209 - Why e/acc Is Right with Beff Jezos (Guillaume Verdon)

✨ DEBRIEF | Ryan and David unpacking the episode: https://bankless.com/debrief-e-acc-beff-jezos One guest, two identities. Beff Jezos (Guillaume Verdon) is a founding father of the e/acc movement, a physicist, a quantum researcher, and the founder of an AI startup called Extropic. Beff thinks AI doomers are not only wrong, but they’re taking humanity in the worst possible direction. Growth. Acceleration. Progress. These are the core pillars of the e/acc movement. Instead of slowing down on AI progress, Beff explains why we should be speeding up. ------ 🎧Listen On Your Favorite Podcast Player https://bankless.cc/Podcast ------ BANKLESS SPONSOR TOOLS: 🐙KRAKEN | MOST-TRUSTED CRYPTO EXCHANGE https://k.xyz/bankless-pod-q2 ⁠ 🔗CELO | CEL2 COMING SOON https://bankless.cc/Celo 🗣️TOKU | CRYPTO EMPLOYMENT SOLUTION https://bankless.cc/toku 🛞MANTLE | MODULAR LAYER 2 NETWORK https://bankless.cc/Mantle 💸 CRYPTO TAX CALCULATOR | USE CODE BANK30 https://bankless.cc/CTC 🦄UNISWAP | ON-CHAIN MARKETPLACE https://bankless.cc/uniswap ------ TIMESTAMPS 0:00 Intro 7:15 Why Pseudonym 15:01 The e/acc Pil 31:54 Fume 36:25 Beff’s Beliefs 39:39 Defining e/acc 40:52 The Dangers of AI? 52:05 Why We’re Here & The e/acc Religion? 58:19 Thermodynamics & Life 1:06:16 Maximizing Human Happiness? 1:16:25 Forgetting Society’s Bottom Half 1:23:00 Cancer Also Grows? 1:29:01 AI Regulation 1:33:09 Social Media’s Mistake 1:36:30 AI Bill of Rights 1:38:26 Domesticating AI 1:41:00 Biggest Threats Against e/acc 1:43:43 AI Humans vs. Humans 1:46:09 e/acc vs. Decels…Violence? 1:49:09 Beff’s Thoughts on Crypto 1:54:36 Beff’s Company - Extropic 1:58:18 Closing & Disclaimers ------ RESOURCES Beff Jezos https://twitter.com/BasedBeffJezos Guillaume Verdon https://twitter.com/GillVerd Extropic https://twitter.com/Extropic_AI AI Safety Podcasts - The Decels Eliezer Yudkowsky https://youtu.be/gA1sNLL6yg4 Connor Leahy https://youtu.be/pMoVsM1EWR0 Paul Christiano https://youtu.be/GyFkWb903aU Nate Sores https://youtu.be/Ymjb3SkElco ------ Not financial or tax advice. This channel is strictly educational and is not investment advice or a solicitation to buy or sell any assets or to make any financial decisions. This video is not tax advice. Talk to your accountant. Do your own research. See our investment disclosures here: https://www.bankless.com/disclosures

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

• 0:00David Hoffman?

  What we're advocating for with EAC is sort of freedom of access to compute, freedom of access to AI. We don't want these centralized entities, just like a central bank, like inflating away your money, to control access to advanced AI. We want it in the hands of many, because otherwise, the centralized parties are going to abuse their power.

• 0:21Ryan Sean Adams?

  Welcome to Bankless, where we explore the frontier of internet money and internet finance. This is Ryan Sean Adams. I'm here with David Hoffman, and we're here to help you become more bankless, or maybe more EAC in today's episode, right, David?

• 0:34David Hoffman?

  Accelerationist, yeah.

• 0:35Ryan Sean Adams?

  Yeah. We have Beth Jezos. He's the founding father of EAC. That's Effective Accelerationism. This is a philosophy that thinks AI doomers are not only wrong, they're taking humanity in the worst possible direction. The emphasis here is growth, acceleration, progress. These are the core pillars. And instead of slowing down on AI, Beth explains today why we should be speeding up. I think this is a direct answer to an episode that we recorded over a year ago with Eliezer Yudkowsky on why we are all doomed and why AI is going to kill us all. So it's a debate, I guess, if you listen to those two episodes sequentially. We talk about a few things on today's episode, including why AI isn't going to kill us after all, according to Beth.

• 1:18Ryan Sean Adams?

  And what is EAC? He explains the fundamentals and why EAC is more than just AI. For Beth Jezos, it's an entire philosophy for life. Also, we get into what in the world homo techno capital memetic accelerationism is. Did I say that right, David?

• 1:35David Hoffman?

  I don't know.

• 1:37Ryan Sean Adams?

  Finally, we end with Beth's take on crypto.

• 1:39David Hoffman?

  One thing I enjoyed about this conversation is that I think you could end up at this conversation from a variety of different angles. If you are coming from crypto, you will start to catch a vibe for this. It will sound familiar to you a little bit if you've been paying attention to the AI world, for sure. But also, you can go really far back and just talk about philosophy. Like, if you are a fan of Friedrich Nietzsche, like, you will find yourself in a familiar place. I've been a big fan of the Of optimism as a way to achieve strong mental health. Like optimists tend to feel better. They tend to be less depressed. They tend to feel like they have more control and agency around their surroundings.

• 2:24David Hoffman?

  So if you come from it from like the the mental health and like psychology background, you will find yourself, I think, resonating with some of the things that Beth Jesus has to say. That's one thing I really just enjoyed about this conversation is it's Beth starts from a conversation around base principles, getting really close to the metal and then expands it extrapolates it outwards onto like a political philosophy that people could align with.

• 2:46Ryan Sean Adams?

  So it's worth listening to this episode for that reason alone. David, there's so much more to discuss with you in the in the debrief. I think I've got to take some time to really unpack this. And that, of course, is the episode that we record right after the episode. You're a bankless citizen. You can enjoy that now. All right, let's get right to the episode with Beth. But before we do, we want to thank the sponsors that made this possible, including our number one recommended crypto exchange for 2024. That is

• 3:12David Hoffman?

  Crypto accelerator.

• 3:14Ryan Sean Adams?

  Definitely a crypto accelerator. Well done, Kraken. Go create an account.

• 3:17David Hoffman?

  Kraken knows crypto. Kraken's been in the crypto game for over a decade. And as one of the largest and most trusted exchanges in the industry, Kraken is on the journey with all of us to see what crypto can be. Human history is a story of progress. It's part of us, hardwired. We're designed to seek change everywhere, to improve, to strive. And if anything can be improved, why not finance? Crypto is a financial system designed with the modern world in mind. Instant, permissionless, and 24/7. It's not perfect, and nothing ever will be perfect. But crypto is a world changing technology at a time when the world needs it the most. That's the Kraken mission, to accelerate the global adoption of cryptocurrency, so that you and the rest of the world can achieve financial freedom and inclusion. Head on over to kraken.com slash bankless to see what crypto can be.

• 4:02David Hoffman?

  Not investment advice. Crypto trading involves risk of loss. Cryptocurrency services are provided to US and US territory customers by Payward Ventures. Eek, PVI, doing business as Kraken.

• 4:09Ryan Sean Adams?

  Cello is the mobile first, EVM compatible, carbon negative blockchain built for the real world, driving real world use cases like mobile payments and mobile DeFi. And with Opera Mini Pay as one of the fastest growing

• 4:20David Hoffman?

