Title: S17 E36: Yonatan Sompolinsky on Kaspa Covenants, DAGKnight & Coordination Markets Podcast: Bitcoin Takeover Podcast (host: Vlad Costea) Guest(s): Yonatan Sompolinsky (co-author of GHOST, PHANTOM/GHOSTDAG and DAGKnight; Kaspa co-founder; Harvard research associate) Published: 2026-08-10 (road-trip interview while driving from the Brașov area toward Bucharest / Otopeni; camera batteries died near the Palace of Parliament and the last stretch was finished on Yonatan's laptop) Duration: 2:44:31 Episode URL: https://bitcoin-takeover.com/audio/?name=2026-08-10_s17_e36_yonatan_sompolinsky_on_dags_stags_the_future_of_kaspa_10_08_2026_16.23.mp3 Audio: https://bitcoin-takeover.com/audio/media/2026-08-10_s17_e36_yonatan_sompolinsky_on_dags_stags_the_future_of_kaspa_10_08_2026_16.23.mp3 Transcript: CORRECTED. Base: the pipeline's local Whisper medium.en transcript (int16) with speaker diarization (Guest A -> Yonatan Sompolinsky), then a local name/jargon pass (gfix) for Kaspa / DAGKnight / GHOSTDAG / PHANTOM / covenants / OP_CAT / BitVM / Stag Hunt / intendos / Bușteni / Sinaia / Brașov / Cake Wallet / Braiins / SideShift / LayerTwo Labs / eCash / zSide / Zooko / Hyperliquid / etc. It was compared word by word against a full local small.en pass (int16, VAD, word timestamps), and disagreements were reviewed. YouTube auto-captions could not be freshly fetched (YouTube bot check), and no independent verbatim transcript of this episode exists in the local transcript collections under bitcoin-takeover-best-quotes, bitcoin-takeover-articles/transcripts or btctkvr-transcripts (only unrelated S15 material; S16 E41 is a different interview). Speaker turns were checked with a full-episode local speaker-embedding scan (two reference voices: Vlad, Yonatan). 80 doubtful stretches were re-transcribed locally from short audio clips with medium.en (int16, Kaspa jargon prompt). Clip-informed turn splits include the mag-mount camera exchange (~00:18), the police / "block rate" driving banter (~01:00), the TTS "Canadian voice" feedback gag (~01:37), the flashlight lighting bit (~02:00), Yonatan flipping the interview (~02:24), and the Palace of Parliament "big brother / distant uncle" goodbye (~02:42). Remaining "Casper" mentions are kept where they mean Ethereum's Casper (gfix already restored those and mapped false Caspa/Casper hits to Kaspa). Names and jargon were checked against the title, the feed keywords/show notes, Vlad's long-form episode write-up, and the prior S16 E41 final. Notes: Vlad guesses the episode number on air and overshoots (this is S17 E36). "World War resistance" is Yonatan's running gag for DAGKnight catastrophe resistance from the prior interview. "intendos" is kept as his coinage for intent-based flows. Doubtful words are marked [unclear: ...] (including the mag [unclear: mix] camera mount, a few [unclear: threads] from prior conversation, and [unclear: gestation] in the uncle joke). ================================================================================ [00:00:45] Vlad Costea: Hello there and welcome to the Bitcoin Takeover Podcast. This is season 17 episode, I guess 37 or 8 by the time I publish it. I'm driving next to Yonatan Sompolinsky in Romania of all places. We met here and we had lunch together. We hang out for a little bit. We haven't discussed anything related to Bitcoin or Kaspa because we saved it for the moment when we actually start recording. And I'm going to be driving you to the airport, I guess we're leaving already. Never stay here. So far, [00:01:18] Yonatan Sompolinsky: Beautiful land, lots of potential. [00:01:23] Lots of unrealized potential I would say. Going back to our conversations. [00:01:30] Vlad Costea: So what are you doing these days? I saw that since our last conversation which took about 7 hours I think. You went to the Oxford Union to give a speech about what you're working on these days. And also you have people still mining and using Kaspa. What's new? [00:01:49] Yonatan Sompolinsky: What am I doing these days? [00:01:55] It's not defined yet, not yet well defined. There's multiple fronts. I would say the development of Kaspa is driving. Michael, Ori — I'm driving too. Thriving. I'm not thriving, I'm sorry. Yeah, you can come to Kaspa and you will thrive. So there are multiple developments in Kaspa. But I'm more witnessing it from the sidelines or watching Michael and Ori and Hans push these things forward. The main things that are happening in Kaspa are, while Bitcoin is debating OP_CAT, Kaspa implemented a very enhanced and well thought hard fork that essentially does covenants in the right way. [00:02:52] Vlad Costea: What does that mean? First of all, what kind of covenants and what does it mean to be done in the right way? So I don't think you're in the shot anymore, just adjust a bit. So I [00:03:06] Yonatan Sompolinsky: Guess there are two aspects to be done in the right way. One is OP_CAT is doing covenants through OP_CAT as an afterthought. [00:03:21] Vlad Costea: OP_CAT is concatenation. You put together strings of data to get a different result. Correct. And [00:03:28] Yonatan Sompolinsky: If you want to implement logic that can inspect all transaction fields, then there are more direct ways to do so. Of course, this enables zero-knowledge verification, zero-knowledge logic, rollups based apps. So one aspect is simply designing an OP_CAT and an opcode from the outset to function as a covenant-oriented. That's the easy part or the more basic and primary part. The more nuanced part is to add to block headers and, in consensus, a richer sequencing commitment. Which means you can get a commitment to the entire sequence of transactions in the ledger. [00:04:46] And the reason this is important is because if you want to implement based apps, which means the app controls nothing about the sequencing, it's fully done by L1, so you need this sequencing commitment. So let me just dive into this nuance because this is something that's sometimes left out of the conversation. You can use covenants to do zero-knowledge logic and verification through OP_CAT, even without this enhanced opcode that I mentioned. But what you can't do is, the prover in those cases will be an entity that views the entire sequence of transactions and can prove that it ran a subset of them and arrived at a correct output. [00:05:55] But in order to further commit the sequencer to have it run all transactions, of course in order, but all transactions and not skip any, you have to have a consensus that L1 needs to provide you, or the opcode, a commitment that you can work with, that ZK can work with. So in that sense, you can do almost based apps in Bitcoin with OP_CAT, but as far as I know, you can't do censorship-resistant based rollups or based apps without this additional feature. [00:06:50] Vlad Costea: I think in Bitcoin they're called sovereign rollups, and they have some trade-offs compared to, I mean, they're sort of like optimistic rollups on Ethereum, but they're not like the complete kind of rollups that you can have when you enable all of the opcodes. [00:07:08] Yonatan Sompolinsky: These optimistic rollups are already without OP_CAT? [00:07:12] Vlad Costea: I think so. I think they're using BitVM. [00:07:16] Yonatan Sompolinsky: Yeah, so with BitVM and without OP_CAT, you can do these optimistic constructions. [00:07:22] Vlad Costea: It's also interesting because there's a jargon concerning it. There's optimistic rollups and there's optimistically verified rollups, and there's a difference between them. Don't ask me to explain it very clearly for you. It's just that we tend to have a lot of terms in [00:07:38] Yonatan Sompolinsky: The space. Optimistic versus optimistically verifiable? Verified, yes. Let's verify. I [00:07:45] Vlad Costea: Think Liam Eagen, he explained them to me, but right now I'm driving. I can't think very hard, but I do want to ask you because rollups and this ZK technology, initially ZK was made for privacy to be able to prove that something exists without revealing what it is. Just like a basic logic of yes, it's true or no, it's not true, but without revealing any more information. And in Ethereum, they started using this rollup technology and also it bled into Bitcoin because they wanted it to scale better. [00:08:22] Ethereum ran into a scaling issue and they tried for many years to figure out ways to make it work. It seems like rollups is their way to the future. And I do want to ask you because Kaspa by design is very different. It mines 10 blocks per second. It has this pruning feature, which makes sure that it's lightweight for validators. Like why does Kaspa need to verify ZK proofs and why you use this data compressing techniques when you're already optimized for scaling or [00:08:55] Yonatan Sompolinsky: Is it for privacy? [00:08:59] Many questions. First, just for the record, we are recording, right? Yes. Okay. So for the record, I'm very bearish on rollups. You know, I'm using it, I'm using it, yeah, sure, or is it, that's all, because, listen. are familiar with it and definitely Bitcoiners when they think those of Bitcoiners, those Bitcoiners who still think about, who do think about apps on Bitcoin are thinking about it in terms of rollups. This is the only reason I'm uttering the word rollup. Otherwise, I would highly recommend against implementing it or spending our time together discussing it. [00:09:46] But to the point and to the historical point, first of all, as far as I remember, ZKs were mainly introduced in... Ethereum as a scaling solution, not as a privacy solution. Of course, you had projects that used ZK as the acronym, hints for zero knowledge. But the term is misleading in most of the...almost all of the heavy load, all the heavy lifting in ZKPs is in the validity rules, not in the privacy rules. Or zero knowledge aspect. And indeed, there's limitations, conceptual limitations to what you can do with privacy and the smart contract if you want the states to remain public. [00:10:42] So even if you are a big believer in privacy, you still need to make compromises if you want to...in the context of smart contracts and shared states. Because you don't want the states to be held by only a few parties. You want it to be computed, not only verified. So the entire concept of...so the whole concept of privacy and smart contracts is in advance limiting you to a very specific or to restricted subset of application and use cases. This is why ZKPs were mainly used or spoken of in terms of scalability. [00:11:27] And as far as I see the current map in Ethereum is that zero knowledge... [00:11:35] Vlad Costea: Ah, it was the scooter that made the accelerator. Okay, I think this one broke the camera. Make this movie. Good for this ride. [00:11:54] There's a cable here that's challenging us. [00:11:57] No, the cable is for the microphones. [00:12:05] Yonatan Sompolinsky: That's what happens when you talk about rollups, the kind of camera flips. As far as I see, Ethereum is shifting away from layer 2s in general. It's assumed, understood now to be an inferior way to run your ecosystem. Again, these insights and intuitions are now...we're discussing something in a great lag relative to the Ethereum community. These things were understood I think already a year ago. The layer 2 roadmap was described...basically was overridden. And now if we're only going towards a layer 1 centric construction, does that work better? [00:13:06] I think we'll do like this. Okay, how are we now? [00:13:10] Vlad Costea: I guess we're fine until the next time I have to turn to the left. Okay, I will just hold it I guess. [00:13:20] Yonatan Sompolinsky: Okay, so the bottom line is rollups should not be the main focus of discussion in my opinion. It should be about what Zero Knowledge is used for. And Zero Knowledge is not used...it can be used for privacy, for a limited set of use cases. It's mainly used for scaling and for separating layers. There is the main point. The main point is you take a verification focus language, L1, a language sitting on an L1 like Bitcoin or Kaspa, and you bridge through Zero Knowledge opcode, you bridge an expressive world of computation to that L1 whilst ensuring that the L1 is thin and not thin in terms of logic and does not at all deal with this complex computation. [00:14:21] This is the gist of using ZKPs today. You can ask why not run all computation on an L1. It's both a matter of scalability and also just a matter of separation of roles. An L1 that wants to secure money shouldn't deal with computation. It's too large of an attack surface or failure surface. This has been understood in Bitcoin and Kaspa by and large follows the Bitcoin design principles. [00:14:59] So even though Kaspa is scalable, we keep it UTXO oriented, we keep it limited expressiveness, and everything is done similar to Bitcoin in terms of the language. But you need to have the advanced opcode in order to ensure that you can do Zero Knowledge in a logical way. Makes sense. [00:15:29] Vlad Costea: Right, so can you tell me about some use cases that you have unlocked with the covenants? What users in Kaspa can do today that they could not in the past? [00:15:40] Yonatan Sompolinsky: Well, these are two different questions. The use case is basically any computation, I would add a reservation in a moment, but any computation that you want to have, any app, that would be the better description. An application that you want to run in a way that interacts with a blockchain, with a payment, can be done with covenants. Let me give an example. You have an AI agent. We discussed before that I have to say the word AI so that the algorithm finds your podcast better. Oh no. [00:16:30] So you have some ephemeral logic that the agent wants to create and at the same time create with the same action, create the logic, define it, condition, and lock the money with so that we can do, essentially we can do a general version, a richer version of atomic swaps. So we need to find a long-term solution for this. Oh, I need to find a USB. [00:17:15] So you're the boss, what do you suggest we do with this? You want me to just keep like this or I'll hold the camera like this? [00:17:23] Vlad Costea: I