{"id":"v_Pcmz_RClQ","title":"Before Agents Control Capital: Security, Risk, and Autonomy","description":"What must be true before autonomous agents are trusted to control real capital, from security and custody to incentives and risk.\n\nNEAR is the unified commerce layer for assets and agents.\n\nWebsite: https://near.org/\nNEAR AI: https://near.ai/\nIntents: https://near.org/intents\nChain Abstraction: https://www.proximity.dev/\nSharding: https://nearone.org/\n\nFor the devs\nGitHub: https://github.com/near \nDev Docs: https://docs.near.org/\nDevhub: https://devhub.near.page/ @NEARDevHub\n\nConnect with socials\nTwitter:   / nearprotocol  \nBlog: https://near.org/blog/\nMedium:   / nearprotocol  \n#NEAR #nearprotocol #NEARAI","channel_name":"NEAR Protocol","published_at":"2026-02-28T00:00:00.000Z","thumbnail_url":"https://img.youtube.com/vi/v_Pcmz_RClQ/hqdefault.jpg","transcript_text":"Hello and welcome everyone also from my side to NEON 2026. Uh I'm Alex Shara from the Neo Foundation and in this panel we're going to talk about what happens when agents move real capital and what it means for security risk and autonomy. Naturally, my team made me moderate this panel because I'm European and we Europeans love talking about regulation and risk. But don't worry because we have two absolute legends joining me on stage today. So, it will be fun. These guys don't need an introduction, but I'll give them one anyway. Let's start with uh the gentleman on the left. Our first panelist is a former poker pro turned cryptov Keshi, currently managing partner and self-described chief hypeman at Dragonfly. Just last uh week, uh he closed a fresh fund uh $650 million at a time when Fortune magazine calls the blockchain VCscape a mass extinction event. Seems like didn't get that memo. So, Hassip, welcome. Thank you. Our second panelist to my right um started coding as a kid in Silicon Valley. Um has founded a bunch of different startups and has never really stopped since. Uh Archan Sethi is currently the co-CEO of Kraken. uh the chairman and co-founder of tribe capital and among many other startups founded terminal AI uh which uses AI uh software to basically do quantitative due diligence. So if anyone is going to replace by AI agents in this room, it's not going to be Archin. He's probably built it. Archin, great to have you. So we only have 20 minutes today. So let's dive right in. I'd like to start with you, Hip. So I believe last year you called during the agent hype uh you called agents basically chat bots with memecoins attached. Y and you also said that you believe uh agents that become real economic actors are probably more than a year away. But also last week uh you published an ex article saying that crypto was never really made for humans but made for machines and you described an envision a future of self-driving wallets. So my question is from your perspective what needs to happen before agents become real economic actors and what would made you uh deploy meaningful capital into agents? Yeah, so it's a it's a great question. Um crypto has a way of getting ahead of itself on every technology. So we got ahead of ourselves many times with respect to how good crypto was. You know in 2017 2018 we were convinced that it was 2026 level good. It obviously wasn't back then. In the same way when we saw you know things like uh AIXPT and truth terminal we were like great AI is already here. It's already using crypto. Everything's ready to go. Let's start trading this stuff. And in reality that was child's play compared to where we can now start to see glimpses of where this technology is going to go. So having agents that can actually manage money in thoughtful and capable ways, we're still not there yet, right? If you if you give your money to an open claw, it's going to do something stupid. It's going to do some smart things, but eventually it's going to do something stupid. And the reality is something that works with money 90% of the time is unusable for actual economic activity because you know if you had an intern that like 10% of the time they'd buy you you know tea and coffee and a donut and then 10% of the time they would go on like go on a shopping spree and you know buy themselves a computer you'd be like okay this intern's an idiot and they're fired. But that's right now where the models are at. So the the way I described in the past is that you should expect there to be a very strict cutoff below which there really is not much economic activity happening. So it's like at 90% efficacy that's really really impressive. But it's still nobody's going to use these things actually within their economic systems. Even at 95% even 95% is still not good enough. So it's a lot of like nothing nothing then something and then everything. And right now we're still in the nothing phase, but we can see it's starting to poke through. And it's only really crazy people or people running very small experiments that are actually going to be willing to give their AI agents money today. Um, but it's also those people who are going to be the first to discover, aha, when you add these guard rails and you create these constraints, now it's actually good enough to be behaving in the way that we might give it that kind of control. So it's like the person who you know um like in SF we are one of the places where Whimo started to operate even before they were at the level of battle testedness that they could be fully autonomous and actually just you know kind of let it rip the way that self-driving cars are starting to be today. Um so the question is like what's going to be the equivalent of SF for agents or for self-driving wallets that are going to have that level of autonomy over your money. Um so you want to be watching for where those experiments are taking place but it's not going to happen everywhere at once. It's going to be in these small pockets of experimental people with very high risk appetite and very good guardrails. Great answer. So basically slowly then all at once, right? Exactly. Um Arjun, you're the co of Kraken. So one of the the stated vision of Kraken is to be a platform where anyone can trade any asset at any time. So I'm wondering does that include uh AI agents trading alongside or on behalf of humans and uh how far is that off? Yeah, I I think the I think the biggest issue when we talk about any of these, you know, paradigm [clears throat] shifts is that we think we know what's going to happen and I think we just take a like a very specific assumption that we have no idea what's going to happen and you know um I know everyone comes from a different perspective but um you know uh he has a perspective that like this is what's going to happen. I I think for us it's more how can we experiment as much as possible to see what this can look like and so we use all these terms but I think what's more interesting is [clears throat] that you now finally have uh what I call a perspective of persist persistent memory for an agent-like capability or hundreds or thousands of them in some cases to do something [snorts] and we've just never done that before it and so a year ago it was like tools and then you had to sort of get involved a human in the loop. Now you can start moving at like you know at a faster and faster pace and and so even even I'm catching up on a daily basis. Even I set up my own Mac Mini. Um we've got you know 70 people at the company and this is their only focus. So I I would argue that uh we we could you know the the speed and the level of innovation and paradigm shift that we're seeing now and the amount of people that are trying to figure stuff out is I don't know exponential compared to what we saw 2 years ago, 5 years ago, 10 years ago across anything that's you know from uh you know data intelligence layer or acting upon it and and making decisions and something that's like linear to exponential and and I'm using these words not to be facicious. It's just that um like crypto itself, we've figured out ways to get permissionless and use it worldwide, right? Similar to how you think about the internet. We've talked