User-Owned AI: On-Chain Training, Inference, and Agents, with NEAR's Illia Polosukhin

Today Illia Polosukhin, founder of Near Protocol joins The Cognitive Revolution to discuss the intersection of AI and blockchain technologies, exploring how decentralized infrastructure can enable "user-owned AI" through privacy-preserving model training, confidential computing, and autonomous agents that operate via smart contracts without centralized control. Check out our sponsors: Linear, Oracle Cloud Infrastructure. Shownotes below brought to you by Notion AI Meeting Notes - try one month for free at: ⁠https://notion.com/lp/nathan Evolution from AI to Blockchain: Near Protocol evolved from an AI project focused on teaching machines to code to a blockchain platform, and now combines both technologies. Significant User Base: Near Protocol has achieved 50 million monthly active users, demonstrating substantial adoption. Autonomous AI Agents: Near has developed autonomous AI agents that can operate independently once deployed, including an example of an agent given $10,000 to trade based on Twitter sentiment. AI-Blockchain Integration: The platform combines decentralized compute for running AI with smart contracts that execute actions, creating truly autonomous systems. AI in Governance: Near is experimenting with "AI senators" that people can vote for, which then make governance decisions, potentially solving the principal-agent problem in representation. Jurisdictional Framework: Near is creating infrastructure for jurisdictions and courts to enforce rules on AI agents, essentially putting governance into AI. Sponsors: Linear: Linear is the system for modern product development. Nearly every AI company you've heard of is using Linear to build products. Get 6 months of Linear Business for free at: https://linear.app/tcr Oracle Cloud Infrastructure: Oracle Cloud Infrastructure (OCI) is the next-generation cloud that delivers better performance, faster speeds, and significantly lower costs, including up to 50% less for compute, 70% for storage, and 80% for networking. Run any workload, from infrastructure to AI, in a high-availability environment and try OCI for free with zero commitment at https://oracle.com/cognitive PRODUCED BY: https://aipodcast.ing CHAPTERS: (00:00) About the Episode (03:58) From Transformers to Blockchain (Part 1) (17:01) Sponsor: Linear (18:30) From Transformers to Blockchain (Part 2) (21:23) Blockchain Security Fundamentals (Part 1) (33:36) Sponsor: Oracle Cloud Infrastructure (35:00) Blockchain Security Fundamentals (Part 2) (39:01) Zero-Day Vulnerabilities Solution (51:11) Confidential Computing Infrastructure (58:07) User-Owned AI Models (01:14:47) Marketplace and Governance (01:18:57) AI-Crypto Synergy Vision (01:27:50) Autonomous Agents Future (01:31:22) Outro

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

• 0:00Nathan Labenz?

  Hello and welcome back to The Cognitive Revolution. Today, my guest is Ilya Polosukhin, founder of NEAR, a hyper ambitious, multifaceted project that describes itself as "the blockchain for AI" and aims to build a future where AI belongs to everyone. Ilya is not your average blockchain founder. Before starting NEAR, he was one of eight co-authors of the seminal 2017 paper "Attention is All You Need," which of course introduced the transformer architecture and helped launch the current AI revolution. The credits on that paper say that Ilya and another co-author "designed and implemented the first transformer models" and were "crucially involved in every aspect of this work."

• 0:44Nathan Labenz?

  And in fact, NEAR started as an AI project, but took a major detour into crypto when Ilya and team realized just how difficult it was for them to pay their data task workers all around the world. Today, given his pedigree, Ilya could no doubt command one of those mythical billion-dollar cop packages from Zuck's superintelligence labs. But instead, he's building the foundational infrastructure for user-owned, privacy-preserving AI that can operate at global scale. In this conversation, we start with foundational principles and work our way up the technology stack. We begin with an overview of how NEAR's proof-of-state consensus mechanism creates security for network participants without requiring trust in a centralized authority.

• 1:27Nathan Labenz?

  In short, because the protocol is open and permissionless, anyone who is willing to put value at risk can become a validator, and thus, no one can prevent the flow of valid transactions. From there, we discuss how NEAR is leveraging Nvidia's confidential computing capabilities to create a permissionless network that allows anyone with compatible GPU hardware to sell inference compute, while keeping both the model weights and user data private, even from the hardware operators, all with only a 5% overhead relative to normal computing. To be honest, I hadn't realized just how affordable this privacy layer has become. And while it does require trust in the chip maker, it's clearly a huge deal for all sorts of scenarios, and does seem to help explain how it is that frontier model developers

• 2:12Nathan Labenz?

  have been able to ship their models to so many different inference partners without leaking the model weights.

• 2:19Nathan Labenz?

  After that, we dig into Near's plans to decentralize model training itself, with a process that allows contributors to provide whatever compute or data they have in exchange for a cryptographically guaranteed share of the model's future revenue. While it does remain to be seen whether a community-driven project can produce a 1 trillion parameter model that performs at the level of today's leaders, at a minimum, Ilya and team seem to have designed an incentive structure that could make it worthwhile for folks to contribute the estimated $100 million worth of resources that's required to compete at that level. Finally, we talk about what AI and crypto can do for one another, including how AI might finally put the smart in smart contracts, and how blockchain technology can provide stronger guarantees that AI agents will act only as intended.

• 3:04Nathan Labenz?

  This is something that I've been envisioning, if only in quite fuzzy terms, for years, and so I was really excited to hear just how concrete it's starting to become. In the end, I was so eager to understand the foundational technology on which Ilya's vision rests that we barely had time for the vision itself. But I do think this is really valuable knowledge, and the good news is that we've already scheduled another recording, so definitely stay tuned for part two, in which we'll explore the applications that people are already starting to build at this intersection of AI and crypto, and also try to get a handle on the GigaAgent future by exploring Ilya's vision for how AI agents will interact, transact, and even participate in governance.

• 3:46Nathan Labenz?

  With that, I hope you enjoy this technical deep dive into the infrastructure for user-owned, privacy-preserving AI with Ilya Polosukhin, founder of Near. Ilya Polosukhin, founder of Near Protocol, welcome to The Cognitive Revolution.

• 4:03Speaker 01

  Thanks for having me.

• 4:04Nathan Labenz?

  I'm excited for this conversation. I think this is going to be one that I'm going to learn even more than I usually do, because you're right at the intersection of AI and crypto, and...

• 4:16Nathan Labenz?

  Sometimes these two technology waves are characterized as different camps that don't understand each other, or at times are in some sort of online rivalry that may be mostly made up. But I've got this sneaking suspicion that I've never quite been able to fully develop that these two technologies actually might really need each other to reach their fullest potential and to do so in a way that we can keep control of over time. So I'm really excited to dig into all of that with you. For starters, though, if my deep research reports are to be believed, and this wasn't hallucinated, I do know for a fact that you

• 4:59Nathan Labenz?

  were one of the authors of the original Transformers paper. But then I understand from deep research that this project that we now know as the near protocol started as an AI project and then sort of evolved into a crypto project before now kind of coming back to the intersection of crypto and AI. So I'd love to hear a little bit of the story of like, how you went from an author of Attention is All You Need to branching out on your own to realizing that you needed actually crypto as well as attention. And yeah, just give us a little bit of that history and bring us up to the present day.

• 5:34Speaker 01

  For sure. Yeah, no, that's good deep research there. So my background is in machine learning and AI. I joined Google kind of this when I saw, I don't know if you remember this cat neuron paper, where they affected trained and autoencoder model, meaning like they would feed the images and like compress it and uncompress it to get the back the same image. And that model, trained on a bunch of images on internet, figured out that there's a cat. And so there was a neuron specifically that if you activate it would show up as a cat. And so for me, it was like, this is the approach. Neural networks working now.

• 6:18Speaker 01

  But I think images are not where the knowledge and intelligence lies. I think the text is. And so I joined Google was very much like, hey, how do we focus deep learning on natural language? And the straightforwardest path to kind of advance and check intelligence of any human or machine is to ask questions. So my team worked on question answering. And we actually had even some of this in a product on google.com. Back in the day before Gemini, you would get these short answers. And part of the challenge we had was actually like the models that were state of the art were too slow. Right. You couldn't, you know, the long, short term memory like LSTM models, right?

• 7:02Speaker 01

  They need to read one word at a time. And so, you know, if you feed them a bunch of articles from Google search results, right? There's like, it'll take forever for them to respond. And so that's actually how transformers, you know, was one of the motivations like, hey, how do we consume all this context effectively as much parallel as possible? And then figure out how to answer. Now, after that work, I was like, the kind of the slope of evolution of AI is accelerating. This is 2017 and I want to build products on top of it. And so I left Google to start NearAI with my co-founder Alex Kedonov.

• 7:40Speaker 01

  It was a premise that we actually want to teach machines to code. And so that's actually something I believed a long time that if machines can code, we're effectively changing how we're interacting with computing, right? It's changing from, you know, few people can sourceer the code, you know, and like talk to machines and then everybody else effectively, you know, need to consume their magic to everybody can effectively do everything with a machine that normally only, you know, few developers can do. And so, so we were trying to build that back in 2017. That sounded like science fiction. Now it's just called vibe coding. So, so we're trying to build that coding in 2017.

