Hi everyone and welcome to another episode of Limited on Blockchain. Today we are speaking to Stan. Stan is the co-founder and CTO of ScaleLabs, a gas-free invisible blockchain network designed for secure Ethereum scaling.
He has been around for a while and he’s actually been around on both sides of the stream, be it academia as well as the industry. He holds a PhD in physics and he has been fairly lauded and given a lot of prestige in the academic world and he has been working as an entrepreneur for several years as well. He’s an active member and contributed to the Ethereum Research Foundation.
This was a very, very interesting conversation just owing to his experience and how experience can really change the way you build as an entrepreneur. So very insightful conversation. I can’t wait for you guys to hear this.
Let’s deep dive right in. Hi Stan, thank you so much for making the time to speak to me today. How are you doing? Good and let me introduce myself a little bit to the listeners.
So my name is Stan Kvatkot. I am the Chief Technology Officer and also co-founder of ScaleLabs. ScaleLabs is a network of blockchains.
I think we’re going to talk a little bit more about it today and I’m actually calling from Portugal, from Lisbon, Portugal. Awesome, awesome. Thank you so much for setting the tone right for this particular recording.
Now, could you tell us a little about your journey so far? How did you get into Web3, Stan? Yeah, that’s kind of an interesting. As everything is in life, it’s a little bit interesting and chaotic. So I’m actually a researcher.
So in my former life, I spent lots and lots of time in physics. I did my PhD in Germany at Max Planck Institute. Then I moved to America.
I was at Los Alamos National Lab in New Mexico for two years and then I was at Stanford University doing research in physics. My boss at Stanford was actually a Nobel Prize winner in physics. He did all of the cool things about quantum mechanics, magnets, you know, all of this stuff.
So I actually never thought I was going into crypto. But then what happened when I was at Stanford, I took a really interesting course from a professor. His name is Dan Bonet.
He is the world’s famous cryptographer. So I took a course from this professor and then this professor, he was actually starting a startup. The startup name was Ingram Network.
It was a database encryption company. Long before the crypto, it was just cryptography. The startup was encrypting databases.
So I basically joined the startup and since that time I was in crypto, in cryptography actually. I founded a startup which was doing cryptography for Wi-Fi. This startup was sold to a telecom company in the UK in London and then I started a cryptography lab which was doing cryptography analysis for the U.S. government, for the U.S. military actually.
And so I was actually doing this thing and then after several startups I was doing an AI startup and then the Ingram Network actually started. So that’s how I actually moved to cryptocurrency. Oh, wonderful.
That’s quite a journey and you have such a varied academic experience as well as with startups. So can you tell me a little from how did you go from an AI startup to crypto and then to scale? Yeah, that’s actually an interesting story. So we were running this AI startup and then the Ingram Network started and basically our venture capitalist told us to look into crypto.
I knew about Bitcoin but I was kind of not so much excited about Bitcoin. But then when I heard about Ethereum it was totally clear to me that was great because Ethereum added the function of actually running programs on blockchain. So with Ethereum you can actually run all kinds of things.
It was so exciting that actually I got a ticket to Cancun and I flew to DEFCON which was one of the early DEFCONs, met Vitalik Buterin, actually met all kinds of people and then started thinking how to do crypto. So then I actually met my co-founder Jack. He was also doing lots of crypto and interested in the new thing and we started thinking about what can we do.
So the idea was that at that time especially Ethereum was really slow and we started to think how we can make it faster. So we came up with this really simple idea that you can run sort of a network of blockchains which are compatible to Ethereum and are helping Ethereum. And these blockchains can be, you can have any number of them.
So you can have many, many blockchains and all of them can help Ethereum and help it scale and help it to have many, many transactions. So that was the idea pretty much and then we actually were really lucky. We got money from venture capitalists, some of the really top venture capitalists in the world and we started building this.
It was a long, long time building. Then we launched the network and now we’re actually running lots of games, running lots of AI applications. So it’s actually pretty successful as a project.
Wonderful. So scale is L2 and you know that is, it’s kind of powering the L2 infrastructure and it is also, I believe it is also like a ZK-Ethereum, right? And it is compatible. So can you tell us a little about what a developer can expect out of scale? Yeah, so basically the idea is that, you know, you have Ethereum, right? And Ethereum runs transactions, but you pay per transaction.
But then you have a game or you have an AI application and this AI application needs to run many, many transactions, way more than Ethereum. So what you do, you actually deploy this AI application or the game, you deploy it to one of the blockchains on Ethereum on scale network and the application runs. But then if you want to have a token or you want to communicate with Ethereum, you have a really fast bridge that connects a scale network to Ethereum and so you can go back and forth.
If you want to actually do something on Ethereum, you can move your token to Ethereum, do things and then you can move back. So that’s the architecture and that’s where we are now. As far as the future is concerned, we are thinking about ZK, but in addition to ZK, we’re thinking about some other really interesting mathematics.
So this year is a big thing for us because we are going to deploy all this new math to our blockchain. In particular, one really interesting mathematical subject is being able to have encrypted transactions. It comes from the point that you actually need to protect your information, basically.
So that’s kind of the big thing for us, this new encrypted transaction. It’s more privacy focused then. Yeah, so basically, I mean, if you look at Ethereum, for instance, or other networks, there are lots of good people, but there are also lots of bad people, right? Like you submit a transaction and someone actually sees the transaction, the information, and then they can front run the transaction, then can exploit.
