Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
#20 Cryptography and Encryption in Blockchain with Jason from Zama
60 Plays7 months ago
Jason is the Ecosystem Lead at Zama, a cryptography company that works to introduce Fully Homomorphic Encryption (FHE) to blockchain. FHE is almost like magic, as it allows computing data without ever decrypting it.
What is Fully Homomorphic Encryption?
FHE is a type of encryption that allows computations to be performed on encrypted data, without ever needing to decrypt it. This might sound like magic, but it's a real technology that's set to revolutionize how we handle data privacy.
Imagine you could send your DNA to a medical provider without worrying about privacy breaches. They could perform tests on your encrypted data and return results that only you could decrypt. This is the promise of FHE — maintaining the confidentiality of your data throughout the entire process.
Why Isn't FHE Everywhere?
Despite its potential, FHE isn't widely used yet. Historically, the technology was too slow and computationally expensive for practical use. However, Zama.ai has been making strides in this area, improving performance by thousands of times since their inception, making FHE more feasible for real-world applications, especially in blockchain.
Zama has in fact been 10x ing the performance of FHE year on year.
FHE in blockchain
Blockchain technology presents unique challenges and opportunities for privacy. Transparency, while a fundamental feature, is also a limitation when privacy is needed. This is where FHE can play a crucial role. By enabling confidential transactions on transparent networks like Ethereum, FHE could help blockchain achieve a balance between transparency and privacy, much like moving from HTTP to HTTPS improved security on the internet.
On chain privacy is becoming a significant topic of discussion in the industry and FHE could provide a new way to interact with on-chain protocols while never sharing personal information.
Jason also highlighted ongoing projects that work with Zama.ai, where they're integrating FHE with various blockchain protocols to enhance privacy without compromising on the composability that makes blockchain technology so powerful. The Cosmos privacy network Secret is in fact in the process of adopting and testing out FHE.
Quantum Resilience
Another significant aspect of FHE is its resilience against quantum computing attacks. Due to the paradigms such as entaglement and superposition quantum computing is capable to easily break many encryption protocols, making it a highly disruptive technology when it hits the market.
FHE offers a quantum-resistant alternative, ensuring long-term security post quantum-computing.
In order to get there, the computational intensity of FHE needs to be addressed through further optimizations and potentially through specialized hardware like GPUs and ASICs designed to handle these kinds of cryptographic calculations.
This podcast is fueled by Aesir, an Algorithmic cryptocurrency Trading Platform that I helped develop over the last 2 years that offers a unique set of features.
Recommended
Transcript
Introduction: Envisioning a Fully Encrypted Internet
00:00:00
Speaker
The endgame is to encrypt the whole internet, you know, to be an internet standard, to be an internet standard. Same way TLS has been added to HTTP in 2006.
Introduction to 'Proof of Talk' Podcast
00:00:20
Speaker
Hello and welcome to Proof of Talk, the cryptocurrency podcast where we invite leaders and builders into the space to come on and talk about their experience in the industry as well as the projects and products that they've been building. My name is Andre and I've been in the cryptocurrency space since 2017. I've also helped co-found algorithmic crypto trading platform Acier that enables users to quickly and easily automate their trades
00:00:41
Speaker
while managing the risk. I'm here today with Jason, who works with Xamarin.AI, a really, really interesting open source cryptography company. Nice to meet you, man. How's it going? Hello. How are you doing? Thank you for having me.
00:00:56
Speaker
Thanks for being here.
Meet Jason from Zama.AI
00:00:57
Speaker
It's fantastic to see companies working on creating new kind of encryption technologies. And I think Zama is doing something pretty interesting. And thanks again for being here to kind of explain what it does. I know we had an initial conversation a while ago when we were planning this podcast. So I know a little bit about FHE.
00:01:19
Speaker
But I think it'd be fantastic to explore that in a bit more detail. But first of all, I wanted to get from you kind of how did you get into cryptography and how did you start working on a product like this?
00:01:30
Speaker
So I went into cryptography in 2017. Like lots of people, I just bought some ETH and Bitcoin just during the bull run, you know, in November, December. Yeah. And that's how I got hooked. And then progressively in the bear market, in the following bear market, I was following people on Twitter and an article coked my eyes.