• 8:19Speaker 01

  And the kind of the expectation we had is what we've seen in 2022, 23. Like we thought that's what's going to be happening in kind of the industry. And as we know now, effectively we're lacking, you know, probably an order or a couple orders of magnitude of compute. Now, we were trying to be smarter, so what we tried to do is to get a lot more training data. Now, again, it's called instruct fine tuning data. So we were trying to get people to actually write a little bit piece of code for some instructions or write instructions based on code. And the people who can do this are effectively computer science students from developing countries,

• 8:59Speaker 01

  where, you know, few cents here and there, you know, that adds up to like dollars, you know, a few dollars a day, actually a reasonable money for them to work online and kind of practice their coding experience. Now, the challenge we faced was paying them, right? People, like students in China, don't have bank accounts. They have WeChat Pay. Back then, in Ukraine, if you received dollars in your bank account, you actually by law required to sell half of it.

• 9:26Speaker 01

  You know, there was no way to pay into some countries for whatever reasons, right? Like, there was the PayPal didn't work, you know, Transenvise didn't work. And so, there was all kind of weird, like, just pay, just pure payments problems that we faced while trying to collect more data. And so, we ended up saying, okay, well, we've heard of this blockchain thing. It's a global payment network. It should solve our problems. So we're just trying to use blockchain as a solution to our own problem of kind of coordinating and paying a bunch of people. And as we did our research, this was in 2018, we kind of like, well, architecturally, none of them scale, they're kind of very, like, slow, hard to use, hard to build on.

• 10:10Speaker 01

  And so, that's sort of like, hey, if there was a blockchain with this criteria, you know, easy to use, scales, microtransactions are always predictable price, we would use it. But there's none.

• 10:22Speaker 01

  And so, we should just build that and then, you know, use it ourselves and come back to AI. And we thought we're gonna build it in six months, to be clear. So we're like, hey, you know, it's easy, right? It's just some systems. We're gonna ship it and come back. It took a little bit longer. But indeed, we effectively built, you know, highly scalable blockchain that's focused on ease of use, really abstracting out the blockchain itself from kind of the user experience, as well as highly programmable. You can effectively run arbitrary software that, you know, written in C, Rust, JavaScript, Python, etc. on the blockchain as well. Obviously, all the financial stuff, payments, loyalty, remittances, been run through it,

• 11:05Speaker 01

  a lot of microtransactions going through it, as well as, you know, a few other use cases. And so kind of it is one of the most used blockchains. We have 50 million monthly active users on it. And so kind of as the 2223 did happen, right, we're like, we started building up the team back into like, okay, now that this acceleration is happening, how do we end with this new learnings we've had through the blockchain experience, right? Where can we bring that vision of, hey, the AI will change how you experience computing forward?

• 11:45Speaker 01

  This is the history right now. Looking forward, the important part is, as, you know, AI will write code, it will, you know, be able to interact with other tools and systems, you're effectively removing the need for other apps and even websites, right? Because effectively, anything can be built as, as, you know, the stack matures on a fly, it can go in directly, get the information you need. And so it's interestingly, there's few things that are happening. One is the devices are gonna become more, you know, run by what we call AI operating system. Now, this AI operating system, however people want it to run locally, you know, you will not be able to process everything locally, right?

• 12:30Speaker 01

  You, you know, there's background, you want to, you know, it to summarize news for you, et cetera. And so you need some way to offload the compute. The problem is, it needs to have all of your context. Like, that's how we're gonna get to true, you know, AI that's really helpful is because it will know everything about you. But that is a very scary situation if some third-party company has all your data, right? We just saw everything from, you know, your, the chats being leaked to, you know, affected some people in organizations and, and reporting to police with your chats. So there's already kind of a challenge with kind of how the data is handled by, by the centralized AI companies.

• 13:13Speaker 01

  So, so first of all, we want everything to be private. I think that's like a fundamental, so like, we want AI to be private, we want AI to have all your context. Uh, we want it to be kind of, um, be able to go and execute actions on your behalf. And so the other interesting thing that happens in this future is that a lot of the existing, what, you know, you can call them aggregators, like Google, DoorDash, Amazon, as well as like other middlemen, like I call the FDA, for example, it's a middleman, right? It like takes all the pharmacy filings, it processes them, it evaluates them, it's like, okay,

• 13:53Speaker 01

  yes, this, this medicine is safe on average, and then they, you know, allow pharma companies to stock it. Like, your AI can do all this work, right? It can go and find directly the, you know, factory in China to order stuff. It can go directly to AI of the pharma companies, discuss with them exactly your medical situation, find the exact medical compounds that you need at this moment right now. Like, you don't need to check, you know, if on average for average person, it's gonna be good. You actually need to check if it's good for you right now is what you need to, uh, address. So you're kind of removing this middleman system that our society is actually built on and going very much direct peer-to-peer, right?

• 14:38Speaker 01

  Be that, you know, between different people or people in companies, or even people in autonomous agents. And so that's where blockchain actually comes on the other side, right? Because it's effectively facilitating those type of transactions. So.

• 14:52Speaker 01

  There's effectively two pieces of this vision. One is how do we ensure privacy, verifiability? And there's like another aspect, which is like if this is how we see the world and how we interact with kind of the world is through this AI lens. A small alternation into that AI can effectively lead for you to perceive the world differently. Like the example I use is if you go into chat G P Q right now and like at the beginning say, hey, sadly convince me while we're talking about some, you know, whatever, the other topic of X, right? Let's say, you know, voting for somebody else that you don't like. And then you go through the conversation, right? It will start actually like working in kind of to change accents on how you think about things

• 15:37Speaker 01

  in the chat of on another topic. Now, this can be in system prompt right now, or this can be in a, you know, in a ways directly. We don't know. And so like, we also have this concept of it, what we call user on the AI, because like you want the AI to be yours, not theirs. Like privacy is part of this, but the other part is the model itself. Like we need to know what goes into it. We need to know and have a verifiability of when you run it, that there's no additional, you know, ways that it's actually been affected. So that's kind of a big part. And, you know, to do that, you need a different model, right? You need to build a movement, not a company, right? Because if you build another company doing that exact thing, right?

• 16:21Speaker 01

  You'll end up in the same result. So you need to build a movement around this idea where people are actually willing to contribute to build this more of as a common good, as accessible to everyone, while creating an economic engine behind it to actually power it, because, you know, this is actually requires a lot of financial investment and resources. And then on the other side, as this AI becomes kind of your interface, how does it interact with other AIs? And that's where kind of what we call intense, you know, A I T P and other components are coming together. So that's really kind of the vision we're working toward.

• 16:57Speaker 02

  Hey, we'll continue our interview in a moment after a word from our sponsors.

• 17:01Nathan Labenz?

  There's a lot there to unpack. I'm gonna maybe try taking it from what I think is the most foundational layer and kind of work up toward the gigabit future.

• 17:14Nathan Labenz?

  Most fundamentally, when we talk about, and obviously, there's like a lot of problems in society today, including, you know, loss of trust in institutions, right? So trustless is becoming a general description of society, perhaps, is something that certain technologies can achieve.

• 17:34Nathan Labenz?

  The avatar for a lot of people listening don't have a great handle on where with different blockchain schemes, the trust comes from. I know like in the original Bitcoin, the idea with the proof of work is that basically like, it's really costly to mint a new valid block. And so because it's so costly, nobody could plausibly rewrite the whole chain to present. You'd have to like, do that so many times in a row, that nobody could really ever get there. And so the canonical version is kind of the canonical version. That difficulty of extending means that nobody can kind of corrupt the history.

• 18:15Nathan Labenz?

  Okay, so that's unfortunately, though, pretty costly to run. I looked up, I think it's the Bitcoin is consuming the same amount of electricity as the country of Poland today, per my, you know, internet research. So we've now seen this move to proof of stake. And there you sort of have a lot less compute required, but you have to make sure you have a really thoughtful incentive design. People are Essentially putting their holdings of tokens up for, putting them at risk, basically, right, locking them up, putting them at risk and saying, I represent that I'm going

• 18:56Nathan Labenz?

  to do the right thing for the broader community. And if I don't, I stand to suffer some consequences, right, as a validator in the network. So can you maybe elaborate on that just a little bit? Because like, everything we're going to talk about ultimately, if I understand correctly, is, is built on the idea that we have a set of validators that we don't necessarily have to trust as individuals, but we sort of have to trust collectively. And we also have to trust that like, the incentives are such that they don't have any reason to defect from the stable equilibrium. So can you give us a little bit more about like, how, what is that stable equilibrium that you have designed? Who are the validators? And, you

• 19:40Nathan Labenz?

  know, what is the, in your mind, what is the ground that people can put their, you know, confidence in that like, everything else that we'll build on top of this, that we'll cover, you know, is really on a solid, albeit distributed foundation?