don't know, I'm focused on driving right now, the road is not friendly. [00:17:28] Yonatan Sompolinsky: Okay, do you want to switch and you'll focus on the camera and I'll focus on driving? [00:17:32] Vlad Costea: No, I'm fine. Okay. I think the audio is more important than the video. So if there are some small kick ups, I can fix them on post, don't worry. [00:17:45] Yonatan Sompolinsky: Okay, I think the video gives context why the audio is there. [00:17:52] Vlad Costea: But also look, the road is like this. It's beautiful, spectacular, but it's not great for recording. I wish I could make one of these. [00:18:03] Yonatan Sompolinsky: Do you have a mag [unclear: mix]? [00:18:08] Vlad Costea: Yeah, I do have a mag [unclear: mix]. [00:18:10] Yonatan Sompolinsky: Can we do this like this? [00:18:13] Vlad Costea: Yeah, but I don't think we're in the shot. [00:18:21] Yonatan Sompolinsky: Yeah, that's not... [00:18:22] Vlad Costea: I mean it works, if it's stable. The obsession with stability in the space. [00:18:31] If it's stable, it works. No more hiccups. [00:18:40] Yonatan Sompolinsky: Okay, so I was saying. So you're asking about, now that everything is stable, let's get to the real deals. So you're asking about what covenants unlock, and specifically what do I imagine people will use it for. In general, if you do covenants in the right way, without leaving room for censorship, then you can do apps. What you won't be able to do is the general DeFi Lego box, where things compose freely. That's not available yet. That's not part of the vision of covenants. It's more related to a separate or an upper layer, or a higher upgrade. [00:19:39] That's called vProgs. We shouldn't talk about it now, but I'm happy to if you want. But first, I'll describe covenants. So with covenants, you can... One nice representative flow through which you can understand what covenant is, or how it will be used, is... In traditional computation platforms like Ethereum and Solana, you need to create a contract, and then in a separate action, you interact with a contract. So a contract is a heavy operation, or an operation that is optimized for lasting usage. With covenants, you can... [00:20:30] Covenants are optimal for ephemeral logic. So you have a one-time interaction between someone, between another person, or another agent, and you want to define logic. You want to define logic that will govern your interaction. Think about atomic swap, but in a more general, you know, vault. So you can, you know, we will swap only if, you know, you pass the coins to this, I don't know, either whitelist, or at least through some other covenant, and only within the next hour. And additionally, I require a certain, I don't know, 3 out of 5 for very... [00:21:22] For integer. Whatever logic you want, that is not... that you can think of as useful for your business orchestration, or for your logic, or for your... for the reason you are using the blockchain. You can do... you can express this in a one-time logic, and in... atomically, within the same transaction, you deploy the logic, and lock... lock the transfer, or lock the coins. This seems like an incremental... it seems like I only say give you one... one inter... you know, one round of interaction, but I argue this is both very representative of what covenants can... [00:22:14] What... what... what flows are typically used with covenants, and I argue this for the agentic economy, you know, for the lack of a... of a more interesting term. For... for these flows, it's very natural. You just interact one time, they say this is what needs... this is what needs to happen. There's no contract deployment. It... the deployment looks one and the same like a regular transfer. Just the scriptPubKey is different. It blocks a logic. [00:22:52] Vlad Costea: I hope that's clear. If not, push back on that. Is there an article that people can read to understand better? [00:23:00] Yonatan Sompolinsky: Technical... technical articles, yes. For the use cases, it's a good question. I think... [00:23:13] I think there's a... I would say an expectation that the most technical people will provide the best use cases, whereas I see it a bit differently. There's a... I think people need to understand what it's... it can... it can do, and you can rely, and you should rely on other people's good ideas. My own... my own... [00:23:58] Like, there are... there are two hats I would say I'm wearing here. There's the use cases that I would push for, and there's the... the good use cases. I'm not sure they are the same. [00:24:11] Vlad Costea: So you're saying that your use cases are not the best ones, or not the good ones? [00:24:18] Yonatan Sompolinsky: I'm saying that I would say they're not necessarily the same level that... something that is... will be used by everyone. [00:24:35] It's... you see... okay, so Vitalik came out with a white paper, okay? Wrote a few use cases. Some hit, some didn't hit. Right. Today, I can list all the use cases that people describe about, you know, the use cases with Bitcoin's covenants. Okay. Why am I not doing that? Why am I not going over the list of use cases that people are describing with Bitcoin? Because crypto in general has now a crisis about what is crypto useful at all for. So when you ask me about use cases for... [00:25:15] For Kaspa's covenants, I'm like... I'm stuttering and hesitant, and the reason is because we should go out of the Kaspa context or about... or out of the covenants context, zoom out for the moment, and ask, okay, what use cases is crypto useful for? And is used for? Okay, within it, we can draw a circle and say, here, Kaspa is optimized for that. That's the right flow for me to discuss. But if you ask in general, what's the use cases for this feature? I'm like, okay, what use cases do you agree crypto is useful for? [00:25:57] So I would push back the question to you, and only then I will engage with you in a useful conversation. [00:26:04] Vlad Costea: Also want to ask you because last year when we did our interview, I think it was in June or something. I remember it was summer, and you showed me your pet turtle named Bitcoin. Yeah. And you mentioned that you were working on a new protocol that's called DAGKnight. Yeah. And you said something very interesting about it, that it's the only one that can survive a nuclear winter, if I'm not mistaken, or something along these lines. Yeah. How's the progress on it? I think it was going to generate 100 blocks per second and also become more censorship resistant. [00:26:45] Yonatan Sompolinsky: It's not more censorship resistant, faster censorship resistant, I would say. So the progress is doing well. There's a secretive meeting somewhere in Europe of a few DAGKnight magicians taking place these days. To accelerate the progress on this. This is also the context why I'm in the area of your country. [00:27:14] Vlad Costea: Oh, interesting. You're hunting for magicians. Exactly. So [00:27:21] Yonatan Sompolinsky: Let me straighten the understanding about DAGKnight. DAGKnight in and of itself is not faster per se. It has more blocks per second. Sorry, it's not faster per se. In principle, you can decide you're running 100 blocks per second even today with PHANTOM GHOSTDAG. And vice versa, you can decide you're implementing DAGKnight and running still 10 blocks per second. The reason DAGKnight is unique is because it is the only protocol which is resilient to 49% Byzantine attacks and does not have a latency parameter. The implication of this is that if you increase the block rate, you can still confirm transactions at the actual network latency and you don't need to protect against the worst case catastrophic latency that can happen anytime in the future of the network. [00:28:36] So in clear terms, if you need to set up in your protocol a latency parameter, then you have to ensure that this parameter represents an upper bound on the latency. Whatever the circumstances of the network are. So you need to protect against worst case latency. But if you have a parameter that is responsive to the actual network parameters, then you are more comfortable accelerating the block rate even if it sometimes will create a lag. But most of the time, you'll just keep accepting transactions at the same pace. [00:29:38] So to give you an example, if today we increase the block rate from 10 blocks per second to 100 blocks per second, then you will need to hard code in the hard fork a change of the parameter k in the network in the consensus, which would imply that confirmation of transactions will take tens of seconds. At least about eight x long. [00:30:02] Vlad Costea: Check out Cake Wallet, which is the place where you find lots of features which are privacy friendly for your Bitcoin. For example, you have PayJoin, Silent Payments. You can even get on the Lightning Network if you're into that. But also there's support for Zcash and Monero and you can buy gift cards, you can do swaps, you can buy debit cards that work at all Visa terminals if you're into that. It's like a super app. You can also buy AI credits. It's quite interesting. Check out Cake Wallet. It's free to download and it's non-custodial as a wallet. [00:30:37] And there's also LayerTwo Labs. They are launching a hard fork of Bitcoin, which is going to be called eCash. You can find more information at ecash.com. They will have a sidechain for lots of popular altcoins like Zcash. It's going to be called zSide, Namecoin. It's going to be called BitNames. They're going to have a quantum resistant sidechain. They're going to have, what else, an asset sidechain, pretty much everything that is popular in crypto, but around the same base layer. And maybe there's going to be a Kaspa sidechain too. [00:31:13] Who knows? It can be part of the plan. You can find out more information at ecash.com. And if you hold any Bitcoin, you're going to be able to claim the same amount in eCash one to one. You can use these coins to dump them on the market to buy more Bitcoin. You can try the new technology, which was designed for Bitcoin but didn't make it because of politics or whatever. And now we finally get to see if this works as promised. Check it out, ecash.com. [00:31:44] If you want to mine eCash next month, no, it's actually this month. We're in August. It's going to be launching around the 21st of August. And if you want to mine it or you want to 51% attack it or you just want to experiment with hashrate, you can go to Braiins. Which has this service. So, Braiins is the oldest Bitcoin under someone taking a picture of us. It's the oldest Bitcoin mining pool. It used to be called Slush Pool. And right now they have a hashrate rental service that's called Hashpower. [00:32:17] The URL is hashpower.braiins.com with a double i dot com. And you can rent hashrate for pretty much anything that you find interesting or useful for yourself. Sometimes it's even cheaper than buying your own share. Or buying your own hardware just to get the hashrate from someone who's doing a good job in terms of electricity cost and efficiency. So, check it out. Orange Rock is the place where you can trade without limits and without KYC. It's very cool. It's like a nicer interface for Hyperliquid. Except that you have everything in one app that right now is available on iOS. [00:32:54] It's coming on Android and desktop soon. And I really like it. Check it out. You can even gamble which means you can use leverage but I don't recommend that. It's a feature if you think you're smart and you can time the market. You can try it. I know that BitMEX is shutting down so maybe you're missing it. Maybe not. Who cares. Check out Orange Rock. And do your due diligence. Don't do anything stupid. And I think I mentioned all of them. I said Cake Wallet, SideShift. [00:33:25] Did I say SideShift? The place where you swap your stablecoins for? No I didn't. I said it in the previous run. SideShift is the place where you swap your stablecoins for something that cannot be confiscated. So if someone wants to pay you in USDT or USDC, you just put your Bitcoin, Litecoin, Ethereum, whatever it is that you want to address into SideShift. And you send the stablecoin address to the person who wants to pay you. What you're going to get once the payment takes place is the asset that you want directly into your wallet. [00:33:57] And you don't have to deal with gas tokens. There's good liquidity, low fees. Check it out. SideShift.ai. Thank you for sending for my ads. Now let's talk about this ZK obsession because it seems like Zcash has me to come back. [00:34:14] Yonatan Sompolinsky: Before Zcash, I'll give you the rundown about our previous diagnosis. And in general, how I would, if that's okay, the mental map that I think you should have. Now that the video is fixed, my brain is slightly more sharp, but not very much. Level my expectations. We had a good hike today and a good lunch. And I also don't have Bitcoin near me, the turtle, so I can't really follow that much. Yeah. Okay. So the mental model you should have is as follows. You like many crypto ecosystems. [00:35:05] For whatever reason, you go to that ecosystem, to that ecosystem. What will you find unique in Kaspa? In what context will you find Kaspa unique and helpful for you? So the day-to-day answer to that, the normal usage of Kaspa, would be in any scenario where you want either a very lean settlement layer that is very fast but lean, right? That you can use, that you can deploy fast logic to, even ephemeral, meaning not lasting smart contract, just a one-time logic. And you want just a ping-pong. [00:35:56] You want to send a logic to lock something and you want to get it confirmed within less than a few hundreds of milliseconds. And you want just to work very fast with logic. Now, there are other chains that provide, that once you deploy the logic, you can interact with it within half a second. But using it in the common manner, which allows you to do thin logic. That was an Aston Martin. That was the model, the 70s model. No, it was newer, but it's not nice. That's why it didn't catch my eyes. [00:36:44] So that's one area where I would immediately tell you with confidence that Kaspa is the