And I think even during the crypto winter last year, there was about a billion dollars actually kind of stolen, you can think stolen or compromised on these blockchains because of these bad people. So with this new technology that we call Byte, blockchain integrated threshold encryption, what we do is when you submit your transaction, it’s encrypted, no one can see it, so you can actually trade or defy, you can use exchange, you can trade tokens. And then when the transaction is on blockchain, it’s actually decrypted and executed.
So it’s only encrypted during this short period of time when you submit. So it’s not really privacy, it’s actually protection against something which is called MEV or front running, where people use AI or bots to actually exploit you. Yeah, that I’ve heard of, yeah.
So that would be perhaps, you know, not something that you’ll have to worry about if you are a user. Yeah, totally. And that’s actually a big thing that we’re working on, because one of the reasons people don’t use decentralized applications is because of this MEV flash exploits.
So to give you a really simple example, right, let’s say you probably use Binance, right? Or Coinbase, you probably use one of those centralized exchanges. So when you trade, when you exchange tokens, you submit your transaction and no one sees it. Coinbase or Binance, they don’t show your information to other people.
That’s why you actually get a pretty good exchange rate on Coinbase or Binance. Now, if you use Uniswap, then when you submit, then there are these bad people that front run your transaction, exploit it. And the way it gets manifested is that you get a worse exchange rate.
That’s why, in particular, the volume on centralized exchanges is way higher. You know, you have Uniswap, but actually the volume is way higher on, for instance, Binance or Coinbase. And with this new technology, the Byte technology that we’re introducing, you’ll be able to trade on decentralized exchanges as secure as you do the centralized exchanges.
So decentralized will become much better, and hopefully you’ll actually get way higher volume and way higher use for these decentralized applications. Right. Yeah, that sounds very interesting.
You call yourself, you know, Steel calls itself basically a gas-free elastic blockchain network. Can you tell me a little about that? And then I would love to understand about the kind of technical challenges, perhaps, that have come in your journey so far and how did your team overcome them? But first, let’s talk a little about the gas-free elastic blockchain network aspect of it and shed some light on that. Totally.
So let’s just take a look at centralized systems, right? Let’s say you want to run a website or you want to run a web store or you want to run a blog or social network, right? The way people typically do it, they use a cloud, something which is called cloud. And the cloud is where you are able to buy computational resources, actually unlimited, in an unlimited elastic way, in the sense that if I want to run a social network, I will probably go and procure computational resources from AWS, from Amazon or from Microsoft or from Google. And initially, my social network may be simple and may not have so many transactions.
But as the network grows, I’ll be able to buy more and more computational power from Amazon or from Google or from Microsoft. And that’s why it’s elastic. It means that when you grow, you’re able to add and add computational power.
Now, the scale is pretty much the same. Let’s say you run a game on scale and what you do, you actually buy or rent a blockchain from scale. And let’s say initially your decentralized application doesn’t have so many transactions, so you just need a single chain.
Then let’s presume that you are super successful, right? And that you have more and more and more users. And at some point, a single blockchain is not going to be enough. So what you can do with scale, you can go and buy another blockchain and another blockchain.
And theoretically, a limited number, you pay scale network, scale network grows, and you are totally elastic in the sense that you don’t have to worry about computational resources. So that’s number one. And number two is that on this blockchain, the transactions are totally free for your users.
You essentially have guests. You probably know that Ethereum has guests and you pay in transactions for guests. But in scale, the guest is actually free.
So as a DApp, as a decentralized application, you have a limited amount of guests on your blockchain and you just give it to users for free. It’s definitely not free for you. Very much like, for instance, if you use Facebook as a user, it’s free, but Facebook actually pays AWS for the computational resources.
Very much the same model we have here, where we actually shift the burden from users to the people that run decentralized applications. They pay scale network, but for the users, it’s actually totally free. So when we say it’s zero fee and zero guests network, yes, zero fee and zero guests, but for the users, for applications, it’s not free.
But for users, it’s super great because as you mentioned, because it’s free for you as a user, you don’t have to have any tokens. You can pretty much just join and play the game, or you can join and use the social network, or you can join and you can use this decentralized application. Everything is free for you, which for users, it’s super convenient because it’s the UI.
If you don’t have to use tokens all the time, the UI may be way, way simpler and actually way similar to centralized applications, such as we’re used to use all of the things like Facebook and Amazon and Google and you name it. This is very, very interesting. I think that for users, it becomes a more seamless experience.
I think we keep talking about how we want to make it easy for the users. We don’t want the users to grapple with the complexities of the Web3 technology, but be able to experience Web3 and scale is providing exactly that. There is no gas.
They are used to a Web2 experience, which perhaps they would get if the application that they’re using is on scale. Absolutely. And you totally got the idea.
And it’s actually ultimately about the user experience because what you want to get, you want to get sleek UI. You want the users not to worry about things. And basically, then the perfect future is that you actually, the blockchain dissolves.
It becomes hidden. So you just use the application. It looks to you just as a regular centralized application, but behind the scenes, you actually have this blockchain, which makes things decentralized and nice.
Wonderful. So I’m sure that there were many challenges that you guys must have faced while building a layer two solution like scale. How did you overcome them? What were these challenges? Can you talk a little about that? I’m sure that our fellow builders would find a great deal of inspiration by all.
Yeah, totally. It’s a great question. So when we started, we basically took a scientific approach.