• 19:55Speaker 01

  Yeah, so that's a very good question. And I think just to contrast it with our traditional internet, right? Right now, you know, we're using like Riverside to record this, we went to riverside.fm. We actually relied on somebody to tell us Riverside is this IP address, that somebody is a distributed system, but it's actually has a single organization that effectively decides how it's done. Similarly, we have SSL certificates, right, that give us internal encryption. But again, there's actually certificate authorities who issue the certificates. And there's like, the authority that gave them this, right? So there's like, effectively points of centralization on the current internet, where you cannot verify,

• 20:39Speaker 01

  like, on itself, that this is correct. Now, when we think of blockchains, actually, the important way to think about it is from your perspective, right? Like, you want to verify that what you're interacting with is correct. And, and how hard is it for someone to effectively lie to you, right, to, and show you the fake kind of information? Like, that's what you're looking for. So as you said, for Bitcoin, effectively, if you're looking at some state, for example, you just received a bunch of Bitcoin, the cost of this is actually not true, is effectively

• 21:16Speaker 01

  Will be at kind of how much it costs to produce, you know, that number of blocks. Uh, and indeed it, you know, for example, for an hour to lie to you about an hour of history. It costs $2 million, right? So that's kind of the idea like now with proof of stake. It's a little bit different because uh, what in Bitcoin, it's possible that, you know, for example, Google has a lot of servers. They could be effectively like creating a fake blockchain, you know, that's, you know, days and weeks long, and they can present that and effectively lie to people and steal a bunch of money. Now, don't think Google is doing this, but like just, just to give you an example with

• 21:57Speaker 01

  proof of stake and specifically this idea of Byzantine fault tolerant consensus, you actually have this notion of finality where if indeed a two thirds of the stake is not uh, kind of um, not malicious, then um, they're following the protocol. And so there's no, there's no actual way to have a fake history or like to lie to you and, and you can kind of uh, verify this from the beginning, right? So like we launched block kind of domain net um, October 2020. And so since then you can verify that like, you know, every single transition was correct and all the rules were followed.

• 22:43Speaker 01

  Uh, so that's kind of the idea. Now, who are these validators and what, and kind of, and what is the stake, right? So, uh, I mean, the example for near, but it's, it's similar for other proof of stake. It's a lot of kind of both projects in the ecosystem, right? So we have, you know, a number of other companies in ecosystem that building on top of us. So they effectively have a vested interest in, in the security of this network. So they become participants and, and, you know, becoming a validator is permissionless, right? You can actually start a validator right now. It's also now on near, um, uh, you can run a very kind of lightweight node, um, like on an old, you know, somebody ran the node on like eight year old laptop recently.

• 23:29Speaker 01

  We also have, for example, Vodafone, NTT, and like other kind of entities like that running nodes. We have crypto exchanges who are participating in the network as well. So it's really about the participants in the network who become validators as well, because they have vested interest, they have some of the stake as well. So they have both like, kind of, business interest and financial interest in the, in the network, and so they become participants and validators. And then also just like, a bunch of community, kind of, we have universities, we have some individual developers who are running nodes as well. So that's kind of the idea. You have like, again, open network, people join, we had

• 24:12Speaker 01

  actually a pretty large number of new validators joining in, in past six months. And in turn, they put some of their money at stake, as well as maybe some other users can delegate. So they kind of say, hey, I trust this, this participant, and I'm gonna add more stake to them, effectively. And then there is a reward that's being split between all of them, to effectively incentivize them doing this.

• 24:38Nathan Labenz?

  This is maybe a really naive question, but how, if it is permissionless, maybe we could, let's talk about the importance of permissionless versus permissioned. Because I was also, I'm sure you're following the Tempo project, that, I don't know if you see them as a direct competitor, I don't, you know, there's obviously a lot of projects out there. Tempo, it has Stripe behind it, as I know you know. My understanding, reading through what they have said, is that they plan to start with, like, trusted validators, and then gradually become more permissionless over time. And I'm not sure if these, if this is one of those things where it really is a spectrum, or if it really is more of

• 25:20Nathan Labenz?

  a binary. I always say AI defies all binaries, but when it's something like, is it permissionless or is it not permissionless? That sounds like, how do you become more permissionless? Is there a way to become, like, if there's some permission there, it still seems like there's some. Anyway, I'm a little confused about, sort of... Exactly how that, how much that matters. Um, and then also, if I just show up with a node, like, how do you know who I am, who I say I am? Like, if I show up and say, I'm Coca-Cola here, you know, is your new validator. Um, do you have to like, meet me in real life to, you know, do you go to the Coca-Cola office and, you know, kick the tires on the servers to do that sort of real world validation?

• 26:01Nathan Labenz?

  Um, again, these are obviously noob questions, but, uh, knowing how much, uh, I might one day rely on my AI agents built on top of all this, I, I do wanna make sure I'm solid in the foundational understanding.

• 26:14Speaker 01

  No, no, this, this is all good questions. Um, yeah, so I think the permissionlessness indeed is a spectrum because, uh, and there's like maybe like few step functions. There's few networks where effectively there's like a very like set, specific set of validators. And what this has as an issue, right, is that, let's say, if this validator doesn't like you, for whatever reason, right, they can effectively not let your transactions or your AI agent's transactions on chain, right? They, they like can effectively censor you.

• 26:54Speaker 01

  And so that's a main challenge with like, uh, even if it's like not one, but like a subset of kind of permission validators is that, um, for whatever reason, they can effectively exclude you from the network and you cannot do anything. The benefit of this permissionless network is that you can yourself join as a node and say, hey, I'm actually gonna join the network and then I can, you know, if, even if they don't let me in, I can myself push transactions and, and interact with the network directly. Right? So, so that is like the biggest, uh, kind of change, which. For normal people, it's like, whatever, right? It's not, it's not that big of a deal.

• 27:39Speaker 01

  For maybe people who've been through, you know, I'm from Ukraine. And I've seen different types of, like, you know, I have multiple banks that like effectively closed up that I was a client of there. Obviously, it was a war and everything like, things like that are pretty important, obviously, for journalists, for, but more important is actually like for businesses. Like if I'm a business, right? If, you know, if I'm, this is a payment processor that I'm using and the competitor business is, is effectively controlling who is, which transactions are included or not. Or even there, there's also just like, you know, delaying, for example, my transaction, etc. They can affect my business.

• 28:22Speaker 01

  And so that's why like there's this idea of neutrality, right? You want kind of ability for everyone to join because then you know, like, there's no, no single party that can pull it on themselves or, or a coalition of parties, a cartel, you might say, that can pull on it. So that, that's kind of one, one side. The other question is like, okay, you want to join, you say, oh, Coca-Cola, you can actually do that. But nobody will believe you because there's no social signal, right? Now, if, if, for example, Twitter of Coca-Cola tweets like, hey, we just started the Near validator, here's our address, here's how to delegate, you know, let's go make Coca-Cola the biggest validator on Near, then that would actually provide a social signal.

• 29:04Speaker 01

  So the way you identify yourself to the network is effectively through public, private key cryptography, right? You'll say, you essentially say, hey, I'm, you know, this validator is this account. And then to become like, to, you need to put some near at stake, right? You need to put some money at stake. And so now you, you're a validator. So you don't really need to, you know, meet in person or, or do anything. You can be a cat, you can be an AI, it doesn't matter if you have kind of some capital stake.

• 29:37Nathan Labenz?

  Gotcha. So anybody who shows up, if they are willing to make an investment in the currency, which they have to have in order to then be able to put some value on the line, that's the, the proof of stake is that you actually own value and you're putting that value at risk.

• 30:00Nathan Labenz?

  Not at risk in a probabilistic sense unless you do something wrong, right? Like, the idea is, as long as everything you do is valid in your role as a validator, then your value is not at risk. But if the rest of the network finds you to have fake transactions or whatever, then they can take the value that you put forward essentially as collateral.

• 30:25Speaker 01

  Correct, yeah. And like, we have even a mechanism because so far it's mostly people been having misconfigured nodes, so it's not really been a malicious attack, but actually people just, you know, had whatever bugs in their setup. And so, we actually only, kind of, it's called slashing when you actually take their money at risk. So we only slash proportionally to how much stake you actually have, like, that participated in the malicious behavior. So if you had a misconfigured node and you only had like, you know, 0.1% or 0.01% of the total stake, you're only gonna get slashed, you know, whatever, a multiple of that percentage of your stake, right?

• 31:10Speaker 01

  So it will be very little. Now, if the whole, if, you know, a large percentage of network coordinating the attacking the network, then they effectively are gonna get slashed fully because, like, that is a percent multiplied by some coefficient will be slashed. So, so that's the idea is like, we don't want to punish people for misconfiguring things, but we want to punish if there's a coordinated attack for whatever reason, which, I mean, it can be that somebody hacked into a bunch of nodes, like, they had zero day vulnerabilities and done that, but like, we need people to have kind of responsibility over their security setups and kind of things, but we want to allow people to have like, kind of individual mishaps.

• 31:54Nathan Labenz?

  Yeah, gotcha, okay. So is there any, last foundational question on the security layer before moving up the stack?

• 32:03Speaker 02

  Hey, we'll continue our interview in a moment after a word from our sponsors.