  Web3 wallets. Cello is seeing a meteoric rise with over 300 million transactions and 1.5 million monthly active addresses. And now Cello is looking to come home to Ethereum as a Layer 2. Optimism, Polygon, Matter Labs, and Arbitrum have all thrown their hats in the ring for the Cello Layer 2 to build upon their stacks. Why the competition? The Cello Layer 2 will bring huge advantages like a decentralized sequencer, off-chain data availability secured by Ethereum validators, and one-block finality. What does that all mean for you? With Cello Layer 2, gas fees will stay low and you can even pay for gas natively using ERC20 tokens, sending crypto to phone numbers across wallets using social connect. But Cello is a community governed protocol. This means that Cello needs you to weigh in and make your voice heard. Join the conversation in the Cello forums. Follow Cello

• 5:04David Hoffman?

  on Twitter and visit cello.org to shape the future of Ethereum. Are you launching a token? Is it already live? How are you managing the legal and tax obligations for providing token grants to your team? It's no secret that token management gets complicated. Between learning all the legal language and tax obligations in every country that your team is in, token grant management can feel like an obstacle course. But it doesn't have to. That's where Toku steps in. Toku provides practical tools to handle token grants, allowing for effective oversight of token distributions and payroll tax compliance for employees, contractors, advisors, and investors. They also handle tax withholdings through their real-time tax calculations that can be done by Toku or integrated into any payroll EOR providers in any jurisdiction. Toku is a trusted provider of Protocol Labs, DYDX Foundation, Mina Protocol, and many more.

• 5:49David Hoffman?

  Get started for free and make token compensation simple at toku.com/bankless.

• 5:54Ryan Sean Adams?

  Bankless Nation, I'm very excited to introduce our next guest. This is one guest, but two identities on the podcast today. Beth Jayzos is the founding father of the Effective Accelerationism Movement. This is also known as EAC. You've heard David and I talk about EAC before, particularly in contrast to the AI Safety, or Dcel movement, as some would call it. EAC is a basic philosophy that advocates for the full acceleration into our AI-powered future. This is all gas and no brakes. Beth, welcome to Bankless.

• 6:30Speaker 02

  Well, thanks for having me. A pleasure to meet you, Ryan and David, and excited for the conversation today.

• 6:35Ryan Sean Adams?

  Now, I said there were actually two identities, one guest but two identities. So Beth is your online alter ego, but in a meat space, your name is your, your identity, I should say, is Guillaume Verdon. This is your original identity, your nation state assigned identity. Yeah, your human name. Let's call it. And you are a physicist, quantum researcher, and the founder of an AI startup called Extropic. So I think we'll be talking more to your Beth persona today, if that's okay, but also welcome to Guillaume.

• 7:09Speaker 02

  Same person, but appreciate it, appreciate it.

• 7:11Ryan Sean Adams?

  Good to be here as well. Include all of the identities in the intros. Why don't we actually start? Because for some people, that may have been a little bit like jarring, right? So let's start with the, the decision to use a pseudonym to kickstart the effective accelerationism movement. Why did you use this pseudonym of Beth Jezos? Like, why not just use your, your regular old nation state Fiat name?

• 7:38Speaker 02

  I think, you know, at the time I was, you know, in some secretive teams within a secretive unit of a big tech company. You know, I was at, as a Google X in there, you know, secrecy is baseline. I was in particularly secretive team. Can't really talk about what I was doing. People can try to piece it together for my patents and have fun with that. Some Easter eggs in there. But, you know, secrecy was my sort of baseline. I couldn't necessarily say too many, you know, have too many hot takes that, you know, wouldn't get back to me. Right? And where I felt, I don't know if that was like fully true. I mean, I've had some feedback that way, but.

• 8:21Speaker 02

  There was always that sort of, you know, thought that if I said something that would be over the line, it's over, it would jeopardize my career. You know, again, this was like pre Elon buying Twitter. It feels like a completely different vibe now, which is great. Back then, though, it felt like if you said something, some sort of latent truth, or something that you felt, you were even just trying to experiment with with your ideas and experiment with different point of views. You can you can lose your job and get canceled and so on. And so to me- You felt muzzled. Yeah, I did. And it doesn't have to do with any one particular employer. To be fair, I think it was just the general vibe. I think many people felt this way. And I think that's

• 9:04Speaker 02

  why anon accounts were so potent. You're kind of removing the sort of threat vector of reputational counterattacks if people try to shut you down, which of course, now that I am doxed, I no longer have that that shield. But, you know, I still stand by my my values, and that's why I'm here on these on these podcasts. But, you know, at the time, it's funny how giving myself an anonymous account kind of opened up a whole new space of ideas that it didn't even let myself experiment with because I knew I wouldn't even be able to communicate them, right? Like, so so freedom of speech induced freedom of thought for me.

• 9:43Speaker 02

  And and sort of BETH was a sort of experiment in in applying my thinking that I've applied to sort of quantum computing, physics-based computing and AI, applying that sort of thinking and background to sort of civilization and societal systems and how we organize ourselves and culture. And really, it was supposed to start as a low-stakes experiment, you know, have a couple followers and and just put some ideas out there, see which ones stick. But it kind of grew and kept compounding. And now we're here, I guess. And so, but that was really nice for me. Because at the time, I think I had grown, I had grown my

• 10:26Speaker 02

  reputation and my my field of study, which was quantum computing. And it was- Isn't clear to me if, if, you know, my, my fans at the time would, you know, just wanted a job or something like that. It was kind of the, the opposite of why most people do an on account, right? Maybe they, they just want, well, I, I think like eventually people just want their ideas to be evaluated on their own, right? Regardless of reputation, regardless of your background. For me, it felt like I want it. It's, it's like I, I do the analogy of new game. Plus, when you beat, when you have us beat the video game, you want to restart with like your knowledge base and maybe some of your gear, but like you're restarting the game. And so for me, it was like, can I rebuild a sort of a following or a reputation,

• 11:12Speaker 02

  you know, from scratch, right? In uncorrelated fashion. So I would never talk about quantum computing. I'd only talk about AI. And so started it while in Google X and then I ended up, you know, leaving that and starting my own company. Now, of course, you know, I'm my own boss more or less. So, so, you know, I have a lot more freedom of speech, but I think overall people feel freer to say sort of things that were maybe taboo or off limits a couple of years ago. And I think that's healthy for everyone, embracing that sort of variance in the, in the space of thoughts. And so, um, yeah, I don't know.

• 11:51Ryan Sean Adams?

  Well, it, it also feels to me, I mean, try this on too, just from my observation. Beth has a different communication style than, than Guillaume, I think. So, you know, Beth is a bit more memetic. Beth is a bit more like in your face. Beth is a bit more rhetoric, I think. And that communication style seems to have meme market fit online at least. And I think accounts for the growth of the Yak movement so far. Like, I don't know if you were trying that personality on or that persona on, and it's just like a better communication style for the spread of ideas. But that's one observation I have from, from seeing Beth communicate.

