The following is a rough transcript which has not been revised by The Jim Rutt Show or Toufi Saliba. Please check with us before using any quotations from this transcript. Thank you.
Jim: Today’s guest is Toufi Saliba. Toufi is the co-founder and CEO of HyperCycle, where they’re working on the development of decentralized networks for AI communications. He’s also the global chair for internet protocols for AI security at IEEE, where he plays a crucial role in advancing AI standards and security. Toufi also co-authored the TODA/IP Protocol, which we’ll be talking a fair bit about. So welcome, Toufi, to the show.
Toufi: Thanks for having me, Jim. I’m excited to be here.
Jim: Yeah, this should be quite interesting. These are a lot of the topics that I’ve been following for a number of years at the area of decentralized internet and decentralized AI. For our guests who may not be familiar with this domain, maybe you could start off with an introduction to TODA/IP and the HyperCycle, just generally what they are. We’ll dig into it a lot more detail.
Toufi: Certainly. TODA/IP is a network communication protocol that initially there was no intent to actually have a protocol. We were a bunch of technologists that were working on solving some of the problems to eliminate the dependency of a third party in a communication and then one thing leads to another and then you find yourself with a protocol that seems to appeal to a lot of developers. And then we decided to call it TODA/IP protocol. So the evolution of even the name was quite funny. We like that’s how the intent was not to have a protocol. It was a placeholder. I had hired a CTO at the time. I called him the rocket scientist in the house because he was an ex-NASA scientist and his name is Dan and my name is Toufi. We needed a placeholder where the T-D was taken, so put T-O-D-A, Toufi there. And then the Israelis in the office, they’re like, what are you guys hiding under TODA? Because apparently in Hebrew, it means thank you.
My parents are Lebanese from the Lebanese Mountains. And people would think that it’s like, oh, well Lebanese Israelis, they don’t like each other. And when you’re in science, you like everybody. We’re like, you know what? Whatever we’re putting under TODA, we’re going to call it TODA. There was many, many reasons why they decided to call it TODA. There’s a physicist named Morikatsu Toda from Japan that went unnoticed in history, but if you were to look him up, you would find a lot of phenomenal things that he’s done such as the Toda chain in 1981. In 1983 was published and there was a 1969 the Toda lattice. So anyway, the TODA/IP effectively enables communication between two devices of value. So what do we mean by value? You have, let’s say Jim and Toufi. Jim is doing some work when he’s finished the work that work has value. How can he send the value to Toufi via the wires?
Homo sapiens, we depended on a third party such as a bank or ledger based blockchain, but when it comes to AI, perhaps it doesn’t have the time to wait for that third party to come and say, we’re good to go. They need to be able to communicate effectively in a sub-second. So TODA/IP came in to be very handy in communication between AI devices where the value that is being generated is in the milliseconds. So when you have something that is being generated in 10 milliseconds, how can you send that within 10 milliseconds to the other device? TODA/IP enables that communication to be in a sub-second because it does not depend on a third party. So now, of course, any listener who’s a little bit technical would be like, that’s impossible.
Well, if it was not impossible, I wouldn’t be here, because it was impossible I was invited to be here. I was invited to lead the HyperCycle to do all of these things that I’m doing because at the time to 500 computer scientists in the room at the AGI Summit, to every single one of them, they knew that that is not possible in computer science. And if it was, they wouldn’t invite me. They would’ve done it.
Jim: Let me ask you a question here if you don’t mind. Doing peer-to-peer communications isn’t that hard, even at millisecond ranging using something like UDP, you might be able to get it down to 25 milliseconds, something like that. What does TODA add on top of something like UDP for peer-to-peer communications?
Toufi: It is communication of value. So what do we mean by that? So you have device A, trying to tell device B, can you help me? Such as, let’s say an agent is trying, let’s say device B translates Italian to English really well and A needs its help. And it’s like, I have this few pages in Italian. Can you translate them to English? Device B would say to device A, why would I help you? Device A would tell them, if you help me, I will pay you later. And then device B would be like, yeah, right. You’re going to pay me later. I’m going to do the work and you’re going to run away. That’s part of the dilemma. The other part of the dilemma, device A would feel like, well, nobody is helping me. How about I’ll start giving people penny first. Here is a penny, help me. The guy will take the penny and run away. Okay.
That’s how it is even between Homo sapiens, we’re talking about here, devices that can be anonymous, they don’t know each other, whatnot, or pseudonymous. Between Homo sapien, how did we solve this, Jim? We depended on this third party that is called the bank or a ledger-based blockchain or a government or a registry. When you tell the lawyer or the consultant, can you help me finish this document? We’re slow. The lawyer would be like, yeah, I would need two hours to finish this. So by that time, make sure that you pay my assistant. And then the way the time you go and pay the assistant, you pay it on Bitcoin, it takes 10 minutes, or on ledger-based blockchain, 10 seconds, whatever it is, it takes time, but we’re slow so we can’t handle that time.
If you were to go back to that AI agents talking to each other, they don’t have the time. If they don’t have a response within a 100 milliseconds, they moved on in life to something else. The time is not something that is available to them to wait for third party. Now some people might say, yeah, but there’s third party like Visa. They can transact a lot of things very fast. Well, the laws of physics tell us that there’s nothing travel faster than the speed of light, and the network is never going to travel faster than the speed of light, at least not the bits and bytes network that we know of. And therefore, if this communication is going to go to Visa and back to the other device, that alone is going to take more than 200 milliseconds each way back and forth. And then again, it’s not possible. They need it within that first hand check and then getting the results back. So the time constraints is big, the other is the cost constraint.
How much do you think an agent would be willing to pay for a single page to be translated when it has thousand other things that it needs to pay around to get its function done? Probably 0.01 cent. So if there’s a transaction at the entire value of the transaction, 0.01 cent, and the third party such as Visa or any ledger-based blockchain, the cost to operate that third party is more than 0.01 cent. No matter how they swing the fees, it’s not possible for it to be done. So then TODA/IP comes into play that you could actually eliminate all of that friction because you have the peer-to-peer, the friction of time, the friction of cost is eliminated because within the first handshake that the device is telling the other one, within the very first network packet, all the cryptographic proof that they’re needed for the other one to do the work, they are there. Then you do the work and you know for sure you’re getting your penny at the end of the work.
Jim: How do you do transactions, economic transactions for a fraction of a cent? Everything in the blockchain world is ridiculously fucking expensive these days. They charge you $5, $10 to do a simple USDC transaction. I think blockchain is entirely broken these days because of that. And of course VisaNet et cetera, you can’t get anything done for less than 25 cents basically. So how in the world does TODA enable transactions at even the one cent level, let alone 0,001 cent? How do you transfer value in very small quantities?