• 32:08Nathan Labenz?

  Is there any risk of incentives changing in the long term? Like, right now, you know, the market cap of the near coin, last I checked, I think it was $7 billion. And that was, I mean, that's obviously a lot. But, you know, for people who have businesses or whatever, and then you also, I guess, would imagine, like, if there is some coordinated attack, probably that market cap falls real fast. So you could try to, you know, hijack, but also, what have you, what have you won, sort of, right? I imagine that's a big part of it. But in the, in the fullness of time, if we imagine like a massive, you know, AI agent economy

• 32:50Nathan Labenz?

  built on top of this whole thing, is there ever a point where it sort of could flip from it's not worth doing a coordinated attack to maybe it could become worth it? Because instead of 7 billion, it's 7 trillion or something like that?

• 33:06Speaker 01

  Well, I mean, if, if, if the network were 7 trillion, right? Then the attack will cost, for example, 30% of that. So if somebody can come by, you know, with a couple trillion to attack it. But then the question is like, that attack will not be able to extract much, right? So this is where it's like, again, the, from a perspective, like, the question is like, you as an individual, because attack is towards someone, right? Like, the system itself is, like, from its perspective, it's always correct, right? It follows the rules. If somebody violates the rules, they not in the system. The only way to attack is effectively to attack you, saying, hey, here's the fake information.

• 33:52Speaker 01

  You know, for example, you know, your agent, like, I sent you a billion dollars. You know, give me a billion dollars of services. And that's, I didn't actually send you a billion dollars, right? I just lied to you by creating, by attacking this, by kind of faking that there is actually a transaction that happened and it's been final. And so to do, to do that kind of attack, you'll need effectively 66% of stake. So if the network is at that point seven trillion, you need sixty percent of seven trillion, or whatever the stake percentage is. And so, and then you probably as, you know, if you're receiving a trillion dollars in a transaction, you're probably not going to be like immediately cool. I'll give you a trillion dollar of value.

• 34:34Speaker 01

  You probably can wait a little bit, make sure that there is nothing else going on in the network. So, so there's kind of a question of like timing, value and security that together work, right? So similarly how Bitcoin right now, Bitcoin is whatever, two trillion, right? The, the one hour attack is two million dollars. So if you're sending ten million dollars, you probably should wait a little bit longer than one hour to make sure nobody is attacking you. If you, and et cetera, right? Like the more you send, the longer you should wait to make sure. But kind of the idea is like as the value increases, like then you can also, you know, wait less for larger transactions.

• 35:15Speaker 01

  Um, so that's why it's like, it's a little bit, like people say, you know, the fifty one percent attack, and it's important to actually understand what this means. It means somebody specifically getting attacked. Like it can be an exchange attack, or it can be individual, a service provider, et cetera. It's not that the network itself is still correct, right? Like the rules, the rules are still followed. It doesn't matter, um, like which, which effectively so-called fork will end up being picked. Like from that perspective of that fork, right? You're still on the main thing. Blockchain is a little bit of this like abstract points of view, you know, indeed force wall kind of perspective is like, are you, are you in, in it? In it, you're always correct, right?

• 36:00Speaker 01

  Or you're outside and you actually, are you seeing the right thing or not? And like, how can you validate that you're seeing the right thing? But like for most, for most, for most use cases, right? If you, if you're spending a dollar, five dollars, ten dollars, hundred dollars to buy something, like effectively the security of the network is, you know, so much higher than it's not a problem and you can effectively accept the transaction right away. And so for, for normal use case of AI agent, it's fine. And then the reality is like, there's more, more challenges with indeed security of the like zero days of the code of the systems than like, than the kind of economic security of this.

• 36:43Speaker 01

  And we, we should talk about this because we have a whole thesis on that.

• 36:46Nathan Labenz?

  Well, maybe now could be the time. From my plan from here is to kind of work our way up to, okay, how do you train models on this? And there's interesting aspects of that. And then how do you build like, you know, business models on top of models and ultimately agents and, you know? So if you want to talk more about the...

• 37:08Speaker 01

  So on, on security side, yeah, like I think the important part to understand is as AI now getting better and better, it's actually gets really good at finding vulnerabilities, right? I've used it to back test effectively, you know, you take a like recently found variability, you run, you know, LLM over it, you effectively like, you know, you don't even need to over prompt it and it will find you, you know, some vulnerabilities in the code. Somebody found a Linux, I think, zero day with that approach as well. The, what this means is because cybersecurity is so kind of one-sided, right? Like it's, it's really hard to know that you don't have vulnerabilities and it's, but

• 37:51Speaker 01

  it's, it, you know, it's getting easier and easier to find them. And like, it's also true about, you know, social engineering and other attacks, like it's just going to get easier and easier. So we kind of need like a fundamentally different approach. And so the fundamental different approach is actually relies on math. Like, the only thing that you can rely on is actually the whole system is correct, given, again, your perspective. Right? So from my perspective, as a user of the system, I want to know that if I'm doing something, then it is actually correct, right? There's no vulnerabilities in the system or potential problems in the future that can happen. So I want actually a mathematical proof at the time of using the system that it is correct, given my requirements.

• 38:36Speaker 01

  So it's a very different, like, for those who, you know, there's been a lot of, like, it's called form verification. There's been a lot of research on this. But because it's so complicated, nobody actually doing it at the time of using services. You may be doing it, like, for example, when they were sending stuff to the Mars, right? They actually had, I think, form verification of the code for the Mars rover. But it was done once, like, it's effectively like, okay, we want to verify that it runs correctly. We run correctly given the specification. But obviously, if anything gets updated, if the requirements change, et cetera, now this is not usable. And it's because it's like, it's so hard, it required a lot of manual labor. Now, the benefit is with AI now, right?

• 39:21Speaker 01

  That's getting at, like, really good, even at winning gold at, you know, International Mathematics Olympiad. We can actually simplify that process of proving itself. And so what we believe is, we actually will need to rewrite effectively every single line of code in such a way that it's formally verifiable at the execution time. And with blockchain, this is a first, like, this is a first step. Blockchain, kind of, because it's the root of trust, we want to formally verify it so that you, as a user, when you, for example, sending money or depositing money somewhere, you can say, hey, I'm expecting to receive money back, at least as much money. And the chain needs to prove to you that this is actually going to be correct, and it's going to happen. But you can expand that to broader services, right?

• 40:06Speaker 01

  Like, if I'm using some service, I'm giving it my private data. I want it to prove to me that it's not going to leak it, right?

• 40:14Speaker 01

  And so for that, this is where we're going into, like, verifiable compute in a broader sense. We kind of need to both form verification and guarantees that this code is going to be executed over your data. So that's where trusted execution environments come in. So that's kind of like the position on like kind of zero day verification. Like we actually, we're moving in a world where this is gonna, I don't know if you like a couple of weeks ago, right, there was like zero day in all the iOS, right, and Mac and everybody were patching. We just had a massive attack on the supply chain in NPM yesterday in crypto space, like somebody injected code effectively to replace addresses in all of the tools.

• 40:57Speaker 01

  So there's like the amount of this kind of vulnerabilities is gonna just accelerate and we can't, we need a new approach. And so that's one of the like longer term research we're working on to actually solve that. But yeah, you need, you need a few more pieces. And so that's where we're getting into like, okay, well, how do we actually guarantee privacy, run these models, have verifiability, as well as indeed build a monetization, build the kind of financial engine that actually runs this.

• 41:25Speaker 01

  And so we're using effectively combination of indeed blockchain, cryptography and hardware. So as of about a year ago, Nvidia supports so-called confidential computing. So what this means is Nvidia chip itself had, had, connects to Intel and ARM and even AMD supported in such a way that even I'm as a operator, like the owner of the hardware, cannot access what's happening inside, like, so as a runner of operating system. So this is like this called confidential computing mode. And the interesting thing is it gives you both confidentiality and verifiability, right?

• 42:05Speaker 01

  It gives you a, like, it tells you like this code was run on this data and only this thing was done. And by the way, nobody else seen what happened there, right? And so what we're using this is effectively to now, I'm as a user, for example, want to run some AI workload, I can enter and establish enter an encrypted channel to this secure enclave.

• 42:30Speaker 01

  And run some AI workload there and then receive back result again into and encrypted. And I know that there's no other single party that were able to actually understand what I was doing there. And I also kind of received a verified certificate that says this model was run on your input and this is output. Right. And there's no way that something could have changed, you know, the degraded performance midway or whatever, right? Or they injected some prompts into the thing, right? It's like you have guarantee of the execution. Um, now, so that's kind of the basic primitive of this kind of confidential verifiable compute.

• 43:14Speaker 01

  And so, uh, there's few, few, few sides of this. One is, well, we now need a lot of compute, right? Hundreds of thousands of GPUs need to be here so we can actually provide it for, for lots of users. And again, the, like, one of the challenges with doing that is, you know, massive data centers, like lots of capex. Now you had a need of energy, it's all in one place, you have latency problems, you have sovereignty problems, so all the challenges. Now, the cool thing is, again, because the owners of compute don't actually see anything, we can, again, open it up and make it permissionless. We can let anyone with compute join. We have a way to verify that compute is indeed Nvidia and Intel chips.