00:01:56
Speaker
And I really dug the Reddit hole in crypto, money, monetary policy, Bitcoin stuff, geopolitical of money, et cetera. And when I was a student, I was not interested at all in money, in economics. And then I found it very fascinating and I read everything I could about it. And then I couldn't stop thinking about crypto.
Understanding Fully Homomorphic Encryption (FHE)
00:02:24
Speaker
So after that,
00:02:26
Speaker
After that, I tried different DeFi protocol during the 2020 DeFi summer and then a while after I decided to quit my job, I was working with a developer agency in Paris and decided to launch a project on stocknet with a friend, didn't work, but I was in crypto and then I got this job for Zama one year ago. Okay, awesome. And what's your current role at Zama?
00:02:54
Speaker
So I'm the ecosystem lead in the blockchain. So basically what I'm doing is meeting all the layer ones, blockchain, all the protocols who could benefit from the technology we're developing, FHE, which I'll explain, don't worry.
00:03:09
Speaker
And i'm doing it's a mix between business development and the event utilization of the technology because very new so you have to explain a lot to do a lot of podcasts just to make people understand because there is so many new stuff in the space that you always have to explain on that so that's normal.
00:03:27
Speaker
100% so you got into into cryptography by first of all being kind of exposed to crypto and like crypto cryptocurrency and then kind of those two jokes at some point. Yeah, yeah, yeah, absolutely. And that's a huge difference compared to all the people in Xama. Most of the people at Xama, they are cryptography researchers. Most of them haven't used the name and haven't trade crypto or do not self-custody knowledge or their crypto. So I'm
00:03:57
Speaker
I'm quite different in this space, but more and more they are like cryptocurrency people coming as we are advancing as a company. But the researchers are mostly focused on cryptography, mathematics, and the underlying tech.
00:04:12
Speaker
Right. Yeah, because I saw it's a relatively small company. It's 72 people, but 36 of them have a PhD supposedly in computer science or cryptography. So that's cool. Cryptography, yeah. Yeah, absolutely. So Xama was funded four years ago by Round India and Pascal Payet.
00:04:30
Speaker
Round is a serial entrepreneur, a French, that already worked in privacy before. Pascal is basically one of the inventors of FHE. The first three years were dedicated to advancing the tech. Now that it's ready, we are going to market. We are selling this to people who need privacy on the blockchain. But first, maybe I can explain what is FHE. Let's start.
00:04:57
Speaker
Yeah, I think that's a good point. 100%. Yeah. So first, FHE means fully homomorphic encryption. So it's an encryption scheme. And basically, the idea is you take some data, your most personal data, the data you don't want to share with everybody, you encrypt them so that nobody can read it. You send it to a server, to a third party, to a cloud blockchain.
00:05:21
Speaker
and the server can do computation without decrypting it. So that's kind of crazy because usually in the normal world, a server always decrypts the information. Information is encrypted during the transport layer, but then decrypted by the server to do the computation. But with FHE, that's kind of magic, magical black box, where you can do the computation without seeing the data.
00:05:44
Speaker
So you do the computation, there is an output, you send it back to the user, and the user, only the user can decrypt it.
Challenges and Advancements in FHE
00:05:50
Speaker
So I'll take a pedagogical example, let's say I take my DNA data, that's very precious, I encrypt it, but I don't trust these companies in Silicon Valley to do the algorithm on them, but I need a service. So I encrypt them using FHE, I send them my data, they can run the proprietary algorithm and their IP on it,
00:06:11
Speaker
to detect if I have a cancer or something. I have a chance to have a cancer. There is an output. They cannot decrypt it because everything is encrypted. They are using FHE. They send it to me and only I can decrypt them. So that's kind of magic.
00:06:26
Speaker
It is honestly kind of magic. It's wild. And there's a lot to unpack there. So it works without decrypting. At any point, the server, it just never decrypts the information. It processes, it computes encrypted information. And I thought that's what's absolutely mind blowing about this kind of thing. I also read an article from IBM. They wrote about homomorphic encryption as well. So it's like, it's really emerging. But how many do you think companies or how many applications are there currently of FHE?
00:06:56
Speaker
So before going in the use case and application, let's start why it's not everywhere yet, because it seems so magical that we could ask ourselves, okay, so why don't we use it everywhere on the internet? Why don't we update HTTPS in HTTPS to make it FHE compatible? That would be the endgame. And the reason is simple is that historically, this technology has been too slow and too computational heavy.