only place where you'll find what you need. I mentioned AI agents, despite the cringe, to mention this. [00:37:01] Vlad Costea: I did see that. There is a trend where people see they're not building for humans anymore. They're building for agents to make transactions. [00:37:08] Yonatan Sompolinsky: Well, even if you build for a human, the human is interacting with whatever they're interacting through their agent. Even if the operator is online, they're still interacting with an agent, through an agent. So it's kind of when someone is building for the agentic economy, the question is whether they are building for autonomous agents, for semi-autonomous agents, or for really a human that is behind the agent and telling it what to do. Anyway, sorry for the distraction. We're on another topic. So one area is just when I want to lock and unlock conditional ephemeral logic, very fast confirmation, very lean operation. [00:37:50] I don't need to deploy a smart contract, and I want most of my uses to be isolated and governed by the UTXO, which is mainly to say I have my own keys, my own coins. I don't need to interact with other aspects. That's one aspect. Another more specific use case or set of use cases is whenever you want to sample the majority of the network, the honest majority of the network, in real time. So you can think of the crypto ecosystem as splitting to two subsystems. You can think of the term real time decentralization. [00:38:48] So most of crypto is either real time or decentralization. No one married these two categories. So [00:38:57] Vlad Costea: You're saying that whatever is decentralized takes time to confirm? And whatever is instant and has quick finality is mostly centralized? Yes. Okay. I think I can agree [00:39:12] Yonatan Sompolinsky: With that. It's kind of how we accept the slowness for the proof of work, heavy chains. And we understand that Solana or Ethereum 2.0 are fast. I wouldn't say Ethereum 2.0 is fast, but Solana is fast and Hyperliquid is fast. But they are more centralized. And to be clear, when I say centralized, we can go into the nuance. I'm not necessarily saying that the Sybil-resistant layer, the miners, so to speak, the stakers of Solana are centralized. You can even provide a softer characterization and say, they are in the short term centralized. [00:40:07] There is a single leader to the production of Solana blocks. So, in that sense, they are fast, but within the time span of the interaction, you are at the hands of one or maybe two validators of Solana. Yeah. [00:40:34] Vlad Costea: I mean, even in Ethereum, it used to be Infura or Geth or one of these large players that were running archival nodes and had every contract stored. They were relying on these. So, it was in many ways, it was maybe distributed, but not decentralized. Correct. [00:40:53] Yonatan Sompolinsky: Though, you can also think about it just in terms of the L1 consensus dynamics. [00:41:00] If you want to ensure that your transaction was governed by a multitude of miners, for instance, to pick a use case that is close to your heart censorship resistance, you need to wait until your transaction was seen by multiple miners. So, in Bitcoin, if your transaction is censored by OFAC and 80% of the network is censoring it, that will take you an hour or so to be mined by an honest, uncensoring miner. That's correct. And in other systems, okay, it's not Bitcoin at 10 minutes. So, even in other systems, it takes tens of seconds or 10 seconds to mine one block. [00:41:51] So, it takes you like a minute or so to be de-censored, so to speak. So, in any scenario where this is too long, where you want real time, but sampling, you need in real time to sample the majority, the honest majority of the network, so that you will get a good treatment. And I'll give another example of a use case. Then you will necessarily use, Kaspa is the only chain you can use. Why? Because proof of work inherently is the only method by which you bundle the writing of the content of the block and being selected within the block. [00:42:44] It's selected to be a miner. In proof of stake, this is almost always, I'll give an asterisk soon, it's almost always separated. So, you are selected as a validator or as a proposer or a builder. [00:43:01] You are selected as a proposer separately from your choice of transactions. So, this separation, I'm not talking about the PBS separation, which is not a real separation. I'm talking about the separation between I have been selected as a proposer and this is the set of transactions I'm mining or proposing. These two roles are separated in proof of stake. In proof of work, they come together because you need to commit to the block you're mining. And then if you want to sample for your use case, you want to get a good sample of the network, but you want your service to be in real time, then Kaspa is the only chain where you can do so. [00:44:01] So, for example, if you want, I'll give an example from the previous world and then from the new world. From the previous world, I feel like crypto was, this is like a long parenthesis, but I feel like the bear market is really, it's not bearish only in terms of the market cap, it's also bearish in terms of the actual usage. Yes. The way crypto is. You see [00:44:38] Vlad Costea: How users disappear once the price doesn't go [00:44:41] Yonatan Sompolinsky: Up anymore. Yeah, and I think this wasn't the case in the past, or at least I don't remember that to be the case. So that's why when we're talking about use case, I'm like, okay, but this is a different type of crisis because now it's unclear what use case is a stick in crypto and whether we are comfortable with these being our, you know, with those remaining use cases being hailed as the pick up of the revolution, so to speak. So, back to the previous flow. In the previous world, one example for that would be any case you want a bridge. [00:45:33] A bridge? [00:45:34] Vlad Costea: Yes, I see the bridge. I wish it was recorded, but... Oh, it's [00:45:39] Yonatan Sompolinsky: A... No, but access it. Oh yeah, you have to stop it. Okay. So, if you want a bridge, or you want an oracle attestation, you want anything that bridges from the outer world into the settlement layer, into the settlement logic, you want any information from the outside. And here, the stark contrast between real time and decentralization, you can immediately feel it, right? Because you can either decide to wait for multiple votes and multiple attestations from different oracles or different entities until you know for sure that this event happened or that event happened. [00:46:24] Okay. But there's a clear choice. You either wait longer in order to sample a larger portion of the network. Yeah. What's this festival? [00:46:39] Vlad Costea: I don't know, I got distracted. There's Labubu dolls and Lilo and Stitch inflatables and actually they're humans and use mascot suits. Anyway, Romania, it's colorful. [00:46:58] Yonatan Sompolinsky: I said that your country has a lot of potential. I mean, the way here was absolutely beautiful. This is where you do skiing? [00:47:06] Vlad Costea: Yes, this is a skiing place. Quite famous. Like, right now I think you're in one of the most popular tourist places in the country. That's why you think it's nice. We think it's nice too. What's the name of this town? Bușteni. Bușteni is tree logs. Tree logs. Nice. [00:47:25] Yonatan Sompolinsky: This actually looks like a very nice ski resort, so I'll accept your invitation that you adjust. This is [00:47:30] Vlad Costea: The same way you actually go to the what you call this telecabin, the place where you have like an elevator for you to ski. I think I did it once when I was 12 or something. I'm very afraid of bruises and breaking my arm or whatever. I did that two or three times in my life and I don't want that anymore. Yeah, but I need to lose a bit of weight before I try extreme sports or at least try skateboarding for a while. Just get used to sliding. [00:48:10] But yeah, what was your point? [00:48:13] Yonatan Sompolinsky: My point was that as a proof of work punk you do come, you know, it aligns with you being anxiety oriented. You're always optimizing for the worst case. [00:48:29] Vlad Costea: No, that's a good approach generally. [00:48:32] Yonatan Sompolinsky: Well, that's the nice thing about Bitcoin the pet because like a turtle always carries all the defense all the time in every block. I feel [00:48:43] Vlad Costea: Like we're getting much more lore right now about your pet turtle. [00:48:47] Yonatan Sompolinsky: No, it's about your pet coin. It's my pet turtle's Ninja Turtle. Like it can do like from the 90s, you know, we spun it out yesterday. Leonardo, Raphael. Yeah, we have like best of both worlds. Like we have all of our protection measures on our back but it doesn't prevent us from doing like 10 or 100 blocks per second. Make sense? Okay. So in the fewest world, any time you want to bridge or attestation from the outer world, you would need to decide whether you want it in real time and then you have to lean on one leader, on one miner or proposer. [00:49:29] Or you can wait for multiple rounds and in every network it implies a different waiting gap in order to ensure that you got an honest and accurate vote or attestation of what happened outside the blockchain. Hence the importance of mining lots of blocks per second. Proof of work enables you, fast proof of work enables you to sample lots of miners. At the same network round, so consider 100 blocks per second, and you have some bridge or some oracle attestation that is material for triggering a certain event on your blockchain. [00:50:24] Thanks to the large population that you mine with 100 blocks per second, within one round, you alleviate the trade-off between real-time and decentralization. This is the holy grail of Kaspa to me. The fact that you were able to implement real-time and decentralization outside a trade-off. Does that make sense? [00:50:58] Vlad Costea: It's like having your cake and beating it to choose anymore? [00:51:05] Yonatan Sompolinsky: You don't have to choose. I'm not saying there are not other trade-offs with this system. You mentioned pruning and you mentioned other stuff, but in the specific context of high police, in the specific context of real-time and decentralization, Kaspa gives you a unique escape from this trade-off. I mentioned this is in the previous world. In the now world, I would succumb to the fact that everything will be used by AIs. I'm not necessarily talking about fully autonomous AIs. I'm talking about semi-autonomous AIs or AIs with people behind them. [00:51:58] I think this is an era where you can easily imagine people optimizing for hit-and-run usages of a blockchain, rather than force the agent into a certain smart contract or a certain logic flow. Because the agents are very eccentric and creative and they will just flow to whatever gives them this hit-and-run experience. This is what I mentioned in the beginning, with covenants that allow you to do this ephemeral logic. Now, all this talk about real-time decentralization is not incidentally related to what you mentioned about World War resistance. [00:53:01] Although this is supposedly on the opposite kind of consideration. Because now we're talking about what protection properties the chain has, rather than what usage properties it optimizes for. [00:53:27] In our regular conversation about crypto, we would expect this speed to be antithetical to hardcore protection from catastrophes. Do you agree with this distinction or are you too focused on the fork? Where's fork? In the road. Oh, fork in the road. Do you agree with your thoughts again about eCash? I caught you again [00:53:58] Vlad Costea: Ruminating about eCash. No, not this time. I do think about that a lot because this could have been the way to scale Bitcoin and get it to do everything and pay the miners for it. Right now we don't know what's going to be the next big use case for Bitcoin and people just speculate that, well, we can use it for this and we can use it for that. But none of these L2s pay the miners that we have and we seem to have trends. And this is a general problem, not just in Bitcoin. [00:54:28] It happened in Ethereum, happened on every chain. You have trends when people do something for a while and then they get bored of it and maybe they want something new, maybe that they find something new somewhere else like in the AI world. And then we're stuck with this situation where Satoshi's biggest bet was that there's going to be enough transaction fees to compensate for the diminishing mining rewards. That's in the whitepaper in the incentives subchapter. And I just don't see at this point how that can happen, how we can get there. [00:55:05] It's not obvious and it takes a lot of work. And unfortunately, Bitcoiners do like to assume that since there is a limited demand, no, since there is a limited supply, the demand is going to go up by itself and people will buy Bitcoin just because it's scarce. And that's stupid. There's a lot of scarce stuff in the world. That's not why people buy something. They buy something when it's useful and when it makes sense for them. And what is Bitcoin used for nowadays? Maybe some payments, but not so much. [00:55:38] Not at scale. It's just some hardcore people who are still using it. And otherwise, most of it is speculation. And the question is speculation on what? Because my understanding was that the speculation was about whether or not this can be the future of money. And when you make it less useful as money and you just want to have a pet rock that's on the blockchain. The question is, what is the end goal? It can't be a store of value forever. By the way, this is like the royal train station. [00:56:11] This is where the king used to come. Oh, this is where you wanted the retiree to be? Yes, Sinai. It comes from Mount Sinai. A bunch of monks retired here after the fall of Constantinople and they built this. It's very beautiful. [00:56:27] Yonatan Sompolinsky: Mount Sinai in [00:56:29] Vlad Costea: Israel here. What's the question? Mount Sinai in Israel or here? Mount Sinai? I