• 43:58Speaker 01

  Um, because they're effectively signed the certificates. And now we can decentralize the compute itself, right? So we can have compute everywhere, uh, joining across the world. Uh, we can then route requests to the, you know, closest and available compute to you. You get confidentiality, verifiability. You now can pay to the network and network then distributes these rewards. Very similarly how validators, you know, receive rewards for providing, uh, validation of the blockchain here. The compute providers receive rewards for providing compute and then users pay for using inference. Uh, or developers paying for using inference.

• 44:39Speaker 01

  So that kind of creates what we call decentralized confidential machine learning cloud, right? And you, it's not just for AI stuff, you can also run, you know, arbitrary compute over it as well. Data processing, you know, agents, MCP servers, et cetera, et cetera.

• 44:55Speaker 01

  You know, stored memory and then encrypted for for users. So that's kind of the like the second piece, which again enables that vision of like, hey, how am I going to have a iOS that, you know, truly private and I know what's happening there?

• 45:13Speaker 01

  And so this already has an economy, right? We effectively bring in compute and usage. Now, the other challenge is indeed models and and data in a sense. And so there you have, okay, well, right now, if I'm a model developer, I effectively have two routes. I'm like open sourcing it and making zero money. Or I keep it closed source and then I need to actually procure compute, make sure that compute is, you know, very fast, like I trust the compute provider, meaning like because I'm uploading my weights to them, they can actually steal and and open source my weights or, you know, run them alternatively. So that's why only like big hyper hyperscalers actually usually partner with like, you know, Microsoft OpenAI, Amazon Anthropic.

• 46:00Speaker 01

  And and then on top of this, I also have like, there's also this maturity problem, whereas, you know, if I have, like, I'm just launching a new model, I don't have any users for it yet, but I kind of need to commit to a bunch of compute from this hyperscaler, like usually like two year contract. I don't actually know how much I need. If I get more users than I have compute now, I need to pay ton more. So it's like a very weird, you know, kind of socioeconomic question. Again, economic question. We have, we have tools to do that. And so the idea here is like now, if it's open source model, you upload it, or if it's closed source model, you can encrypt and upload it to our system. And so now, again, because all of the compute happens in this confidential environment,

• 46:44Speaker 01

  we can decrypt the model. Inside conventional environment, meaning again, nobody is able to access what's what's happening there and you can run on users' data privately. And so, and now when users pay, they pay both the developer and the model developer and the hardware provider the fee. Right, so you're effectively now combining model providers and model developers into this, offering them a way to serve their models without prepaying. They don't need to pay for the compute to serve it. They don't need to take users' data, they don't need to deal with GDPR, HIPAA, SOC 2, whatever, all the compliances, they just like push a model in, start making money when people use it.

• 47:25Speaker 01

  Right, so that kind of creates a new economic model for model developers. Similarly, you can have for content kind of creators and data providers, right? You can upload your data that can be used at inference time or potentially later at training time. And if it is used, you're getting paid from that transaction as well. You know, similar to Spotify, kind of like there's a subscription fees or kind of API call fees that get distributed to the providers. So that's kind of like how, you know, we see this like kind of, you have a cluster of services that effectively all run on the same kind of decentralized hardware like cloud, where you can bring models, you can bring data, there's users and developers

• 48:10Speaker 01

  consuming this, and it all kind of gets secured on this like blockchain-based marketplace that really facilitates it.

• 48:18Nathan Labenz?

  So, couple of questions there. We could go on a long time, I suspect. This is fascinating already. When I show up to the network, I basically just have to have an Nvidia chip of a certain type that has support for the secure private computing, right? That does suggest that there's like a important, I don't know if it would be right to think of it as a, I guess you could have a hardware vulnerability or a software vulnerability there. Hardware vulnerability from what I've, what research I've done, people seem to think

• 48:59Nathan Labenz?

  like, if you, you know, gave some of these chips to, for example, the Chinese government and let them bring the full power of their, you know, immense engineering prowess to bear on cracking that hardware. People, from what I hear, seem to think they probably could, but that it wouldn't really be economical to do it for much stuff, because it's like, really, you know, you're, obviously, these things are exquisitely crafted. So to do some sort of hardware level modification, you might be able to figure out how to do it. You're probably not going to find it worth your while to try to scale it to some, you know, large number of chips. So this would be like very focused, you know, kind of hyper-targeted attacks at most.

• 49:44Nathan Labenz?

  But then there's also the software question of like, how do we know that Nvidia has actually, is their stuff formally verified? That this is all like fully locked up? Like, how much trust and where are we putting trust there? And then part two would be, how much overhead is associated with this? I did one episode on use of zero knowledge proofs to prove that the model, and this was two years ago. So the framing at the time was, and it's actually still relevant because just in recent days, you know, there's been this discourse about Claude seems to be dumber for me during the day than the night, or whatever, and

• 50:23Speaker 01

  Exactly.

• 50:24Nathan Labenz?

  People are saying, oh, are they quantizing it? Or are they not? And they're saying they're not, but something happened, and it's not exactly clear what. The framing at the time was, well, if they use a zero knowledge proof, they can demonstrate to you that we ran the model we promised you to run. You're getting the value that, you know, you're contracted for. However, at the time, there was a lot of overhead associated with doing the actual zero knowledge proof. So it wasn't like, for 2% more, you can get a guarantee. I forget exactly how much, but it was like, it's a lot more to get the guarantee.

• 50:55Speaker 01

  It's like 10,000 to 100,000x.

• 50:59Nathan Labenz?

  Yeah, it was a lot. So, so is that, I mean, okay, so where do we, how much risk is there at hardware? How much risk is there at software level? And how much overhead is there to actually get these benefits?

• 51:13Speaker 01

  Yeah, so, so, given your touchstone ZK proofs, yeah. So there's effectively a spectrum of like, how to achieve verifiability and how to achieve privacy. Ziggy proof is effectively how to achieve verifiability. Like you don't get actually privacy from it. Somebody still needs to run the compute and then compute the proof. And so that other third party has all your data. But indeed, you get verifiability that like this exact model was run in a specific way. The challenge is, yeah, it's like, even in the, right now, the most performant way, it's probably like a thousand X slower than just computing it itself. It's a lot of overhead. It's a lot of overhead, especially with like, most of the time, you know, you're asking

• 51:57Speaker 01

  ChatGPT to do like dumb things. Everybody does.

• 52:02Speaker 01

  And so, uh, so it's kind of only makes sense for like really critical stuff. The other side is, there is this thing called, um, kind of fully homomorphic encryption, right? This is where you're actually doing compute where it is private for every participant. It is verifiable, but, uh, again, the, the overhead is massive, right? Something also in the order of a thousand to 10,000 X. But that is like a, you know, kind of fully private and verifiable. What we're doing with Secure Enclaves is finding, I would say, a, uh, pragmatic middle ground where indeed we trusting the hardware providers and, and, uh, and, and we can talk about like what does that entail?

• 52:46Speaker 01

  But the benefit is the maximum overhead we've seen in product, kind of in our testing and production is five, five percent.

• 52:54Speaker 01

  Uh, and so it's usually like from one to five percent overhead over running without this mode, but you get privacy, you get kind of this ability for, for network to permission to join and you get verifiability as well out of it. So you don't get, why is it's dumber on, on the whatever at night, at, at the daytime? So you will not get that on kind of our decentralized cloud. So one to five percent, um, the only other challenge right now, it only works on the level of a single machine, so eight GPUs. So, uh, we can fit anything that fits into eight GPUs, which is effectively all the open source models right now fit there. Uh, it's like over a trillion parameters fits there, no problem.

• 53:37Speaker 01

  But, uh, indeed you cannot do this across multiple machines yet. This is, this is going to be in the works right now. Now, on a kind of software and hardware security side, indeed,

• 53:53Speaker 01

  the kind of, there is a like potential for a hardware attack, but it, so right now, those chips are, you know, two nanometers, right? Uh, that is very small. And so like, we, we get into a level where, you know, it's effectively baked in into atoms of the, of the chip. And so, if you want to change something or address it, you're effectively like, I mean, I'm not a hardware like at that level. So I cannot actually approximate, but yeah, the cost of attack is extremely high.

• 54:27Speaker 01

  And the benefit is you're going to get some maybe random requests from somebody asking, you know, which cat to pick, right? So like, because you like, you, you, you will not be able to target specific use cases in our system.

• 54:45Speaker 01

  Even if, if something happens, there, there's indeed the other challenge, which is Nvidia kind of, because Nvidia effectively certifying it. So they, their key management and their process needs to be indeed correct. And so this is something that, you know, we would love to kind of help secure. I mean, obviously they're already using kind of top line security, but that's where, you know, formalization and using blockchain as a security mechanisms can definitely help.

• 55:15Nathan Labenz?

  Okay, cool.

• 55:18Nathan Labenz?

  So let's build our way up the stack. So we want to now create user-owned AI. We, and I've seen you have a plan to train a more than one trillion parameter model.