about this multiple times. And so, I'm not saying something that's new. And if you can do that now with an army of folks or agents or agent-like capabilities to do a lot of what you couldn't do before, um it just opens up a very different path. And and that's how we think about it on behalf of customers that are either extremely well um technical in nature all the way down to an average everyday consumer and how do you sort of bifrocate and build products for them and and so I I don't think it's that far away like I I actually think like you know based on what we're already building we're you know weeks and months away we're not years away. Great. So, so when you say we're weeks and months away from what exactly? Are you saying we're building, you know, agent-like capabilities for our customers in a regulated and safe fashion, right? So that they can do it on their behalf or hey, you know, the capabilities that Citadel has as an institution, you start breaking those pieces apart and say, you know, there we talk about breaking, you know, large banks and say we're going to have uh decentralized banking and financial services. Well, you can do that with institutions, too. You can do that with what smart traders do or you know how people want to save their um their assets or strategies that they want might want to have employed. And there's no reason to have an agent. You can have agents, thousands of them. There's no there's no reason why we couldn't do it. Um [snorts] I think it's more about like obviously how do you do in a safe way, capable way, regulated way, but those are, you know, those are things that already exist. You just have to build against it, right? Sure. Fair enough. Quick quick follow-up question on that, Archer. So basically um you're saying we're already you're already moving towards agent deployments. Now if we move towards a future where agent hold wallets, they execute trades, manage portfolios, um obviously there's a lot of exciting stuff that can happen for the users. But doesn't it also increase the attack surface? Um so basically with more autonomous trading agents being deployed, doesn't the attack surface grow quicker than we can secure it? Um how do you think about that? or do you think there's already solutions to to fill that gap? Look, I think the attack surface grows as much as the security surface grows, right? It's more about like what are the use cases that are going to be scared about. So, we just started hearing about OpenClaw and like, you know, things are going crazy or they're only going crazy because you're allowing it to go crazy, right? It's kind of like the wild wild west. If you go out into the wilderness, [snorts] there's you know, spiders, poisons, disease. It's just it's just, you know, what it is. And so as as you have more and more guardrails as you move forward um or understanding what safety looks like or where you're willing to experiment or what you're willing to lose then you have a you know a better pathway and I think that's no different than what we've seen in crypto. Sure it's happening at a faster and faster pace but that also means that you can secure yourself at a faster and faster pace. Hi Brad where do you land on this? Yeah I'm I'm skeptical of the story. I think that um like the reality is that you want to be very cautious of trying to ingest your worldview of technology by reading like Twitter hype people and like watching Twitter demos. Uh there was there was a story that went viral I think just yesterday of the meta like head of safety at the super intelligence lab uh who their open claw started deleting all their emails and she was like frantically trying to like please stop stop and it just like it queue was filled up and it wasn't responding in time and so it just blasted through her email inbox and she had to run to unplug the Mac mini that was running all this. Now this is a very evocative story but I think we all kind of understand that like look the it was like openclaw to say you know ironclaw is a different story. Open claw was literally vibe coded. Okay it has been widely described by everybody who's actually looked at the code as being just an absolute security nightmare and every update is like thousands of other PRs to try to fix security issues from earlier PRs. Um, so the reality is that this is really wild west shit, right? When when you're talking about self-driving agents and that that are purely autonomous in the way that openclaw is. Now, all that said, this stuff will get better really fast because there's now so much interest, there's so much economic demand for this kind of thing, but it's also important to have clarity about what where the capabilities actually are today, right? So, when you read, you know, I don't know, these like uh these Twitter influencers who talk about I gave my open claw my entire life and my email and now they're just like running my job, like I no longer even involved in my company. Um, you should be very skeptical that these are realistic stories. The reality is that we don't know of anybody basically at this point who actually has a valuable company that's been built entirely on vibe coding, right? Uh, and that's not to say we won't. We almost certainly will at some point, but the reality is that these things are not that easy. It's easy to fall into uh into traps. It's easy to make big fundamental mistakes. And that's why I think what you'll see, you know, OpenAI when they acquired or acquired, you know, hired the the OpenClaw guy, uh Peter Stein Steinberger, I think is his name. Uh they said, Sam Alman said in that blog post that the capabilities of OpenClaw are fundamental to the future of OpenAI, right? He signaled very clearly, this is where we want to go. All the big labs want to go in this direction because they realize this is what people want. They love this story. This story is so compelling. Everyone wants to believe this. But the reality is that when you see a story like, you know, this woman at at Meta having her AI agent delete all of her emails, one story like that will tank you. If you are a if you are a major company offering a consumer product, you cannot do that shit. So the reality is that the risk appetite of people who are playing in open source land who are doing crazy Twitter demos is going to be way far ahead on the risk spectrum compared to what what people are offering you within OpenAI or with anthropic right in the same way you go on chain you go on pumpf fun you're having a very very different experience than somebody trading on Kraken and to be clear that means a lot of rugps a lot of brain damage a lot of hacks and scams but uh like if you want to be on the Vanguard you're going to be taking a lot of risk and that's going to be true for a while. It's not going to be true that in a short period of time, these things are going to be battle tested enough that you're going to want to use it with your grandma and just say, \"Great, plug this into your Gmail and like your life is going to be taken care of.