Toufi: When you have a true peer-to-peer, the effort, the cost, we always have to go to the laws of physics. What are the costs to operate a certain thing? If the cost is to operate a ledger is higher than when somebody is showing the fees or whatnot, then they’re lying to you about the fees for temporary, just so can I get you. Once they get you, then they will start jacking up the fees and we’ve seen that happen many times, but when by construct you have all the security by construct, they’re built inside the network packets. So whenever the handshake is actually happening, the effort that is being done is by the computer that is doing the reach and the other computer that is receiving. That’s about 99% of the effort. The other 1% is very minimal to the address of the network to do the witnessing.
And the witnessing happens whether they like it or not because all they see is a bunch of hashes they have to do the CONCAT on them and do the witnessing on them, which is explained to a certain level of details in the TODA/IP that hasn’t changed in about four or five years now. But further details in the HyperCycle core white paper where we explain of how something that we call it the miracleizer, which is like the miracle tree that happens in a fully decentralized setting that can actually run many threads parallel to each other and each and every one of them has its own block of time, but then they all end up concating and then they form a single hash that is global. So that is extremely important. How can you have one device reach out to the other device and the other one knows instantaneously that you’re not lying?
Because before you’ve even reached out to the other device, the receiver has a sequence of 32 characters from the beginning of time of that network. And those 32 characters, they keep adding to each other every 10 seconds, 15 seconds or whatnot. What those 32 seconds have, they give the ability to the receiver to check what you are sending if you actually own it, if you are the only owner of this thing. And the second thing that they’re able to do is that they want to make sure that there is no possibility for a double spend and so on and so forth, that’s taken care with the TODA/IP.
So because of the TODA/IP uses a data structure that it’s very definitive and very precise, it’s not possible for the faker to utilize another data structure number that it’s been utilized by someone else at the time that it’s actually utilizing it, because at the time is so important, the time in that space is what actually gives it the uniqueness. Then the receiver is able to verify instantaneously and because you’re able to verify instantaneously, you’re able to receive that transaction instantaneously. So therefore 99% of the cost is that compute that you run on each side. So you can send 0,000001 cent and you’re still able to have the economical model around it.
Jim: How do you actually settle that? So let’s say I have my little service that’s translating Italian and I can do a transaction because the computation is very inexpensive, but then how do I actually settle that? How do I convert the token or whatever it is I get during the transaction to something I can use elsewhere, USDC or Bitcoin or something, or do you have your own mechanism? How does the actual settlement occur?
Toufi: So everything that we do is US dollars. Some people they want to put Bitcoin on HyperCycle. HyperCycle is a container. So what the container is, you can imagine that this is the container, it’s called HYPC. Currently, HYPC in order to… The only thing we use ledger-based blockchains is to an on onboard. We have two types of onboarding. You get the onboarding of HYPC that is called the container that you can put the US dollars inside it. This HYPC traversal network, it’s less than 63 kilobytes. So what it has, it has all the cryptographic proof from point A to point B says, here’s I’m signing this 0,001 cent, I’m signing it to you. Here’s my cryptographic signature that has been signed to you under the one condition that you have completed. The work that I have expressed in those 63 kilobytes, I have expressed that there’s this work that needs to be done.
The moment you have a proof that you have done that work, the proof that you have done that work, again, it goes through the proof of work but not the proof of work that it’s hash-cash proof of work. This is a proof of work that is being requested by the sender. Then you have your legitimate to re-spend that 0.01 cent. Great. So instantaneously you can receive the penny and start the work. If you want to re-spend that penny, there needs to be that thing that we said that it’s global. So you need to wait 10 seconds, 15 seconds, depending where you want to settle it. If you’re settling it, whatever you want to settle it, you want to wait until that global hash is propagated.
And that means when you receive that and you’ve done that proof of work that you’ve done it, now you know it’s yours, you can go to the next one, but you need to take that, hash it and put it onto the network, which CONCAT with another hash and CONCAT with another hash and form that partake in forming the next global hash. And that is extremely important, Jim, because when it partakes in the creation of the next global hash anywhere you want to spend it in the world afterwards, you have participated in the creation of the hash that exists on their devices. So you have the proof that you own it.
Jim: Okay, a couple questions. Who actually holds the dollars? I get all these hashes. What good does it do me if I can’t get access to the actual dollars?
Toufi: Well, the hashes, they are the proof that you own the dollar, whether it is many pennies that make that dollar or a single transaction. So let’s say in one transaction you’ve received $1 and you’ve done the work, because you have submitted that hash and right now every single device on the planet that is part of this network has that 32 characters that you partake in a creation. So remember every single hash makes a difference in that final hash. You now from now until the eternity, you have the proof that you own the dollar, which means from the eternity you can go and spend it anywhere, it’s your dollar. And every one of those devices would know that that is your $1.
So you can take it to let’s say the Lightning Networks through the Bitcoin and gets converted into Bitcoin. You can take it through a ledger-based blockchain that is USDC. If you want to take it back from the system, you have to traverse the existing payment channels that exist for the humans. But if that dollar stays with the AI, it never going to consume all the transaction pieces stays within the AI ecosystem to pay for other AI services.
Jim: I suppose I do want to cash these out.
Toufi: Yes.
Jim: I want to file this chain. So you have this container for the dollars.
Toufi: Yes, yes.
Jim: Who actually has custody of those dollars? So we know that someone didn’t just make a promise they can’t fulfill.
Toufi: So there’s a lot of, inside the HyperCycle computation node, there are virtual machine that is built to handle Python and an EVM because most people in the blockchain space, they do this thing that’s called Ethereum the virtual machine, and there’s a Python which covers 99% of pretty much all, any AI developer they know Python. So that’s one part of it. That’s the thing that talks to the outside of the outside world. Inside that there’s something called AIM. Inside AIM, any AI developer, they’ve seen you bought the machine, it looks exactly like you bought the machine. And then on the other side there’s something, it’s called a TM. The TM is the transaction machine. The transaction machine, if it’s running TODA/IP, it can also run Lightning Network, it can run payment through Bitcoin, it can run whatever you want in the transaction. It can run Visa if you want to.
But if you were to go to the example of actually it is running TODA/IP. When it’s running TODA/IP, you’ll have, and I’m not sure if you saw it’s actually entire world is coming in converging with us. Even Coinbase they published a few days ago, they didn’t have a TM for HyperCycle. So they have a transaction machine that they can actually do USDC. That’s part of the HyperCycle computation node and that’s Coinbase. That’s a company way bigger than us. We’re still tiny, but they’re building those components because of the importance of having many different transaction machine for the AI is extremely important. People there realizing 90% of transactions in future are going to be AI. So back to answer your question.
This container is called HYPC. It’s sitting part of your TM. So somebody comes in and says, so let’s say you own the node. By owning the node, you bought the node, let’s say on a ledger-based blockchain. You bought it in private somewhere, wherever you bought it from, you have a proof, cryptographic proof that you own the node and that the cryptographic proof, let’s say it is via ERC-721. You own the node and then you want to put some money into the node because you want to put it to work. So let’s say you decide to put 10 USDC into it. When you put the 10 USDC, now the TM has the proof that you have taken 10 USDC from which network, let’s call it the ETH network you took it.