00:07:21
Speaker
So it would require hours of computation and millions in servers to make it work. So that was true a few years ago, but we at XAMR are dedicated to this mission to accelerate it and make it usable in the real world. So we have already increased 1000x improvement in FHE since we started, and we are gaining 10x improvement every two years.
00:07:50
Speaker
So right now, the technology is ready for all the blockchain use cases. Why? Because the blockchain that's the only industry in the world that is already used to slow stuff on that desperately need privacy. And that is ready to implement a new crazy cryptographic schemes, magic.
00:08:10
Speaker
They did it with ZK five years ago, and it worked pretty well. They accelerate this technology like crazy, even Vitalik couldn't believe it. And so we are doing the same with FHE, bringing it where it's the most needed. And then to all the use cases as the health data that I just described with my DNA data, the Web2 use cases,
00:08:32
Speaker
We will need more improvement in computation time. We will need hardware accelerators. So right now our libraries are running on CPUs, but progressively we are moving to GPUs, FPGAs, and ASICs so that they can run on the best possible hardware. But it will come next year for the ASICs.
00:08:51
Speaker
Right, right. So you've said that it has been historically quite computationally heavy to perform this kind of encryption. Yes. How does it compare, let's say, roughly, if you have a figure in mind, how does it currently compare with something like PGP, for instance, which is super basic encryption? Like in terms of code? Code.
00:09:13
Speaker
It's like I can give blockchain numbers if you want. So we can do only a few transactions per second in the blockchain while in the very centralized blockchain, I don't know, 10,000 transactions per second. So we are like 100,000 times slower than plaintext computation. That's the order of magnitude. Right. I see. Okay. And that you said it kind of goes 10x every two years.
00:09:42
Speaker
Yes, 10x, so we are improving the software layer and the hardware layer to improve it now.
FHE in the Context of Quantum Computing
00:09:49
Speaker
But that's roughly the order of magnitude. Right. So you're improving the hardware layer by, like you said, making it compatible on GPUs and on ASICs eventually, and that'll give you more computational power. Are you guys ever planning to support quantum computing? Because it seems like a problem that quantum computing would just be able to solve the speed issue like that.
00:10:11
Speaker
So this is a very good question and of course we thought about it when we started the company four years ago and the FHE itself and the FHE scheme we are using, TFHE, FHE over the Taurus, there are many different schemes who are using TFHE. TFHE itself is based on a primitive scheme called LWE
00:10:31
Speaker
which is quantum proof. So the good news is, FHE is quantum proof. So we are okay. We are okay. If there is a quantum computer coming out of nowhere, we are safe.
00:10:43
Speaker
That's awesome. That's really cool. I feel like this is also like a very emerging conversation, creating like quantum proof encryption and quantum proof applications. I actually came across a couple of articles last week. It was pretty silent on that side, like the tech, but now it seems to be a discussion that's happening. So it's cool to see that you guys are looking into that as well.
Privacy in Blockchain: FHE vs. Monero
00:11:04
Speaker
Yeah, absolutely. So it would be crazy today to launch a cryptographic company without thinking about quantum proof because we know they are coming. So we need to update all the system, all the TRAT5 system, all the blockchain systems to it.
00:11:23
Speaker
Right. So going back to the technology itself, how does it work in simple terms? How can a machine compute information that is encrypted without it having to be encrypted?
00:11:36
Speaker
So to be honest, I'm not a cryptographer myself, not even developer, even if I take saving. So going into details would be quite hard for me. What I can give you is give you just an overview, basically what you need to do. So you take your plain text data or the data you want to encrypt. Then you add some noise.
00:11:57
Speaker
Okay, so you have some noise to do the encryption. And then you run the FHE algorithm. So you run FHE. And historically, when doing the FHE computation, the noise you were adding to your ciphertext was growing too fast. And so your message was getting squeezed.
00:12:23
Speaker
And so one thing we are doing is like we have a way of doing FHE where the noise doesn't take all the place. I'm sorry, that's the only level of underlying explanation I can give on this. That's cool, man. That's cool. I'm not a cryptographer myself, so it's not like if you were to give me a super technical explanation.
00:12:45
Speaker
I will probably not be able to understand it, but I would try. So obviously there are a lot of advantages to FHE compared to any other kind of encryption. What are the main advantages of using it and why should AI and blockchain use it?