So when we started, there were so many blockchains, but then you would kind of try to read the white paper of a particular blockchain, and you couldn’t understand it. And so we kind of struggled initially. And you were reading white papers from different blockchains.
They were all different. And we were trying to understand the mathematics, and it wasn’t so easy. So we kind of decided to start from the scratch and understand the mathematics first, understand the computer science and make it simpler.
So what we did, basically, we took academia. We started reading research papers from academia. What we realized is that you actually use proper mathematics.
Things actually become way simpler. So then we said, okay, what we’re going to do, instead of inventing our own thing, we’re going to read all of these papers from scientists and try to compile something meaningful. First, it has to be meaningful.
It has to be something also fast, but it’s also provable, right? Like if you have mathematics, so you have computer science, a great thing is to have something provable, so you can show that the thing works. And if it’s mathematical, then you don’t have to care so much about bugs or crashes, because mathematics actually protects you against all of those things. So we started looking into computer scientists, and we realized that we can actually create a protocol, scale protocol, which is pluggable.
And it looks like a Lego game, where you have different pieces, you can combine the pieces, and then you have a working blockchain. But as you combine the pieces, you can have different algorithms from research. And when you use a different algorithm, you get a different property.
So you can get things faster, you can get them more secure, more stable, depending on what computer science algorithms you use. But whatever you do in this case, it actually remains provable, because you only use kind of the clean things, only the things which you’re proving by mathematicians and computer scientists. So we actually compiled this thing and built it.
And because it’s all based on research and academia, we’re able to use a way smaller team, and actually have way more bugs or crashes, because it’s all based on mathematics. So our network has been running now for about four years. We never had any crash, or any stall, or anything getting stuck, because it’s all mathematical and provable.
We are probably secure under the assumption of two thirds of our validator nodes running well. And so basically, the only thing we actually have to get is we have to make sure that the validators, two thirds of them are doing well, and then the network actually can run. And if in the worst case, there is something really bad happening, like a crash, for instance, on AWS, or some other cloud provider, if there is a huge crash, what’s going to happen, we’re going to actually pause until the crash completes.
And mathematics actually helps us to recover. So there’s a recovery mechanism, which is also probably secure, where if it crashed, the network crashed, and then it runs again, the blockchain will actually magically resurrect itself, so to speak, and actually again run. So the entire thing is then, because of all the things, we can sleep well, we don’t crash, we can develop slower, we can improve things.
And at the moment, because of all the things, we have really great performance. We are probably the most performant blockchain algorithm now running. We are run on the very low power machines in the cloud, so our validators don’t have to pay a lot.
You can run scale on a $200 per month machine, and it’s all because of mathematics and computer science. This is wonderful that you kind of went back and you perused academic case studies, and you basically derived your problem from there, and you went from there. I think that is a good way to build something that actually the users will end up using.
So now, when you’re talking about users, can you tell us a little about the traction that you’ve seen so far, and the kind of apps that are being built on scale? So we have about 2 million transactions per day currently. We have, I think, about 15 blockchains now running on scale, and the number is increasing. So we have over 2 million transactions per day.
Partially it comes from the fact that it’s free, so users can do lots of things. So we have 2 million transactions, but these transactions are also sometimes very complex. We’re running some of the most complex, probably, blockchain applications ever.
As an example, we are running a network, an AI network, which is really, really cool, I think. That’s one of the interesting applications of AI that really works. So the network is called Xorde, and what they do, they solve a really important problem where, let’s say, you have Twitter, or now X, right? And you want to analyze Twitter or X using AI.
Now, Twitter or X, they won’t actually allow you to download data. You try to download tweets, you download a thousand tweets, and they block you, essentially because you are centralized. But what Xorde does, it runs these millions, thousands and thousands of decentralized nodes, and all of them pull data from Twitter, from social networks, bit by bit, and each particular node just pulls a little, but because there are so many of them, there are thousands and thousands of nodes, they can actually analyze the entire Twitter.
As we speak, they actually, in real time, are able to pull this data from Twitter, and then what they do, once they have the data, you actually be able to analyze it using AI, the way you want. But the particular thing they do, they do things like social sentiment. For instance, you have a particular cryptography, crypto token, or you just have a regular stock like Microsoft or Google, and you want to analyze sentiment.
Maybe people like Tesla, or maybe people like Google, or maybe they don’t like it on a particular day. So when you get the sentiment using AI, you are able to plug it into trading, and then you are able to make a decision to buy or sell a particular stock or particular crypto based on this AI thing. So I think that’s one of the probably really interesting applications of AI and blockchain.
People talk a lot about AI and blockchain, it’s definitely interesting, but sometimes people will just use basically two words and say, okay, AI, blockchain, but then you have to find something really meaningful. It has to be some meaningful interaction, and I think what the guys at Exordio do is really meaningful. Another really great thing we have, we have lots and lots of games.
We are probably number one now in terms of games, like top games running on blockchain, exactly because of the UI. You can actually have way nicer UI in the game if you have zero fee transactions. So we have lots of games, we have lots of new games coming.
We have a partnership with Unity. I think we were the first project that actually had this partnership where there is a particular plugin to Unity. Unity is kind of an engine, you know, nowadays people don’t actually build games, they use Unity.
And then basically we have this Unity plugin, and then you can have all kinds of fun stuff and blockchain crypto stuff inside of the game. This is exciting, the fact that you guys are partnering with Unity is very exciting, I think, because especially, you know, it is a platform that has been around for so long, and I think games really do help, you know, with any kind of a layer to show traction, to see how it kind of can scale. So this is wonderful for you guys.