don't know. In Israel, I guess. That's the one. There's only one. The one from the Bible. Yeah. Between Israel and Egypt. And they came here and they called this place Sinaia. So, yeah. [00:56:51] Yeah, anyway. I gotta ask you about what's your latest... I don't know. Do you think about Bitcoin at all? Do you follow the conversations, for example, with people wanting to filter some transactions? Or do you follow the eCash conversation where there's gonna be a hard fork for trying this sidechain tech? [00:57:14] Yonatan Sompolinsky: Bitcoin is dead boring to me at this stage. I would see Zcash as... [00:57:27] I agree with that. It's my favorite version of Bitcoin. Yeah. I would say if something becomes protection heavy and not usage heavy, I would bet on Zcash, not Bitcoin. At the current stage, Bitcoin is both boring as a tech technology and the discussions surrounding the Bitcoin devs engage in are very outdated, very lagging behind the burning questions. So the eCash thing could have been a good idea a few years ago. It feels a bit too little too late in the sense... I'm not saying that it's not the right path for Bitcoin, but since eCash is not branding itself as a Bitcoin, nothing interesting would come out of Bitcoin as I feel. [00:58:54] If I'm following what's happening in other chains, I'm following Zcash. Of course, I'm following anything interesting that I encounter. But mainly, other than Kaspa, Zcash following Solana. [00:59:18] Vlad Costea: I want to say that one or two weeks ago, I lost count of the days. I've had this finance guy who created this one Bitcoin exchange. His name is Cory Klippsten. And he was on my show and he was saying, look, even the VCs who were scamming in the past were able to revive Zcash. And I was thinking, how do I explain to this guy how big of a deal it is that you can have, first of all, zk-SNARKs in a shielded pool that can have a verifiable supply. [00:59:46] And this is a major breakthrough that they were able to do with Ironwood. And secondly, the Tachyon project, which is about synchronization in a way that sort of does client-side validation in order to be able to scale this technology, which previously was considered to be unscalable. You're looking at zk-SNARKs saying, well, it's a nice experiment, but it doesn't scale. And now they have a way to go forward to scale to millions of users. [01:00:14] Yonatan Sompolinsky: Yeah, police. Another arrow. I think they will arrest you for driving too slow. They take pictures. Your blocks are like driving me nuts. [01:00:28] Vlad Costea: My what? [01:00:29] Yonatan Sompolinsky: Your block rate. [01:00:31] Vlad Costea: My block rate? Am I slow? [01:00:33] Yonatan Sompolinsky: You're protective, I would say. [01:00:36] Vlad Costea: I'm talking and driving, so I'm trying to not make mistakes with either of them. So yeah, I'm driving conservatively, not very fast. But yeah, I don't want to get stopped by the police. It can be interesting. Maybe we can get more views if we actually get pulled over. [01:00:55] Yonatan Sompolinsky: But I just prefer your drive. Like, you know, in Kaspa, we can teach you how to do things faster. No, in Bitcoin, we wait for six confirmations. You definitely drive like that. [01:01:13] Vlad Costea: So I was talking about Zcash. [01:01:15] Yonatan Sompolinsky: Yeah. So you snuggled in there, an interesting feature or trait about the verified supply. You were referring to the formal verification of the new pool? Yes. Yeah. So this is like a tectonic movement in crypto for me that has went under the radar. And I felt like I was almost the only one who noticed it. And I didn't understand why it wasn't spoken. Was it just me? Or like it? The fact that the biggest cryptographers and the most hardcore cryptography chain has now shrunk the role of cryptography from verifying the entire state transition to verifying the supply to me is a very, very big deal. [01:02:48] How come this didn't make it to the headlines? [01:02:55] Vlad Costea: Well, Sean Bowe is a genius and I think he's very much ahead of his time with lots of his ideas. And it takes time. Like, I think even Halo 2, which he used for removing the trusted setup in Zcash, that sort of flew under the radar. But once people made sense of it in a couple of years, it was everywhere. Everyone wanted to have the Orchard pool. [01:03:19] Yonatan Sompolinsky: Sure. But you understand that there was a bug in the state transition, right? Yes. And this bug was not fixed, cannot be fixed. Any Orchard in Orchard, right? Or getting in Zcash a historical verification of the state transitions is unfixable, right? You can't have it, right? Yeah. Now, I would expect this topic to become much more of a debate how now the biggest and most active cryptographers are promoting the role of proofs and formal verification on a strict subset of what used to be, you know, zero knowledge proofs or validity proofs were used to be hailed for it, right? [01:04:30] Right now, no one is talking about verifying the entire state because we can no longer verify the state of Zcash, right? That is a historical state. Yes. So now the verification is narrowed to a strict use case of the supply, not even to other aspects, just to the supply. And to me, this is a very big shift. And a few days after this, I spoke with Zooko about it and he said, yes, his approach was always pragmatic in the sense of using proof for whatever we can have them for, I guess, and for now, if we can have it only for the supply, so be it. [01:05:29] The reason I'm mentioning this is because in Kaspa, we don't have archival nodes, we are pruning everything, right? And this means that Kaspa does not verify, does not, you know, the ordinary flow of Kaspa, of Kaspa consensus does not verify at all historical correctness. It does not provide you proofs and does not attempt to give you, to provide you proofs of historical correctness. And at the time, people felt that this is, you know, this is a bug and the fact that we didn't save archival historical data from the initial months of Kaspa launch is a big stain on the correctness. [01:06:17] And I met a Zcash dev, a core dev, which kind of pushed this criticism on Kaspa. And I asked him, what would you say about the, you know, the lack of historical correctness in Zcash? And this was before the bug. This was regarding what's happening. But, you know, and it kind of gave it an honest thought. And, you know, reconsidered his stance. So from that perspective, I think the rise of Zcash is, first of all, highly justified. It is much more refreshing than what is happening with, you know, what's happening or not happening in Bitcoin. [01:07:13] And for Kaspa ecosystem, it is big news because Bitcoin's launch was pristine, has no flaws and everything is auditable and there are no historical bugs and there's also no usage. But beside that, the growth of Zcash, the recent growth of Zcash is very, is a very good sign and paving the way for Kaspa in the sense that it gives room for proof of work coins that are proof of work projects that provide, in some aspects, better properties than Bitcoin. Of course, not without a tradeoff, but there are concrete formal aspects in which you can show this is a, provides a Pareto improvement over Bitcoin. [01:08:19] And I feel that up until Zcash's rise, there was a very, very restrictive glass ceiling on Kaspa's growth potential and adoption potential in the, in the hardcore community, if it makes sense. [01:08:46] Vlad Costea: You mentioned the name Zooko and I know that you share a very interesting story with him. At some point, he was broke and he was sleeping in your car. I'm not mistaken. [01:08:57] Yonatan Sompolinsky: I don't know if he was broke and I wouldn't, I wouldn't say he was broke and, and [01:09:03] Vlad Costea: Even if he was... I said he was broke and he was sleeping in your car. I don't know about now and maybe we just broke. Oh, [01:09:09] Yonatan Sompolinsky: Broke. No money. Um, I wouldn't, I wouldn't... But was it your car that he borrowed for sleeping? I wouldn't, I wouldn't know if Zooko was broke, but, but I don't, but I wouldn't rule out that even if he wasn't, he was, he wasn't still is a very low key. So I don't know if he was broke, but yeah, he, he's, he, he came to Israel, I think if, if it was in one of the... And I can't recall why he came to, it was some gathering or conference, but it wasn't scaling Bitcoin, it was years before. [01:09:54] And, um, yeah, so I have, uh, I know Zooko for, for, for some time. He was, uh, he actually gave me a very nice, uh, my, my, my best memory from Zooko is like one day before the, um, the sampling Zcash was announced, the sampling bugged. I met him in Stanford and we spoke about Kaspa and, you know, the different trade-offs that, uh, that I'm, I'm contemplating this was before the launch. And he took, he, he was very chill and told me just, uh, just, uh, just launch and, and, uh, see what reality gives you. [01:10:44] Could be, could be considered today, uh, I'm not, I'm not, um, I'm not presenting the, the conversation in the right, in the right light. It, it meant a lot to me to his push. And then one day after, um, the, the, they announced, uh, one year later, you know, it took like almost a year to patch it. [01:11:10] Vlad Costea: Yes. I know. And they had to keep it secret because of the time. Yeah. As they were afraid someone was going to exploit it. [01:11:16] Yonatan Sompolinsky: So it was a very nice, um, gesture, I would say to speak like in a very chill manner at someone, um, to contemplate with me the, um, the hardships of my project when he's less than 24 hours before, uh, I would say the biggest upset, um, of Zcash up until recently. The, uh, I think there's some downplaying of the, um, okay. So, uh, of the trust issues with, with this bug, it's not that I do think it was exploited. It's more like for people from the outside, to me, it seems the logical thing I would do if I would, would consider the bug, if I were discovered that the bug, you know, I would, again, I'm not, uh, I'm not, uh, justifying it. [01:12:16] I just know what I would do. I would just print for, in the cigarette, an infinite amount, just, you know, to be sure. And then maybe someone lost their coins. Right? You can, you can, you can, uh, you can realistically say that if someone didn't, uh, migrate to Ironwood within, I don't know, a year or two or three, wouldn't you say they probably lost their coins? [01:12:46] Vlad Costea: I don't know. Maybe that they're waiting for it to be more tested and that they're gonna, yeah, you're somewhat right about calling this an Orchard, which makes no sense. You got to get out of Orchard. If [01:12:58] Yonatan Sompolinsky: Someone told you there that they lost, if someone told you in confidence they lost this and this amount of coins, would you then feel that you can use their amount with, with your bug? [01:13:10] Vlad Costea: I don't know. Maybe they're lying and then you're gonna end up hurting the system in a way that's reversible. I do agree that people lose coins all the time and maybe that inflation, clear or hidden, can make up for that. I think Peter Todd is a champ, you know, saying that 2% demiurage is not inflation because it basically compensates for the coins that were lost and he has a paper about that. I just want the system to be fair and I want the coins to be mined fairly, not printed in a way that exploited a bug. [01:13:51] I mean, it gets to a point where it's sort of voluntary central banking at this point because you want to adjust the supply to make up for something that was unfortunate and might have happened. You're not sure, like how do you even verify that the person has lost their coins? You have to trust them. And I do hope that the Orchard bug was not exploited and no extra coins are gonna show up. So far I think the migration was double digit. I'm not sure if it went up to 20% in the first week, which I think is a bit disappointing. [01:14:26] I was expecting people to be more excited about this. I was one of the people who transferred his Zcash from Orchards to Ironwood on day one. And I've actually tried the various services and wallets that do the transition. I think they're very cool in the way that they handled this. But yeah, I have high hopes for Zcash because it actually dares to be different. And I also have high hopes for eCash because last year when we were doing our interview and around the time when I also interviewed Vitalik, nobody thought that Zcash was gonna make a comeback. [01:15:04] They were saying, well it's kind of dead. It's like a zombie coin. It started out this interesting technology, but now the same technology is used for Tornado Cash and Railgun and what else. There's also the Fence, your Maya protocol, which I think uses it. Aztec, I think too. So it seems like people were talking about Zcash, like this proof of concept which gave birth to lots of other protocols using the same technology. And then suddenly Zcash came back because people figured out how to make it scale. [01:15:39] I think that was the big reveal. And then they also found a better way to do shielded pulls. And the market reacted to this. Was there a Silicon Valley involvement? Yes. But is there some actual hope that comes from the technological progress? Also yes. In Bitcoin we haven't had much hope in a long time with something that's actually gonna scale and actually gonna bring off a bullish use case. Since the ETFs it's been stagnating. And you get Michael Saylor saying in tweets very directly that he doesn't like covenants. [01:16:12] He doesn't like increasing the block size. He doesn't like a lot of stuff that can potentially get more users to use Bitcoin. Because his understanding of it is that it should stay this institutional grade asset. And this is not really what I signed up for. And I think even the most toxic maximalist guys have to face themselves in the mirror and ask why are they even here? Do they still hope that the project is gonna get turned around and become cypherpunk again? I mean I don't blame them for that. [01:16:43] I used to be in that camp. But when you see so much information that gives you the contrary