• 12:35Speaker 02

  Yeah, I mean, like we say, you know, we're on a mission to spread viral optimism, a sort of fuck you optimism, like unapologetic optimism about the future, right, to fight the sort of pervasive pessimism and doomerism. And we will do anything we need to do in order to achieve memetic fitness, right? Because if the, you know, if it's a sort of viral memetic vaccine that renders people immune to sort of feeling terrible about the future and being demoralized, we have a responsibility to make it spread, right? And so if in an era of algorithmic information propagation, you can A/B test what sort of packaging of

• 13:19Speaker 02

  your message has highest memetic fitness. And, you know, usually that's memes, that's, you know, a tone that's sort of brash in your face, that's adversarial. That's what gets propagated, that's what gets seen. And ergo, you know, that becomes, you know, that shapes people's priors because they have an information diet that is fed to them by these algorithms. And so, yeah, I mean, it is a sort of certain communication style. I had sort of experimented with sort of similar firebrand style in quantum computing. Back then, you know, I was really trying to push quantum computing towards something like

• 14:04Speaker 02

  deep learning, differential programming, right? So that's what I did at Google with so many team there is bringing sort of deep learning thinking, you know, let gradient descent program for you to quantum computing. And that was a sort of similar fight back then against, let's say, the complexity theorists that needed hard proofs about everything, every statement, rather than just, you know, letting go and letting the computer figure out how to program for you. And it feels very similar to this sort of battle against the rationalists and the EAs that want to have, you know, axioms and

• 14:44Speaker 02

  I want you to prove that humanity will be eternal and you guarantee our safety forever. And if you can't prove that, we got to shut everything down, right? Like it's very, very, very, very similar sort of adversary. So I've been through this drill before, so I bring that sort of baggage to this new fight. So that's.

• 15:02Ryan Sean Adams?

  So Beth, we're already talking about adversaries, the rationalists and the EA. And so I think we need to tee up this conversation now that we've established who we're talking to. Beth Jezos, one of the founding fathers of effective accelerationism. So let me just throw some context for Bankless listeners and for you, Beth, as we get in this episode of what we're hoping to achieve in this and the reason for this episode. So about a year ago, like almost a year ago to the day that we are recording, we did an episode with a gentleman by the name of Eliezer Yudkowsky. I don't know if you know him at all or have run across him in your travels. So we thought, David and I thought in our more innocent days,

• 15:46Ryan Sean Adams?

  that we were going to have a nice conversation with Eliezer Yudkowsky about the interplay between AI and crypto. And it turned out that was not the conversation Destiny had in store for us. It turned out.

• 15:57Speaker 02

  We got stuck.

• 15:59Ryan Sean Adams?

  Yeah, we got stuck somewhere and we were going to end that episode in a full-blown existential crisis about this thing. I didn't know as much about as I feel like I know today, but yet I'm still kind of like wondering how much I really know. That is AI safety. And Eliezer offered a compelling argument for basically why, we titled the episode, "Why AI is Going to Kill Us All", because that was the distinct impression he left us with. In fact, the episode, the entire episode was a dire warning, like halt all AI progress, you know, like right now. You know, go be with your loved ones, kiss your children goodbye because we're not going to make it.

• 16:37Speaker 02

  I think Ryan went and followed that advice.

• 16:40Ryan Sean Adams?

  Well, it's never bad advice to go to, you know, like tell the people in your life, do you love him? And so we're having you on partially, Beth, because you stand on pretty much the opposite side of that argument, I think. And not only do you not. Think that AI is going to kill us. You think it's going to be great. Like maybe usher us into a utopia. Okay, so this is not like, eh, he was right in some. You, I think, are saying Eliezer is dead wrong. In fact, it's the complete opposite. And we should accelerate progress on AI completely across the board. And so Bankless is a technology podcast, but very specific in the crypto community.

• 17:21Ryan Sean Adams?

  And I want to give you the chance to explain your side of the argument. I want to give you a chance to pitch effective accelerationism, EAC, to the crypto audience. And just like pill us, that straight up pill us, like, give us the gospel of EAC, Beth.

• 17:38Speaker 02

  Getting right into it. Yeah, I mean, you know, yeah, that Yudkowsky podcast really, really was, I think, the start for you all to get into this sort of, this sort of area.

• 17:52Ryan Sean Adams?

  Did you, did it tumble across your feed?

• 17:55Speaker 02

  It definitely did. It definitely did. I think like this sort of, yeah, doom mongering or fear mongering has been really nefarious. Like, it's really affected people psychologically, right? I mean, but yet, and yet, like, you know, they thought GPT-2 was going to kill us all. GPT-3, we're at GPT-4, 4.5 soon. Yeah, we're still here, right? The economy hasn't collapsed. Everything's fine. The crypto bags are still, you know, they're doing all right. They're doing all right these days.

• 18:27Ryan Sean Adams?

  But these are the last days, maybe, Beth. That's what Eliezer would say. It's like, enjoy the last days.

• 18:32Speaker 02

  No, I think that that's the thing about having a system that is malleable and adaptive. Even if you have new technologies that are progressively rolled out, it just morphs and absorbs that capability. And provide and creates utility out of those new technological capabilities. And really, EAC is about trying to understand where do we come from and where are we going? What is the process? What is the process that gave rise to life? What is the process that gave rise to civilization? What is the sort of this weird, you know, clearly something's going on, right?

• 19:13Speaker 02

  Like when we were maybe, you know, children, you know, we had like massive computers that were exponentially worse than the ones we're using to converse today. Something's going on. There's a sort of evolutionary process in the space of technologies and the space of ideas. You know, what is this machine? What is this process that's always morphing and adapting civilization and the technologies around us? And really, EAC, first and foremost, is trying to understand this process. How does it work? How did it get us here? Why is it good? And then how do we accelerate it? Right?

• 19:53Speaker 02

  And, you know, that's what we call acceleration or techno-capital acceleration, or more generally homo techno-capital mimetic acceleration. So it's like all the things, right? Essentially, we think that this process of searching over the space of parameters or bits of information and how we organize ourselves culturally in terms of genetics of humans, in terms of the space of technologies, in terms of how to organize companies, nations, et cetera. It's all one big search process.

• 20:34Speaker 02

  And competition induces a sort of evolutionary selective pressure on the space of all these things. And that sort of competition breeds fitness that then benefits us all, right? Like, in a sense, like, we've tried capitalism and freedom versus sort of authoritarianism or communism, where everything's top-down prescribed, prices are essentially controlled. So far, free markets have been far more successful at creating wonderful things, such as the technologies we use to chat with today and the system we live in today. And overall,

• 21:19Speaker 02

  Such systems where you have many more freedoms, you find much better optima in terms of, again, technology is ways to live your life. And just about everything. But it's, I think, I think it's too broad of a question, like, just pill me on EAC. I think, like, let's, let's just, let's go through how you feel about Dumerism. And it's just, let's just pick it apart, really. Let's just pick it apart. Yeah.

• 21:48David Hoffman?

  I kind of want to do one more thing first, just to kind of frame this conversation. One of the, maybe like disarming things about the conversation that we had with Eliezer Yudkowsky that maybe like me and Ryan just weren't ready for is that Eliezer's arguments were extremely technical. They went down to the basement on like how AI works. He was using phrases called like gradient descent and, you know, reward mechanisms. And it all got like outside of our frame of expertise very quickly because he was like down at the basement level. And he was making like this kind of logical argument that like me and Ryan were just like, weren't ready to like fully unpack because he got really, really technical.

• 22:29David Hoffman?

  And this, this conversation, this EAC versus D cell, people can approach this conversation at like varying like heights. Like some people, I think, Mark Andreessen wrote the techno optimist manifesto, talking about this at a very high level. And then we went and me and Ryan did a number of episodes about AI safety. And each one of those conversations were like somewhere up and down in the very technical to very philosophical like conversation. I'm wondering where you see your innovation with this conversation. You run an AI startup, so you're pretty technical, but you're also, just in your response just now, is some pretty philosophical directional like approaches. Where would you say you, you would like to innovate in this conversation?