\" Yeah. You know, if you go back far enough in the internet, open source, moving money, being illegal, when we started crypto, you know, Rails, um it's like, oo, scary, you know, you know, this is going to fuck everything up. It's kind of the same thing I'm hearing now, which is it's going to ruin everything. And the reality is is that yeah, there's going to be use cases that are scary, but then we just improve and make it better again and again. Um, and so like a my grandma is not going to use any neo banks or crypto stuff now anyway. So of course she's not going to use this stuff. Um, so it's more about like what as society evolves obviously the world is global and then we you start thinking about building for a permissionless society. I think we kind of forget about that in that if you want to build for that, if you want to move at that speed and innovate at that speed, you have to be willing to play with toys and experiment, sometimes those break. Well, great. Do that in a controlled environment. It's kind of saying like should we build for nuclear power or not? Well, of course there's like hot stap there's a there's concrete walls, right? Like deep mixture. It's not it's not like we're openly just saying like let's just throw it out there and blow everything up. What I am not saying do not experiment with these tools. Absolutely you should and you should do it in a way that like I mean everybody here is at a fucking crypto conference like you people are all on the frontier but there's there's there's going to be a gradual process by which like in even with the internet right the degree to which people who got onboarded onto email like you know you were on the you were on email much much earlier than the average person and back in those days there was it was it was a lot harder genuinely to to interact with but eventually everyone got it there was only like 10 million people available that had the but my whole point here is Now you've got billions of people that have access. So the speed at which something can be adapted, secured, innovated on top of, built on top of just different. And so now we're seeing like agents talk to each other, fight with each other. It's it's kind of cool to watch in the first place. Absolutely. Absolutely. But I think it's also important to realize, look, it's been three years since ChapT launched, right? Um it's not that long. It's not that long. That's right. But it also is longer than I thought it would take for this technology to disseminate throughout society. So far there was a there was a report released recently that showed that uh only about 1 billion humans have used any kind of AI tools out of 8 something billion total human beings. Okay. Um only less than 1% of those have paid for it. Less than 1% have paid which means that most people are using the free tier of all this stuff which is crazy to me right. basically using, you know, GPTO uh what what is it? Uh GP5 mini or using, you know, anthra quad mini, right? Or Gemini Mini or whatever these models are, which are very far from the frontier. Most people don't even know that it's worth paying for these things. They think it's worth more paying for Netflix than it is for paying for an AI tool. So the reality is that there's always going to be this jagged frontier of adoption, especially when it comes to things that require nuance and risk calculations like something like OpenClaw there. I I completely agree with you. Open cloud will get more and more robust, easier and easier to use, but almost certainly there's going to be a spectrum of people who adopt it early and people who adopt it very very late. Yeah, I think you're talking about like infrastructure in the back be able to support a consumer use case. Um, look, Facebook, their newsfeed is all algorithmic and you know, machine learning and linear regression for the last, let's call it 10 years and they spent billions of dollars perfecting that. we can probably do this stuff now with a fraction of the cost for what we want to do for financial services or start a new sort of Facebook. Of course, I'm not saying that network effect is not important. It's the fact that we're able to move at a speed that we haven't been able to for all of the financial services that we're talking about. So, if you want to make something safe, you could probably do it faster. If you want to innovate, you could do it faster. If you want to proliferate it out there, you can do it faster. It's not about people saying that I have a open claw or iron claw, etc., and that I'm going to be able to do anything with it. It's more of that what do those things do in the background? Similar to when you go around email, no one talks about IMAP, POP 3, SMTP. No one talks about that. Like that's the infrastructure in the back. It's more like what can we start building on behalf of a customer experience. And that's where most companies win is at that application experience level. So I totally agree. But but look, draw the line out. Okay, let's say you're open AI, you've acquired Peter Steinberger and you're like great, we've got our own, you know, Open AI claw and open AI claw is going to work for normal consumers, right? with OpenAI claw almost certainly it is going to be optin for every financial transaction right like today OpenAI can already use money to uh buy things on Shopify on Etsy right but every time you do you click the button you always approve every single transaction okay now if you're on OpenClaw you can just give your AI model money and just have it buy stuff while you're asleep okay now the how long do you think it will take before open AAI is willing to say the user doesn't have to uh approve every single transaction. I I think you should Sorry, man. I apologize. No, no, please. I I think you should flip it around, which is when you've got a registered investment advisor, you are giving it to them. Obviously, there's rules and regulations and portfolio management services, but like look, if you are, you know, [clears throat] building out your own, let's call it trading um uh platform. Yeah. Or or you know, you want to go against Jane Street or Citadel. Look, all those systems, those bots trade when they sleep. It's built to do that. Of course. Yeah. you can do the same thing if you start enabling it. So I I think the point is like who cares about the specific amount of money that you want to test with. The whole fact is that you're going to bifurcate what uh is already distributed out there in a way where you can have a large set of your own army. You can call them agents, whatever you want to call them to be able to act on your behalf, but you're able to employ them in a way that you want to based on a strategy that you think might be reasonable or that you want to learn. and they're doing it for you without un without having to build the technical infrastructure or the technical application layer or the bots that most people use today. Why not why not give that ability to everybody? And so I think like that that's actually I think what's more important is that everything a traditional infrastructure that spends millions if not billions of dollars uh investing in to deliver to a consumer but then there's so many middlemen in between taking on all the economics. If you can do it yourself, why wouldn't you? It's it's the whole reason why would you trade a self-custodial wallet or a regulated any? Well, maybe in the US you want to do the regulated any, but if I'm in Argentina, I'd rather have a self-custodial wallet. I love I don't think we're Oh, sorry. Go ahead. No, no, sorry. I I love the debate, but we're almost out of time. So, I'd love to end I would have loved to talk more about regulation and and the liability gaps, but maybe let's end the panel with a few rapid fire questions. Um maybe 10 15 seconds each. Archin, let's start with you. So, name one area in crypto where autonomous agents uh will surpass human performance within the next 12 months. Trading, I mean algorithmic trading 100%. What percentage of your own portfolio could an AI manage better today than you? 0% is also okay. Um I don't know, 5%. 100 for me. You mean 100? Like we're not as smart as these systems when you put them into play. Like I would rather have my own citadel at home trading for me all day long. I would I don't trade all day long. So like I I do VC and I think AIS are not very good VCs. Yeah. Like all right, let's take a step back. I know we're out of time. This is rapid fire. You don't you can't rapid fire. But but think of it this way. If you have if you give your money to a wealth manager, it is the same thing. Do you want to give a wealth manager or someone who's working on your behalf personalized to you 24 hours a day? Okay. So, so let me let me follow up on this with one more rapid fire question. What's the maximum amount of crypto you would trust an autonomous agent today? No kill switch. 