00:13:00
Speaker
The simplest idea is just for confidentiality. We are sending a prompt to OpenAI, a cloud, and all the new AI tools, and we are revealing what we want to prompt those systems.
Balancing Transparency and Privacy in Blockchain
00:13:18
Speaker
Ideally, we'd like to live in a world where we can encrypt the prompt, run the AI, send it the results and only you can decrypt it because that's like your natural rights. So the first and most obvious one is like just privacy and confidentiality. This is for AI.
00:13:36
Speaker
And then for blockchain, you know, the goal in the blockchain world, the ideals is to replace the financial infrastructure of the world. It's to update it to something that is much better, much faster. But the main trade off so far is that we have so we have decentralization. That's great. It's not controlled by anyone. That's everything is transparent.
00:14:02
Speaker
And this transparency is like you couldn't use an Ethereum layer 2 tomorrow or I don't know Solana or any huge infrastructure blockchain because it's transparent. I could even pay my bakery or the supermarket because everybody on AI could know what I'm worth.
00:14:22
Speaker
So the infrastructure is not ready. So we are adding this confidentiality layer on top of the blockchain because they have great properties, but they don't have confidentiality as a property. And transparency, in some cases, is a feature of the blockchain, but in most cases, this is a negative side effect. We want to get rid of. So the same way you had HTTP to HTTPS, so your first
00:14:49
Speaker
solve the scaling of the web and then you encrypt it. Now we are scaling blockchain but then we'll encrypt it and we are ready to encrypt it.
00:14:57
Speaker
Right. Well, I mean, you still have, you have Monero, which is a privacy-focused solution right on blockchain. So you have that. But then that brings an interesting conversation because you would need, I guess, like you mentioned with your example with AI, you would probably need all of these companies like Claude and OpenAI to implement FHE, right? To like take your solution, implement it on their side. Do you think they would? Because it seems like they would most definitely want to take that data.
00:15:26
Speaker
feed it back to the AI so the AI can keep learning. Yes, so you're right, but before answering this one on AI, you mentioned Monero, and I have to react on this. So this is a good point. So Monero, for those who don't know, like Zcash is historically the privacy coin.
00:15:44
Speaker
We have to define something in privacy. When you do a blockchain transaction, there are different things you can encrypt and hide. If you are hiding the sender and the receiver, the regulator do not like it. And that's why they are trying to take down Monero.
00:16:00
Speaker
But that's not what we do. Okay. Let's take, for example, Tonadocash. On Tonadocash, you are basically mixing the sender and the receiver, but the amounts, they are public because it's only a bunch of 1Es or 10Es on Tonadocash. We are doing the exact opposite of Tonadocash. So we are not encrypting the sender and the receiver. What we call privacy in the blockchain is encrypted the amount. So if I'm transacting with you, I'm sending you tokens.
00:16:27
Speaker
Only the amount being transferred is encrypted. The graph, the sender and the receiver is fully available and that's why we believe that this is compliant privacy because we can see who is transactive with who but we are not the devil like Monero or even if I don't consider them the devil, I consider them great project but regarding compliancy and regulations, at some point we have to work with these guys.
00:16:55
Speaker
Right. Yeah, no, that makes sense. That seems to be the consensus. Like Solanas, they've started offering token extensions and one of the extension is a privacy-focused token extension. And that's exactly the way you've described it, is the way it works. It hides the amount or the object transacted rather than the parties involved in the transaction.
00:17:17
Speaker
You're right. And to answer the second part of your questions regarding often AI are strong or other like Web2 or other applications, why would they use privacy? So we believe that user shouldn't care about privacy, but developers should.
00:17:38
Speaker
Okay, so that's our big statement. And look at what happened with the messaging. At first, end-to-end encryption didn't exist. It wasn't a thing in messaging, and that wasn't something most of the people were requiring. But what happened is, at some point, there was a signal, and there are WhatsApps who did end-to-end encryption.
00:17:57
Speaker
it became a standard among the developers and it spread to everybody else. So we want to do the same. Our target people are not the, our target audience is not the people in the street, the everyday Joe, it's the developer. And we think that if developer consider it a priority, then it will become a standard. So that's how we approach this thing.
00:18:20
Speaker
Got it. That makes sense. That makes sense. Are there any current limitations of applications incorporating FHE at the moment? It's easier to do computation over simple numbers and reals. It's harder to encrypt
00:18:43
Speaker
texts and to do a weird computation on text than is it to do on small integers. So that's a limitation, of course. That's why the go-to-market is like simple settlements.