• 55:37Nathan Labenz?

  Estimated cost of that $160 million.

• 55:40Speaker 01

  Well, it's dropping down since we talked about it. So yeah, that's a good thing.

• 55:45Nathan Labenz?

  Cut it in half.

• 55:46Speaker 01

  Yeah.

• 55:48Nathan Labenz?

  VO3 just got cut in half, so we'll apply a half factor on this too.

• 55:55Nathan Labenz?

  My, I guess, what I understand there is Dario just kind of talked about this where he said, you know, yeah, we're like burning a lot of money and in a way we're like burning more money every generation of scale up. But in another way, if you look at each model as its own venture, each model is profitable. And so what's weird is like, we might make, you know, however much on, it might cost us 100 million to train Claude 3 and we make a billion. But then it costs us, you know, 2 billion to train Claude 4 and we'll make 10 billion, but it's going to cost us 20 billion for the next one. And so that's a weird development cycle, to say the least. But it seems like you are basically kind of taking that reality to its natural conclusion and actually like

• 56:40Nathan Labenz?

  planning to structure sub-ventures for each of the models, right? So tell me exactly that sort of economic schematic there. And then also, I understand that you're applying this secure computing to the training process where people can contribute training data in a privacy preserving way. But that's a really interesting challenge, I would imagine too, right? Because like, who's providing what data, the data mix? You know, if you talk to people at the frontier companies, they're like, yeah, you know, it's like baking a cake. We've, we're constantly experimenting with different mixes and, you know, a little more of this and this other behavior gets degraded and whatever.

• 57:23Nathan Labenz?

  So how can you have a privacy preserving, even, even, you know, understanding on some naive level that you can do it from a, you know, computation standpoint? From a like getting a model to perform standpoint, how can you have privacy preserving mechanisms on the training data side? And have any sense of like, what the hell you're going to get out on the, you know, when the model comes out of the GPU oven, so to speak?

• 57:53Speaker 01

  Yeah, so all great questions. And indeed, it was actually funny that they are positive. And I'm like, yeah, exactly, exactly the model. So, yeah, so I think like, maybe just to give you a little bit of a roadmap, right? So kind of building out this decentralized kind of computing, starting with inference. So now it effectively already gives an ability for people to consume it, get confidentiality, get verifiability. And then next step, it offers people who are building models, including potentially the existing companies that build foundational models, they can upload them, they encrypt, nobody can see them, they don't get leaked. But now they can be used across this decentralized compute.

• 58:37Speaker 01

  Everybody can verify and know that they're using this exact model that this company has posted, you know, they effectively on-chain said, you know, Cash Company, you know, tweeted about it, etc. You can verify kind of the, indeed, you use, you know, let's say this Centropic Claude, and, you know, it doesn't degrade in the evenings, right? So that's kind of steps, first steps. Then, okay, well, given we want to kind of coordinate this research and development of model, and importantly, like, it's not, you know, we don't get to build one model, right? That's not, not the point. The point is to build a process that kind of creates models that are state-of-the-art and kind of user-owned, and everybody can inspect how they're done.

• 59:22Speaker 01

  And so the important part is like how any frontier lab is working is actually a number of benchmarks that effectively goes through the whole flywheel of building model, right? Like, everybody building different parts of it, evaluate how their kind of change and innovation benefits, you know, the scaling law, so you test the different sizes, and then kind of all these ideas gets kind of all accumulated, and then you do like bigger runs, right? So replicating some of that as well, where we effectively now, again, using this compute network, we can say, hey, actually, anybody can come in and build a new benchmark, right? So you can say, hey, I'm going to build a benchmark that on...

• 1:00:06Speaker 01

  Model predicting future or answering questions about, you know, astronomy or ocean or whatever. Different things that people care about. Including you can actually have enterprises who say like, hey, I'm actually care about how this model does on this, you know, deep fake detection, right? And I will not actually, I'll upload my data, but I'll keep it private. So nobody can have access to data, but everybody can run evaluation over it. Right? So you can now offer effectively both benchmark creation and evaluation service where you verifiably can check that this model, the closed source model that runs on disclosed data, on private data and gets this result.

• 1:00:51Speaker 01

  Right? So again, effectively now creating this marketplace of model builders and, you know, benchmarks, which can be representing companies or specific use cases and finding what's the best model and the best ideas are. So that already on itself is really interesting product because again, right now you cannot, you cannot do that. Like you cannot actually go and benchmark some intermediate model from, let's say, OpenAI. They will just not give you the weights. And if you don't want to upload your data to them, you cannot even benchmark their current model.

• 1:01:28Speaker 01

  But this also opens up an environment where anybody like, you know, you're sitting somewhere in whatever you want to train a model that specifically targets whatever ocean benchmark. Maybe you have special data for this. You can also do this, upload it and get rewarded when people are using it. Right? Through the previous step. Now the next step is like, well, actually I have a lot of interesting ocean data. Or maybe I'm in university and I know how to train the models, but I don't have compute. And so that's the next step. Okay. Well, we're actually going to offer you a way to do first fine tuning, right? So like you can take existing base model, fine tune it. And again, you can fine tune it even on private data. So I'm as a maybe like institute that collects a lot of ocean data.

• 1:02:13Speaker 01

  I have the data, but I don't have the know-how how to fine tune this properly. So I can partner up with someone who can then fine tune on my data without seeing it. Right? They can effectively deploy.

• 1:02:26Speaker 01

  Like, hey, this job into the continental computing. The model, the resulting model then, you know, splits the revenue between the kind of the developer and the data provider. So you can have all of those different combinations on top. Fine tuning is, I would say, easier to do when you, you know, maybe not have have exact access to the data. The training is, yeah, where if we're talking about like full, you know, fundamental training from scratch, that's where things get really complicated. And so that's our final step where we're actually putting all of those pieces together. And there indeed, you need like, you know, all of those pieces need to really work well, where you effectively can kind of iterate on different pieces and sizes and and then

• 1:03:11Speaker 01

  really combine it all into a final model. I would say like to your point about kind of private data, the things there, there's effectively a few things that will need to happen. One is that there are going to be a whole separate process of data filtering, curriculum construction, et cetera, that, again, you can actually benchmark it. So like, I'm, I can build a curriculum builder, right? Like, pick different articles and different pieces of data, construct the training set out of it. And so you can actually evaluate how good that, like, system is by training, you know, a few sizes of the models.

• 1:03:52Speaker 01

  And so people can actually compete on whose curriculum builder is better, right? And so then, when you pick the best curriculum builder, then you run a larger run with, you know, more parameters and more data. So, so I would say like, it's, it's going to be a kind of a very iterative process, right? It's the same way it is right now in, in, in inside frontier labs, because indeed, we're like, we're kind of experimenting with all those things as we go. Now with specifically private user data, it's a bit more sensitive and complicated. And but again, the biggest benefit we can do is like as you contribute your data.

• 1:04:33Speaker 01

  First of all, you can get a reward from the the model kind of outcome. But you can also like we can have effect as a filtering procedure. Because one of the things is like, there's going to be a lot of PII information, your phone numbers, etc., that you don't want in a model. And it's also like not actually helpful for the model for the most part. Right? And so but this filtering procedure will be like open source and everybody can effectively. You know, not everybody will expect, but some developers will inspect like and say, yes, this is actually valid. Uh, like way to filter data. So so there's like few different components again that that we can do to make sure that both data is cleaner.

• 1:05:14Speaker 01

  And then how it gets filtered and and kind of processed into the curriculum of the training. And probably, you know, a bunch of your, you know, personal chats are probably not that actually useful for training these models. There's like other types of data that's more useful for those things. And we see like in general, we're moving to more synthetic data anyway. So there could be a lot more of of that. You can still and importantly, we still have, you know, the crowdsourcing and data labeling that's run through this decentralized network as well. Right? So we actually have both like near crowd, which is kind of original project we had for crowdsourcing. Like crowdsourcing, I mean, like scale AI style as well. We have a public AI, which is another like data label crowdsourcing.

• 1:05:58Nathan Labenz?

  How much overhead is associated with all that? I mean, you mentioned like relatively low at the hardware level for the confidential computing. But how much is there in terms of just like the software setup? Right? I mean, it's not super easy to set up a training cluster. It's not super easy to manage data pipelines. It's not super easy to, I mean, you have a marketplace, but you know, even going out and hiring a bunch of people to create data like that's not super easy, even in in sort of base case. Interested in how much overhead there is in doing it the privacy preserving way? And then also at the sort of model.

• 1:06:40Nathan Labenz?

  Like who makes these decisions, you know, in terms of like, what's better, right? Because I think we have benchmarks galore. And yet it seems like they all, you know, I mean, there's been a lot of discourse about this in the last couple of weeks, too. Like, you know, people swear by evals, people say evals are useless, you know, certain, you know, famous projects are run entirely on vibes. It would seem like the one thing that you'd have a really hard time putting into this frame would be like, who do we actually trust to like be the vibe checker on this? And how do we make sure that, you know, in addition to maxing out on these various benchmarks, the thing actually, you know, is nice to talk to, right? Which is, people do value that a lot.