100%. No, but how much? Like everything that happen. I think this is not true. I think this is not true. It's I would do it. You would do it. Okay. Will will you commit now to putting all of your crypto into an open a claw agent to trade it for you? I would. You would. You would or not? And uh and I think we should just check the receipts on when we do that. I I'd say in the next 6 to 12 months I would do it. in the next 12 months. Everything that I have personally in crypto, I would just roll out. Okay. I will tag you in 12 months to see whether or not you have publicly put all of your crypto into a claw. I love that. I think that's a good way to end the panel. Thank you, Ar. Thank you, Hib. Thank everyone.","transcript_synced_at":"2026-11-03T04:39:19.569Z","segments":[{"start_seconds":10,"text":"Hello and welcome everyone also from my side to NEON 2026. Uh I'm Alex Shara from the Neo Foundation and in this","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=10s"},{"start_seconds":18,"text":"panel we're going to talk about what happens when agents move real capital and what it means for security risk and","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=18s"},{"start_seconds":25,"text":"autonomy. Naturally, my team made me moderate this panel because I'm European and we Europeans love talking about regulation and risk.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=25s"},{"start_seconds":34,"text":"But don't worry because we have two absolute legends joining me on stage today. So, it will be fun. These guys don't need an introduction, but I'll","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=34s"},{"start_seconds":41,"text":"give them one anyway. Let's start with uh the gentleman on the left. Our first panelist is a former poker pro turned","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=41s"},{"start_seconds":48,"text":"cryptov Keshi, currently managing partner and self-described chief hypeman at Dragonfly. Just last uh week, uh he","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=48s"},{"start_seconds":57,"text":"closed a fresh fund uh $650 million at a time when Fortune magazine calls the blockchain VCscape a mass extinction event. Seems like didn't get that memo.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=57s"},{"start_seconds":69,"text":"So, Hassip, welcome. Thank you.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=69s"},{"start_seconds":72,"text":"Our second panelist to my right um started coding as a kid in Silicon Valley. Um has founded a bunch of different startups and has never really","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=72s"},{"start_seconds":80,"text":"stopped since. Uh Archan Sethi is currently the co-CEO of Kraken. uh the chairman and co-founder of tribe capital","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=80s"},{"start_seconds":87,"text":"and among many other startups founded terminal AI uh which uses AI uh software to basically do quantitative due","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=87s"},{"start_seconds":95,"text":"diligence. So if anyone is going to replace by AI agents in this room, it's not going to be Archin. He's probably","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=95s"},{"start_seconds":102,"text":"built it. Archin, great to have you. So we only have 20 minutes today. So let's dive right in. I'd like to start with you, Hip. So I believe last year you","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=102s"},{"start_seconds":111,"text":"called during the agent hype uh you called agents basically chat bots with memecoins attached.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=111s"},{"start_seconds":117,"text":"Y and you also said that you believe uh agents that become real economic actors are probably more than a year away. But","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=117s"},{"start_seconds":126,"text":"also last week uh you published an ex article saying that crypto was never really made for humans but made for machines and you described an envision a","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=126s"},{"start_seconds":135,"text":"future of self-driving wallets. So my question is from your perspective what needs to happen before agents become real economic actors and what would made","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=135s"},{"start_seconds":144,"text":"you uh deploy meaningful capital into agents? Yeah, so it's a it's a great question.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=144s"},{"start_seconds":149,"text":"Um crypto has a way of getting ahead of itself on every technology. So we got ahead of ourselves many times with","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=149s"},{"start_seconds":156,"text":"respect to how good crypto was. You know in 2017 2018 we were convinced that it was 2026 level good. It obviously wasn't","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=156s"},{"start_seconds":163,"text":"back then. In the same way when we saw you know things like uh AIXPT and truth terminal we were like great AI is already here. It's already using crypto.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=163s"},{"start_seconds":172,"text":"Everything's ready to go. Let's start trading this stuff. And in reality that was child's play compared to where we can now start to see glimpses of where","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=172s"},{"start_seconds":181,"text":"this technology is going to go. So having agents that can actually manage money in thoughtful and capable ways, we're still not there yet, right? If you","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=181s"},{"start_seconds":189,"text":"if you give your money to an open claw, it's going to do something stupid. It's going to do some smart things, but eventually it's going to do something stupid. And the reality is something","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=189s"},{"start_seconds":197,"text":"that works with money 90% of the time is unusable for actual economic activity because you know if you had an intern","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=197s"},{"start_seconds":205,"text":"that like 10% of the time they'd buy you you know tea and coffee and a donut and then 10% of the time they would go on like go on a shopping spree and you know","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=205s"},{"start_seconds":213,"text":"buy themselves a computer you'd be like okay this intern's an idiot and they're fired. But that's right now where the models are at. So the the way I","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=213s"},{"start_seconds":221,"text":"described in the past is that you should expect there to be a very strict cutoff below which there really is not much economic activity happening. So it's","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=221s"},{"start_seconds":229,"text":"like at 90% efficacy that's really really impressive. But it's still nobody's going to use these things actually within their economic systems.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=229s"},{"start_seconds":237,"text":"Even at 95% even 95% is still not good enough. So it's a lot of like nothing nothing","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=237s"},{"start_seconds":244,"text":"then something and then everything. And right now we're still in the nothing phase, but we can see it's starting to poke through. And it's only really crazy","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=244s"},{"start_seconds":252,"text":"people or people running very small experiments that are actually going to be willing to give their AI agents money today. Um, but it's also those people who are going to be the first to","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=252s"},{"start_seconds":260,"text":"discover, aha, when you add these guard rails and you create these constraints, now it's actually good enough to be behaving in the way that we might give","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=260s"},{"start_seconds":268,"text":"it that kind of control. So it's like the person who you know um like in SF we are one of the places where Whimo started to operate even before they were","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=268s"},{"start_seconds":276,"text":"at the level of battle testedness that they could be fully autonomous and actually just you know kind of let it rip the way that self-driving cars are starting to be today. Um so the question","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=276s"},{"start_seconds":286,"text":"is like what's going to be the equivalent of SF for agents or for self-driving wallets that are going to have that level of autonomy over your money. Um so you want to be watching for","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=286s"},{"start_seconds":294,"text":"where those experiments are taking place but it's not going to happen everywhere at once. It's going to be in these small pockets of experimental people with very high risk appetite and very good guardrails.