FHE's Role in Confidential Financial Transactions
00:18:55
Speaker
I'm sending you tokens. You subtract my balance. We add some value to our balance, a very simple computation.
00:19:02
Speaker
And so once the computation become more complex, it's more costly, but it will be tackled with hardware accelerators. So on the short term, yes. On the long term, we believe that hardware acceleration will be so huge that you'll be able to do all the FAT computation like you don't even think about it.
00:19:24
Speaker
That's awesome, man. That's seriously awesome. So going back to use cases, what are the current use cases that you guys are seeing? I'm sure there's some cool use cases that are happening right now in FHE, in encryption, in AI and blockchain. What are the cool ones that you've seen?
00:19:41
Speaker
For most of the AI use cases, so first I'm not a specialist because I'm mostly working on the blockchain team at XAML. But the rough idea is we have to wait for the hardware accelerators in this space. So we won't have broad public use case where this is available, but they are like stronger use case in credit scoring, for example, in finance, traditional finance.
00:20:08
Speaker
And for us data, there is an amazing amount of stuff. But I can talk maybe more about the blockchain use cases. So as I was saying, we are encrypting the data. And there is one big narrative right now in
00:20:22
Speaker
the blockchain world is the tokenization of real world assets. So basically, bonds, stock, private equity, asset-backed securities, all those derivatives products, they are moving on chain. Very progressively, the dollar already moved on chain with stablecoins. The treasuries are moving also on chain with T-bills. More and more projects are building on this.
00:20:45
Speaker
And a lot of these assets will require some level of confidentiality. You won't be able to put all the data on chain. Because that's normal in the business when two enterprise and two companies transact. Lots of data is not public. That's not the norm. The norm is privacy. And the more and more you put this data on chain, for example, for private equity, think about tokenizing mutual fund, for example. That's not private. That's not public.
00:21:14
Speaker
So you have to encrypt it. And this is all go-to-market today. So we are working with financial institutions that want to tokenize assets. So what they did five years ago is they tried to tokenize them in private blockchain. It didn't work because private blockchain are just bad databases, as most knows. And so that's just a bad issue. You should take your SQL database.
00:21:36
Speaker
But the public blockchain had great properties, composability with the rest of the ecosystem, but they didn't want to use it because it was too transparent and too public. So now you can have the best of the two worlds, public encrypted blockchain using FAT and the composability with the rest of the ecosystem and you can hide all the data you don't want to share. So that's perfect. So that's this narrative that we are pushing.
00:22:00
Speaker
Another one is ID. Sorry to interrupt you there for a second. What data would you be encrypting if you were tokenizing a company? What kind of information would you use? I'll give the most simple example. You know what a NERSC20 is? It's a smart contract of a token.
00:22:23
Speaker
On a near C20, the balances are public. So you can go on the USDC or C20 smart contract. You can see how many tokens is owned by each address. So you can have a private year C20 where each amounts, the balances are encrypted data. So you can have the same year C20, but everything is encrypted. So if you need to tokenize your company's stocks, equity, or a mutual fund, or any other financial instrument,
00:22:53
Speaker
You know who are the owners, but you don't know how many tokens they own. And that's the most simple use case of something you can encrypt.
00:23:02
Speaker
Oh, that's a very good point. Yeah, for sure. I was talking to one of the former Core Ethereum developers a couple of months ago, and he's actually gone from... He was also the guy that built Kristoff. He's also the guy that built the first smart contract for the DAO back in 2016. Oh, okay. Oh, legend.
00:23:24
Speaker
Right, yeah, a pretty well-known guy in Ethereum and everything. And he's moved away from this, and he's been building, in the last couple of years, he's actually been building, the reason I bring it up, an asset tokenization solution. And he's basically tokenizing GMBH companies, giving them the option to tokenize assets and allowing people to basically almost like crowdfund by buying these tokens.
Implementing Privacy-Focused Blockchains with FHE
00:23:53
Speaker
I didn't ask in terms of privacy, but he, when he spoke to me, he said that he's doing it using ERC 20. So I don't think there's a privacy element to it, which, uh, it's, it's really cool. If I going to be in touch with them, I might just mention, uh, you know, to give you guys a shot. Yeah, absolutely. That would be amazing.