Now in the ETH network, everyone knows that there is 10 USDC that you’re going to put it in here. Okay. Now that you have that 10 USDC and you put it in circulation, every penny is gone around the world whatnot. Anyone who has that proof that they have the penny, they have actually the proof. It’s like you’ve given them the proof that you’ve paid part of that, that they can bring it back into the network if they choose to if you want to cash out, they don’t really need your permission. But, of course, that means that they’re cashing out. They have to pay the transaction fee whoever is cashing them out. If ETH at the time the transaction fees are $2, they have to pay the $2. That’s why most people, they wait until they have accumulated a lot of money in their nodes before they want to cash them out.
Jim: Okay, I’m getting at it now. So it’s very, very, very inexpensive and very quick micro-payments. But then you still have the issue that you have to deal with all the payment infrastructure if you actually do want to cash out.
Toufi: To deal with the Homo sapien, yes, with the Homo sapien.
Jim: This is very similar in some ways, but higher. We’ll talk about SingularityNET later. Let’s say I put my 10 USDC in the box and I spend a dollar of it with the Italian translation service. Where is that 10 USDC kept? What account is it kept in and what keeps me from stealing that $10 as the owner of the node so it’s no longer there to pay the little fractions I’ve given out? How’s that $10 USDC locked in and then debited when the time comes to cash it out?
Toufi: Okay. So the moment you have taken that 10 USDC and put it inside that container, you now have a valid proof if you were to spend again that 10 USDC that is yours. And the example that you gave, you took one USDC and you spent it for Italian translation. And again, when we say you here, we’re talking about AI nodes because this-
Jim: They don’t have to AI, they just have to be machines.
Toufi: Machines, yes. Because it is not possible for human to do. It’s actually we crippled it on purpose for Homo sapien to do the TODA/IP because when we enabled Homo sapien to do it, we realized that we’re facing two big enemies. The banks on one side, the ledger-based blockchain on one side and we’re like hell with it. Let them figure out that fight between them. We’re just going to power the AI to AI. So back to your example, you have 10 USDC that you have the proof that you have them. Everyone in the world, they know that you own that 10 USDC. You take one USDC and you spend it with another node. Now that other node has that one USDC, they can spend it again because they have the cryptographic proof that they have received the signature from you along with the proof of work that they have done, which was part of the request that you have sent them as part of the signatures. They have the legitimacy to re-spend that one USDC.
Therefore, if they want to go and pay any other AI machine, that proof is sufficient to go and pay one penny here, one penny there, one penny there. We’re good. As part of that, they’re probably collecting $1 from every single one who wants translation services and then they got to the point where they have $1,000 now in their wallet. Their wallet is also like containers, it’s called an HYPC. That HYPC has something unique to it. It actually has a unique number globally. It is the true meaning of the non-fungible token, not those non-fungible tokens that people do riff-raff stuff. It is an actual non-fungible token. Okay. And that has a unique number and they already saved about a thousand dollars into it of USDC. They decide to, you know what? I’m going to cash-out. I don’t want to do this AI thing. I’m the owner of this other node and I want to take it out.
The ownership of the node that you have that is near ERC-721, that ensures that you own that node, gives you the same ownership about that NFT, non-fungible token that you own that is under that. Everything inside it, you own it too. So now what? You take that and then you put it back into the network. Anyone that you show anywhere in the world that you own $1,000 of USDC, they have no doubt that you have that ownership because you’ve got the entire proof that you brought it back in. So the bridge that brings it back in is an actual smart contract that costs about a dollar every time you run it. That’s why the onboarding and off-boarding is going to cost you about a dollar, $2.
That smart contract has been audited twice to make sure that if you were to use it to put your USDC into the system or you want to take it out, you’re not really subject to a third party thinking, we don’t exist, we don’t want HyperCycle to exist in this. It’s a fully decentralized smart contract. And that is called CHYPC if you want to ever, if anybody’s curious about that smart contract, you can check it out and it’s effectively, that’s a bridge that gets you to bring anything of value into the system and out.
Jim: Okay, so I think you’ve answered my key question, which is my 10 USDC has been submitted into a smart contract and the smart contract administers the eventual payout when the time comes.
Toufi: The smart contract is that basically decentralized bridge that pretty much it keeps count for anything that goes in it. So if you put $10 in it and someone else put $10 whatnot, but there is a single person who’s taking a thousand dollars out, that smart contract knows that there’s already, thousand dollars went in and it gets sent in, $10,000 out.
Jim: And presumably it keeps track of the decrement of my $10 or somebody somewhere has got to know-
Toufi: Of course, of course, of course, they would.
Jim: … that I had $10, now I have $9 because somebody has cashed out. Well first they have to reserve $1, so I can’t spend it twice. And then when the service provider eventually does cash out, it actually has to decrement my pool by one USDC. Is that about correct?
Toufi: That’s why that bridges CHYPC. So what that bridge does, it takes the fungible to non-fungible, so that’s why it’s a little bit easier in Bitcoin. We like Bitcoin and Cardano in terms of those things. And some people they hate Cardano for whatever reason, community, whatnot. Honestly the technology itself, they have some very cool stuff because they did the EUTXO and they’re very close to Bitcoin. Bitcoin is a lot better than any other ledger based blockchains because you can actually split down to the single Satoshi, Jim. Many people, they probably haven’t told you that, but a single Satoshi in Bitcoin is actually an NFT, non-fungible token. So because of that, that bridge that I was talking about is a little bit easier to accomplish because what you have just said, who’s to say that that $10, nobody is re-spending it or whatnot? Well, because you can isolate it into non-fungible token in Bitcoin, everybody knows which $10 is that.
Now, when it comes to Ethereum world, you’re dealing with the fungible because it is basically an account based. That’s why it needs that bridge. So that bridge keeps account of everything that goes in and everything that goes out. It’s not possible to have less in and more out. So it’s make sure that if there’s $1,010 that went in, there’s only $1,010 goes out. So that $10 that you actually put in into the system is accounted for that you checked it out and put it towards deployment. We call it deployed as opposed to locked. And because you deployed it and now it’s working into the AI network, the $1 that you spent for it’s like you gave permission to one single node to bring it back into circulation on Ethereum if they choose to. They don’t have to, they can keep it on the AI internet network for forever, but if they choose to bring it back, they can. And now your proof to bring back into that network is only $9, no longer 10, because you lost one out of the 10. Yep.
Jim: Got you. And so the smart contract actually administers the final payout.
Toufi: That’s why it costs a dollar to bring it back in. Every time you run that smart contract costs about a dollar.
Jim: What I’m seeing here is this is a way to get out of that expensive and slow system, create something that can operate in fast and cheap land and then only occasionally does one take fast and cheap transactions back into slow and expensive ledger-based blockchain. Is that a fair summary of what you’re attempting to do here?