But now I want to, you know, play the devil’s advocate a little bit. Some people at times, I have also done that argument that, you know, becomes Ethereum’s dominant scaling solution. Do you think modular blockchain, like scale, can coexist? Or will one model dominate? You know, obviously, this is a more complex way of asking that so many L2s are getting created every other day.
How do you foresee the future? Like, how do you think this space will evolve, especially vis-a-vis L2s, and especially vis-a-vis modular blockchain? That’s a great question. I think I’m going to come a little bit from the history point. So let’s take a look at blockchain, right? So initially, there was Bitcoin, and it was kind of a really fun time, because everyone loved blockchain, everyone loved Bitcoin.
Then there was Ethereum, but then there was crypto winter. And during crypto winter, the growth of blockchain was a little bit not so great, right? I think one of the reasons was the governments, in particular, the US government, was actually, they were putting lots of restrictions on blockchain. And, you know, it was hard to start a game.
It was hard to start a DeFi application. Now things changed. Now you have a new administration in the US.
And, you know, whether you voted for Trump or not, or you like Trump or not, I think it’s definitely true that for crypto in particular, it’s going to be good, because the US government actually removed lots of restrictions. And because the US matters a lot in this world financially, you know, when the US removes restrictions from blockchain, all other countries are probably going to follow. So my prediction is for the next four years, what you’re going to have, you’re going to have lots and lots of growth.
And also, you’re going to have lots of new people coming to blockchain, because during the crypto winter, you know, especially in the US, there were engineers that were literally scared to do blockchain, and people were actually trying to leave and do something else. Now it’s totally different. So crypto winter, during the crypto winter, you had lots of experimental things, but they did not grow so much, right? You don’t really see people now, like in the subway, or if you’re on a bus, you don’t actually have people playing so much of blockchain games, for instance, right? So it’s still experimental.
The entire field of blockchain, many things are still experimental from the point of view of users. You don’t have so many users. But I hope during the next four years, because restrictions have been removed, we’re going to have lots and lots of users.
And it means that lots and lots means actually, you know, you’ll actually be in the subway, or in a bus, and people around you will play blockchain games. Now, if you want many, many people to use blockchain, let’s say you want a billion people to use blockchain, right? You need billions of transactions per day. And then if you look into AI, if you look into some meaningful things, for instance, AI to AI payments, right? Like if one AI agent pays to another AI agent, it could be trillions of transactions per day.
It could be things like micropayments, where one AI agent is using the computational power or the API of another AI agent, and maybe pays, you know, like one US cent, or maybe pays one hundredth of US cent. So that’s kind of the perfect future where people actually start using blockchain and actually moves from being experimental to being a real thing. Now, if this kind of perfect future actually works, then if you want to have billions and trillions of transactions per day, you want to have a network like scale because of the elasticity.
We actually can have millions of blockchains if people need it. With rollups, because rollups post on Ethereum mainnet, the bandwidth will still be limited. You rollups, you have basically base from Coinbase.
I think they have like 10 million transactions per day now. And these 10 million transactions, they actually take half of the Ethereum network because of the bandwidth. So with all of the rollups currently, you can only run 20 million a day.
20 million is not one billion and is not one trillion. So our hope is that with scale, you can scale way higher. And I think definitely AI to AI payments are going to be huge.
Also, games, you know, any technology, the first thing that happens is games, because gamers are pretty adaptive of new technologies. So games, AI to AI payments, things like Xorg that they mentioned, like analyzing things using AI. I think if we’re going to have these billions and trillions of transactions per day, then scale is going to actually grow really fast.
And what we see, actually, with all of these new things, all of the new excitement, we see lots of projects actually coming to scale. So not only we have this 2 million transactions per day, but the great thing that we’re starting to see in this year, in 2025, there are many people coming back to blockchain and starting new projects. So we actually have lots and lots of cool games, for instance, which are just being built, or venture capitalists are actually now starting to invest in early, early stage startups that are building this new AI to applications, for instance, for AI to AI payments.
So in general, I’m actually very optimistic. I think that, I mean, one thing which is kind of interesting about blockchain, usually new technologies actually grow really fast. But with blockchain, there was this period, as I mentioned, the crypto winter, which was kind of a staggered growth, and people kind of got back a little bit to being experimental again.
But I’m actually really optimistic about the future, because the governments, I think they won’t be able to go back. In a sense, they opened this floodgates, and I’m actually really hopeful. I think the way you’ve kind of put it is one of the better ways that I have heard this particular question being defended.
I think, like you mentioned, that there was a bit of a crypto winter, and now there are more kind of solutions bringing up. You’ve highlighted a few very good points that gaming is one of the easier, I think, fruits that can be on the path to scale into solutions. So those are all very, very valid points.
You did mention that, you know, you guys are, there are more AI applications. And I think my research tells me that, you know, you also have announced an AI inference on chain. So can you tell me a little about that? How does that work? Why is it a big deal for blockchain adoption in general? I mean, I can understand it, but perhaps in simpler words, it’ll be good for our listeners to get how you know you are really working on the crux of AI as well as that. Yeah that’s a fantastic point and AI inference I think is going to be one of the driving engines. I’m just going to tell you a little bit how it happens with Exordio because it actually runs you can actually use Exordio today to get all of the sentiment and they have a pretty nice website actually where you can actually try this AI on blockchain.