00:24:11
Speaker
And so once you tokenize an asset that require confidentiality, okay, now it's on chain, that's good, you can leverage another benefits of the FHEVM. And so contrary, so maybe let's step back a little bit and say, how can you do confidentiality on chain today? You have four different solutions, okay? One of them is FHE, okay?
00:24:35
Speaker
Another one is ZK-based privacy. ZK is mostly a scaling technology, but it's a little bit a privacy technology. Great company like Aztec or Alio are trying to solve this problem with ZK. We hope that they will succeed. And so ZK is a solution, MPC is a solution, untrusted execution environment are the last solutions. So you have those
00:24:57
Speaker
four different solutions. FHE is the only one that keeps the composability and the data availability on chain. So contrary to ZK, or maybe you can have it in T, but they are not secure. But with FHE, you can keep this amazing property, which is composability. And I'll take an example. Once you have tokenized your privacy or C20, your mutual fund, or your companies, or anything, your assets, and the information is private,
00:25:26
Speaker
Then what you can do is have an identity smart contract. For the first time, you can put your private identity on chain because they're encrypted. So I'm Jason, I'm French, more than 18, I'm not American, et cetera, et cetera. I put it in a smart contract on chain. And then you can have this private year C20 talking with my
00:25:46
Speaker
So ID contract and to send a token, it needs to check that I'm compliant with the rules, with the third smart contract where all the rules of your local government are written in a smart contract, very transparently. And so all of this works together and you don't need to compute proof, ZK proof. You don't have all this overhead that you have in ZK based privacy. And it's very elegant. It's very simple and everything ran on change.
00:26:13
Speaker
That sounds like a really cool vision, and I'd love to see it happen. It sounds amazing to have an ecosystem that's transparent but secure at the same time on Chain. Yes. So that's what we're all trying to build this year. And we have three amazing companies. So we don't run a blockchain ourselves at XAML. We don't have a token. We are not a blockchain company. We're a cryptographic company. But we are selling this tech.
00:26:39
Speaker
two people launching their own blockchains. And so right now you have three companies, Phoenix, that's the team behind Secret Network also, the Cosmos, privacy layer on Cosmos, you have Inco, and another one is Shiba Inu launching an FAT blockchain called Treat.
00:26:59
Speaker
Yeah, that's crazy. That's crazy. So we have all the spectrum, the most serious guys in tokenization to see by a new great Memcoin community. And the three of them will launch an FHE VM before the end of the year, maybe beginning of next year for the mainnet, but the testnet are coming. And you'll be able to, most of them are layer ones today. We cannot do optimistic rollups using FHE, but it will also come.
00:27:21
Speaker
but mostly layer ones and you'll be able to have smart contract on them and try all these use cases that I'm seeing on. Like always in blockchain, mostly we'll be surprised by what people build. So there will be hackathon on everything and people will come with great ideas. Of course, casino games is one of them. For the first time, you can have an encrypted state
00:27:42
Speaker
to play poker, to play blackjack. We know that gambling is a great product market fit for with all these memcoins and that like the gen stuff you have on chain. So it will work for sure. For sure, yeah. And that's the cool thing when you build something which is more of like an infrastructure kind of technology because it allows new companies to come in and use that technology in a way that you haven't even thought about. Like they're going to be building things you would not even expect.
00:28:08
Speaker
There was a really cool example we had in one of the previous episodes talking about a similar topic. Imagine you're building the Apple iPhone. It's 2007, you just built the iPhone. Would you be able to predict Uber? You probably wouldn't.
00:28:25
Speaker
No, of course, because there are billions of people more intelligent than you that want to solve other problems. They will see your technology and they will make it their own. And I'm really excited. This is only the beginning.
FHE's Potential Across Industries
00:28:40
Speaker
I've been in Xama for like one year. It was mostly research, but now we are putting this stuff in production, confronting it with the real world, with real use cases, with developers. And so this is an exciting time.
00:28:55
Speaker
We know that this is a core problem to solve in blockchain. We have everything. Scalability is being solved. EAP4844 has been great on Ethereum. UX is being solved with the kind of abstraction on many other new technology. And now privacy is the last big challenge that remains to be solved.
00:29:19
Speaker
That's fantastic. What was the solution you said you're working with Cosmos, a privacy-focused solution for Cosmos? Oh, so Phoenix is one of the companies launching an FHDM blockchain, and they used to run a secret network. Secret network is a Cosmos blockchain using a trusted execution environment as a privacy tech, and they are like OG of privacy in the blockchain space.