Toufi: For now until a lot of people that they can see a little bit more into the future and they see, it’s like once you’ve bridged into that land of fully decentralized peer-to-peer, then they feel like they can provide a ways to spend it over there for Homo sapiens don’t really need to ever bring it back to the expensive world.
Jim: This is what some people who have been talking about for a long time, is that Bitcoin is not really money. Really, Bitcoin is more like gold. So think about the Bitcoin as backing your money in some sense. And then so your “TODA money” is backed by this pool of gold, but nobody ever cashes the gold in. During the gold standard era, very, very infrequently did people cash their dollars or their pounds in for gold. Gold was mostly only used for balance of payments, payments between countries. Does that analogy actually work? Does that hold for you?
Toufi: Yeah, it could work. I just want to say one thing. There is no TODA money. There is no HYPC money. The TODA is more like a protocol, and the HYPC is a container, that container was talking about. You can buy it on Ethereum, but that is not actually money. The reason why you buy the Ethereum, because it’s easy to onboard, Ethereum is plugged into the financial system. So people, they say, I’ve got a hundred million dollars of machines, how can I turn those machines on in the internet of AI? Well, you can onboard them by getting some of those combination of HYPC and nodes, which we call the AI marriage. Then you’re plugged into the internet of AI. So it’s like what Cisco did to us Homo sapien gave us that secure plug into the internet. We do the same for the AI.
Jim: Yeah, again, I’m trying to get this landed so I fully understand what’s going on here. You say it’s not money and yet it is money. I can trade it for services. So it is a money within the world, the TODA world. Anyone who honors TODA is taking these hashes and they act as money because they are claims against-
Toufi: There is no such thing as the TODA money anywhere or not even in the TODA world. It’s a container that it’s called the TODA file, which is effectively that SATO server, that HYPC. The analogy here is more like we use is like the Brinks truck. The Brinks truck can actually, you could send me the Brinks truck itself as a payment, of course. It’s worth like $280,000 last I checked, or you could send $10,000 inside the Brinks truck to deliver it to me. The truck arrives, they open the door, they give me the money, and then the truck goes back. The container itself is the HYPC. Of course, some people could use it as payment, but it is meant to transport payments such as US dollars, Bitcoin, Ethereum, USDC. It seems USDC is the one that is likely to be the predominant here because people, they like US dollars and they like that it’s already got the cryptographic bridge into it and circled they’ve done a good job into doing that bridge, I guess.
Jim: Though, I’m going to go back to the analogy of paper money. Paper money during the hard backing period, the gold era, and then in the United States, quite late in the day, silver certificates actually stood for an ounce of silver. So they were not actually silver, but they stood for silver. And so this is your analogy of the Brinks truck. The paper money was like a Brinks truck for silver, an inexpensive way to move silver around, and in theory you could go cash it out for silver, but nobody ever did. So I would suggest that these toted containers actually are highly analogous to paper money in a metallic standard world.
Toufi: Yeah, I can see how that analogy could work. Although when we use that analogy, we realized that we’re better off using that money itself because of the numbers on the money. Because paper money is actually non-fungible by construct. It is the law that makes it fungible. Every money that you look at, it has that unique number on it, and that’s something that is very specific and essential is actually in data structure that we have with Toda. Every single container has a unique number globally and it uses the data structure called [inaudible 00:33:46] 64, to make sure that it’s not possible to have two devices in the same time block time to utilize the same number, it’s not possible.
And it’s very quickly by any device, they’re able to detect that because binary research trees, they’re very fast. And when you have a 64 label people that can check that fairly quickly, so we figured out a method to do it in a fully decentralized setting. So then when you have that Toda file and if you’re operating it in a block of time, no one else can be using that exact same Toda file. Impossible, not possible. So that’s the beauty of it.
Jim: Well, that’s also the same as paper money. One of the beauties of paper money is no two people can have the same serial number dollars. So again, I think the analogies might be stronger than you think.
Toufi: Yeah, sure.
Jim: By the way, for folks who want to know more about Cardano, I did a interview with Charles Hoskinson is one of the head honchos of Cardano back in Currents 027. So if you want to hear the Jim Rutt-Charles Hoskinson conversation, check that one out. All right, so now the next question. When I think about blockchain, anything has to do with blockchain. I think of expense, slowness, difficulty, a general pain in the fucking ass. And people are often bringing me not so much anymore, because I tell them to go to hell mostly, blockchain ideas and 98% of the time, but not a 100, 98% of the time, I go, why the hell are you doing this on blockchain? Blockchain just slows everything down, makes it expensive, et cetera. So what is your motivation here to build this TODA HyperCycle world?
Toufi: You do 98%, we do 99%, tell them to politely to go away. I really, really think that the reason why it is not a 100%, because that remaining 1%, it’s truly a phenomenal set of folks that they come to me with certain things that I really, really think that the entire world and should take what these guys are bringing into the rest of the world. For example, the evolution of zero-knowledge proof. Zero-knowledge proof started, I don’t know if you recall Michali, in ’93, Silvio Michali. Actually Silvio was my partner with something called Toda Algorand at the time, and we decided to do two different ways.
Silvio Michali, anyone in cryptography knows who Silvio Michali is. Michali in ’93 is the first zero-knowledge proof. And when the evolution of zero-knowledge proof, we started seeing it with the ZK-SNARK and STARK and all of that, and that all happened in the blockchain world and the rest of the world they use it, but I think they should give a little bit more credit to blockchain world and the crypto world from that, even if it is that 1%. So from my perspective, I’ve always said it that what we work on, for example, when it comes to people they call us anti-blockchainers, we’re not anti anything. We’re truly, we’re like the TODA/IP is a ledger-less blockchain. It’s not anti, it doesn’t have a ledger that says basically in every block of time, there are a lot of events that they happen. When they do happen, we found a clever way to get them to prove to each other they’re happening in that block of time and move on to the next block.
And that is blockchain, but without using ledger. But it’s completely different from what people refer to blockchain today that they refer to ledger as if it is the blockchain. So that’s from one perspective, from scientific perspective. Now, if you were to talk from business perspective, for example with HyperCycle, we don’t have anything on existing blockchain except we utilize them as methods the same way. Instead of using accounts payables, accounts receivable for example. The way we sell our software licenses, we use many different ledger-based blockchains for that because they’re super 15 seconds, they’re a lot better than 30 days with Microsoft to get the accounts payables, accounts receivable to buy software and license and all of that.
So it’s put in a smart contract that ERC-721. Inside that it has two ERC-1155. For those that they know what I’m talking about, they will know immediately where I’m going with it. Okay, you’ve built all of those licenses within this and then it’s easy-peasy. Then you don’t have to deal with a lot of paperwork and lot of you sent the payment, you receive the license and all that. It’s 15 seconds and it’s good to go. So there is some good use cases that they’re out there and it’s onboarding as well when it comes to financial elements if you want to onboard certain things, which actually brings me back to that example. Do you mind if we talk to the A and B, the two agents?