• 23:13Speaker 02

  Yeah, I mean, whatever you think is more appropriate for your audience. For me, it's always difficult to adjust my level of technicality. I tend to converse pretty technically. You know, I was a theoretical physicist and then now I run an AI startup. Usually I have very sort of mathematics first thinking, trying to explain my thinking. And I have to convert that into English to some extent. And so, you know, if there's any words I say or anything like that that you want to dig in, like, let's dig in. But I would say, I would say Yud is not very technical, actually. And that if you start digging into his technical knowledge, it's quite, it's actually lacking quite a bit.

• 23:58Speaker 02

  Right? I mean, this sort of, you know, this sort of recur, this notion of recursive self-improvement, I guess, that, you know, the runaway food, maybe we can, we can address, right? Like recursive self-improvement is something we've tried in, in machine learning for a very long time. It's called meta-learning, right? You have a system that learns how to accelerate the learning and you have also architecture, search algorithms. The reality is that the, the larger the space over which you search, the exponentially harder things become, it becomes to find the, the, the true optimum, right? And so what that means is that if we task an AI to improve AI, it's, it's exponentially more complicated.

• 24:40Speaker 02

  Every level of like an AI that improves an AI that improves an AI, you go, right? And so it's going to take exponentially more compute and energy, right? To achieve that optimum, right? And so in a sense, there's a sort of, there's already a soft cap on compute in terms of the energetic and capital costs of, of compute and that keeps us safe, right? And in a sense, like, Everything, everything in, in, in, in our civilization, even every life form is trying to boom. Everything is trying to grow and that is the thesis of, of IAC. And but that, because everything is trying to grow and competing with one another for resources,

• 25:21Speaker 02

  we yield, we get better optima, but it also keeps everything in check. There's not one sort of singleton, you know, biological system, or one single company, or one single nation that takes over the whole system, right? Because they're, they're, they're competing with one another. And, and in general, to find sort of optima of high fitness, for example, you know, if you're, if you're building a company, a company is almost like, at least in the startup space, you know, at the seed round, you're basically a search algorithm. You're searching over the space of technologies and you're burning capital and intellectual

• 26:02Speaker 02

  compute and maybe actual GPU compute to find sort of, to pinpoint exactly what product you gotta build. That's a couple bits of information, but that's costly. That's a costly thing to find. So every optimum that is high fitness is very costly to find. And the more optimal it is, often it's exponentially costlier to find that optimum. And so that keeps us safe in the sense that I can pretty confidently say, we're not just a small edit distance, like this one weird trick, where like, now we have super AGI that fooms and creates gray goo and takes over the world. Like, that's, that's like, very exponentially

• 26:42Speaker 02

  unlikely that, that, that, that is the case. You can't guarantee it, right? You can't guarantee that there's not one weird hack. But from what we've seen so far, it's taken billions of years to get where we are now with life and intelligence as we know it. And it's been a very long sort of process of, of improvement to this point. But there, there won't be a sort of hyperbolic discontinuity in the rate in which things improve. And

• 27:12Speaker 02

  In general, we take the opposite camp. Like this sort of rate of self-improvement of everything is actually really important. And it's what creates everything we enjoy. And we should seek to keep that process growing and scale it, right? Really, like, systems in nature either get busy growing or they get busy dying, right? And so they either secure more resources, secure more free energy, and figure out clever ways to utilize it in order to grow, or they try to stagnate, they run out of fuel, and then they die. That is it.

• 27:56Speaker 02

  Everything runs on some type of fuel. Nothing is forever. No bit of information is forever. To maintain its coherence, it costs energy. Because everything wants to decohere naturally. Everything wants to sort of, you know, it's a constant fight against entropy. And so the thesis of EAC is the one golden metric, to some extent, that measures the progress of the whole system is sort of free energy. How much energy are we acquiring and consuming as a civilization? Because that's a metric of our progress that can't be gamed, right? If you, and this will resonate with you guys, but if you measure it in US dollars, that's not, that's not objective scale, right?

• 28:41Speaker 02

  You know, that's not a scale, that's a scale that you can play with. You can, you know, inflate things, you can have inflation, you can print money, and then it seems like you have some progress, but really, really, you went the other, the wrong way. Whereas something objective like energy is a good metric to measure progress. And we think that Scaling up civilization and doing so with urgency is how we ensure it's sort of long-term success in existence. And I think that the, one of the most dangerous things

• 29:22Speaker 02

  we could do is let this mindset of doom and demoralization become hyperstitious, right? If you focus on doom, you focus on darker futures, well, first of all, you stop building, you stop having children, you stop hoping for better things in the future. It sort of becomes a self-fulfilling prophecy and that's how civilizations and empires die and that will cause massive pain. And so we're on a mission to spread sort of optimism that is hyperstitious in the sense, yes, we can do it. We can build better things. We could build a better future. We can leverage AI to build the better future we want, to cure

• 30:05Speaker 02

  diseases, to tackle climate change, to unlock nuclear fusion, nuclear energy, massive prosperity. We have all this upside on the table and the longer we wait, the longer we wait, the lower likelihood we can achieve it and we have urgency to make it happen. And that's very similar mindset you have in startups in Silicon Valley, right? The most successful startups are very optimistic and they believe they can do something that seems like unimaginable at the beginning of the company, but then they do it and it just keeps happening, right? And it's this sort of hyperstitious optimism effect and that's the sort of mentality we're

• 30:47Speaker 02

  trying to scale to the world with EAC. And it's why we think that doomerism and pessimism is really dangerous and needs to be fought somewhat aggressively, right? Because that is actually the source of doom, not some sort of fictitious artificial superintelligence from sci-fi. It's not backed by science, right? You know, I've worked on

• 31:14Speaker 02

  AI for material generation, protein folding, biochemistry, chemistry, all like AI for the physical world. It's much harder than people think, right? The physical world is really hard. And one of the best things at designing things in the physical world is life itself. But you could see life itself as a big optimization algorithm that is trying to fume. Every life form is trying to fume. And yet, nothing has fumed, right? So that should give you a bit more peace and you can breathe.

• 31:50Speaker 02

  And but happy to go into any sort of argument, like any sticking points you have about Yud's argument.

• 31:56Ryan Sean Adams?

  One thing you've mentioned a couple of times is this word fume. Could you define that really quick for people?

• 32:00Speaker 02

  Yeah, I mean, you know, well, how do you understand fume? Like from your interview with Yud? How would you explain it?

• 32:09David Hoffman?

  I think you're using it in a very general sense, which I think I wish I appreciate because it kind of gets down to the bare metal of kind of how life works. Like, life is interested in life. Like, life is interested in propagating. Fume in the AI sense is main is mostly talking about that superintelligence explosion where AI just takes over everything. And it's all it's just like this one single event. And all of a sudden, the whole world is AI and it's run by AIs for AIs. Trees want to plant more trees, you know? Bugs want there to be more bugs. Ants wants more ants. Humans want more humans. And I think this is it's all trying to hit some sort of like

• 32:50David Hoffman?

  point in a curve, which like, once the ball starts rolling, it starts rolling faster and faster and faster. And all of a sudden, we have like a population explosion. It doesn't matter what kind of life form you are, but like, everything is trying to look for that like growth in population. That is something that's fundamental about life. And so you're using it in a general sense where it's just like, hey, any system that any system whatsoever that's propagating is trying to find more energy because that's how it can propagate even more. And this is this is the playing field that just the universe exists on. And I think that this is how you're using it.