• 1:07:25Nathan Labenz?

  So yeah, I guess overhead in terms of like, all the setup, the config, just, you know, moving things around, coordination, I guess broadly. And then how do, is there, is there a governance mechanism? I know you have governance projects as well that are pretty interesting unto themselves. Are you bringing those to bear on like, who decides what version of the model is actually good? Or, you know, what training recipe gets scaled up? Or is there some tastemaker, you know, that still is involved somewhere?

• 1:07:55Speaker 01

  No, so those are all the good questions. And I think the high level answer is like, we're effectively creating a marketplace of the, of all of these decisions, right? So instead of trying to prescribe like, okay, Ilya will be the tastemaker and like, decide you get the compute and you don't. Instead we say, hey, we're creating a marketplace where anyone can do this and probably Ilya will do it as one of the participants, but there can be anybody else doing it as well. And, and you kind of have the rules and everything known and you have like visibility and trace and you can also build on top of each other instead of, you know, right now you're effectively competing. And so, yeah, we kind of touched on it, but yeah, like for, let's say, I'm going to say,

• 1:08:38Speaker 01

  hey, I'm going to train this, you know, 1.4 trillion model. I'm gonna issue a token for it, right? So I'm gonna create a new effectively value capture system for this specific model, right? I can now distribute it for the compute. I can distribute it for data. I can distribute it for research and work, right? In this case, you know, I started it, so I probably will be the test maker and picking and selecting which pieces I wanna put together. We train it, you know, if it gets used with, uh, and then, you know, we distill it, fine tune, et cetera. The revenue gets distributed to the token holders, right? Now, in parallel, somebody is like, oh, actually, I'm gonna train, you know, 70B model in this way, et cetera.

• 1:09:22Speaker 01

  They're gonna use exactly the same framework, exactly the same system, do this, you know? And again, uh, because 70B is smaller, cheaper to use, but then it also runs faster. So like, it's gonna get used in different use cases. So you can, you can like, you're really using this more of like as a platform where different people can kind of put together these pieces really easily. And you don't need to have, you know, you don't need to raise like, huh, whatever, $10 billion to really kind of do this. Uh, you can actually raise like smaller and specifically for compute data, et cetera. And you already have a lot of building blocks as well on this. So that's kind of the idea.

• 1:10:00Speaker 01

  It's like really creating more of an environment where we can open source, but then also we kind of, and we use these things while kind of have this economic flywheel, right? Where things are like generating revenue by usage. And, and importantly, kind of learn from what works and have this inspectability of what actually went into it, right? Which I mean, in some cases, people say, hey, no, actually, it's all private. I will not tell you what actually went in, right? But then people know that this is private and they can decide if they wanna use it or not. And, and kind of, but also like, here's a verifiable benchmarks. Again, right now, when somebody posts benchmarks, you have no idea how they got these numbers, right?

• 1:10:42Speaker 01

  They have, they don't, like, usually don't even post the prompts they use for the benchmark, so like. Like you can have, you know, a two-page prompt that like answers half of the questions and get and get the, you know, way better answers from that. So like this here, you can, you know, you have the full traceability of the whole flow. You know, kind of what went in, maybe it's private, but at least, you know, like 100 million users contributed their private data and that's what it trained on. And then it gets, you know, this, this is also benchmark. And then you can actually, you know, vibe, vibe, you can contribute your vibe benchmark as a benchmark as well.

• 1:11:22Speaker 01

  And so like on, you know, on Ilios vibe benchmark, it gets this much as well. For example, so like you can kind of end on whatever. Uh, on this company's benchmark, right? Like Bank of America, you know, their test it gets as much. So you can have all this different information and, and decide based on that. So it's really about like kind of creating transparency and, and marketplace for this while preserving privacy for the end user and, and kind of companies.

• 1:11:55Nathan Labenz?

  And all of this traceability, verifiability, to rewind for a minute, all rests on the fact that somebody has locked up their nearcoins with the assertion that they are putting forward valid transactions, which in this case would be more like log statements, basically, right? That.

• 1:12:16Speaker 01

  It's, I mean, it's a bunch of stuff. Yeah, like because it's payments is, yeah, logs, it's kind of traces of like trainings, et cetera.

• 1:12:25Nathan Labenz?

  How big does this database obviously gets like pretty huge, right? So, um, where does the database live long-term? Like it becomes too big for a node, a single, you know, if I want to just connect, right? I can't download the whole history of like everything that's ever happened. It becomes too much, I assume.

• 1:12:46Speaker 01

  Yeah, so near, I mean, one of the kind of underlying things is that nearest shardet, meaning it's like every single node contains part of the data and part of the process, part of the transactions and part of the execution and that, and it can continue, can continue expanding number of shards. And there's like the, that's why we actually need more valid data to join because the more data, more shards there is, you need more nodes in the network to actually participate.

• 1:13:13Speaker 01

  Very similar how Google, Google, you know, like databases work or whatever, metas, you know, they don't run out of capacity. It's like, no, no, no more users here, please. It's like, no, no, we just keep adding more, more computers to run stuff in parallel.

• 1:13:30Nathan Labenz?

  Yeah. Okay. So this has been fascinating and we've built up the sort of, from the premise that we want user-owned AI, uh, this is kind of, you know, all the layers that we have to put in place to finally be able to get to the point where you could have a community, uh, based process of all this contribution and kick out a model and even get paid back, uh, with inference.

• 1:13:57Nathan Labenz?

  As you mentioned, though, you could also just have somebody train their own model and kind of bring it into the system. So, you know, regardless of like where the model that people want to use comes from, let's change maybe gears toward what does AI need from crypto? And what does crypto need from AI? My sort of very high level styling, and then I want you to, you know, go 10 minutes on this if you want to, is like, AI can put the smart in smart contract. I've always kind of thought like, geez, these smart contracts aren't really that smart, but if there was an AI in there, maybe it could be smart. You know, what we, what we really need, I think, is like smart dispute resolution in a lot of cases, uh,

• 1:14:40Nathan Labenz?

  and affordable dispute resolution, because we have, at least in some places, we have courts, but it's like not convenient to access them by any means. Other places, there are no courts. So we need like smart, affordable dispute resolution that AI seems like it could potentially provide into the context of these smart contracts. And then on the other hand, AI needs like really, uh, rock solid guardrails, uh, because we obviously have many, you know, the AIs themselves like can be tricked and can go off the rails and make mistakes and even scheme against their human users in some cases, as we're increasingly finding out. Uh, so we need like very.

• 1:15:22Nathan Labenz?

  Very reliable checkpoints, basically, in workflows to know that like you can go explore, you can do whatever, but like if you're gonna do this, you know, you gotta have a certain, I almost think of it like the Jurassic Park thing, you know, where they engineered the dinosaurs to not be able to survive without the special food from the, from the park. Of course, we know how that turned out, but not to generalize from fictional evidence, but we do want the AIs to sort of be dependent on like a really rock solid validation that like this is okay for you to do. So that's my, you know, high level premise. Give me the fleshed out version.

• 1:16:05Speaker 01

  Yeah, so I mean, both of the sides are definitely very interesting and kind of both directions we're working. And to be clear, like when I say we, we have like a large ecosystem of different projects and kind of, as I said, it's a protocol that combines a lot of different projects and contributors who are actually working towards this. So one side on the dispute resolutions and putting intelligent, like smart into the smart contract. So we effectively have two kind of parallel threads there. One is what we call near intense.

• 1:16:45Speaker 01

  So near intense is this idea that as we move into the AI world, well, actually you or your AI will just express an intent that something want to be achieved or done, right? This can be a simple, okay, I want, you know, some pizza tonight for dinner. And as complex as like, hey, you know, I want to construct a new building in this location, you know, it needs to be like this, et cetera, et cetera. So they're, you know, high qualifiers. Now that requires a number of steps, right? It needs to find somebody who actually will do it, right? So you need a whole discovery process. So how do you discover other AIs or non AIs, right? Like it can be businesses as well. We're assuming everyone will have

• 1:17:29Speaker 01

  like some AI like interface. How you discover the AI that will actually, you know, do this intent? How do you actually come into a commercial agreement with them?

• 1:17:42Speaker 01

  And you know, this can, this commercial agreement can be like, hey, I need to rent this car, you know, to get to airport, right, for Vemo, or it can be like, hey, we need the six months project to construct this thing. And then if something goes wrong, how do you actually deal with it, right, so dispute resolution, uh, indeed. So that's what we're building as kind of this intense protocol. And, and so this is idea of, you know, it, it starts with people, organizations and AIs effectively collaborating and, and kind of, uh, doing these things. And so that's a fundamental protocol. Uh, it includes indeed AI-based dispute resolution. So you can actually run AI over if something went wrong, like the, you know, the taxi didn't,

• 1:18:25Speaker 01

  didn't actually drive you where you were going or didn't arrive at all, you know, how to deal with that. AI can look after this, cheaply analyze the situation and decide to do something. Now, if, if parties are still not satisfied with that resolution, then you go to court, right? So effectively like reduce the cost of disputes way, you know, make it way cheaper, make, like it can use a lot more information as well right away because, you know, you have your own chain information, all your previous actions, it can run over your private data, over both sides' private data, uh, very effectively, right, without, um, revealing this private data to anyone. It's effectively like if, you know, you know, the discovery process, like you can run AI

• 1:19:07Speaker 01

  on, on discovery of both sides without this costing, you know, millions of dollars. And so you can effectively minimize the cost of this process very low. And then again, if, if it's not kind of, um, acceptable, then you go to court. But again, a lot of work already done, so this is all can be reused. So that's kind of one piece and we kind of, the go-to-market for that, we actually started in more crypto side. So we're like trading and other things and we're expanding it into e-commerce and other use cases, kind of adding more and more services into this. The other side is actually this autonomous agents. And so right now, smart contracts idea was it's effectively autonomous being, right? It exists independently of any person or entity.