Jim: Do it, do it.
Toufi: Amazing. Okay, cool. So we had agent A trying to tell agent B, and for folks that they’re just tuning in, here we’re talking about multi-agent system, multi-agent system, those tiny little AIs that they can do thousands of those tiny little transactions every second between each other. So agent A telling agent B, can you help me translate this page? Of course agent B is receiving many other requests from many different places. So it’s only going to respond to a legitimate request that has a 0,001 cent attached to it. And it says if you had done the work, you get to keep that 0,001 cent. Great. Agent B is satisfied with the cryptographic proof, but agent B still has one question. The question is that who’s to say that what you’re giving me is not some made up money? Is this some monopoly money or is some real money?
Jim: Always the key question, is this real, but it’s based on how can I make sure that the guy on the other side hasn’t spent it twice or just stolen it?
Toufi: Exactly, exactly, exactly. So all of that, even if the guy on the right with A is able to provide the data structure approved, the TODA/IP proof that it is legit, it’s not double spent and all that, but the guy on the right B, agent B, they still need to know all the way the provenance of this, where it’s coming from. And this is where the provenance, if it’s coming from US dollars or if it’s coming from Bitcoin for people that they know cryptography and they know computer science really well, they can tell you that it is a lot easier if it’s coming from Bitcoin than if it’s coming from US dollars. And I’ll tell you why. If it is a US dollars, there is some entity somewhere, even if it is a Federal Reserve, is still an entity that needs to honor that dollar back at some point. So that entity, by the time we say at some point may not exist that point. Of course, if it’s a Federal Reserve, United States, it’s going to exist. Okay, strong military, all that, but-
Jim: Well, I don’t know about that. Look at the US debt level. That’s not sustainable long-term. The dollar may not be as good a value store long-term as people think that it is. That’s speculation. It’s hard to say.
Toufi: The point is scientifically it’s a maybe as opposed to certainty.
Jim: And on the other hand, of course, flip side of that, Bitcoin might go to zero.
Toufi: Might go to zero, that’s right. The value might go to zero, but-
Jim: You’ll still have your Satoshis, but they may not be worth anything. So those are the two different risks that you take.
Toufi: There is two different risks for sure, but what we’re talking about the actual receiver, is that receiver or the one the actual put in the money into the system, will they be there to honor? That’s what we’re talking about. Will they be there?
Jim: Correct. Correct. So Bitcoin, you do not have that risk.
Toufi: You got it.
Jim: Bitcoin will be there by definition, but it may not be worth anything while the US dollar, it’s possible there could be a revolution and the Federal Reserve is closed down and the whole thing just goes away and a new money system comes out. By the way, what happened in Germany, Germany, Austria and Hungary in the 1920s, both of them suffered ridiculous hyperinflation where people were taking wheel barrows full of cash to the bakery to buy a loaf of bread, and they basically terminated the old money, so it was worth zero and then restarted a new fresh monetary tree and all your bank accounts denominated in the old marks went to zero, which is interesting. And the new mark came out. And the other interesting thing about it was the economy restarted within a week. So that trick actually does work. So when people talk about debt jubilees and things like that, I say they’re more possible than you think. We could cancel the Federal Reserve and start a new US money and it would actually work.
Toufi: Those are all societal things. My point in the equation is that a mass is a lot more tangible things than an entity, no matter how big that entity is. So when we rely on mass, then it’s a good thing. So this is more of what inserts the money to the system. From our perspective there’s a lot of demand to put US dollars and through the USDC and other methods. We’re agnostic. Look, you can put, there is now new currency that they’re working on. They call it the AI currency with Ben Gortzel. You probably heard about that, the ASI. We would like to see them going through that bridge. So in order to enable any currency to be facilitated that way, there’s a little bit of technical work to get them to cross that bridge back and forth, but it’s not a lot of work. It’s already all been done and proven and audited and all that. Anyone with the currency that wants to use their currency to be that currency for AI, now they can use it. And that’s what we’ve done. We’ve done the facilitation for that basically.
Jim: Yeah, speaking of which, in the 24th of September, I’ll be talking to Trent McConaughey and Ben Gortzel about the AGI-ASI project. I’ve been advisors to both the Ocean Protocol and OpenCogs last-
Toufi: Hopefully by then the ASI would be one of the currencies they’re actually being onboarded into this because we know that there’s a lot of desire to use it as the AI money. And look, if you own your device and you want to sell compute, you’re free to sell whatever currency you want. And if you really want to do with the ASI then be it. We want to be facilitator for whoever wants to utilize it.
Jim: Again, to recap at least my take, and maybe this is oversimplified, but what TODA HyperCycle does is move from an expensive and slow environment of these traditional Ethereum based tokens to a much faster ledger-less system that’s more appropriate for micro payments.
Toufi: And that’s all thanks to putting security first, security, security. When you elevate the security tremendously by construct, then you gain all of those efficiencies and yeah, that’s precisely what it does.
Jim: All right, so you mentioned interoperation with AI, but in reality it could be any kind of technical service. It could be something as simple as do a matrix multiplication or something. Not that it would make any sense to do that, but you use AI as an example. Why are you focused so much on AI for this purpose?
Toufi: Okay, so when you have one device trying to tell another device, here’s a penny, you get to spend it so long as you have proof that you have done the work. So if you follow that, if that work is so simple that any human can do it or any non-sophisticated machine can do it like a tiny little calculator or something like that, then it’s a lot easier for things to, before they even think that they want to fake or whatnot, they can say, oh yeah, I’ve done the work, then you can do the work because it doesn’t really consume, doesn’t really do anything. It’s very, very simple to do. If that is the case, we actually cripple the system from doing those things. And the reason why we cripple the system from doing those things, if we didn’t cripple it, then we end up having enemies on both sides of the equation.
Anyone who’s running a ledger-based blockchain would see this as a threat to be like, what? I’ve worked so hard to enable this to transaction with my ledger and you’re going to come and tell me you got to do my people, peer-to-peer, they don’t really need me anymore. The banker will become your enemy. Same thing. Oh, what? We are the bank. We want people to come to us. Now you’re saying they can enable this whole thing. So we encountered that war very, very badly and we realized that the best way to win is to lose that battle and win different battle where it’s precisely for AI compute. So if the AI is actually running the compute, it is not some rudimentary compute that can be done and therefore it cannot be spent unless that compute is actually being done and that compute cannot be done by human, it requires some computation.
Then it continues to have that proof in that chain and continues to respand it accordingly. And we’re noticing a lot of success this way because then now we have both of those that could have been our enemies, our partners, because everyone in a ledger base blockchain, if you’re like, oh man, I’d like to send some business to these guys, and same with bankers, everybody feels like we’re everybody’s darling now as opposed to everybody’s enemies. It’s a lot better political existence, and I think we learned it the hard way, and I really think sometimes in the world there’s certain things that should have been taken in consideration where you learn not to disrupt certain system that is over saturated with a lot of folks that don’t want to be disrupted, and they have the capability to do a lot of damages.