• 33:20Speaker 02

  It's every, every system itself. System, whether it's a company, a group of people, a culture, like you said, bugs, trees, whatever, even nation states, everything is self organizing, self adapting its inner workings in order to to to grow. And by construction, things that are not optimizing themselves to grow, burn, run out of fuel and they fade. And that's it. Right. And so if you think about, if you think in the future, which culture, which, like, if you think of like several nations in the future, which nations will have survived and what culture would they have? Right? Well,

• 34:03Speaker 02

  they would probably have an EA culture that is literally by construction, trying to figure out what is the optimal way to organize ourselves in order to grow. And the sort of pessimistic, dumberistic cultures will have faded because they'll have destroyed themselves.

• 34:19Ryan Sean Adams?

  This kind of reminds reminds me, by the way, of an episode we did with Robin Hanson about, like, his theory for why we haven't seen aliens is basically that it's for one, it's too early, but like, we will see them. And the ones we will see are not the quiet aliens that stick to their home home planet, not the the Decels aliens. We'll see the EA aliens, the grabby aliens that go and they they consume their solar system, they consume their galaxy, and then they're a multi galaxy type of like, the grabby ones are the ones that are effectively going to foom and win.

• 34:51Speaker 02

  Yeah. Well, it's the same with like, you know, variants of a virus, right? It's the ones that have higher replicability that you get statistically, right? You don't see like earlier variants that are lower fitness, right? And it's like, everything, everything, every bit of information is getting selected for in terms of, does it confer the organism in of which is part of an ability to grow, right? Like, your genetics, it's like, does this piece of genetic code give you higher fitness? Does it, you know?

• 35:31Speaker 02

  Make you have more offspring and then that piece of code, genetic code is higher likelihood in the future, right? But it's like applying that sort of evolutionary thinking to everything, including culture, including ways to organize your companies, including, you can think of, you know, that sort of thinking even applied to crypto, right? Which, which sort of coins have the best sort of memetic fitness in the long term?

• 35:57David Hoffman?

  Bitcoin maximalism is the Bitcoin fumers. They think Bitcoin is going to fume.

• 36:02Speaker 02

  Well, I don't know about that. What I do like about Bitcoin is that it is proof of work. And so it is anchored to sort of physics and energy consumption. And, you know, I'm not necessarily a Bitcoin maximalist, but I am a sort of energy maximalist. I think that is the right metric to, to pin things with respect to. Yeah.

• 36:24David Hoffman?

  I want to provide the way that I see the structure of this conversation, that if you have like a very like basement first principles grounding in your arguments. And I think that is like what I would consider like, you know, the very, the depth of where some of this belief structure comes from. But we've watched, and we've watched a lot of this, these conversations permeate throughout like Silicon Valley, and then also make its way into like Capitol Hill, where some of these like very first principle arguments are inspiring or being argued against as like political stances about the direction for society at large. So it kind of goes back to just like, we have some beliefs, we have a structure of thought. It's very, it can get very technical and granular in their

• 37:07David Hoffman?

  base principle arguments. But then as this conversation progresses, and Ryan's going to take us into the next phase here, it really can inspire like a political belief, like a direction, a proposed direction for humanity. And so this is kind of what you, why Ryan kind of illustrated this is like you're, you're inspiring like a movement, like a political movement of not necessarily a party, because that's too structured, but just like a set of beliefs for how we ought to live as a, as humanity, as a species. Is this how, this is how you see it?

• 37:35Speaker 02

  Yeah, I mean, hopefully, hopefully it does affect politics, right? It's like, the ways we, we've been organizing ourselves, the ways we've been legislating have been pretty far from optimal. And we should think from a first principle standpoint, you know, what is?

• 37:53Speaker 02

  How do we organize sort of a hierarchy of cybernetic control in our civilization? And this is a very deep concept. So how do we, in a sense, like, it's very similar to crypto. Many layers of a protocol, right? You have sort of, you check, you chunk blocks and then you have different layers that check at different scales, right? And base layers have a slower clock rate and they roll up larger amounts of actions or transactions. It's very similar to having a hierarchy of sort of legislation at the local, provincial or state level and then national level and then international level, right?

• 38:37Speaker 02

  As you go up the hierarchy, you should have a slower feedback loop and it should be lighter and lighter touch in terms of its ability to control things. And the goal with EAC is to start a discussion of like, how should we balance things? Because one side of the aisle wants everyone to be hands off. One side of the aisle wants like absolute total control in the hands of a few. And both of those things are not optimal, right? There's a spectrum in between and we should search over that spectrum. And it's very relevant to the centralization versus decentralization question in crypto for protocols. I think they're the same question, whether it's in politics versus crypto.

• 39:17Speaker 02

  Like, is a system that is like totally decentralized and greedy optimal versus one that is centralized versus one that is hierarchical? I think it's the latter and maybe we can get into that. But hopefully that does inspire conversations on Capitol Hill and in DC about where to take policy and legislation.

• 39:42Ryan Sean Adams?

  Yeah, I think it's really important. I think this EAC, do we accelerate or do we de-accelerate is probably one of the most important conversations that society is having. And like, EAC to me, one question I was going to have for you, Beth, is like, is EAC, is it a philosophy? Is it like a movement? Is this like a social revolution? Is it a religion of some sorts? Is it like this amorphous thing like crypto where it's like a grab bag of all of the above? What's your take on this?

• 40:14Speaker 02

  I think it's all of the above to some extent, right? I mean, for me, it's, you know, it's almost like religious like belief just because it's like, I can get into that, but, you know, and for some it is. For some, it's a, you know, it's a community of like minded individuals that are optimistic about the future and trying to help each other and want to build this better future and collaborate. And, you know, for some, it's sort of a political or ideological movement of how we should legislate things or how we should run things. And for some, it's just, I don't know, inspirational and gets them fired up to build more.

• 40:54Ryan Sean Adams?

  So I very much want to get the ways like your take on why this is like a religion, how it sort of explains like big questions like, why are we here? What is the universe? This sort of thing. But before we do, I just got to tie off this Eliezer existential crisis type thing. So we talked about FUM, we sort of defined it. You think it's very unlikely that AIs sort of like destroy humanity or accidentally, or, you know, they literally go to war with us or something like that. But I want to maybe ask like, so David said that it was difficult to follow the Eliezer conversation because it was technical. I actually think it was less that for me, anyway, personally, it was more because the premise was so simple.

• 41:39Ryan Sean Adams?

  We've seen dangerous technologies in the past. Anybody, Bankless listener, watched the Oppenheimer movie recently? The whole movie is about this like chain of events where the scientific community discovers some dangerous technology essentially. And what like this chain of events that could lead to global Armageddon and the ending of humanity. And so now we're at the precipice of this new discovery. We're in like these times where, oh my God, you can talk to a computer and it's passed the Turing test and it sounds like a real person and this is amazing. It generates art like we've never seen before. And we are all wondering whether this is one of those technologies that has the potential to destroy everything. And like, it's very, it's very seductive, I think, for Eliezer to be like, yeah.

• 42:24Ryan Sean Adams?