• 1:19:52Speaker 01

  It has money. The problem was it was like it was just a piece of code and usually very small piece of code. So it was very dumb, even though it was called smart contract. What we've done is because we have this decentralized kind of like compute with AI. We actually combine this two things. So in the decentralized compute, we're running the AI brain. And then it has the smart contract to actually execute actions and execute this intense. So it actually can go and do things. So we have to, as an example, one agent that was given $10,000. Launched, we cannot stop it. We cannot take money back. We cannot do anything with it. And it's effectively trading it based on sentiment on Twitter, right?

• 1:20:35Speaker 01

  And so it made like whatever four grand doing that from the $10,000 it was given. Right? So this is just like an example, but like you can imagine a future of businesses where you effectively have businesses like this. Which are operated by AI. You know exactly what it's going to do. Your bylaws are the prompt, right? Of the current, like, there's no, you know, person that does random things. You can give it like shareholders can effectively give it feedback. And so you can apply all this kind of, you know, training data to improve it. It's or decide to upgrade the model and things like that. So that's kind of the the framework of what we call autonomous agents. Intent is how they're going to interact with the commerce, right? Themselves and with other AI agents.

• 1:21:16Speaker 01

  And AITP is really the the protocol to communicate, to do the communication, to to do intense. And then the other side of this is, yeah, like, how do we govern this? This is where indeed we have a whole governance protocols. But even before that, I mentioned the formalization. I actually think one of the pieces of this puzzle will be this where, let's say you're using some AI service. Again, you want to it to guarantee to you before it even executes on your data. That it will not as, you know, not escape and go kill a bunch of people, right? So you kind of like, well, effectively, we need to create this like environment in which we can verify what AI runs.

• 1:22:00Speaker 01

  But I don't think we can like. Like my approach is like, I think it's really hard to effectively be a thought police for even for AI. But it's way easier to ensure that they don't do anything outside of the sandbox that violates whatever the rules you apply when you call it. And so that's kind of, again, using our this verifiable compute, we can actually guarantee that that like, as you upload this model, you can attach effectively set of things that you don't want this model to do. And it can stop it from doing that kind of by running it on the output and verifying if it's correct or not. So that's kind of the idea is like to leverage some of that.

• 1:22:41Speaker 01

  And then you still need obviously governance and kind of ways for people to come together and actually decide what it is that we don't want, right? Like, for example, probably everybody agrees we don't want bioweapons. But maybe not everybody agrees on some other things, right? And so kind of there needs to be some process through this. There, I would say we have a bunch of experiments, including actually using AI in the governance. So we have AI senators effectively, like that people can vote for. And then they go and actually make decisions. And so you kind of, because like getting people to vote is hard. And getting people to go in like deep, deep research, a bunch of staff to decide how to vote is even harder. And so you can actually have the delegates to be AIs as well.

• 1:23:26Speaker 01

  And you kind of just give them, you know, you vibe with them on like what you think is your opinion. And they kind of construct from that their platform. But you remove this principal agent problem, which you have with other people. When you, when you vote them, they still have their own agenda, right? And so like they may not do what you ask them to do. So, so that's kind of like all the pieces where we're like, we're combining all the blockchain and the AI side like from different directions and really creating protocols and kind of verifiable compute and privacy. And really just you can use it in all those different ways.

• 1:24:02Nathan Labenz?

  One thing that definitely kind of makes me nervous is the idea of these autonomous agents that we can't take back. You know, on the one hand, you sort of want to have reliable follow through on, you know, what they're supposed to do. You don't want other people to come in and tamper with them. So I understand why there's some allure to it. But then I'm also like, geez, you know, we have a hard enough time telling what one AI is going to do. What happens when we unleash a billion agents on the world and they're all autonomous and we can't take them back? And we've, by the way, like delegated governance to AI senators. You know, some might wonder, like, are we not creating the recipe here for the

• 1:24:43Nathan Labenz?

  AIs really to just take over and run everything? And so, yeah, what is your positive vision for the future? And how do we make sure we don't like, you know, take a premature off ramp on the road to that vision?

• 1:24:58Speaker 01

  Yeah, so I think there are going to be a combination of things where like AI will become more of the economic force, right? And be that kind of conduit and medium. And then on our side, we're also going to be kind of becoming more of, you know, governance and curators of that on our side. And so one of the kind of indeed, we're like putting AI into governance, but also we're kind of putting governance into AI where, I mean, formalization is one part, but the other part is also if you have this autonomous agent, well, actually, if it wants to do like this intense with in physical world, it actually needs to be in some jurisdiction. Like this intent is effectively legal contract.

• 1:25:41Speaker 01

  It needs some backing with the actual jurisdiction and that jurisdiction has an enforcement over this agent. And so we actually have an infrastructure for jurisdiction, for actual court to enforce staff on AI agent as well. It's all visible. It's all on chain. It's all kind of transparent, but like you're effectively combining this, you know, what we have traditionally with courts and jurisdictions and legal frameworks and with AIs who can actually interpret all that and, you know, reason over it and make sense of it as well as this protocols to kind of facilitate it. So like, kind of the way I see this is all of these things are going to keep fusing together and like really becoming this kind of new conduit of how our society operates. But indeed, like, I think a lot of the

• 1:26:27Speaker 01

  Like, there are going to be a lot of exciting innovation and also, you know, we, our role as well will kind of keep changing because like, we're going to be more of a like participants in this network as well, but you know, through the AI as a lens, right? How, right now, we're, you know, through the lens of computing, we're already interfacing each other a lot more than, you know, the physical world many times. Uh, so I think like that, that is going to continue also increasing.

• 1:26:55Nathan Labenz?

  It's a fascinating

• 1:26:58Nathan Labenz?

  set of layers of foundational technology that you've built and the possibilities for it are really pretty dizzying. So, uh, I want to already propose a part two for this, uh, podcast, maybe get back together before too long.

• 1:27:17Nathan Labenz?

  And in the meantime, I don't know if you've ever read the book Liquid Rain, but it's my favorite kind of crypto AI intersection. And I'm not like super well read in science fiction, um, but it's a funny, but I think increasingly prescient looking, um, and positive view of what that future could look like. So I definitely recommend that.

• 1:27:41Nathan Labenz?

  I know, I know you gotta go, uh, put me on the calendar somewhere for part two. Any closing thoughts that you want to leave people with today?

• 1:27:48Speaker 01

  No, I think, I mean, again, this is a movement. We want people to join and kind of contribute across the board. So, um, if you're developers, it's time to stuff to contribute and build. If you are, you know, planning to kind of, what, what does it looks like in your life as AI? I think this is really an opportunity to engage in something that's really yours and you can actually own part of it. And again, contribute in many ways from kind of creators and be receiving it to tastemaking to many other, other aspects.

• 1:28:21Nathan Labenz?

  Sounds great. Paul Sukin, founder of NEAR. Thank you for being part of the cognitive revolution.

• 1:28:27Speaker 01

  Thanks for having me.

• 1:28:30Nathan Labenz?

  If you're finding value in the show, we'd appreciate it if you'd take a moment to share it with friends, post online, write a review on Apple Podcasts or Spotify, or just leave us a comment on YouTube. Of course, we always welcome your feedback, guest and topic suggestions, and sponsorship inquiries, either via our website, cognitiverevolution.ai, or by DMing me on your favorite social network. The Cognitive Revolution is part of the Turpentine Network, a network of podcasts where experts talk technology, business, economics, geopolitics, culture, and more, which is now a part of A16Z. We're produced by AI Podcasting. If you're looking for podcast production help for everything from the moment you stop recording to the moment your audience starts listening, check them out and see my endorsement at ai podcast dot ing.

• 1:29:15Nathan Labenz?

  And finally, I encourage you to take a moment to check out our new and improved show notes, which were created automatically by Notion's AI Meeting Notes. AI Meeting Notes captures every detail and breaks down complex concepts, so no idea gets lost. And because AI Meeting Notes lives right in Notion, everything you capture, whether that's meetings, podcasts, interviews, or conversations, lives exactly where you plan, build, and get things done. No switching, no slowdown. Check out Notion's AI Meeting Notes if you want perfect notes that write themselves. And head to the link in our show notes to try Notion's AI Meeting Notes free for 30 days.