00:29:47
Speaker
Previously it was called Enigma 2. Some people know it as Enigma, but it's rebranded as Secret Network. And so now the same team is launching Phoenix. They are like super guys, great focus on privacy.
00:30:02
Speaker
That's awesome. Yeah, I'm going to look them up because I like the Cosmos ecosystem. I think they have their idea of bringing together the entire Web3 and crypto ecosystem is fantastic. So it's great to see other developers building on top of Cosmos. That's really cool too. So your solution currently is focused towards, I'd say, institutions or foundations or businesses rather than
00:30:27
Speaker
individuals. Would there be a use case for an individual or a small team of developers that might want to adopt that?
00:30:35
Speaker
Of course, there is always place for such developers. So our focus is on layer one, layer two. People want to integrate FHE as a new module in their blockchain. So that's one. The second one is for big institutions that wants to launch something like maybe an avalanche subnet using FHE or their own blockchain.
00:30:57
Speaker
And for all the developer or people or individuals, what they can do is first participate in Akatan of Phoenix Inco Treat Shiba. And they can also develop on our DevNet. We have a DevNet, which is kind of one-server blockchain just for people to play with it. They can do that. But mostly, once you have to wait for developers to build applications for the end user to use them, of course.
00:31:27
Speaker
Yeah, that's a lot of the time what happens in crypto because it's still a highly emerging space. You need to rely on the developers and teams building up the infrastructure that then allows the end user to make use of these in a nice and seamless way. And it seems to be a topic that, again, I see a lot of companies building towards this or talking about it, which is, again, it's really good to see.
00:31:53
Speaker
So apart from AI and blockchain, which is obviously your focus right now, are you planning to be part of other industries or are you looking at other areas where FHE could be added?
00:32:11
Speaker
Yeah, finance, just traditional finance outside blockchain is a big, big market for us. We are already in talk with lots of big banks, institutions, research team within banks are really interested in FHE, as it could solve many problems for them. Maybe one of the most simple use case for banks is like credit scoring.
00:32:35
Speaker
You know, right now, imagine you have two banks, or one banks and one insurance, or let's say two banks. Both banks have their customer databases with lots of private information. This is highly regulated and they cannot do anything with them. They can just keep it and use them internally. So it's very hard for this institution to collaborate together.
00:32:58
Speaker
and to share some data because too many regulations, too many lows, and it's too risky for them. What you could do is encrypt both databases in FHE, and you could run computation on them just to check if there is some common patterns, if there is some, I don't know, lots of additional intelligence of running computation of both blockchain and unrerevealed
00:33:21
Speaker
what can be revealed and you can work like this. So that's an interesting use case for banks. So you'd be able to cross-check accounts or information between accounts? Yes.
00:33:36
Speaker
Yes, quad-checked or apply common rules without decrypting your data. Because right now, these guys are not sending their data at all. They don't collaborate with anybody. It's hard for them.
The Future of Internet Privacy and Encryption
00:33:49
Speaker
Yeah. Yeah. No, that's interesting. And they could collaborate without trust. So that's what we'd like to offer. Yeah. Is that something you're currently working on or you're just on your roadmap in the future?
00:34:03
Speaker
So it's on a roadmap, but the talks are already engaged with these people, but they are waiting for improvement in scalability. I see. Yes, because of the highly expensive computation that is still a bit of a... Yeah, I get that. So what are your guys' plans for the next couple of years then? So next couple of years.
00:34:26
Speaker
Hardware acceleration will be a big thing. We'll have to update our libraries to make them perfect for the hardware acceleration. And then the end game is to encrypt the whole internet, you know, to be an internet standard. To be an internet standard, same way TLS has been added to HTTP in 2006.
00:34:47
Speaker
We we have this this name of an endgame which is HTTP Z as a zama encrypted but we don't see any limits in this technology we can see you could rebuild.
00:35:02
Speaker
All the current applications using FHE, if your timeline is long enough, why would you give your personal data to Facebook? You just need Facebook to run a huge server, running an encrypted data, but you don't want them to know anything about you, and you want to control your private key with all your data.
00:35:21
Speaker
and you could decide who you can reveal to. And you know, FHE could work perfectly well with other cryptographic primitive like zero knowledge. I won't go into details, but there is great way to make all these technology collaborate together to have a more encrypted, secure, private, confidential internet.