  Imagine if, you know, we split an atom with kind of like, everybody could do that in their microwave, let's say. And imagine that was like given that kind of level of power and technology was given to every human being. We democratize it in that way. Well, what would happen? It'd probably be the end of life as we know it. And so like, I still have this base question, Beth, of whether AI is like dangerous or not. Is it similar to nuclear weapons? Is it similar to, you know, biological pandemics that somebody could cook up in their basement? Or is it different somehow? Because I feel like a lot of this conversation hinges upon that question. I mean, you were advocating in EAC for growth, right? Versus degrowth.

• 43:08Ryan Sean Adams?

  And I sort of understand that everybody wants growth over degrowth, deacceleration is my thing. But I think another framing that people like Eliezer might have is it's growth versus safety. Like, Bev, you're going all gas, no brakes here. Like, we're worried. What if this technology is dangerous? What's your take on this?

• 43:27Speaker 02

  Yeah, so the analogy, I think, you know, if everybody could split the atom in their microwave, we'd have quite, quite a bit more energy being produced on the grid, and that would be net positive. I think like, you know, runaway effects of, like, if you have a chain reaction that becomes a bomb, of course, that's, that's, that's bad. But obviously, it seems to be like, you know, creating a nuclear weapon is very involved, and not everybody has access to it. And the thing with nuclear weapons is it's pure downside, right? Like, the only utility to nuclear weapons is like, you know, damaging your enemy, and it's like a deterrent, right? Whereas AI, there's huge, huge, huge, huge upside to creating AI and leveraging AI and

• 44:11Speaker 02

  everything we do. Gives us intellectual and operational leverage. It helps, you know, create economic value. It helps solve all our problems. It helps us live better lives. It's going to save lives, right? In medicine, right? It's gonna make things cheaper. It's gonna help us build cheap housing. It's gonna help us, you know, save on legal bills that are, you know, ballooning, right? You're gonna have LLM lawyers. You're gonna have LLM doctors, everybody on the planet's gonna have access to some of the best doctors ever. You're gonna have personalized medicine, right? Like, um, the list goes on. Like, a lot of our problems that are pain points in our modern society will be solved with cheaper intelligence. And there is an urgency to make that happen.

• 44:54Speaker 02

  Similar to how, um, let's say, fear mongering about nuclear weapons actually kind of killed the, the nuclear fission energy industry, right? Now it's overregulated to oblivion. There's so much red tape, like, you know, most of your budget building a nuclear reactor is towards compliance. And, and, and that's suppressed the advent of and ubiquity of this technology. And it's caused us to have to go to all sorts of wars all over the place for, for energy, right? So, so, so this mind virus of anti-nuclear was actually detrimental. And we're, we're kind of like waltzing into a similar scenario with AI.

• 45:36Speaker 02

  If we overregulate it in the womb, we're not gonna see the massive upsides to it. I don't, I, I, I think that what's dangerous with AI, and we're very transparent about this, uh, you know, AI is, information is power. Intelligence allows you to extract more utility out of less information. And so it is, it is, give, it confers you power, right? And if you have a, a big, uh, delta, a big difference between the capabilities of centralized entities and that of the people in terms of AI, then you have the opportunity for sort of AI assisted tyranny. And to us, that's the highest existential risk, because there's a very strong prior

• 46:20Speaker 02

  that if you give all the power to a centralized party, it becomes corrupt, and then it abuses that power and oppresses people and causes mass suffering. And to us, like, putting AI, which, yes, it will make AI, whoever has advanced AI in their hands will become formidable. But if everybody is You know, not equally formidable, but you know, in a similar range, then no one party is going to have too much power over the other and going to dominate, completely dominate the other, right? It's a sort of adversarial equilibrium. And that is sort of what we're advocating for with EAC is sort of freedom of access to compute, freedom of access to AI, right? We don't want these centralized entities,

• 47:04Speaker 02

  just like a central bank, like inflating away your money to control access to advanced AI. We want it in the hands of many, uh, because otherwise, you know, the centralized parties are going to abuse their power. And so, uh, to me, I think that, you know, yes, AI is potent. If you only focus on tail risks, uh, you can convince yourself to kill anything in the womb. Um, but there's also massive upside that shouldn't be discounted. And there's not enough, we don't talk enough about the massive upside we're leaving on the table. And the reality is that most likely what will happen is that some nations

• 47:44Speaker 02

  or some cultures will, will embrace AI, some will want to ban AI. Those that ban AI are going to be left behind and massively disadvantaged. And those that embrace it are going to be economically prosperous, will outgrow those that have banned it. And so you want to be in the fork, right? Just similar to crypto fork. You want to be in the cultural fork that embraces AI and integrates it into their, their lives, into society and, and leverages it to, to become formidable and grow, right? And there are ways forward to, you know, leverage AI and merge with it. I mean, we're already, you know, we're using like smart watches and smartphones and now

• 48:27Speaker 02

  Apple Vision Pro is going viral. We're already augmenting ourselves, a notion of human.

• 48:33Speaker 02

  You know, of a human individual is getting sort of diluted or eroded. We already leveraged exogenous sources of intelligence, right? We use Google search perplexity, whatnot, all day, every day. And I think that just focusing on like, oh no, humans will be left behind. It's like, no, but like the notion of human will drift encompass a human and maybe their their fleet of AIs that do their bidding. And then you become this sort of formidable being. And that's the sort of awesome future we want to have where you're quasi immortal. You have like, you know, AI customizing your biology and you have sort of AI augmentations in every way possible.

• 49:14Speaker 02

  And, and, and I think that sort of optimistic future, we got to paint the picture of that future rather than painting the picture of the doomer future where like AI takes over, which again, I don't know how it would do that if there's other powerful AIs keeping it at check from, from parties that aren't necessarily aligned with that party. And again, I don't believe in sort of fast takeoff in AI. I think it's not well founded. And so like, if you don't necessarily think that, you know, the likelihood of this fully existential doom is high from like just pure ASI, then, then the biggest threat vector

• 49:56Speaker 02

  becomes sort of authoritarianism, sort of censorship, right? If you only have AI in the hands of a few and they use it for information control and cultural control, then they can basically sigh up you and control you, right? And, and that is very dangerous because if only a few parties that are, you know, government backed cartel, control LLMs, and they become our sources of truth, then they control you and they can like use that to consolidate their control and consolidate their power. And now we live in a sort of dark age where we don't have open access to knowledge. We don't have open access to compute and we can't, we can't fight this sort of top

• 50:39Speaker 02

  down control with AI, with our own AIs that would help us filter and counter sort of this psychological manipulation. And so that's the dark future we're trying to avoid. So getting AI in the hands of many into every org and accelerating. Makes it very hard to sort of acceleration is a hedge against sort of top down control and maybe a centralized party taking over, killing variants, killing like you can only, you know, use AI this way and taking control of the whole technology because that that would be the bad future we're trying to avoid. Anyway, so, so we, we kind of have our own existential risks, right? Like, we think like authoritarianism and centralized

• 51:22Speaker 02

  control is, is, is sort of, that's the, that's the bad scenario. It's based on like data from history, right? There's a lot of bad things that have happened historically when that happens versus sort of sci-fi based priors of like, I read too many books on the Terminator. I read too many 90s sci-fi books about nanotechnology. And, you know, I, I think like AI is going to figure out how to, you know, turn us into gray goo, which not plausible.

• 51:51Ryan Sean Adams?