scenario, I'm not as bullish anymore. I'll let you talk because there's a motorcycle in front of me and I need to be extra careful with this guy. [01:17:00] Yonatan Sompolinsky: I guess my question is do you think if another coin had the same bug and patch, would it survive it? [01:17:12] Vlad Costea: You mean like Zcash? [01:17:15] Yonatan Sompolinsky: Yeah. My hunch is, and maybe it's unavoidable, but nonetheless we need to speak about it. My hunch is [01:17:24] Vlad Costea: That... I don't think any other coin would have been abandoned and forgotten for a while. But maybe that the precedent that there was such an incident before in Zcash helped it. Maybe the fact that this was discovered by one of their developers and not by some Kimi K-free user also helped. [01:17:48] Yonatan Sompolinsky: Yeah, definitely. [01:17:49] Vlad Costea: I mean it came from inside of the team. Twice. Like the times when the inflation bugs were patched. And they cannot say with 100% certainty that there was no exploit. But also their code base is so ahead of the curve that there may be 5 people in the world who even understand what's going on there. And I think that's a problem. But it's a problem that's going to get fixed as there's going to be more education and more academics entering the space and learning how this works. [01:18:22] It was the same with Ethereum not so long ago. [01:18:28] In what way? A lot of people... I mean I still have a hard time defining what Ethereum is. It's not a world computer. It's not whatever. [01:18:41] Yonatan Sompolinsky: So I think Ethereum has the same sense of... [01:18:49] Vlad Costea: Identity crisis. [01:18:52] Yonatan Sompolinsky: Yeah, a lot of excitement or something very tame about the future. I forgot the new acronym that they provided. What was this? [01:19:14] The supposed cypherpunk acronym. I forgot the acronym. But it feels more like a virtue signaling list of traits. And I don't think it's repairable. Because the Ethereum ecosystem has fragmented too much around their tools that now that we know how to scale... Oh, I'm stepping on it. Now that we know how to build Layer 1 zk and it became more feasible for consensus protocols, I don't know if they will use a DAG version or not. But at least there's knowledge how to do faster consensus. It still feels that the L1-centric part of Ethereum, and I'm saying L1 not because L1 is the most important, but because L1 is the only area where social consensus is to coherence and cohesion, it feels that in Ethereum the L1 sub-community is much much smaller than in other ecosystems. [01:20:50] Ethereum is much more fragmented. I agree with that. And so now that some core devs or researchers are laying down the vision, laying out the vision of, let's go back to L1 principles, again, feels too little too late. It doesn't matter if you'll code it with AI or not. The question is, the culture of Ethereum is not around a lean center compared to Solana, which is a very cohesive community. You can feel it from the outside. So in some sense, Solana is similarly more interesting than Ethereum, is more active similarly to Zcash being more active than Bitcoin and more interesting. [01:21:46] And it's not an incident that they have, you know, the intersection of leaders or influencers or however you want to call it, the people who push forward Solana and Zcash intersect. [01:22:00] And so in Ethereum, I don't see it. I don't see the L1, back to L1 roadmap relevant anymore. It smells like virtue signaling. I don't know. That's how it smells. I hope it's more than that, but I'm not sure. [01:22:22] Vlad Costea: I want to ask you about something a bit more personal, because I've seen that you went to the Oxford Union to give a speech recently. And you also mentioned something about a new political movement that you want to start. Can you tell me more about that? [01:22:37] Yonatan Sompolinsky: I wouldn't say I want to start a political movement. I mean, I don't mind being a president, but at least in my own house, I still don't have that authority. Oh, no. So my political aspirations are for now to conquer my house and then we can move forward. [01:23:02] Vlad Costea: That's how you scale. [01:23:03] Yonatan Sompolinsky: Yeah. For zero to one and then we can think forward. [01:23:15] The winding down of the spirit of the revolution of crypto is not something too unique to this revolution. It happens time and again in revolutions. And in that context, I had a long, I would say, multi-year thought, what should crypto go towards? And I came up with a thesis or maybe a manifesto, what the direction should be in my own humble opinion, from my own limited view. It's more about, so cypherpunks are hailing the individual. [01:24:17] Vlad Costea: That's a good way of describing it. [01:24:20] Yonatan Sompolinsky: And I don't know if you're familiar with Hayek's True Individualism versus False Individualism. He has a nice article. It's not a book. It's more like an article. But excerpts of this appear also in The Will to Serve Them and other books where he lays out the two traditions of individualism, the French one and the English one. The TLDR is that individualism as we are used to recognize it, and I'm speaking in broad strokes, of course. But the main voice of cypherpunks is emphasizing the atomic individual, which is not only the beginning of society, the atomic particle society, but is also isolated. [01:25:36] I'll give you an example. The leaders of Bitcoin, at least when I remember them, when I encountered them first, were very isolated. They live in undisclosed areas. Sure. Can you give a few examples just for the record? I mean, Jameson Lopp comes to mind, right? As a very characteristic person, representative of the movement. I don't know. Andrew Poelstra comes off similarly, and the Blockstream guys, which are great guys, but they're still talking in very individualistic terms in the sense of centering the principles around the isolated individual. [01:26:44] And this is where Friedrich Hayek draws the line between an individualism, which centers around the individual as an isolated player in the scene, which he calls false individualism. Is that French or English? That would come from the French tradition, yes? And Thomas Paine, Rousseau, that kind of tradition. And versus the English tradition, which is more conservative, more gradual, didn't develop as a revolution, which still centers society around the individual. But the assumption or the context is not that the individual remains isolated and that all of our efforts are directed towards the individual that trusts no one and connects to no one and lives in the woods, which is, or at least was, the hardcore vibe of Bitcoiners, at least in the first decade, before MicroStrategy became the face of Bitcoin. [01:28:32] It's centered around, so the true individualism, the English tradition of individualism, is still centered around the individual as the atomic particle, but that individual, the context is that that individual creates ties and connections with, you have 7.6 kilometers ahead. [01:28:53] Vlad Costea: Can you put it here, just so I can see it better? So far, it was just straight ahead for a long time. No, there will be more turns. But yeah, I do understand that there is a certain type of individualism that is guaranteed, for example, by Magna Carta, which is a fundamental document in the English tradition, which describes that every man's home is his castle, for example. It's something that's also part of Timothy May's writings when he published the Cyphernomicon. [01:29:30] Yonatan Sompolinsky: So it's not to say that everyone is, yeah, it's not to say that this necessarily describes the entire cypherpunk movement. Of course not. Yeah, I'm sure many... [01:29:42] Vlad Costea: There are a lot of left-wing anarchists, there are a lot of right-wing anarchists, schizophrenics, conspiracy theorists. We have quite a nice family in this space. [01:29:54] Yonatan Sompolinsky: Yes, but still, the leading voice, the temperament is around the atomic individual, so I think there's a difference between... there isn't necessarily a one-to-one match between the ideologies and the temperament of people. [01:30:46] I'm not necessarily claiming that this is a divide of values that I can identify clearly. Cypherpunks so far took only the isolated individualist part of the ideology, but the leaders definitely shared that temperament and radiated these values whether they believed them fully or just in their personal lives. [01:31:53] To be clear, the English version still believes that the individual is the atomic particle or maybe the family, but the point is that the individual is directing the values and the information and their choices shaped society around them. So it's not collectivist at all. It's not that the antithesis to the isolated individual is a collectivist. It is the connected. Hayek has this word, the enthusiastic individualist. [01:32:37] It's a passage in Hayek. I don't want to ruin it, but it's around the enthusiastic. You are cutting that guy. So the enthusiastic individualist creates ties and connections and communities, which you understand why this kind of sentiment or approach was pretty absent in crypto for the first decade. And in that context, what I think crypto should evolve into is this more coordinated usage that is appealing to people who are not disconnected from society and living in the undisclosed areas in the forest rather to ordinary people that create scenes, create communities, create connections, create trust. [01:33:42] Now, the fact that people have trust connections does not mean that we should introduce trust into the backbone of society, of the settlement layer or the crypto layer. The hardcore layer should remain trustless and paranoid. But the kind of applications that we should ideate and design and enable should be, while they should be built on trustless rails, they should be optimized for individuals that are connecting. Under that umbrella term, I proposed the coordination market thesis or the project Stag Hunt. I'm not sure if you're familiar or you read the short documents that I published. [01:34:51] But are you familiar with Stag Hunt in general? Stag Hunt? Stag Hunt, yeah. I'm not sure. So you're familiar with Prisoner's Dilemma? [01:35:01] Vlad Costea: Prisoner's Dilemma, yes. It's a classic lesson in economics and game theory. [01:35:08] Yonatan Sompolinsky: Yes, and I hate that lesson. [01:35:12] I mean, it's dark. [01:35:21] You don't like prisoners who turn on each other? [01:35:27] I don't know if I don't like that, but I don't like the fact that this is presented as the problem of humanity. Of course not. It's just a simple [01:35:41] Vlad Costea: Example of understanding incentives and the cost of collaboration versus the cost of being adversarial. [01:35:50] Yonatan Sompolinsky: I'm not sure it's not because, for instance, in Scott Aronson's Moloch writing in his blog where he describes Moloch as the demon of defection. I think it's a literate way to describe something that many ideologues, a mistake that many ideologues make. They assume that the problem is the Prisoner's Dilemma in the broad sense of people defect, people betray, and people assume that society is limited. And this is explicated. [01:36:43] I'm not guessing this. People are explicating their beliefs that society is doomed due to Moloch. [01:36:52] Vlad Costea: Okay, we're back with the second or third part of the podcast. We had our conversation about the different philosophies of freedom and individualism. And what else do you want to talk about, Yonatan? What's exciting in your world? [01:37:09] Yonatan Sompolinsky: So, I have a few [unclear: threads] from the previous conversations. Let me just turn off the voice. [01:37:18] Vlad Costea: Yeah, I'm right. Right. [01:37:24] Yonatan Sompolinsky: Gives me feedback. Canadian, very Canadian voice. Okay, so I wanted to complete the picture regarding the connection between the speed of DAGKnight and the World War resistance. And elaborate on or mention a certain flow, a flaw or a flow that can utilize real-time decentralization. But first, let me elaborate or concretize the coordination market thesis, what it actually needs from the base layer, and how it would actually look like. So, Stag Hunt is essentially the claim. It's a contrast to Moloch perspective or paradigm, which views defection and betrayal as the main human challenge. [01:38:45] Or society challenge. And instead, Stag Hunt is a game, a less known game, but still classic game theory. Where the failure is coordination, it's not defection and trust. Meaning, so the... I'm both in charge of the camera and in charge of uttering words. Okay, so the idea of Stag Hunt, the Stag Hunt game, in contrast to Prisoner's Dilemma, is that coordinating is still an equilibrium in game theoretic terms. Which means, in Stag Hunt, in contrast to Prisoner's Dilemma, coordinating is still an equilibrium. Which means, we are still, even if we are working with self-interested individuals who are rational or pursuing their own utility and the values and goals, in a Stag Hunt matrix or game, coordinating is an equilibrium. [01:39:57] So you don't need to give up on your interests. You don't need to be altruistic for Stag Hunt to arrive at a coordinated outcome that everyone is happy with. However, there are also cases where we are stuck in inferior equilibria. So, contrast this to Prisoner's Dilemma. In Prisoner's Dilemma, the only equilibrium is that where we are all worse off, and no one is cooperating, and everyone is worse off. That's the only equilibrium in Prisoner's Dilemma. In any scenario you incorporate Prisoner's Dilemma, you're incentivized to defect. In Stag Hunt, you either work alone or you work in coordination with others. [01:40:45] Both of these modes is an equilibrium point. And an efficient functioning society spends more time in the superior equilibria, where coordination happens, versus the inferior equilibria where everyone works alone. So far, is this a useful contrast between Prisoner's Dilemma and Stag Hunt? [01:41:10] Vlad Costea: Okay, yeah. That's a good [01:41:12] Yonatan