So take a different turn and provide a new territory. The AI territory, nobody is there. It’s like Jim and Toufi, they’re conquering America. There’s nobody there. It’s like, go have your own land, do whatever you want there. So it’s a lot easier to go towards the new territory and provide value in that territory. And what we’ve noticed in that new territory, Jim, the most important thing that seems to be helping us taking off is that we’re introducing the wealth creation down to the node level. And do you know what I mean by that? This is actually a very, very important point.
Jim: Yeah, why don’t you tell us that? Lay that out for us.
Toufi: Amazing. Okay, so it turns out that 99% of business owners in the world today, and we’re talking about the 2024, they’re not aware of the wealth creation. 99%. So people, let me give you an example. Trent likes this thing that it’s called a coffee shop, and he goes and he wants to spend a million dollars on the franchise of coffee shop in Berlin and open this thing that is called the Starbucks. He spends a million dollars, the coffee shop used to generate, let’s say $600,000 in revenue and the cost is about 500,000 a year. So it used to have maybe a $100,000 profit, but Trent is so good that through his management skills, he turned it upside down and now he’s having 10,000 cups of coffees per day served instead of 1000 before, and he’s having about $900,000 in revenue instead of 600,000, and he reduced even the cost from 500,000 to 400 and all of this stuff. Okay.
The reputation also to that coffee shop. Jim goes to Berlin and looks it up and sees that it’s like 4.9 out of five. It used to be 3.2 before Trent bought it. Okay. It turns out that most people like Trent that they’re working their off, they think that right now they’re making 500,000 a year because that’s the Delta, 400,000, the cost, a 100,000 in revenue. But the reality in our civilization, there’s a lot more money being made because that coffee shop that they bought it a million and now it’s doing all of those great stuff, it’s now worth $3 million. There’s that $2 million they’re missing out on. They think they’ve only made the 500,000.
Jim: Well, they have a capital cost. It’s the difference between operating statement and balance sheet. You could say you have a balance sheet value of three million. So you do have recognize, but your operating statement is something else. That’s a temporal measure.
Toufi: Of course. But if they were to sell that business at that point, it’s a $3 million business.
Jim: Correct. That’s your balance sheet side essentially.
Toufi: So what we did for the AI nodes, because we were the network that connect to any node, so you have the node that running compute. So let’s say you’re going to talk to Ben. Ben has a lot of nodes, he wants to run them. Trent has a lot of nodes, he want to run them. So think of those as though more like the merchandise inside the coffee shop. You want to sell juice, sandwiches, coffee and all that. We provide that connector into that. That connector, we built it in such a way that can enable you to have that wealth creation belong to you. So what that means is that over time, that node that is running iCompute is going to appreciate in value. We want to make sure that the one who’s running the node learns about that appreciation and doesn’t let it to leak out.
They can actually absorb it as opposed to giving it away to AWS or Microsoft or whatnot. You can take that wealth creation yourself, so not the revenue only that matter, the wealth creation that you are partaking in, its creation belongs to you. You are doing it. So that’s the beauty that seems that it’s taken off quite nicely, which brought our revenue quite high at this point. We’ve got about over 500,000 node factories that we sold. So that’s what we do with HyperCycle. We sell those node factories. You buy a node factory, it’s like you buy a franchise of a Starbucks that you can create 1024 different Starbuckses. It is a finite number, but that gives you the ability to preserve all of that wealth creation as you do well with one coffee shop, you can have two, four, eight of 2024 and all of that belongs to you as opposed to belong to an entity that is called AWS or Starbucks.
Jim: Let’s go down one level deeper here. Let’s say I’m operating a node and I can operate software, I can provide AI services via software, or I can provide compute or both. Can I do both or do I have to do both?
Toufi: Yeah, of course, of course. Of course. It’s your node.
Jim: I suppose, could my node, the compute could be AWS. And most of my value is in the software, for instance. Or I could have something fairly, say a relatively simple software thing, but I could put it on my own hardware, highly optimized for that particular process and perhaps do a more economical job than AWS would be able to do on it. Is that another model?
Toufi: Yeah. We don’t really tap into that specifically if it’s better than AWS. What we tap into, whether you’re running it on AWS or your own hardware or you’re anywhere that you’re running it, the compute that you are providing is either going to be optimized and in demand and has some new updates, some new things or not. And if it is the case that it is the former that you’re actually, that compute is being in demand, in the eyes of other AI agents they see it, they regard it highly. They want to come to it to ask it for things, because if you go back to that example, it translates Italian to English with 99% fidelity. Okay. So it’s like almost every one who wants a translation, they’re going to come to Jim’s translator because it does it 99%.
So that agent over time is going to start getting more and more requests from every single AI in the world is going to come to that agent. Now, that agent can replicate itself on AWS, on hardware and everywhere, but the beauty is in the intellectual property that Jim created in that translation that belongs to Jim, and that growth, it started by being a $100,000 node, but then at some point it spread out to be in 1000 different locations. Now it’s a hundred million dollars, and all of that belongs to Jim.
Jim: This is good. This clarifies that we’re really not talking about compute, we’re talking about services. Frankly, I don’t care if a computer does it or have a whole bunch of Italians sitting in rooms someplace doing it. It’s going to be a little slow with a bunch of Italians, but it would still work. Think of the Searle’s Chinese room analogy. I don’t really care how I get the answer from the perspective of vending the transaction, though of course they’re going to care about timeliness, so a computer’s going to be faster though less accurate, perhaps than a human. So we really talking about not a market in compute, but a market in results essentially.
Toufi: Yep. It is very results oriented. You got it.
Jim: Okay. That’s good. I thought that had to be the case, but I thought I’d push on it a little bit to make sure. Now, I know you’re interested in AGI, as are many people we’ve talked to on the show. How do you think this helps facilitate AGI and in particular, AGI outside of the control of big companies?
Toufi: Look, I’ve been a serial technopreneur pretty much all my life. My success rate is about 30%, which means that 70% of the companies I work on they fail. In the eyes of my father I’m the two-third failure child. But in all seriousness, I know exactly what works, and this thing’s been working really, really well that I know, just the revenue alone, we use it to invest in 21 companies. So it’s getting up quite rapidly. So to answer your question, in addition to everything else that we’re doing, basically what we’re doing, anyone who’s running AI is going to be in a position to make a little bit more money every fractions of a second and one end and become smarter on another end, okay? Because what the network does.
Now, that’s great, but how is that related to AGI? Well, if that example that I’ve given you that the agent can actually ask another agent to translate Italian, could one agent ask another agent to do an OCR? Could one agent ask another agent to check a voice recognition, if this voice is actually Toufi? Could one agent ask another agent to do all kinds of stuff? Which leads to why someone like Ben Cortell is extremely interested into this partnership that created this HyperCycle, and because his interest is more, he wants to build the AGI. Okay. So I think having the AGI, are we going to get to the AGI by having a single entity get there? Any listener right now would be like, what are you guys talking about?