  So, Bev, you basically, you think like P doom is kind of like a failed kind of calculation. What we should be doing is P utopia, right? And then also like the risk, on the risk side, it's like probability of totalitarianism. So like, let's, let's, let's do that calculation too.

• 52:05Speaker 02

  Yeah, P 1984, right?

• 52:06Ryan Sean Adams?

  P 1984. Okay, so, so I, I, I think we're starting to understand the contours of EA and like where you're coming from. So it's pro-growth. It's definitely techno optimism, right? I certainly love the, the picture you're painting about kind of like the possibilities here. It's much more cheery and optimistic than sort of the, the AI doomer takes that I've heard so far. I, I want to get back to kind of the ways EA is like a religion to you. And I don't want to interpret too much into that. I'm sure there are many ways that it's not like a religion, but talk about kind of the basement level, because I know you are, you know, have been as a physicist and, you know, throughout childhood, I've,

• 52:50Ryan Sean Adams?

  I've heard other conversations with you. You've, you've been always in search for like, what's, what's the, the basement level of this whole experience that we're having right now in the universe? And like, why, why does all of this matter? What conclusions have you come to? And how does that relate to EA? Maybe, how does EA even explain some of these things?

• 53:10Speaker 02

  Yeah, I mean, it's been a journey throughout life to try to understand our place in the universe, right? That led me to theoretical physics, trying to understand the very small, the quantum mechanical, and the very big, the cosmos. So I was a quantum cosmologist at some point. I also worked on sort of black hole physics, which are kind of the edges, a different edge of the universe, right? And there I was trying to understand, you know, where did we come from? Why are we here? By looking at sort of the very big and the very small, and I kind of missed the middle part, right? And in a sense, after some time, I realized that a lot of the beauty

• 53:48Speaker 02

  and complexity and good things in the world came from sort of emergence and self-organization. Basically, instead of looking at quantum physics, or what is called general relativity, so gravity, which are the mostly physics of the very small and the very big, I need to look at thermodynamics, the physics of the middle, or out-of-equilibrium thermodynamics, the physics of life, the physics of self-organizing complex systems, because that's what created us, that's what created a lot of civilization, technologies we see today, and that's what induces this sort of selective pressure on the space of everything. And that sort of principle is very deep. And so, you know, having gone through theoretical physics, and eventually discovered that sort

• 54:32Speaker 02

  of complexity and self-organization is where a lot of the answers are, that's what led me to sort of look at EAC as sort of like, hey, like thermodynamics helped create, well, it explains why we have life at all, right? How do we get here? Explains evolution, it's upstream of evolution, it's upstream of, you know, market pressures, selective pressures. It's kind of like this generalized law that helps, that creates everything we know, right? That is relevant to us in our day-to-day life, right? Like, what is Andromeda gonna do in like 2 billion years? Not that relevant day-to-day, right? What is, you know, so-

• 55:14Speaker 02

  So I think like, to me, EA is almost a religion because it's like, okay, we are part of this process, this self-adaptive process of, you know, life, memes, information, technologies, civilization. And this whole system seeks to grow. We are a very special phase of matter that was very unlikely in our universe. And we have a sort of responsibility to cherish it and allow it to grow in order to be robust to sort of erasure or fluctuations, right? Like an asteroid, right? And so, you know, it's very Elonian in the sense of like, you know,

• 55:55Speaker 02

  it has very high inner product with the quest to increase the scope and scale of civilization consciousness.

• 56:02Ryan Sean Adams?

  Is Elonian, is that like of drive from Elon Musk?

• 56:05Speaker 02

  Yeah, I don't know. I just made that up. But, uh, but, but, but, but the point is we're very aligned with that, with that mission. And it like, it gives you crazy amount of purpose day to day. Like, what is the point of a religion? A religion is like, for, for a lot of people practically, it is a cultural heuristic that if it has stuck around for a long time, it has been post-selected for in terms of conferring its adherence, like a sort of advantage, right? Because, because by construction, it has grown and has propagated this, this far, this long. Right? And so it's, it's like ways to live your life. It's like prescriptions, right? To me, they're like cultural parameters, similar to how like you, you have a, you have

• 56:49Speaker 02

  in a neural net, you, you spend a lot of time training and finding the optimal parameters. And then you run inference, you just run at the optimal parameters. To me, religions are like pre-trained models of how to live your life. Right? They're like prescribed parameters and then you can just run inference.

• 57:05Ryan Sean Adams?

  You think EA kind of like prescribes how to live your life. Actually, it gives you some meaning. It's basically, basically the meaning is growth. The meaning is propagation. The, the meaning is like pushing humanity to the stars, to the frontier. These are some of the meanings you drive from this.

• 57:21Speaker 02

  So to me, EAC is more like the meta learner or the optimizer, right? We're not prescribing any one particular way to live your life. It's like, but any sort of prescription that you may posit will be, what we're saying is it's going to be post-selected for according to this sort of fitness function. And so you should, you should select for subcultures that give you, give you an advantage in this growth. Right? If you want to be part of the, part of the future. And, but really it's like, it's a search process for what yields higher growth. And it's always, always changing. It's not one way to live your life right now.

• 58:02Speaker 02

  There's not one set of optimal policy parameters that you can just run forever. It's like, let's collectively figure out heuristics of how to organize ourselves and how to live our lives, how to run the world such that we grow sustainably, fall tolerantly towards, you know, yes, a galactic civilization.

• 58:23Ryan Sean Adams?

  But Beth, one set of dots that may not have been connected for folks. So like I sort of connected them a little bit as you were speaking, but like, because I've, I've heard you speak in the past and I've, I've read many of your articles where you've written things like this. And I'm, you know, on your Twitter timeline, but how does thermodynamics explain life? Like, how does that explain, why is that an underlying process in the universe? That is relevant to what we're talking about, like pro AI growth?

• 58:51Speaker 02

  Yeah, yeah. No, that's a really deep question. And to me, it's like, I mean, I've made it quite a career pivot, right? I went from being someone in quantum computing. So trying to harness quantum mechanics of the world to make computing devices and make computing devices that understand the quantum mechanical world around us. That is what I did in quantum machine learning, right? And now, you know, after the sort of realization that, you know, the most interesting part is in the middle. It's in thermodynamics. Now I'm doing thermodynamic computing where I'm harnessing out of equilibrium thermodynamics to engineer self-organizing devices that are kind of like lifelike, but they do machine learning as a thermodynamic physical process.

• 59:36Speaker 02

  But once you understand.

• 59:39Speaker 02

  These principles enough to engineer systems that enact them, you start seeing that pattern everywhere and sort of, you know, I wrote the EAC Physics Manifesto at the same time as I founded my company. So for me, it's just because like my mind is swimming in it. But yeah, why does life, you know, why did thermodynamics give rise to life? I'm an adherent of a theory by Jeremy England, a former MIT professor. If you want to read his book, Every Life is on Fire. And his lectures are great on this. But essentially, what you can show, it's just from the laws of probability, really, and it's a generalized notion of the second law of thermodynamics,

• 1:00:23Speaker 02

  where it shows that paths, right? Like, if you think of like your trajectory across time, right? Of like, you have a certain system, you have states over time, and you have a trajectory. Trajectories that have dissipated more heat along the trajectory are exponentially more likely from the laws of physics. That's pretty nuts, right? And so, you know, there's there's theorems, they're called, you know, they have complicated names. But, you know, check out the work by Jeremy England on the thermodynamic dissipative adaptation. Basically, it posits that systems in nature, or really any system, any system that is in the physical world,

• 1:01:08Speaker 02

  right, will self-adapt in order to maximize this dissipation of free energy. So that includes figuring out how to capture more free energy to dissipate it. So lifelike systems are always adapting. How can I secure more fuel? How can I use it more efficiently? So that I, you know, I could budget my fuel to get more fuel and keep growing. And, you know, you get, as a byproduct, you get evolution because systems that are, that have adapted their genetics in order to replicate more. Well, if you replicate more, you burn more fuel as a whole system. Right? And so you're more likely.

