeah, thanks for that. Hey everyone, Popp here. If you liked this episode of Off the Chain and want to help us take crypto to the top of the Apple, Spotify, and other podcast charts, please do us a favor and rate, review, and subscribe. To review, simply go to the Off the Chain homepage, scroll

• 55:03Anthony Pompliano?

  down until you see the five blank stars. Taking 15 seconds to fill those stars in and leave a quick review goes a long way in helping us take the entire crypto ecosystem to the top of the charts. I appreciate you listening and see you next time on Off the Chain.