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=294s"},{"start_seconds":304,"text":"Great answer. So basically slowly then all at once, right? Exactly.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=304s"},{"start_seconds":307,"text":"Um Arjun, you're the co of Kraken. So one of the the stated vision of Kraken is to be a platform where anyone can","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=307s"},{"start_seconds":316,"text":"trade any asset at any time. So I'm wondering does that include uh AI agents trading alongside or on behalf of humans and uh how far is that off?","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=316s"},{"start_seconds":326,"text":"Yeah, I I think the I think the biggest issue when we talk about any of these, you know, paradigm","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=326s"},{"start_seconds":334,"text":"[clears throat] shifts is that we think we know what's going to happen and I think we just take a like a very specific assumption that we have no idea","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=334s"},{"start_seconds":342,"text":"what's going to happen and you know um I know everyone comes from a different perspective but um you know uh he has a","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=342s"},{"start_seconds":351,"text":"perspective that like this is what's going to happen. I I think for us it's more how can we experiment as much as","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=351s"},{"start_seconds":358,"text":"possible to see what this can look like and so we use all these terms but I think what's more interesting is","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=358s"},{"start_seconds":365,"text":"[clears throat] that you now finally have uh what I call a perspective of persist persistent memory for an","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=365s"},{"start_seconds":374,"text":"agent-like capability or hundreds or thousands of them in some cases to do something [snorts] and we've just never","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=374s"},{"start_seconds":382,"text":"done that before it and so a year ago it was like tools and then you had to sort of get involved a human in the loop. Now","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=382s"},{"start_seconds":389,"text":"you can start moving at like you know at a faster and faster pace and and so even even I'm catching up on a daily basis.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=389s"},{"start_seconds":395,"text":"Even I set up my own Mac Mini. Um we've got you know 70 people at the company and this is their only focus. So I I","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=395s"},{"start_seconds":404,"text":"would argue that uh we we could you know the the speed and the level of innovation and paradigm shift that we're","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=404s"},{"start_seconds":412,"text":"seeing now and the amount of people that are trying to figure stuff out is I don't know exponential compared to what we saw 2 years ago, 5 years ago, 10","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=412s"},{"start_seconds":420,"text":"years ago across anything that's you know from uh you know data intelligence layer or acting upon it and and making decisions","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=420s"},{"start_seconds":428,"text":"and something that's like linear to exponential and and I'm using these words not to be facicious. It's just that um like crypto itself, we've","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=428s"},{"start_seconds":438,"text":"figured out ways to get permissionless and use it worldwide, right? Similar to how you think about the internet. We've talked about this multiple times. And so, I'm not saying something that's new.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=438s"},{"start_seconds":448,"text":"And if you can do that now with an army of folks or agents or agent-like capabilities to do a lot of what you","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=448s"},{"start_seconds":456,"text":"couldn't do before, um it just opens up a very different path. And and that's how we think about it on behalf of customers that are either extremely well","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=456s"},{"start_seconds":466,"text":"um technical in nature all the way down to an average everyday consumer and how do you sort of bifrocate and build products for them and and so I I don't","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=466s"},{"start_seconds":474,"text":"think it's that far away like I I actually think like you know based on what we're already building we're you know weeks and months away we're not years away.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=474s"},{"start_seconds":480,"text":"Great. So, so when you say we're weeks and months away from what exactly? Are you saying we're building, you know, agent-like capabilities for our customers in a","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=480s"},{"start_seconds":489,"text":"regulated and safe fashion, right? So that they can do it on their behalf or hey, you know, the capabilities that Citadel has as an institution, you start","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=489s"},{"start_seconds":498,"text":"breaking those pieces apart and say, you know, there we talk about breaking, you know, large banks and say we're going to","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=498s"},{"start_seconds":506,"text":"have uh decentralized banking and financial services. Well, you can do that with institutions, too. You can do that with what smart traders do or you","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=506s"},{"start_seconds":512,"text":"know how people want to save their um their assets or strategies that they want might want to have employed. And there's no reason to have an agent. You can have agents, thousands of them.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=512s"},{"start_seconds":523,"text":"There's no there's no reason why we couldn't do it.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=523s"},{"start_seconds":525,"text":"Um [snorts] I think it's more about like obviously how do you do in a safe way, capable way, regulated way, but those are, you know, those are things that already exist. You just have to build against it, right? Sure. Fair enough.