Sompolinsky: Description. So, I argue that we overlook the Stag Hunt. That people with enthusiasm to change everything and to radically change and override the capitalistic framework will always go and or will tend to overclaim the flaws of a capitalistic society or free society. Whereas the Stag Hunt is a more moderate, is a softer claim. It doesn't claim that the system is broken. It just says, it claims that we are lacking coordination devices for efficient coordination devices. So how would that look like in a blockchain setup? I call this Project Stag Hunt. [01:42:08] And this is another thing I'm working on. A major thing I'm working on. [01:42:16] Consider the following flow. You know [01:42:21] Vlad Costea: What intents are? Intents? Yeah. I know about NEAR Intents, but maybe you should define what you mean by intents. [01:42:28] Yonatan Sompolinsky: Yeah, so NEAR Intents is an example for a ephemeral intent. It only happens in a limited context, but let's use NEAR Intents as a starting point. You sign essentially a transaction that permits someone to take your funds or to more generally use your funds or swap your funds with someone else, provided they meet a certain set of conditions or constraints on the virtual machine. So you can say something of the form, you can specify, I'm willing to spend, you can take out of my wallet one ZEC, an exchange of at least 50 ETH or something like that. [01:43:25] And then if someone can satisfy these conditions, they are permitted to act on them. How [01:43:33] Vlad Costea: Is that different from a smart contract? [01:43:39] Yonatan Sompolinsky: It's usually more friendly to cross-chain, where you can define things not as function calls rather as constraints on the virtual machine. So it's more friendly to multiple smart contracts because you don't necessarily, in smart contracts you force a certain function call. If you direct the transfer or the modification of your funds to a specific contract call. Here, you're just stating what you want to happen with your funds and under what conditions and you're not specifying how the solver, so to speak, satisfies them. So it's a more flexible framework. [01:44:24] In a smart contract, if it will be through a smart contract, it will force the counterparty that's satisfying your intent to flow specifically to a function called that smart contract. And this is inefficient and there's no way to do this. So this is a more general framework that allows you to say, this is what needs to happen with my stake. [01:44:51] Vlad Costea: We are back for possibly part three of this interview. You're talking about what you're working on, which is, what's the name of it? Project Stag Hunt. Stag Hunt. [01:45:05] Yonatan Sompolinsky: Go ahead. Then I'm going to have to... Oh yeah, take your time. The name stems from a classical game, I think by Uso, who was the first to, of course, he didn't afraid it in game theory terms. But the game, the raw game, is an ancient game where two ape men can either hunt a hare or a stag, a large deer. And you can hunt a small hare and get a small reward, or you can coordinate together on hunting a stag and then you get a larger reward. [01:45:55] And the reason this game is interesting is because, you want to turn on this light for me? The reason this game is interesting is because, in contrast to Prisoner's Dilemma, in this game coordination is, I think I ruined the light for you. [01:46:31] It's fine. Okay, so in contrast to Prisoner's Dilemma, this is a, coordinating together our actions is an equilibrium point, which means you do not need to be altruistic in order to improve society. An efficient society of players in this case coordinates on hunting a stag together, and there's no risk of defection when both parties decide together to hunt a stag. This is the main contrast point to the Prisoner's Dilemma talking point, or the Prisoner's Dilemma view of society, where we have to have some regulator or some super intelligence to govern us because we are unable to reach the high welfare on our own. [01:47:35] I view this tangent as more representative of a free society and more optimistic at that, because it argues, implicitly argues that the only quote-unquote problem of society in the context of, in a specific context of utilitarian players or agents, national utilitarian agents, is lack of coordination, not lack of trust or not lack of shared interests. If you adopt the stag hunt mental model, then you no longer view society as an organism that needs external intervention or a regulation in order to arrive at efficient welfare or efficient status. [01:48:41] This is the more abstract part of the tangent thesis. The concrete realization of this is in the form of coordination markets, and we started describing this as an enhancement of intent. So the way I like to view these new primitives that I'm designing as intendos, which you can think of as... intendos? It's intendo. Just like in intendo, like an enhancement of intent. And the idea behind intendo is it's a persistent, richer form of intent where, and in the context of stag hunt, it carries logic in the form of if many people will sign this same intent or intendo, consider mine as signed as well. [01:49:57] So the way you coordinate with people is you sign a commission to move your funds or move or trade your funds in a way that is conditioned on many other people coordinating the same move with you. Right. And I argue that this primitive has far-reaching consequence, and we can speak either about the consequence first or about the actual cryptographic flow or consensus flow. So let me start perhaps with the latter, with the actual concrete flow. Does that make sense? Sure. So again, what do you need from the base layer in order to support this coordination? [01:51:04] So first and foremost, maybe let's take maybe a concrete example of a use case. OK, so consider you want to move your liquidity from a certain liquidity pool to another one. But you will only do so if many other users and liquidity providers, but also users, will sign this transition with you. You don't want to move alone. Moving alone doesn't make sense. Then you will sign a intendo, and the intendo will say, I am moving my liquidity if at least a thousand users or x amount of capital moves with me or signs the same intendo. [01:52:07] And you can decide freely what's the threshold that you are comfortable with. You can say at least three people and a hundred dollars or at least three million dollars or thirty million dollars. You can decide your own threshold for how much coordination you are requiring in order to hunt this stag, so to speak, together or act in a coordinated manner. OK. You sign this and you send it to the operators of the stag hunt layer. And importantly, in many use cases, you will want these operators to not know anything about the state of this proposal. [01:53:01] So I call it a pack, the pack formation. In many cases, you want it to be opaque so that the operators, you can think of the operators, for instance, as the Kaspa miners. They won't know what the state of the pack is. So you need to use some form of threshold homomorphic encryption for that. There are other use cases where you don't mind so much. But in many use cases, this is very sensitive. The reason this is sensitive is you don't want the operators to have the advantage where they can either front run or back run the coordination, the coordination initiative. [01:53:50] You send it to that layer and that layer computes the state of the pack. And the state of the pack can either be no one is shifting to the other pool, no one is jumping to the other pool, or it can identify a large subset of intendos or people whose conditions were met. And then you want the coordination to snap immediately. So consider the liquidity pool example. If many, many people or lots of capital has all crossed the threshold that they defined, that they chose, for sufficient coordination, then you want it to snap immediately. [01:54:41] And you don't want to wait. Again, for two reasons. One reason is just it opens a surface attack for the operators. If you have this lag between when it is discovered and when it is executed, that's reason number one. And reason number two is this system should allow users, even if they signed a intendo, should allow them to move their capital freely. So the fact that you signed a intendo saying, I'm moving one million Bitcoin from here, from this pool to that pool, or one million USD, whatever that system, whatever stablecoin that system uses, from this pool to that pool, doesn't mean that you lock the capital. [01:55:34] You can freely move your capital, and the operators just need to continuously evaluate the largest subset whose conditions were met. This is a concrete DeFi example. What I was alluding to is that this has also more societal, ideological, political consequences. Because in many cases, you can argue that there's a lack of coordination in the sense that many people would be happy with... Many people's preferences and views are not getting represented because they're just scattered and not aggregated correctly. Right. So take for example, I think one example I mentioned, I'm not sure if I mentioned it in the six-pager or the Oxford address, was take for example a case where many people are upset with a certain policy of a stream [01:56:58] Vlad Costea: Service, let's say the Netflix. Okay. And that happens all the time, right? Yes. [01:57:04] Yonatan Sompolinsky: Now, there is some level of competition. You can arguably find other content in other places, but less so for families and small children. So with canceling your subscription by yourself, okay, you feel good with yourself, but you won't really find an alternative. And this happens a lot that it could be that the silent majority or the scattered majority has no voice and is not able to move together. So it could be the case, I don't know that for sure, but it could be the case that many people are wanting to move away. [01:57:51] Vlad Costea: What happened there? That's too dark, so it's closed. [01:57:56] Yonatan Sompolinsky: Yeah, turn off. It could be the case that many people want, many users want to shift away from that service, but they're not doing it for the more technical fact that they're miscoordinated. They're scattered. There's no discovery layer for the shared preference and the shared interests. So it's not the fact that they don't trust each other or that there's bad intentions. It's more that people are scattered, their preferences are not aggregated in an efficient manner, and there's no device for moving together from here to there. [01:58:40] So how would this look with coordination market? You would sign a intendo and intendo will say, if a new service is spun with, I don't know, commitment to these and these values, or flat tweets that this service is amazing, or some other set of specials or conditions, if this service is, if a new streaming service is spun and here are the conditions that, and it meets these and these criteria, then immediately and atomically I join if, I don't know, if at least 10,000 people join, or at least 1 million people join, or maybe, and it secures funding of at least $100 million, so it has a chance. [01:59:38] I can condition these things, and then I have no restrictions on my funds. [01:59:48] I can take out, I can empty my wallet, I can add more money to my wallet, I can retract that intendo. There's no limitation on what, there's no limitation for me on this primitive. It's actionable, but it's not committed. That's the distinction we should have in mind. The operators of the Stag Hunt layer, once they prove that they have, that my conditions were met, can snap it and they should snap it, but there's no commitment. I can increase my thresholds, decrease them, change them. I [02:00:31] Vlad Costea: Actually have a smart idea. Can you open this? There's a flashlight inside. And we can use that for putting it in the background for lighting the camera better. Spooky. Does it have battery? It has the button at the bottom. [02:00:50] Wow, that's very spooky. Is it good? [02:00:54] Yonatan Sompolinsky: It's good. I mean it's much better than before. [02:00:57] Vlad Costea: Yeah, unless it's modern. More than that? [02:01:01] Yonatan Sompolinsky: No, it's exactly the dark atmosphere we needed. [02:01:07] Vlad Costea: It feels like a summer camp with a fire in the middle and we're telling stories about hunting stags. I was going to ask you, what are the stags that you think we need to hunt in crypto? Because obviously there's this libertarian cypherpunk ideology which is at the core. But do you think it's BlackRock? Do you think maybe it's some nation states? Who's the stag that we need to coordinate to hunt? So hunt [02:01:34] Yonatan Sompolinsky: In this context would be capture and utilize, not take down in this specific context. [02:01:44] Vlad Costea: So would you say you're more of a minarchist than an anarchist? What's that? Minarchist means you want a minimal state. So you don't want to destroy the state, you just want to shrink it. [02:01:56] Yonatan Sompolinsky: What would be the difference between an anarchist and a libertarian? [02:02:00] Vlad Costea: Well, libertarian is more of an umbrella term. It means you just like freedom. It's usually associated with the American movements and libertarians. It's also associated with left-wing French anarchism and it carries a lot of weight. classical liberals are libertarians by some definitions. But in the case of minarchists, it's more of a clear term which says, well, this is about minimizing the extent of the states, making it smaller. [02:02:33] Yonatan Sompolinsky: Sure, but I'm not sure I understand what's the difference between a minarchist and a classic liberal. [02:02:40] Or it's just that the terms have got mudded and we need a new term to describe the same things. [02:02:44] Vlad Costea: It's the same as vagarist and anarchist. I mean, you have a lot of overlap. But my point is about the stags that we need to hunt. And you said that in this context. Hunt means take over, not destroy. [02:03:03] Yonatan Sompolinsky: A stag is a