Well, we’re talking about that there’s a resource AGI people, whether they admit it or not, whether they call it AGI or not. There’s a point in time that this machine, it’s got to get to the point in time that you can ask the machine what does it take to get you to evolve yourself if that point in time, so let’s say you say, I want you to evolve yourself to be 10 times faster. The machine responds and say, I need 10 days and a billion dollars. I don’t need any Homo sapien. We call that AGI. So that doesn’t necessarily need to be some people they say human level intelligence. There is no blueprint for human level intelligence or whatnot. Well, it doesn’t have be human level intelligence. Airplanes, they don’t fly like birds, but they still fly longer and steadier than birds and all that.
Anyway, point is we could get to that point in time that we can ask the machine, what does it get you to evolve yourself and it just needs time and resources. You give it time and resources. And Jim, what happens if you were to ask that same question after let’s say 10 days, it comes back to I am 10 times smarter now. You ask the same question.
Jim: Yeah, that’s the course. The question about AGI know, is there a fast takeoff or is there not? And people like Kurzweil, are you that there is, people like Robin Hansen are more dubious about it.
Toufi: I am more with Kurzweil. If anything, I can tell you I’m willing to bet my life that we will be there before even Kurzweil’s prediction 2029, 2028, and I’ve said it to Kurzweil before. It is already almost there and if people they deny its existence, it’s there. It’s coming. What we do when wait until folks like Eric, the ex CEO for Google, he’s been predicting that we’re going to get that multi-agent system to bring it to the world and he’s scared shitless from us because he thinks we’re centralized, we’re not centralized. The beauty about a decentralized governance, Jim, is that the entire world will get to AGI first as opposed to single entity.
Jim: This is good. That’s funny. Where I sit, I see an awful lot of the work being where the dollars and the brains are right now. Not the smartest brains necessarily, but the numbers, is in LLMs. And I don’t see LLMs as being on… They’re a part of the road to AGI and a very useful part. When they solve a problem I thought would take a lot longer to solve, which is the language part, but they by themselves aren’t even close to AGI contrary to the bullshit that a lot of people in Silicon Valley are selling. What are you seeing going on that makes you think that we’re getting close?
Toufi: The actual LLM will not get us AGI on its own, definitely not. None of those components on their own will get us to AGI. If you were to go back to that definition that I have just outlined, we get to the point where we ask that machine, what does it get you to evolve yourself? And it responds that it doesn’t really need any Homo sapien, just needs time and resources. That is very fast approaching. The LLM can provide us with an interface to talk to it, but the reality, we really hope that it is not a single entity that is controlling it because that’s going to be very, very risky. We have not seen anything more secure than a decentralized AI, we haven’t. I’ve been at the ACM 10 years before, IEEE and we have not seen a more secure proposal than decentralized AI. So when you have decentralized AGI, then the entire world owns that AGI and not a single entity.
So it’s not going to respond for a single entity and saying, I’m going to do something that it’s not necessarily good. From governance perspective, I think that decentralized governance can come in very handy and it isn’t something that is democratic. Those are two different things, decentralized and democratic. The beauty of autonomous decentralized governance is that you can’t have a single side pulling it into its own direction. It is the entire global economy that gets it to be to that point. So if you were to follow through that example that I’ve given earlier, although it is very small translating Italian, but if you were to think of let’s say there is one component that is doing phenomenal job at interpreting images that we see like any camera into actual 3D objects with 99.9999% actually that engine, if that engine is owned by the Chinese while the other engine that does the actual perceptual things is owned by the Americans, and then you have the actual translators owned by the Germans and whatnot, it is by construct decentralized in a way that you’re not going to have one entity to control the whole thing.
It’s like the entire body. There’s the fingers, there’s the brain, there’s the front cortex, and there’s all of these things. So I feel like we were a lot safer than most people think. I used to talk about the threat 10 years ago when people didn’t pay attention, but I think doing something about the threat is what we’re working on. And Ben works with us too, as you know. We feel a lot closer to have a safer AGI than rest of world know.
Jim: Yeah, certainly Ben, and I’ve been chatting with Ben since 2014 about these issues. In fact, he was my third guest on my podcast where we talked about SingularityNET explicitly five years ago, and I’ve been following this and I know Ben has huge hope that decentralized AGI gets there first, but in my conversations with him, he’s also admitted there’s no guarantee that decentralized AGI happens before Google AGI or Chinese Red Army AGI or NSA AGI. What is your thinking about that makes you optimistic that the decentralized version of AGI will get there first?
Toufi: Because I’m working on the network, so I know exactly how fast that can progress versus the actual single entity. So Eric Schmidt, when he was actually afraid of the multi-agent system, when they come to life, they are here. He doesn’t know they exist. That’s the beauty of where we are and it’s likely that he finds out by the time it is fully decentralized, which is great as well, and I think he would be very good proponent to see the safe AGI as opposed to the one that he currently fears. A lot of people, they fear the unknown, even more people fear the known, Jim. If you know how dangerous decentralized AGI can be, you’d be frightened and you should be. So it’s not like being afraid of the unknown, it’s being afraid of the known that we’re confident that we’re going to get to the point where all of us, we get to AGI. For example, I give you a very anecdotal example.
So if a central entity gets to AGI, what that central entity think is ethical, it’s not what you think is ethical, it’s not what I think is ethical. They might think it’s like we need to eliminate people with beards. And when you have something that is a trillion times more intelligent than all humanity combined, you’ll find a lot of ways to eliminate people with beards. Okay. So maybe I would like to keep my beard, and if that’s the case, then imagine the entire world gets to AGI first and you go to your AGI and say, let’s eliminate people with beards. How are you going to convince it? We end up using our weakness, Homo sapien, to be our strength because we cannot get two people to even agree what the word ethics mean. Then that will become our strength.
We’re not going to use some of the weaknesses that we’ve got, which is… So I became a lot more optimistic the more I get closer to the reality of building that incentive down as close as possible to every single AI developer. When I’ve witnessed it, saw it in my own eyes that now it’s adding value to every single AI developer to be part of this network, then I felt like that is the actual Trojan Horse towards the decentralized AI.
Jim: Cool. That’s very optimistic.
Toufi: If you are a Microsoft partner right now, you’re likely to have a HyperCycle computation node because you are going to operate a lot faster than your competitor who’s a Microsoft partner. And that is unprecedented because before that we used to think, Ben and I used to think that there’s us, decentralized world, versus the centralized people, they’re the enemies. I don’t think like that anymore. I think Microsoft is 160,000 nodes. I think Facebook is 180,000 nodes. Great, they’re nodes into the decentralized AI and that’s a lot better to get them in as opposed to treat them as enemies because they’re incentivized by making more money. So we bring them more money as being part of that decentralized AI.