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=525s"},{"start_seconds":535,"text":"Quick quick follow-up question on that, Archer. So basically um you're saying we're already you're already moving towards agent deployments. Now if we","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=535s"},{"start_seconds":543,"text":"move towards a future where agent hold wallets, they execute trades, manage portfolios, um obviously there's a lot of exciting stuff that can happen for","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=543s"},{"start_seconds":551,"text":"the users. But doesn't it also increase the attack surface? Um so basically with more autonomous trading agents being","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=551s"},{"start_seconds":558,"text":"deployed, doesn't the attack surface grow quicker than we can secure it? Um how do you think about that? or do you think there's already solutions to to fill that gap?","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=558s"},{"start_seconds":568,"text":"Look, I think the attack surface grows as much as the security surface grows, right? It's more about like what are the use cases that are going to be scared","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=568s"},{"start_seconds":576,"text":"about. So, we just started hearing about OpenClaw and like, you know, things are going crazy or they're only going crazy because you're allowing it to go crazy, right? It's kind of like the wild wild","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=576s"},{"start_seconds":584,"text":"west. If you go out into the wilderness, [snorts] there's you know, spiders, poisons, disease. It's just it's just,","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=584s"},{"start_seconds":592,"text":"you know, what it is. And so as as you have more and more guardrails as you move forward um or understanding what safety looks like or where you're willing to experiment or what you're","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=592s"},{"start_seconds":600,"text":"willing to lose then you have a you know a better pathway and I think that's no different than what we've seen in crypto. Sure it's happening at a faster","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=600s"},{"start_seconds":608,"text":"and faster pace but that also means that you can secure yourself at a faster and faster pace. Hi Brad where do you land on this?","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=608s"},{"start_seconds":615,"text":"Yeah I'm I'm skeptical of the story. I think that um like the reality is that you want to be very cautious of trying","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=615s"},{"start_seconds":622,"text":"to ingest your worldview of technology by reading like Twitter hype people and like watching Twitter demos. Uh there","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=622s"},{"start_seconds":630,"text":"was there was a story that went viral I think just yesterday of the meta like head of safety at the super intelligence lab uh who their open claw started","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=630s"},{"start_seconds":639,"text":"deleting all their emails and she was like frantically trying to like please stop stop and it just like it queue was filled up and it wasn't responding in","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=639s"},{"start_seconds":647,"text":"time and so it just blasted through her email inbox and she had to run to unplug the Mac mini that was running all this.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=647s"},{"start_seconds":653,"text":"Now this is a very evocative story but I think we all kind of understand that like look the it was like openclaw to say you know ironclaw is a different","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=653s"},{"start_seconds":661,"text":"story. Open claw was literally vibe coded. Okay it has been widely described by everybody who's actually looked at the code as being just an absolute","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=661s"},{"start_seconds":668,"text":"security nightmare and every update is like thousands of other PRs to try to fix security issues from earlier PRs.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=668s"},{"start_seconds":676,"text":"Um, so the reality is that this is really wild west shit, right? When when you're talking about self-driving agents and that that are purely autonomous in the way that openclaw is. Now, all that","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=676s"},{"start_seconds":685,"text":"said, this stuff will get better really fast because there's now so much interest, there's so much economic demand for this kind of thing, but it's","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=685s"},{"start_seconds":693,"text":"also important to have clarity about what where the capabilities actually are today, right? So, when you read, you know, I don't know, these like uh these Twitter influencers who talk about I","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=693s"},{"start_seconds":702,"text":"gave my open claw my entire life and my email and now they're just like running my job, like I no longer even involved in my company. Um, you should be very","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=702s"},{"start_seconds":709,"text":"skeptical that these are realistic stories. The reality is that we don't know of anybody basically at this point who actually has a valuable company","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=709s"},{"start_seconds":718,"text":"that's been built entirely on vibe coding, right? Uh, and that's not to say we won't. We almost certainly will at some point, but the reality is that these things are not that easy. It's","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=718s"},{"start_seconds":726,"text":"easy to fall into uh into traps. It's easy to make big fundamental mistakes.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=726s"},{"start_seconds":730,"text":"And that's why I think what you'll see, you know, OpenAI when they acquired or acquired, you know, hired the the OpenClaw guy, uh Peter Stein Steinberger, I think is his name. Uh","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=730s"},{"start_seconds":739,"text":"they said, Sam Alman said in that blog post that the capabilities of OpenClaw are fundamental to the future of OpenAI, right? He signaled very clearly, this is","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=739s"},{"start_seconds":748,"text":"where we want to go. All the big labs want to go in this direction because they realize this is what people want.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=748s"},{"start_seconds":753,"text":"They love this story. This story is so compelling. Everyone wants to believe this. But the reality is that when you see a story like, you know, this woman","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=753s"},{"start_seconds":761,"text":"at at Meta having her AI agent delete all of her emails, one story like that will tank you. If you are a if you are a","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=761s"},{"start_seconds":769,"text":"major company offering a consumer product, you cannot do that shit. So the reality is that the risk appetite of people who are playing in open source land who are doing crazy Twitter demos","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=769s"},{"start_seconds":778,"text":"is going to be way far ahead on the risk spectrum compared to what what people are offering you within OpenAI or with anthropic right in the same way you go","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=778s"},{"start_seconds":787,"text":"on chain you go on pumpf fun you're having a very very different experience than somebody trading on Kraken and to be clear that means a lot of rugps a lot","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=787s"},{"start_seconds":794,"text":"of brain damage a lot of hacks and scams but uh like if you want to be on the Vanguard you're going to be taking a lot of risk and that's going to be true for","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=794s"},{"start_seconds":802,"text":"a while. It's not going to be true that in a short period of time, these things are going to be battle tested enough that you're going to want to use it with your grandma and just say, \"Great, plug","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=802s"},{"start_seconds":810,"text":"this into your Gmail and like your life is going to be taken care of.