So basically the question is why do you want to actually have inference on blockchain. The reason is that if you run AI in a centralized matter it’s always tuned so to speak in some way. You know you can chat GPT and you ask some question it doesn’t answer the question or if it answers the question it answers in a particular way.
You have a competing thing like Grok from Twitter from Elon Musk right. You ask the same question from Grok and the answer sometimes will be different. If it’s a question about mathematics you know the answer is going to probably be the same but if it’s a kind of a hot political question or if it’s a question where there are several opinions you know chat GPT will give you one answer and Grok will give you another one.
But the point is that you don’t know how it has been tuned so it’s actually better for you to have diversity. It’s actually better for you to have open source AI where you go to a particular you know AI chat and you get a particular answer but at least you know. For instance maybe you’re politically left and you like left questions and you go to a left board or right or maybe on a particular subject.
Maybe you’re more like a scientist and you want a question about vaccines and you like vaccines. Maybe you don’t like vaccines but in general I think people have different opinions. So there’s a question about how you resolve different opinions for AI and probably the best way is just to have many diversity.
At least when you go to an open source agent you know how the agent has been trained and what’s kind of the bias of the agent. So that’s why I think on blockchain what you can do you can actually have many different things and having different opinions but you know how a particular thing works. So AI for blockchain actually in some way it means open source and it also means diversity.
You know there are different things, different opinions, at least you’ll know what how it’s going to answer. So the way AI inference is going to work on blockchain it actually works now on Xorder. The data goes to blockchain with scale and that’s a really big point.
You have to be able to store lots and lots of data on blockchain because with scale we have this elastic network. We’re able to store literally terabytes of data on blockchain and then the data is going to be on blockchain. The actual network, the neural network is also going to be on blockchain and it’s going to be many many different networks, open source networks.
You’ll be able to deploy your own network if you like a particular way to answer things. So you’ll be able to deploy it on blockchain and then AI agents need lots of computational power. So for inference you need lots of hardware so the agents are off the chain.
So with Xorder what we do, what they do on our chain is that basically the agents pull data from the blockchain. They do the inference and then they push the result back to the blockchain. Now there’s one particular technical question.
What happens if an agent becomes malicious and answers in the wrong way? Well this is actually easily solvable. You can have five agents answering the same question. If one of them becomes malicious and doesn’t use the right network then the four of them will do the good thing and the bad guy is going to be punished.
So there are ways actually to protect against malicious agents. But the future I think is going to be diversity of opinions, open source and blockchain. And people actually, I think they’re coming to this because people are like asking chatGPT questions and they’re like, why isn’t chatGPT answering this particular question? Or sometimes they don’t like the answer, right? These two things are going to be actually solved with blockchain.
Interesting. So you’ve kind of explained to me how AI inference kind of on chain works, but do you think that this is something that you’ve announced recently, right? So how soon can a user or a developer use this? So with Exorda you can actually use it today. You can actually go to their website, Exorda.
So this is one of the examples and you can actually use this thing. I believe there are actually, we have lots of our pipeline, we have lots of early stage AI startups. So I believe that you probably by the end of 2025 you will be able to use lots and lots of these things.
Definitely they have lots of development. The good thing about AI and blockchain is lots of venture capitalists actually like basically doing the thing. So there are lots of money now flowing into this particular subject.
People are actually working around the clock, I think, on all of the things. There will definitely be some kind of like a market like Darwinian extinction. There are a little bit too many projects, but I think overall the next year, 2025 is going to be huge for all of the things.
You actually have lots of things which I actually don’t like so much. You have things like AI agents, having some crypto tokens and trading. Actually, it’s a little bit of a hype application.
I believe when you have some new technology that you have always this kind of a bit of a hype thing, but they will probably subside. So the real use is, as I mentioned, in diversity, in computer science and being able to actually run multiple neural networks, because with ChargePT it’s just a single network, with Grok it’s a single network, but with blockchain it’s going to be thousands, maybe millions of different networks or neural networks. Okay, this is very interesting.
Actually, you kind of touch a little upon my next question that I was about to ask you, but I’ll ask you nevertheless. There’s basically, like you mentioned, there’s a lot of talk about AI-generated smart contracts, AI agents dealing with crypto. What is your take on it? How do you think this can scale? There is also certain specific group of people who seem to think that AI will replace developers in the blockchain space in terms of coding.
I don’t believe that. I think, like you mentioned, any network is just as good as the data it is being fed. So I don’t think that the human element can ever go away entirely.
I might be being very optimistic about this, but I would love to get your take. What do you think about AI-generated smart contracts, developers, AI, and in general, AI agents dealing with crypto? Absolutely. It’s a fun question.
I mentioned when I started that I was actually a researcher in physics. So I’m kind of a weird guy because I’m not a computer scientist, a physicist, and I have different opinions on things. So my boss at Stanford, he got a Nobel Prize in physics, and he had lots of opinions about computers, about AI, about quantum computers.
Basically, if you ask physicists, not computer scientists, you look at the human brain, there is this neural network, which is kind of similar to charge GPT. It analyzes, for instance, language. Now we’re talking, I understand you, you understand me.
It’s all this really fast neural network. But in addition to that, probably inside of the neurons, there are these quantum computers, which are really, really tiny. And at the moment, we don’t know so many things about quantum computers.