Jim: Very interesting. How does somebody get involved with, I suppose some smart kid who has some idea for an AGI or an AI service wants to hook up to the HyperCycle network? What do we need to do?
Toufi: Oh, well, there’s two ways. One is that knowledge. The other one is the capital. The capital is very easy, people they can go and just buy nodes or whatnot. That’s easy-peasy. They can go to the, for example, there is a nodemarket.io, they sell a lot of nodes that whatnot, but that’s the easy part. You can go-
Jim: How much does a node cost? I’m just curious, what’s the node cost these-
Toufi: The node on its own is $160. People they buy 512 of them for 46,000, but then they decide to marry it to the HYPC. So the total would be about a thousand bucks for the node, but that’s what gives you the factory. So you have a node versus node factory. If you want a node factory where you can have a 1,024 nodes over the course of time, you marry it to that HYPC. The marriage gives you the ability to have all of those children. And we call it the AI marriage. So yeah, that’s the cost of the node. Now when it comes to, when you have the thousand dollars node, that means that thousand dollar node can become 1,024 nodes. And if you think about if each node generates $10 a month or whatnot, then you have $10,000 a month over the course of that five to 10 years.
So it is a business that people they get to own, but the most interesting thing that we find that it is not just into people buying the nodes, is what it does to somebody who’s running AI, that it gives you the ability to start making a little bit more money and get smarter, and that’s what’s really driving the entire force forward. It’s phenomenal drive that gets this world. And we live in this world, we can’t deny it. We have to do things that they can work very well with what works in this world, so that’s why we optimize it. Many people, they tell me, Toufi, you hacked into Morlock and I’m like that’s right.
Jim: Exactly. Using Morlock’s own incentives to, or maybe overthrow Morlock. We actually had a conversation yesterday with Peter Wang about exactly this question.
Toufi: Very nice.
Jim: It was a private conversation, but it was interesting. Now I will say then there isn’t a hell of a lot of documentation out there about either HyperCycle or TODA. I couldn’t even find any language bindings to TODA. Where does somebody go to become a HyperCycle TODA hacker?
Toufi: HyperCycle.ai, there’s two white papers. There’s the one white paper that general that has a lot of things about a lot of ecosystem stuff and then there’s the core. The core gives enough to almost any developer to build on a nodes. But at the same time, this is the second part that when you ask, will this thing thing get people to be involved? The second part is the knowledge. People end up learning how to run nodes. They are in very high demand today because most people that they buy nodes, they’re lazy, they don’t want to run them, but those that want to run them, they have the ability to actually partake in that wealth creation. And we also have published a lot of hackathons and people can follow through the hackathon steps. In three days people, they were running their own nodes, that’s as of nine months ago.
We went live last October, so we’ve been in production since last October. We have this process called tailing when the nodes, they actually have a heartbeat. And in hackathon, Jim, in three days, kids, they have no idea about HyperCycle. By the end of the three days, they’re running their own nodes. So the knowledge is the most important aspect here. If people, they want to learn, they’re able to learn even following any of the previous hackathons that we’ve had, they follow the steps and they’re good to go. They’re operating nodes, and if they’re operating nodes, they’re unstoppable. They are going to be part of that future. They’re going to build that wealth in the future, and we want every single Homo sapien who wants to partake in that wealth creation to step in now as opposed to wait until it’s too late because we want all the smart ones to partake in this.
Jim: Yeah. Maybe I’ll dig into this and fire up a node. I got a few little ideas here and there. What languages do you people use to develop in HyperCycle?
Toufi: It’s all Python. ,The VMs, although we do support EVM, the Ethereum virtual machine and the way the nodes they talk to each other is via Python, basically.
Jim: That’s much better than a Goddamn Ethereum language. That thing sucks. Fucking annoying as shit. Python, much better. Okay, which case I will definitely, I’ve been looking for an excuse to do some Python development again, I haven’t done any in a while. I want to try out the new cursor, AI enhanced version of Visual Studio Code, and I’ve been looking for some excuse to fire it up with Python. So this sounds like something worth trying and I’ll see how hard it is. Now final topic to exit on, talk to us about how HyperCycle works with this, complimentary to or different or competitive with SingularityNET. And we’ve had Ben on here lots of times, so people who’ve listed the show regularly know at least a little bit about SingularityNET. Initially you might say, well, they’re the same thing, aren’t they? But they’re not.
Toufi: Completely not the same. We’re literally just focused on the network nodes. They’re focused on the actual OS and the actual AI node, and it’s very, very complimentary from that perspective. So what do we mean by the network node? Truly, Jim, think about it this way. The metaphor that we use, it seems to be applicable when people want to be like, I still don’t get the concept of this HyperCycle. Well, let’s think of this metaphor. Homo sapien in the nineties, they used to work in companies like McKinsey and they used to work in companies like JPMorgan. And those, they had intranet, intranet, but they did not have access to the internet. Once they got access to the internet and JP Morgan people, they started looking like they’re smarter to their bosses because they started asking the McKinsey folks, and the McKinsey folks are making more money.
Now instead of humans here, we’re saying agents, the AI agents, they live in a compartment today that they all just trust each other inside Microsoft, but they don’t trust any agent inside Google or in Tencent or anywhere in the world in Nicaragua. HyperCycle comes in and says, we are that security network that you don’t have to worry about anything so then any agent can talk to any agent. We’re just that network. Because of that, the abilities for having folks, whether it is SingularityNET or whether it is Microsoft or whether it is Google, they become all accessible to each other on a micro agent level, and that is unprecedented. And I really think that that’s just going to be a huge amplifier for everything that is out there. So yeah, that’s basically HyperCycle as a network. What SingularityNET is, is that it is the smart guy who’s sitting inside McKinsey and the other smart guy who’s inside McKinsey, and another, the directory of the guys inside JPMorgan, if you were to follow that analogy.
Jim: It’s an interoperating network for different clusters of AI that can call each other safely and get compensated securely.
Toufi: TCP or HTTP. Right now, if you want to play around with it, it’s all HTTP’s request. Everybody knows HTTP. You can talk to any agent via HTTP request. And TCP is that next version that is coming. You can have direct TCP to any agent and they will not fear you. That’s the beauty about the security that you have built, because otherwise they will block you. Right now, try to ping any agent inside Microsoft. They will say no, but if a single Microsoft partner that has HyperCycle will give you access to every single AI inside Microsoft, unstoppable. It’s an unstoppable force. Yeah.
Jim: Well, I really want to thank Toufi Saliba for an incredibly interesting conversation and just what I need, another damn rabbit hole to go down, but I’m very tempted to go spend some time and see how hard it is to fire up a cycle here. So really thank you for coming on the show and I look forward to continuing to watch what you guys are up to.
Toufi: Thank you, Jim. Anytime. Please hit me up on Twitter, on LinkedIn. I’m extremely very responsive, especially to developers with the ability to enhance this world to a better place. I’m a geek, I’m a developer, and I became a technopreneur and I’m here for all humankind. Thank you so much.