\" Yeah. You know, if you go back far enough in the","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=810s"},{"start_seconds":818,"text":"internet, open source, moving money, being illegal, when we started crypto, you know, Rails, um it's like, oo,","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=818s"},{"start_seconds":825,"text":"scary, you know, you know, this is going to fuck everything up. It's kind of the same thing I'm hearing now, which is it's going to ruin everything. And the reality is is that yeah, there's going","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=825s"},{"start_seconds":833,"text":"to be use cases that are scary, but then we just improve and make it better again and again. Um, and so like a my grandma","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=833s"},{"start_seconds":840,"text":"is not going to use any neo banks or crypto stuff now anyway. So of course she's not going to use this stuff. Um, so it's more about like what as society","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=840s"},{"start_seconds":849,"text":"evolves obviously the world is global and then we you start thinking about building for a permissionless society. I think we kind of forget about that in","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=849s"},{"start_seconds":858,"text":"that if you want to build for that, if you want to move at that speed and innovate at that speed, you have to be willing to play with toys and experiment, sometimes those break. Well,","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=858s"},{"start_seconds":867,"text":"great. Do that in a controlled environment. It's kind of saying like should we build for nuclear power or not? Well, of course there's like hot stap there's a there's concrete walls, right?","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=867s"},{"start_seconds":875,"text":"Like deep mixture. It's not it's not like we're openly just saying like let's just throw it out there and blow everything up.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=875s"},{"start_seconds":882,"text":"What I am not saying do not experiment with these tools. Absolutely you should and you should do it in a way that like I mean everybody here is at a fucking crypto conference like you people are","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=882s"},{"start_seconds":890,"text":"all on the frontier but there's there's there's going to be a gradual process by which like in even with the internet","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=890s"},{"start_seconds":897,"text":"right the degree to which people who got onboarded onto email like you know you were on the you were on email much much earlier than the average person and back","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=897s"},{"start_seconds":905,"text":"in those days there was it was it was a lot harder genuinely to to interact with but eventually everyone got it there was only like 10 million people available that had the but my whole","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=905s"},{"start_seconds":913,"text":"point here is Now you've got billions of people that have access. So the speed at which something can be adapted, secured,","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=913s"},{"start_seconds":920,"text":"innovated on top of, built on top of just different. And so now we're seeing like agents talk to each other, fight","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=920s"},{"start_seconds":927,"text":"with each other. It's it's kind of cool to watch in the first place.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=927s"},{"start_seconds":932,"text":"Absolutely. Absolutely. But I think it's also important to realize, look, it's been three years since ChapT launched, right? Um it's not that long.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=932s"},{"start_seconds":939,"text":"It's not that long. That's right. But it also is longer than I thought it would take for this technology to disseminate throughout society. So far there was a","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=939s"},{"start_seconds":947,"text":"there was a report released recently that showed that uh only about 1 billion humans have used any kind of AI tools","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=947s"},{"start_seconds":954,"text":"out of 8 something billion total human beings. Okay. Um only less than 1% of those have paid for it. Less than 1%","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=954s"},{"start_seconds":963,"text":"have paid which means that most people are using the free tier of all this stuff which is crazy to me right.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=963s"},{"start_seconds":969,"text":"basically using, you know, GPTO uh what what is it? Uh GP5 mini or using, you know, anthra quad mini, right? Or Gemini Mini or whatever these models are, which","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=969s"},{"start_seconds":977,"text":"are very far from the frontier. Most people don't even know that it's worth paying for these things.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=977s"},{"start_seconds":981,"text":"They think it's worth more paying for Netflix than it is for paying for an AI tool. So the reality is that there's always going to be this jagged frontier","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=981s"},{"start_seconds":988,"text":"of adoption, especially when it comes to things that require nuance and risk calculations like something like OpenClaw there. I I completely agree","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=988s"},{"start_seconds":996,"text":"with you. Open cloud will get more and more robust, easier and easier to use, but almost certainly there's going to be a spectrum of people who adopt it early and people who adopt it very very late.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=996s"},{"start_seconds":1006,"text":"Yeah, I think you're talking about like infrastructure in the back be able to support a consumer use case. Um, look, Facebook, their newsfeed is all","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1006s"},{"start_seconds":1014,"text":"algorithmic and you know, machine learning and linear regression for the last, let's call it 10 years and they spent billions of dollars perfecting that.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1014s"},{"start_seconds":1022,"text":"we can probably do this stuff now with a fraction of the cost for what we want to do for financial services or start a new sort of Facebook. Of course, I'm not saying that network effect is not","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1022s"},{"start_seconds":1031,"text":"important. It's the fact that we're able to move at a speed that we haven't been able to for all of the financial services that we're talking about. So, if you want to make something safe, you","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1031s"},{"start_seconds":1039,"text":"could probably do it faster. If you want to innovate, you could do it faster. If you want to proliferate it out there, you can do it faster. It's not about people saying that I have a open claw or","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1039s"},{"start_seconds":1048,"text":"iron claw, etc., and that I'm going to be able to do anything with it. It's more of that what do those things do in the background? Similar to when you go","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1048s"},{"start_seconds":1055,"text":"around email, no one talks about IMAP, POP 3, SMTP. No one talks about that.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1055s"},{"start_seconds":1060,"text":"Like that's the infrastructure in the back. It's more like what can we start building on behalf of a customer experience. And that's where most companies win is at that application experience level.