The only thing that we know about them is they can potentially run way faster, dramatically. They have dramatically higher computational capacity. So physicists, not computer scientists, physicists think that the human brain has a quantum computer inside, in addition to just a regular neural network.
So this quantum computer is able to do things like imagination or creativity. So with charge GPT, it can only answer questions that have been already answered. Charge GPT is basically a machine learning thing where there’s lots of data, and you ask a question, it just answers based on this data.
It does not have self-awareness, it does not think, and it’s not creative. One way to understand that is that actually asking a question is a harder thing than answering a question. So charge GPT is able to answer a question if you ask a question, but it is not able to generate its own creative questions, and it’s not able to create new things.
So I think the reality is that there’s a little bit of a hype about AI or machine learning. Actually, machine learning is a way better thing because AI would probably mean actually doing a quantum computer. So what we have now is not actually AI.
We have machine learning, we have lots of data. Based on this data, it actually answers questions. But this is actually very useful.
So my prediction is we’ll never, until we actually construct a quantum computer, we’ll not be able to have AI in terms of self-awareness and creativity. And this thing, physics is not a very highly funded science at the moment. So quantum computers, when they design them, I don’t know, maybe 10 years from now, 20 years from now.
But in the meantime, what they call AI now, the machine learning actually, is going to be incredibly useful. So AI agents will be saying to replace mundane things. For instance, when you answer emails, there are some mundane emails you answer.
You pay, like I pay for my cell phone, I pay for apartment, for my utility. It’s very painful. AI agents will be able to have your money and do things like really mundane, boring.
Whatever boring in life is going to be replaced. So you’re going to be running an AI agent on your computer, it will pay for your mobile network, it will pay for your utilities, for your water, for your heating, whatever. And this is also going to be great.
You’re also going to be having robots that will also be doing mundane things. They will make coffee for you. They’ll probably be able to do dishwashing or do prepare food.
All of the mundane things in the world are going to be replaced. And it’s going to be happening really fast. The impact of agents on the industry is going to be great.
I think AI hype coins, this is just like a really fast element. They’ll subside. There’s really no point in having them long term.
But AI agents doing mundane things are going to be super. And also, I mentioned, these agents will be able to pay other agents. So they’ll be able to have micropayments.
As an example, if you read the New York Times, you will be able to pay one cent per article. And it’s not like you will be paying, your AI agents will be paying one cent. So AI agents, if you give a little bit of money to an AI agent, let’s say you give your AI agent $20 per month, you’ll be able to do really, really small things that will make your life way better.
And then as they improve, you’ll be able to give them $100 per month. And basically, they will remove all of this boring stuff, like painful stuff we have in our lives. And blockchain is great because if you’re talking about micropayments, there is no way to do micropayments on centralized networks like Visa.
So with blockchain, you’ll be able to really, really pay for small things with AI agents. That’s why I actually, as I started, we’re really hopeful at scale that if these real things take off, then we’re going to be doing really well with AI. So AI is actually pretty useful, although I totally agree with some people that say there’s a little bit of a hype, but there’s always hype, right? Like any new technology, you have a lot of hype things initially.
Yeah, that is bound to happen. I think with any new tech, as you said, there’s going to be a lot of hype. And I completely agree with you.
I think AI will help us get rid of the mundane, and that will leave us with a lot more time to actually pay attention to what an individual really wants, and they can decide how they want to spend their time. But in terms of, I still feel a little curious, and I think this is more of a philosophical question than anything else. Because we always, when a technology is growing, and when it’s scaling, at times, there are folks who perhaps don’t see their potential, and they don’t see how it can perhaps change the way the existing status quo.
So do you think that you and I might be on that side of where we are not realizing the kind of potential these AI agents, the kind of power they might eventually exert in terms of tokens or micropayments, or just generating smart contracts? Do you feel that is going to happen, and it’s a possibility? I think it’s definitely going to happen. One thing which one has to be careful about is that, you know, we actually are looking into the future, where lots of things are going to be replaced by AI. And the question becomes, what happens to people, right? It’s a big societal question.
So if everything is going to be replaced by AI, all the mundane things. But the problem is that many people in the world, when they get paid for work, they do mundane things. So we need to start thinking about the society.
What will people do, right? We don’t want to have people which are going to be very lazy, and basically not developing their brain, and relying on just all of these AI things. So we need to build a model of the society where, yes, all the mundane tasks are going to be totally replaced. There’s no way to avoid this, probably like one, two, three years from now.
But what people are going to do is a big question. And I think that’s the other side of blockchain versus AI. Because blockchain can expel people to you, and other things like decentralized science, where you pay people tokens if they do science.
Science funding around the world is totally bad. You know, people don’t really invest in science. There are really important questions like cancer, for instance, biology, right? That unfortunately, the society does not really solve these problems.
You know, with cancer, for instance, for the last like 20 years, how much progress we had, not so much. So a perfect society would be if we somehow using blockchain, we are able to, maybe using tokens, whatever, crypto, AI agents, we’re able to pay people to do science and to solve problems that are really fundamental and really hard. So that would be a perfect society.
And then we see some of the things, you know, like pay-per-play games, where people play, just get paid to play games, or there are people that get paid to do fitness, you know, to run. I think it’s a good start. What we want to have, the perfect world would be where you are living in this new shiny world, and you get paid every time you improve yourself.
Maybe you do fitness, maybe you learn new things, maybe you get a new degree, a university degree, right? Then you get paid. Maybe do you participate in the research? But it’s a huge societal question. And with this every new technology, there will be a period of turbulence in the societies.