metaphorical term describing here a shared utility that we are unable to grab if we're just coordinated. It's a metaphorical category that describes anything that we would want to do together. For instance, in the liquidity pool, I would say the stag is simply the superior liquidity pool. In crypto, there are lots of capital locked in dead, cruel or stale areas or dead chains. I wouldn't say dead is not the right way. It's not the right description. Inactive, fluency-colored up. Yeah, but I don't mean that capital exists in inactive pools. [02:03:59] I mean, it exists in shitty technology, for the lack of a better description, or clunky technology, where people are not moving out of there because the users are also stuck in the same clunky technology. And for instance, take the entire Ethereum with the L2, with the entire mess over there. That would be moving away, coordinating moving away to a superior and more cohesive ecosystem would be a stack in this terminology that I'm adopting or proposing. And in that sense, if we can utilize Stag Hunt for coordinating migration to superior tech, that would be a big win. [02:04:58] Although I wouldn't limit my focus in Stag Hunt or in general Project Stag Hunt to a crypto night. It's more about ... I wouldn't limit it to the crypto field. It's more about general use cases that could be brought about. It's similar to the prediction market, I would say, in the sense, not in any sense other than it not being necessarily associated with crypto. It could be understood by normies as just tech that doesn't necessarily associate with crypto. So in that sense, I would view it as a more general endeavor. [02:05:53] Now, in the case of ... the interesting and cool part is that these intendos and Stags can compose. So you can imagine something like, someone says, okay, if there are 10 million people that want to purchase a subscription to a local streaming service, then I want to purchase as well. And then another person says, if there are at least 10 million people or 5 million people that want to do so, I'm investing, I'm seeding or I want to join as a seeding pool investment of that service. [02:06:42] And then a third person can provide other conditions. So you can sort of compose these coordination initiatives in a way that isn't available in traditional infrastructure. So I view it as something that can, first and foremost, be exciting and very expressive and cool to play with. And I'm sure many people will find good and exciting ways to utilize it and propose very creative flows. [02:07:28] So one set of use cases is, I would say a category of use cases, when you think about Stag Hunt, is exiting network effects or hopping between networks. [02:07:42] One main characterization of cases where we as a society, we can say we agree that we are all suffering from similar pathologies is when we are stuck in network effects. We can't hop to better technology, to better consensus systems, to better social organizations, because there is consensus around a certain focal point that everyone agrees is inferior. But moving alone doesn't make sense. That's the broad category where that's one broad category where Stag Hunt will shine. [02:08:34] Vlad Costea: I'm not sure if I have another question about Stag Hunts. I definitely want to read more about it if you have something to send me. But I also observed over time that there is this weird overlap between Ethereum and Kaspa, where the Ethereum whitepaper mentions your block that I designed. And also, for a while, the scaling solution for Ethereum was called Casper. [02:09:02] Yonatan Sompolinsky: Yeah, that was a... And you [02:09:07] Vlad Costea: Also deal with GHOSTDAG and stuff, so there must be something there like GHOST and Casper, Kaspa. [02:09:15] Yonatan Sompolinsky: It's just that there's no alignment there. It's just that if you have come up with cool or just new ways to do consensus or to do other primitives, you didn't really have the attention of the Bitcoin community. And Vitalik and, at the time, Vlad were more receptive to these ideas. So the GHOST protocol ended up in Ethereum whitepaper because Vitalik was more open to consider other consensus systems. At the time when, I mean, Casper, I think Vlad coined it, [02:10:13] Vlad Costea: If I'm not mistaken. Vlad Zamfir, yeah, he hasn't been very active lately, but he was on my podcast like four or five years ago. [02:10:21] Yonatan Sompolinsky: Yeah, so I think GHOST gave inspiration to... to Vitalik and I think also Vlad at the time. But they used, then the LMD GHOST, I think was the name of the staking version of it. GHOST didn't end up being implemented in Ethereum, and they ended up implementing a version of Inclusive, which is, I guess, our first formalization of the DAG framework. I mean, with time, these ideas just have their own inertia. So it's not that because we came up with the first DAG formalization or messaging that, you know, this means that everyone is copycatting in some sense. [02:11:21] Not at all. People have their own creativity around these ideas. I would also say that I refuse to see DAG as the main contribution. I don't like DAGs. At least I don't like their... I don't like the impression that a DAG is an invention or some novel idea. DAG is just a data structure. You can use DAG in a very stupid and void manner, vacuous manner. And you can use DAG in a sophisticated manner. I prefer thinking about... I'll give you the properties of the consensus systems, of the consensus system we're building, and then if you ask me how this is done, I'll mention DAG. [02:12:10] But really, when I describe DAGKnight, you know, for the irony, DAG is there. But if I describe a DAGKnight, I wouldn't describe it as a DAG protocol, although it is. I would describe it, as I described it to you, as the first partially synchronous protocol. synchronous or parameterless, 49% Byzantine protocol. That's how I would describe it. It happens to use DAG because DAG is the most general framework. Is this a good time to describe why DAGKnight is World War resistant? [02:12:50] Vlad Costea: Sure. I mean, it's a bit late in the interview, but it's night outside, and we do like DAGs, so... [02:12:58] Yonatan Sompolinsky: You can broadcast it in the morning if it's better. [02:13:02] Vlad Costea: No, I will probably broadcast it maybe around noon for Eastern time in the U.S. so that my American viewers can watch this live. [02:13:12] Yonatan Sompolinsky: Okay, at their peril. So, in short, in very concise terms, the fact that you have a very fast consensus that is not limited by any hard-coded latency parameter already implies that you are resistant to network hiccups or network catastrophes. So these properties are dual, and in fact you can't implement one without the other. So let me repeat that. Any consensus system that has the property that it can confirm transactions as fast as the underlying network permits, and there's an asterisk behind permits, but it's just an advanced asterisk that I will mention only later. [02:14:07] Any consensus network that is as fast as the network, as the underlying network, is necessarily resilient to that network breaking apart. And counterintuitive as it may sound, it's because if the consensus protocol does not know what the network latency is, then necessarily it's either as fast as the network is or as slow as the network is. Which means that if the network breaks apart, the protocol does not know that the network necessarily, that something bad necessarily happened. It will just aggregate in an oblivious manner and will just confirm blocks, confirm transactions much slower. [02:14:52] Does that make sense? It does. So this property is called Responsiveness or Partial Synchrony. Michael Sutton is basically the informal CTO of Kaspa. I haven't met him yet. So you should bring him on the podcast. He also, together with Ori, created a new programming language for covenants. It's called Argent on top of the Silverscript compiler. You should definitely be invited to speak about this programming language. I won't belabor on that because it's not mine to celebrate, but it's definitely an achievement that will interest anyone that's programming, I think, in the new era. [02:15:44] Because it's just something that's very natural to work with, definitely if you know what covenants are and you understand their value. So we spoke about DAGKnight. So DAGKnight is a work with Michael. And when Michael came up with DAGKnight a few years ago, five or six years ago, with the kernel, this was a big excitement for us. But it took years to really build it into a secure protocol. [02:16:28] The problem we have with selling this idea is because distributed systems folks, they are convinced that there's an impossibility theorem stating that no protocol can be responsive and at the same time be resilient to 34% Byzantine attacker. So there's a famous impossibility theorem in the consensus protocol. Protocols that supposedly prohibit a consensus protocol from satisfying these properties that we mentioned, together with being 49% Byzantine resilient. [02:17:07] We describe in the paper how we bypass this impossibility theorem. But the bottom line is that if you want to remove any trust from the network layer to minimize your assumptions on the network layer, then DAGKnight is the pinnacle of what you're going to achieve. And it's theoretically optimal. [02:17:36] Vlad Costea: Okay. I do want to ask you, do you ever feel like you're too ahead of your time? Because obviously there are not too many people who understand this or who care about this. We still have designers who are trying to find ways to make the Bitcoin Satoshi design scale and add features. And you're coming up with something which you call a generalization of the Nakamoto consensus. And maybe it flies over many people's heads because it's just so different. It's not even using a blockchain switch, which an entire industry has been used to. [02:18:15] It tries to be so different in lots of ways and finds new workarounds for solving problems. Like, did you feel you're ahead of your time by a few years? You think in a few years you'll be just like Zcash and you will finally get your flowers? I'm counting on you for that. That's why I'm here, I guess, right? [02:18:36] Yonatan Sompolinsky: Yeah. I'm not sure what's the difference between ahead of my time and a poor communicator. But definitely these ideas are... [02:18:51] Actually, with respect to DAGs, these ideas did not in the form of DAG, but the idea that the focus on BPS versus TPS... [02:19:13] Vlad Costea: Which is blocks per second as opposed to transactions per second. [02:19:16] Yonatan Sompolinsky: Yes, that focus, that shift of mind happened I think already two years ago, or one year ago maybe, when all the projects started. I think Solana pushed this idea or this trait more and then there was the base on optimism that also started with 100 millisecond blocks and all that. I think Kaspa is a bit isolated in the reach, in the sense that people are just tweaking their classical BFT parameters. Okay, let's move it from two seconds to 200 milliseconds and then their marketing department just celebrates it. [02:20:24] Kaspa, I think, had difficulty with penetrating the more serious, the more hardcore circles. So I guess there's, in some senses, we had the challenge of falling between the cracks, because the hardcore community is very focused on improving incremental stuff on Bitcoin and other networks, and the non-hardcore crypto is just increasing the performance of their blockchain for whatever trade-offs and celebrating them without any accountability on the trade-offs it creates. I hope that with Zcash rise, as I said, it's like a revival of the innovative Bitcoin. [02:21:32] So recently Bitcoin core devs are starting to, have started speaking about speeding up Zcash, right? And shortening the block times, and they're definitely aware of DAGs. So maybe they didn't reach yet the timescales that require DAG, but definitely the time of Kaspa will come, I'm sure. I'm sure because of two things. I don't think anything interesting will come from the Bitcoin community, in terms of technology at least, and I don't think you can achieve the properties of Kaspa without the protocols that Kaspa is operating with. Add to this the frenzy of AI that will require everything to be very fast. [02:22:30] People are using things with very low levels of patience, which is good. We hate patient people. I think this recipe is a good, optimistic case, but definitely we need, I think, more voices and a better penetration of crypto circles, I guess. [02:22:55] Vlad Costea: I agree with that. What do you think is going to be the next big narrative? Because every cycle we have something that really gets the normies involved, or at least what you hope is going to be this time around, because it seems like we sort of ran dry in terms of use cases for the blockchain or whatever is going to become popular in mainstream media. It was tokens for a while, it was DAOs, it was DeFi. What do you think is next? Is it prediction markets? Is it something more interesting? [02:23:25] Is it intendos? Coordination markets. Coordination markets. Yeah, coordination markets. So that's what you're working on, and that's what you hope is really going to catch on. [02:23:38] Yonatan Sompolinsky: Yeah, I'm not sure trend is a good word in this sense, but yeah, I think coordination markets definitely has the right grammar and the right intentions. [02:24:04] Vlad Costea: Also, for the purpose of documenting this trip, we left from Brașov in Romania, which is a city very beautiful in the mountains, and now we're in the capital, Bucharest. So, it's been a very long drive with stops, and when our cameras ran out of batteries and we ended up using Yonatan's laptop as backup, because we just had more stuff to talk about. We couldn't just have a one-hour interview like most people. That's not the kind of people that we are. [02:24:39] What do you want to say in closing, Yonatan? We're almost close. Oh, you forgot anything? Yes, in like 15 minutes. [02:24:51] Yonatan Sompolinsky: I would want to interview you. [02:24:53] Vlad