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1060s"},{"start_seconds":1068,"text":"So I totally agree. But but look, draw the line out. Okay, let's say you're open AI, you've acquired Peter Steinberger and you're like great, we've got our own, you know, Open AI claw and","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1068s"},{"start_seconds":1077,"text":"open AI claw is going to work for normal consumers, right? with OpenAI claw almost certainly it is going to be optin","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1077s"},{"start_seconds":1085,"text":"for every financial transaction right like today OpenAI can already use money to uh buy things on Shopify on Etsy right but every time you do you click","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1085s"},{"start_seconds":1094,"text":"the button you always approve every single transaction okay now if you're on OpenClaw you can just give your AI model money and just have it buy stuff while","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1094s"},{"start_seconds":1103,"text":"you're asleep okay now the how long do you think it will take before open AAI is willing to say the user doesn't have to uh approve every single transaction.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1103s"},{"start_seconds":1112,"text":"I I think you should Sorry, man. I apologize. No, no, please. I I think you should flip it around, which is when you've got a registered investment advisor, you are giving it to them.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1112s"},{"start_seconds":1120,"text":"Obviously, there's rules and regulations and portfolio management services, but like look, if you are, you know, [clears throat] building out your own, let's call it trading um uh platform.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1120s"},{"start_seconds":1130,"text":"Yeah.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1130s"},{"start_seconds":1130,"text":"Or or you know, you want to go against Jane Street or Citadel. Look, all those systems, those bots trade when they sleep. It's built to do that.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1130s"},{"start_seconds":1138,"text":"Of course. Yeah. you can do the same thing if you start enabling it. So I I think the point is like who cares about the specific amount of money that you","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1138s"},{"start_seconds":1146,"text":"want to test with. The whole fact is that you're going to bifurcate what uh is already distributed out there in a way where you can have a large set of","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1146s"},{"start_seconds":1155,"text":"your own army. You can call them agents, whatever you want to call them to be able to act on your behalf, but you're able to employ them in a way that you","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1155s"},{"start_seconds":1162,"text":"want to based on a strategy that you think might be reasonable or that you want to learn. and they're doing it for you without un without having to build","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1162s"},{"start_seconds":1169,"text":"the technical infrastructure or the technical application layer or the bots that most people use today. Why not why not give that ability to everybody? And","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1169s"},{"start_seconds":1177,"text":"so I think like that that's actually I think what's more important is that everything a traditional infrastructure that spends millions if not billions of","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1177s"},{"start_seconds":1185,"text":"dollars uh investing in to deliver to a consumer but then there's so many middlemen in between taking on all the economics. If you can do it yourself,","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1185s"},{"start_seconds":1194,"text":"why wouldn't you? It's it's the whole reason why would you trade a self-custodial wallet or a regulated any? Well, maybe in the US you want to do the regulated any, but if I'm in","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1194s"},{"start_seconds":1202,"text":"Argentina, I'd rather have a self-custodial wallet.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1202s"},{"start_seconds":1205,"text":"I love I don't think we're Oh, sorry. Go ahead.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1205s"},{"start_seconds":1207,"text":"No, no, sorry. I I love the debate, but we're almost out of time. So, I'd love to end I would have loved to talk more about regulation and and the liability","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1207s"},{"start_seconds":1215,"text":"gaps, but maybe let's end the panel with a few rapid fire questions. Um maybe 10 15 seconds each. Archin, let's start","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1215s"},{"start_seconds":1224,"text":"with you. So, name one area in crypto where autonomous agents uh will surpass human performance within the next 12 months.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1224s"},{"start_seconds":1232,"text":"Trading, I mean algorithmic trading 100%.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1232s"},{"start_seconds":1235,"text":"What percentage of your own portfolio could an AI manage better today than you? 0% is also okay. Um I don't know, 5%.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1235s"},{"start_seconds":1244,"text":"100 for me. You mean 100?","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1244s"},{"start_seconds":1246,"text":"Like we're not as smart as these systems when you put them into play. Like I would rather have my own citadel at home trading for me all day long. I would I don't trade all day long. So like I I do","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1246s"},{"start_seconds":1255,"text":"VC and I think AIS are not very good VCs.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1255s"},{"start_seconds":1257,"text":"Yeah. Like all right, let's take a step back. I know we're out of time.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1257s"},{"start_seconds":1261,"text":"This is rapid fire. You don't you can't rapid fire.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1261s"},{"start_seconds":1264,"text":"But but think of it this way. If you have if you give your money to a wealth manager, it is the same thing. Do you want to give a wealth manager or someone","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1264s"},{"start_seconds":1272,"text":"who's working on your behalf personalized to you 24 hours a day?","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1272s"},{"start_seconds":1276,"text":"Okay. So, so let me let me follow up on this with one more rapid fire question.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1276s"},{"start_seconds":1280,"text":"What's the maximum amount of crypto you would trust an autonomous agent today? No kill switch. 100%. No, but how much?","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1280s"},{"start_seconds":1288,"text":"Like everything that happen.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1288s"},{"start_seconds":1290,"text":"I think this is not true. I think this is not true. It's I would do it.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1290s"},{"start_seconds":1293,"text":"You would do it. Okay. Will will you commit now to putting all of your crypto into an open a claw agent to trade it for you?","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1293s"},{"start_seconds":1301,"text":"I would. You would. You would or not?","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1301s"},{"start_seconds":1303,"text":"And uh and I think we should just check the receipts on when we do that. I I'd say in the next 6 to 12 months I would do it. in the next 12 months.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1303s"},{"start_seconds":1309,"text":"Everything that I have personally in crypto, I would just roll out.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1309s"},{"start_seconds":1312,"text":"Okay. I will tag you in 12 months to see whether or not you have publicly put all of your crypto into a claw.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1312s"},{"start_seconds":1318,"text":"I love that. I think that's a good way to end the panel. Thank you, Ar. Thank you, Hib.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1318s"},{"start_seconds":1323,"text":"Thank everyone.","url":"https://www.youtube.com/watch?v=v_Pcmz_RClQ&t=1323s"}],"youtube_url":"https://youtube.com/watch?v=v_Pcmz_RClQ","url":"https://near-pedia.xyz/videos/v_Pcmz_RClQ"}