And I think the governments and democracies, they will have to make this solution because the kind of a dark future, we don’t want the dark future. It would be that everything is done by AI, people become very lazy. But another really bad thing which can happen is the aggression, right? We see things around the world where people like, you know, get paid, but then they go and get the war with other people, right? And the really bad thing would be that people that have lots of time would be engaging in wars.
We don’t want to have it. We want them to develop themselves. We want them to become more intellectual.
And I think, so the blockchain, I think, needs to mature. The problem is that also was that during crypto winter was kind of a negative selection, in the sense that people that cared a lot, they were not actually leaving the space while lots of hype and speculative things were happening. Hopefully, it’s going to change.
Hopefully, we are going to actually have having these new technologies where these new tokens improve the society. And it’s the question that’s very important. We cannot avoid the question because it’s an incredibly turbulent time for the humankind, right? And so we need to make sure that the future is good.
Right. I think this is a very interesting take. And I think it gives me and I’m sure it’ll give our listeners a lot to think about as well.
So thank you for sharing that particular perspective. And now I would like to ask you about, you know, the security and decentralization of scale, because of the way that kind of structure, what are you doing to ensure the long term security and decentralization as a domain ethos for this particular blockchain network? Oh, totally. It’s a great question.
And actually, you know, if you compare us against rollups, we are blockchain, so we have independent validators, we don’t actually control our blockchain in any way. One thing is that, you know, sometimes people don’t do open source. So you have a blockchain, and then you’re asking a question, how a particular thing works, or how a bridge works, or how a transaction works.
And some of the people that don’t really do the open source, so you don’t know. The great thing about scale is that everything we do is open source, all of our code is on GitHub, you can actually check every single thing, how scale works. And then we are trying to make SIG secure.
So one thing which is actually very different with scale is that we have hardware security. Intel processor has this hardware security protections called SVX. It’s pretty much like a small tiny processor inside of Intel chip, and it’s super protected.
So all of our crypto keys, all of our crypto operations are inside of this Intel SGX chip. And we are pretty much one, or there are only one or two other blockchains that do this. So for instance, if you want to hack scale, even if you hack into a scale computer, a scale node, you will still be able to get the crypto key, because the crypto key is inside of this protected Intel chip, inside of this SGX environment.
Because of this, one thing which was really good about scale, we were running like now for about four years, we have many, many blockchains. There was not a single accident with scale where we got hacked. And I think we actually had some of the governments like North Korea, I think you probably know this, that North Korean government attacks many blockchains, and they have people that are actually regularly trying to hack you.
So we had some attempts, but they never succeeded. And I think they actually hacked lots of other projects, there were billions and billions of dollars going into them essentially. But with scale, really what happens, I think when they learn about SGX, essentially it becomes really hard investment for them to try actually hacking Intel SGX, because it requires lots of computer science understanding and actually knowledge.
So they’re just trying other networks. SGX was really an amazing thing for us in terms of hackers. We never actually had so much of hacker attacks, and never had a single successful attack.
We never had even an attack where the chain will actually stop. So you probably know some of the other chains have this high profile events where they crash, and then they don’t work for like a day, and like they fix something and they start working. We never had something like this.
And one of the things is open source. Actually, we actually have this bug bounty program, and some really cool people are trying to analyze our source code. And we did pay some bug bounties in the past, because it’s all open source, people can analyze it, and they can submit the bug bounty submission, and then fix the source code.
So that actually was a pretty interesting thing for us. We never had to pay a huge bounty, so we never paid lots of money. But many people found some small things which we could improve in terms of security, and they actually fixed those things.
So if anyone is listening to this podcast, if you guys want to earn a little bit of money, scale source code is all on GitHub. If you are actually willing to find something in us, you know, some small security vulnerability, you’re welcome. And we always welcome all of the open source developers.
So to answer your question, again, I think open source is key. It’s better to actually reveal your source and let people attack it. Hiding things is bad, I think.
And that probably is number one for us. Number two is actually doing computer science and going deeper in terms of hardware. And in this case, in the example of Intel SJX, it was a really, really helpful thing for us.
So this is great. I think this will put a lot of folks’ minds at ease that you are taking security and decentralization, those two main concepts as something very interesting to your way of scaling your platform. This is a question that I have actually received from one of our team members.
And they were talking a little about how, you know, your gas fees are reduced. Does this have an impact on the validator incentives and the kind of, again, the decentralization ethos that scale with also daily? That’s a great question. So in our model, it’s a little bit different.
So the network is actually collecting fees from DApps. So every DApp pays some amount of money per year for the chain. So let’s say you rent the chain from scale, you pay this money.
So this money we’re paying to validators. The validators are actually getting the fees, and the fees are actually increasing. So our goal is to become self-sustainable.
When the network started, pretty much every network starts with inflation. And so basically that happens because your network is empty. No one is running anything, but you still need to pay validators.
So we started with inflation, but our inflation is getting reduced dramatically each year. And ultimately, there will be no inflation. So we are like Bitcoin.
Our token supply is limited, and the end game is have zero inflation. So the end game is to have everything paid by DApps. It has to be, as I say, the sustainable business model for us and for our network.
So currently, we’re kind of in the middle, I think. Many chains are already paying. And actually, this year, everyone will be paying for chains.