00:35:43
Speaker
That's a big goal and I do hope to see it happen because I think data is being monetized to shit right now, which is crazy. It's also going to be very disruptive as well because if you're suddenly blocking monetization for companies like Facebook, Instagram, TikTok, stuff like that, they're going to have to adapt to it.
00:36:05
Speaker
Yeah, but we have a way to match user data with people who want to make advertisements that could match them together. They could be paid the right way without Facebook being the middle and seeing everything.
00:36:20
Speaker
And that would be great. So everybody would be very happy. Of course, you want to keep some advertisement. So it's a legitimate industry, just too big now. And that's the only way to monetize most of the social networks. But there is good intermediary where the central period doesn't see anything. And we think it could be great. But we don't want to get rid of this model. I don't think it's plausible.
00:36:46
Speaker
Yeah, well, I mean, it depends how how you see the internet developing in the next few years, because it looks like it's more moving towards a subscription base slowly. Like people kind of had a shock when Twitter moved to the subscription, like eight dollars a month thing. But a lot of people have accepted it.
00:37:02
Speaker
And it looks like it's gonna be moving in that direction. So maybe, who knows, maybe we live in a world where advertising on the internet is just gonna be a thing of the past. You're just gonna have subscriptions on top of subscriptions that you will pay for what interests you as a user.
00:37:19
Speaker
And on social networks, I'm really interested into what's happening in the forecast or ecosystem right now. That's so great. I'm really optimistic about decentralized social network where you can choose which is your clients. So you don't like webcast, you can use supercast, you can use like many of the different clients. I've become optimistic about decentralized social network recently.
00:37:47
Speaker
Me too, me too. I didn't know much about it up until two months ago and I went on Farcaster. I've created a Farcaster account. I've gone into like looked at how it works and it's so cool to have the data on the chain and then anybody could build a different client. Let's say you want to show that data differently. You want to create different kind of taxonomies for your groups or you have different algorithms to compute that data and to display to the users. There's no limitations on where it could technically go. It's really fun.
00:38:18
Speaker
Absolutely. Imagine if you had like something like Twitter or like a big social network at that scale, just have a completely decentralized index and then anyone could kind of take it and create their own little, you know, application client for it.
Community and Innovation in Blockchain
00:38:35
Speaker
It's fantastic. I love to see more of that stuff.
00:38:55
Speaker
Even people who have made dozens of millions, they are still participating, at least in the Ethereum community, and that's something I really like. People are really happy to be here. It's one of the few really, really optimistic communities in the world. There are lots of tensions, there are lots of fears, but I love this space for being so future-driven and optimistic.
00:39:02
Speaker
Do you have a Forecaster account?
00:39:19
Speaker
Yeah. And the cool thing about Farcaster, if you look at the different, like Twitter is very price focused. Like people talk a lot about charts and price. And on Farcaster, people seem to be talking more about the tech and building it. Yeah, because it's more like more researchers, more like high quality content for people. Yeah, I agree.
00:39:41
Speaker
Yeah, it's a good space. I think it's a good space for anyone that wants to learn more about crypto and kind of understand more than just the price hype that's happening every day now.
Zama's Achievements and Future Plans
00:39:52
Speaker
But yeah, this has been great. Thanks a lot for coming on. Do you have anything that you want to share with the audience, any links or anything you want to announce?
00:40:03
Speaker
Nothing special to announce. Last week, we announced a fundraise of $73 million at Zama. Congratulations. That's just the beginning for us. You can go on zama.ai. You can follow me on Twitter at abedjazi on Twitter, A-B-B-E-Y.
00:40:26
Speaker
ZZY, I'm not sure I pronounced it correctly but we'll put it in the description probably. And yeah, very happy to be here and follow me if you want to follow our previous stuff on the blockchain. Nice one, man. Well, listen, it's been a pleasure and let's do this again sometime. Maybe we get a one-year update or something about Zama, see where you guys are at. Whenever you want and you'll be welcome if you come to Paris.
00:40:51
Speaker
Will do. I might take you up on that. I might just go to NFT Paris next year. I'll be there. Perfect. Just
Conclusion and Future Collaborations
00:40:57
Speaker
write to me. Thank you very much. Awesome, man. See you around. See you everyone. Bye.