Costea: Really? Why do you want to know, sir? [02:24:58] Yonatan Sompolinsky: What would it take, one question I have, what would it take for you to become an XYZ maxi? You spoke about narratives and trends. What would capture general, it could be Stag Hunt, it could be Kaspa, but it could be a third thing. What are your criteria when you are, when you are, like are you, what needs to happen for you to say, you know, I'm committing my work or my interest or part of my interest, large part of my interest? [02:25:40] Vlad Costea: I do like stuff that stops Big Brother, so that's why I'm interested in Zcash. It's the fork of Bitcoin that has the best privacy that this design can possibly have. It has the kind of cryptography that was considered to be mostly theoretical before they actually made it happen. That's also something that the Monero fans don't understand, that they didn't, at least not before the full team membership proofs, which are not even live yet, they didn't write their own cryptography in Monero, but in Zcash they started from scratch, from just an idea they had to come up with something. [02:26:17] So I do respect this and it's kind of the Bitcoin that I wish we had, but we did not for various reasons, some of them understandable, like the inflation bugs. Maybe we didn't want to have that in Bitcoin and we wanted to test this tech before. But my point being is that my background is that I studied political science. I was very much interested in human liberty. And I remember having my mind blown when I was reading John Stuart Mill, even if I find him silly nowadays with the utilitarianism. [02:26:52] And I think it's kind of a naive concept. I still can relate to that feeling of thinking how you can maximize your freedom, of course, with some responsibility. But still, like that's why I got into Bitcoin. I was also an advocate for Internet freedom in Romania. When I turned 20 years old, I went to my first protest and it was against a law that was international. It was called ACTA, A-C-T-A. I started an online petition that was like the peak of collectivism, which got one hundred thousand signatures in a week. [02:27:32] And that was a star in a newspaper for a while. And I kept pursuing this. If you look at my academic background, you will see that I did my bachelor thesis about online elections. And then I got interested in Internet privacy and freedom. And that's also the lens through which I discovered Bitcoin. And whenever I discover something that adds more privacy, gives people more freedom, I sort of like it. The tipping point, usually, I think for everyone, is when it feels kind of obvious that it's happening. [02:28:11] And it's like I think Paul Sztorc, the guy who's launching eCash, presented this metaphor. It's like when you're inviting people to a party and the first people are usually not the coolest kids. They're just your close friends. And they are coming because they like you. And they know that they trust your competence to do a good party. But you also want girls to your party. You know, you want drinks, you want funny people. [02:28:38] You want guys who are able to... The big weather is chasing you. Yeah, we're being chased right now by the police. No, I don't think so. But obviously, I'm not one of the cool kids. I'm obviously not. But I'm also somewhat ahead of my time by a few years. I discovered this recently. I got into privacy before it was cool. I was telling people to buy Zcash when it was still $30. And I sort of saw this narrative shift coming that people are going to ask for more privacy in their protocols. [02:29:16] And now I also sort of foresaw what was going to happen with the Coldcard hack, which is really shitty. I think it puts the entire security industry to shame. And also the whole Bitcoin only movement. Because for the sake of following the dogma, they're willing to look the other way in so many regards. Oh, I used to go to university around here. Was just one of the buildings that was annexed to the University of Bucharest. Kind of crazy. Used to be the guy who takes the tram, now I drive an Alexis. [02:29:53] Life gets good, man. The heart of Zcash. No, it's not Zika. I mean, if Zcash hits like $10,000, I'm going to get a much nicer car. But until then, I'm happy with this. My point being, so your question was, what does it take for someone to be a maximalist of XYZ? It's always the club, I guess. You got to like someone who's into that and then pursue it. [02:30:23] And that's why you're on my podcast, because I do believe that you're one of the underrated innovators in this space. And I like this. And I know that in a few years, people look at this and you're like, oh, look at this guy. He was interviewing this person who's super famous now. Wait, am I am I one of the cool or the girls or the weird ones in the beginning? You're the close friend or the researcher. You're the academic. You're not quite the party person, or at least that's not. [02:30:53] You are the party person. I'm the party person. I'm the one who brings people. But yeah, I think what Kaspa needs, if it's more specific about Kaspa, is more wallet support. Exchanges, honestly, I've come to the conclusion that, of course, they're nice for the liquidity and they give you some degree of legitimacy because they acknowledge that, OK, it's good enough for them to list it. But they're shitty also, and they're going to delist you if you don't have a certain amount of volume. [02:31:26] And sometimes they also suppress the price, like you saw in the case of Monero, where the fact that it was listed on Binance was it turns out it was detrimental to price discovery. They were shorting it like crazy to keep the price down. Of course, you do want to have multiple sources of truth. You don't want to have just a couple of places where the price is this, because that's what gives the price more legitimacy. It's not just one place where it's trading for this amount, because obviously you can do a lot of wash trading and manipulate markets if you don't have much liquidity. [02:32:04] But I think these wallets nowadays have really turned into exchanges in and of themselves. In and of themselves. So you can see how, for example, Cake Wallet, Edge wallet, they all have a swap feature. And some people will just never use it, look at it like, oh, it's just something near intense. It's also super big, possibly one of the most interesting things in crypto altogether, because nowadays it doesn't matter anymore what you own. You can turn that into anything for payment. And that's something that I really wanted to see. [02:32:45] It's like Uniswap on steroids. Uniswap started on Ethereum. It just swapped between tokens on a smart contract. [02:32:58] Do [02:33:00] Yonatan Sompolinsky: You imagine a world where, or a reality, that's the bad word, a reality where the cypherpunk movement is broadening its values from the protective to the active? And by that, I mean we are used, and when someone listens to you and to your podcast, we are used to identify cypherpunk with, as you said, fighting the big brother, fighting, fending off institutions, choking on the network or on the social and economical and economic ecosystem. It's a protective set of values. [02:34:11] Vlad Costea: It's also negative. It starts from the state action, and it derives its legitimacy from violence and oppression, and says let's invent the tools that make this not possible. It's not about destroying the state, it's just making it less potent in terms of a crazy bus driver. [02:34:34] Yonatan Sompolinsky: In that sense, there's an a priori limit to how much the movement can affect normies' life. Yeah, I agree with that. And that's the context in which I'm like, okay, so this is when we achieved all the dreams of cypherpunks. How far, how deep into reality will this be integrated? So in Ethereum, you have the... [02:35:10] Someone's way, way behind us. Let me [02:35:12] Vlad Costea: See if I have any cash for him to go away. At least they used to watch your windshield or something. Just give him this. You are the giver. Hey, look, I won't take a commission on your charity. Thanks, boss. You see, they speak English. He's a street beggar, but he's a professional beggar. [02:35:40] Yonatan Sompolinsky: So there's a limit to how much the cypherpunk movement can integrate into the flows and lives of ordinary people if the values are exhausted by protective and negating state power or centralized choke points. I'm not saying, of course, I'm not alluding to this is not important, but this has a limit and we may have reached this limit. It's not that the job has ended and completed and there's no further tasks, but I'm not sure there are further goals. [02:36:36] Vlad Costea: You can say this goal is a little bit defeating. And I also look at myself sometimes, I feel, you know, I look at myself in the mirror. I'm older, I don't have the energy to fight as much. I haven't gone to a street protest in, I don't know, eight years or something. But my point is that it's very legit to ask yourself, what have we really accomplished here? Because a bunch of guys got really rich in the space and they live a good life. And they seem content, like they're not willing to fight as before. [02:37:10] And I ask myself, am I willing to fight as before? Sometimes, yes. I look at guys like Amir Taaki, for example. I think he is still uncompromising in some ways, but in other ways, he sort of, you know, he made a deal with the Swiss government so he can stay there and whatever to build DarkFi. I don't blame him. He does need some sort of protection or safe environments to be able to do what he wants. Just a second, let me take a left here. [02:37:43] Yonatan Sompolinsky: Paradoxically, I would say that Amir is more on the, you know, despite the DarkFi can signal, Amir's values and his statements are more aligned with also a positive... So we need to... [02:38:17] You're fucking here, right? Yeah. [02:38:20] Vlad Costea: Okay, good. Yeah, we're good. So, I'm getting a bit sloppy here. Oh, you're good. [02:38:51] Okay. So, what was I saying? Oh, we're gonna see the Palace of Parliament. We can actually park there and take a nice picture. Glad to see you. Yeah, that's... It's one of the nicest places at night in the city. [02:39:08] But yeah, I do question a lot what we accomplished because in some ways maybe I agree with you that there wasn't much more that we could do or there isn't at this point. But I also feel like we haven't really tried as much as we could. Maybe we didn't give this our best. [02:39:30] Yonatan Sompolinsky: My point was not necessarily deflating. I was more giving context on why either coordination market or something of that form, of the same... That rings similarly and of the same tone should be promoted by cypherpunks. And by that I mean any positive vision for how crypto can enable positive solutions or positive organization of society. Not only or of people, not only in the way of, okay, you are... you are protecting against the big brother, but also some positive... positive... solution is not the right word. Positive vision for how people operate. [02:40:41] It could be coordination markets through Project Stag Hunt, it could be someone else coming up with a better or more accurate vision. But it needs to be, I feel it needs to be of the same... ring rather than... I think we covered the protective goals. We didn't necessarily achieve them all. We can improve that with Tachyon, we can improve that with more wallet integration, but broadly I think we covered, at least in principle, the cypherpunk protective ideals and goals. And we have broadened it to something that's more... [02:41:22] That's also positive, not only negative. [02:41:27] Vlad Costea: I can relate to this and I'm happy that we haven't entirely given up. You know what I mean? Like we could just say let's settle for this institutional great asset, which is BTC. And maybe that this is what it's meant to be and no much more than this. And I'm super happy that we're still fighting, we're still building. And there are new ideas that come along and they try to change the industry from the foundations, not just refine on making blocks a bit larger or, you know, a bit faster. [02:42:05] So yeah, that's where I'm coming from. We just arrived at the Palace of Parliament. Can you film this? Is it a popular ride? You want this? Yeah, it's crazy. It's been a long, long, long drive. You need to stabilize it a bit, but yeah, this is the second biggest administrative building in the world after the Pentagon, which is really crazy. We had a crazy dictator in our country who started our nation just to build this and 80% of it is not being used. They actually rent it for weddings. [02:42:43] And not too many people do their weddings here because it's expensive, but that's besides the point. [02:42:50] Yonatan Sompolinsky: Yeah. So this is your big brother? [02:42:53] Vlad Costea: It's not my big brother, no. It's like my distant uncle that I don't like, like a very old one, the one who smells, the one who comes to you and always gives you bad gifts and you're like, how the hell do you always mess up? [02:43:11] Yonatan Sompolinsky: Thanks for the ride. [02:43:13] Vlad Costea: The one who doesn't remember your birthday. [02:43:16] Yonatan Sompolinsky: You want to go out and celebrate your [unclear: gestation]? [02:43:20] Vlad Costea: Okay. We can use our phones for recording that. [02:43:23] Yonatan Sompolinsky: Okay. Sure. [02:43:25] Vlad Costea: Yeah. I think we recorded enough in here. 58 minutes. It's crazy, man. We were like, oh, let's do 15, 20 minutes and that's it. And then we drained the batteries. [02:43:36] Yonatan Sompolinsky: Thanks for the ride and for the opportunity again to see you. [02:43:42] Vlad Costea: Yeah, it's been fun. Yeah. You got to take a picture here, post it on Twitter, but maybe post it after you leave. That's good for OpSec. Yeah, yeah, yeah. Definitely post it tomorrow after your flight. Definitely. Please do. Thank you very much, Yonatan. This was not recorded, so just move a bit. So yeah, this handshake right here, like secret societies, you know, secret handshake.