So we expect that in 2025, actually, these payments are going to be much more important for validators than actually having inflation. And the payments are like, actually, validators are paid two months. Now we have something which is called a decentralized quality of service protocol, which is another really cool thing.
Basically, the validators are constantly getting analyzed by, again, by decentralized network, by decentralized agents, where every second, every minute, if a validator is not doing well, for instance, the validator is slow, the validator is not able to produce blocks, anything bad can happen to the validator. There’s a reduction in the fees. The fees actually depend on the quality of the node that the validator is running.
And as I mentioned, we are pushing really hard to become self-sustainable in terms of the fees. What we want to have, we want to have people pay, DApps pay, and validators getting this money. And the huge push for us is 2025 is to increase the number of chains, because we believe that in this new environment, we’re going to get way higher growth in particular applications, having more users.
And then as they have more users, they’ll be able to run more chains and actually pay for the existing chains they have. Wonderful. That’s excellent, I think.
And there’s a good explanation. I think this answers the question very well. And I’ve just seen the time, we’ve completely run out of time and I completely got lost listening.
This has been such a wonderful conversation. But before we wrap this up, I’ll ask you two more questions. And the penultimate question is basically, what are the kind of common mistakes users have three founders making while building any decentralized application? And I’m asking this one macro question, because we’ve talked a lot about it, but with your extensive experience, I’m sure that you have some good advice for our two brothers as well.
Yeah, that’s a fantastic question. So I think the mistake that we made was because we come from scientific backgrounds, we wanted to make everything perfect. So it took us a long time to build the blockchain.
We had to read all of the papers, we had to build a mathematically clean protocol. And probably it will be better for us to just release fast and kind of iterate. So that was the biggest mistake.
I think it’s better to actually release something imperfect and then have crashes and maybe even some people complaining. So we actually took a different approach. Now, actually, it’s good for us because the network is running, we’re actually making incremental changes, making way higher performance.
But my advice for everyone would be to actually release faster. Now, to answer your second question, I think, as I mentioned, there’s a new environment, and there are more people coming, entrepreneurs coming to blockchain with tiny startups. I think I would actually recommend, if you have a tiny startup, I would recommend not to go to VCs, to venture capitalists, because it’s usually pretty hard to get money from venture capitalists, especially if you’re not from the US or you’re somewhere in the world where it’s actually harder to get access to them.
So my advice would be that if you have a tiny startup, first of all, just build it yourself, right? Especially with AI now, you don’t have so much money, just build it yourself, run it yourself. But another thing which is good about crypto is that you have alternative sources of investment. So basically, if you have a perfect early stage startup, first, it actually saves money, it doesn’t really need much money.
But the second thing is that you can go to different projects, there are foundations that actually are able to invest to provide grants, and people have been using this. But the funny thing I’ve been seeing is that you have people that are building really important startups early stage, and they don’t apply for grants, but those also have grant seekers, which kind of jump from one network to another and just build, I think, meaningless things just to get money, this grant money from the foundations, and they don’t actually building anything. So what I have been pushing with scale, when we do grants to people, we’re trying to actually filter out the spam and the hype, and actually try to understand really important things that actually are going to improve the world, or actually have lots of users.
So my advice would be to come up with something that actually can have lots of users, build things without actually getting any investors, and then you can probably get a grant from different foundations, from different blockchains, but use this money to actually build a meaningful thing, not just build the thing which is going to die next day. So that’s actually a pretty good environment. The great thing about blockchain is a great investment environment.
Do computer science. I would actually encourage computer scientists, people that are deep in technology, to actually go and do startups on blockchain. And actually, I think it’s pretty good.
I think there’s a pretty good chance of success. I think that’s wonderful advice available faster, and build something that is actually creating value for the end user. So Stan, this has been such a wonderful conversation, and kind of answered so many questions of a scale about this advice for entrepreneurs.
There’s this one question that I ask everybody who comes on the show, and I think I’ll be remiss if I did not ask you the same question. You’re a man of science, you made a leap from, you know, you were doing CS, and then, you know, you studied physics, and then, you know, you worked in AI, and you moved to Web3. What would you say to the skeptics? So what would be your one line, perhaps, suggestion for the skeptics and people who are still very skeptical about this industry, so that they can truly start living? You know, there’s like, I think on Facebook, they had this thing on the wall, you know, that you have to build things and break things.
So essentially, all things in the world are experimental. It’s actually totally okay to fail. And in society, in the new society that we have, there will be a new generation.
This new generation will use computers in a way we don’t even imagine how they will use them. And blockchain will be definitely native to them. So the reason why blockchain will ultimately get successful is because of people that are now like teenagers.
For them, it is going to be the technology they will use every second, you know. So we are building something for the children, essentially, they will use it. And no matter what happens in the meantime, there may be some crypto winters, crypto summers, I think the result will be fantastic.
And I also would like to thank you so much for inviting me to this podcast. I think it was a very meaningful conversation. I also think we need to have like slower conversations, deeper, and anytime I’ll be happy to get back to this podcast.
Also, our listeners, please, if you have any additional questions, tweet these questions to me. I really love when people ask interesting questions on Twitter. So please, let’s continue this conversation and I’ll be always happy to come back to this podcast.
Thank you so much, Stan. This has been a really insightful conversation. And yeah, let me just stop the recording once and then, you know, we can talk about it.
You can perhaps come back because I would love to have a deeper conversation as well. I feel like I have so many more questions to ask you. So thank you again.
Thank you so much. Great. Thank you so much.