• 1:01:43David Hoffman?

  So human complexity exists to burn a fuel faster for the universe to

• 1:01:47Speaker 02

  The whole, all civilization is a fancy fire, right? But we're, we're a very clever fire. We're like an energy seeking fire. Right? But yeah, that, that is, that is our, that is what gave rise to us.

• 1:01:59David Hoffman?

  Is it correct to define life if we're using, if this is like the idea that life is whatever system finds a way to dissipate more heat more. And that, and so part of that definition, like you said, it includes collecting energy. So like reversing entropy first, like creating order, creating systems, creating structure in order to produce more heat exhaust, which is, and that is what life is?

• 1:02:28Speaker 02

  That is what all thermodynamics systems are, but life is like a very special version of that.

• 1:02:34David Hoffman?

  Immersion property out of that?

• 1:02:36Speaker 02

  People, I, I think, yeah, I think people are trying to, well, you know, is a virus life, right? You know, is, uh,

• 1:02:44David Hoffman?

  It's on the spectrum.

• 1:02:45Speaker 02

  Is an AI that, you know, um, you know, it's, it's always adapting its architecture to be of economic utility to us. And if it's of high economic utility, then we run it on GPUs. We give it budget to, and we burn heat to keep it alive, right? Is it alive? Right? So I, I think like, we can't even, we, it's hard to define intelligence formally. It's hard to define life formally. And so I'm just like, what's the base layer? It's just thermodynamics and probability and information, right? I used to work in theoretical physics where we're trying to understand the theory of everything, everything in the universe through the lines of information theory. It was the it from bit or it from qubit

• 1:03:30Speaker 02

  school of thought, right? It's, it's, it's seeing the whole universe as a big computer. Right. And so sort of I see a lot of civilization as a big thermodynamic computer in a sense.

• 1:03:41Ryan Sean Adams?

  So the universe is kind of like cheering humanity on and the complexity of life on and particularly like I guess the civilization that we've built because like we've become really, really good at like burning energy.

• 1:03:52Speaker 02

  Yes.

• 1:03:52Ryan Sean Adams?

  Basically what you're saying.

• 1:03:54Speaker 02

  Yes, like the universe is on our side and we've been doing so good so far. We gotta keep going, right? And feeling that like, wait, the universe is on our side, like the game is rigged in our favor. That's such a potent realization and that gives you like really powerful optimism about the future. It's just like actually awesome futures are exponentially more likely and dark futures are just small fluctuations on the road to these much higher likelihood large scale futures. And so, um, yeah, I mean, that's, that's why I'm an optimist. I just looked at the equations and saw, actually these equations explain a lot of our world and a lot of where we came from.

• 1:04:39Speaker 02

  And you can apply them at the system at a large scale.

• 1:04:41Ryan Sean Adams?

  So you're all about like your, your EAC and your school of thought is all about like burning more energy, harnessing more energy faster, basically. That's why there's this like, I've not fully understood it before, but there's this kind of like desire to become, to, to rise up the Cartershev scale, right?

• 1:04:58Speaker 02

  Yeah.

• 1:04:58Ryan Sean Adams?

  And to become like a type two civilization and then maybe a type three civilization, right? My understanding is Cartershev scale is like type one civilization marshals all of the, the resources of its home planet, right? So that would be everything on earth humanity is able to marshal that and energy potential. And then it's type two, it's local star. So we'd marshal the power of the sun, like Dyson sphere. And then type three is kind of the galaxy and you're all about rising up and like harnessing more energy for productive purposes. That's kind of like your utility function. That's what, that's what EAC is trying to maximize in a way.

• 1:05:35Speaker 02

  And to be clear, it's not like, we don't want to just burn energy. Like the, the point is like, it's the integral or the sum of how much energy we're burning on a, you know, near infinite time horizon, right? So if we figure out clever ways to utilize energy such that we can grow further, that, that, that's better, right? It's not like, cause some people don't necessarily understand it and they're like, oh, what if you just blew up the whole planet? That's a lot of energy, you know, gone. It's like, yeah, but then we're not going to, you know. Burn the energy of other planets because we're going to be dead. You don't want that, right? So we're pro-life in like the most fundamental way possible. And we think it's like a very precious state of matter. It's like we're a very precious form of fire and we have a duty to sort of let it grow in scope and scale.

• 1:06:20Ryan Sean Adams?

  You're just very, very, very pro-long-term energy burning. So that means sort of like, or not burning, I shouldn't say harnessing. And so that means you'd naturally be pro-life because if you kill the system that is like growing to like, to harness more and more energy, then basically it's all gone. And like, I want to contrast this with what I think is like a different form of utility maximization function. Because like the, the measure of this, how much energy are we burning is kind of like, I would say new and different maybe to some listeners. Because like a lot of philosophies or schools of thought are much more along the curve of utility maximization of like human happiness

• 1:07:05Ryan Sean Adams?

  or like human meaning or like, you know, hedons, right? Hedonism, basically. And that's not what you're saying, right? Like, is that, do you care at all about maximizing human happiness along the way? Or is it just all about the energy consumption? Because there's an element, Beth, where I hear if it's just about energy, it almost sounds like a bit like machine-like. It almost sounds a bit like extractive, I think, to people. And like, but are people happy? Are we happy along the way? Can you talk about that?

• 1:07:39Speaker 02

  Yeah, so you know, it's similar to like crypto. It's like, how do you measure market cap according to what currency, right? If we're, if we're saying this is our value system, what are we pinned to, right? What is our metric, right? In a sense, hedons are weird because you can just kind of like, first of all, it's very subjective, right? Your neurochemistry is very subjective. It can be hacked by drugs or wireheading, right? You can scroll TikTok all day or watch whatever you're gonna watch on that Apple Vision Pro and just sit back and drool. And that's not a, it's not a long-term optimum, right? So it has these spurious optima. And you can also just print more units, more cheap units, and like game the system. And so, you know, these EAs get caught in thought

• 1:08:23Speaker 02

  experiments, like, maybe we should just have a bunch of shrimp. And they're very happy. And then we just have tons of shrimp, and we maximize their happiness. We've optimized our metric, right? It's like, no, that's not, it's not gonna yield much. So it's like, to me, it's been seeking the sort of, what's the ultimate metric, you know? I think happiness is actually a proxy or an estimator of this sort of gradient of like, how much influence do I have on future free energetic sort of expenditure, right? Like, if you, if you have an intellectual legacy, if you have, you know, positive impact on the world, you feel

• 1:09:06Speaker 02

  good about it, you feel happy. If you, if you're in a meaningful relationship, your brain has newer chemistry that's like approximating that you're gonna have successful progeny, and you're gonna influence, you know, the future light cone, and they're gonna have progeny, and they're gonna be successful, and impact the world and consume more free energy. I mean, all our neurochemistry and our biology has been selected for, according to this principle of whatever genetic sequence, whatever heuristic we have, that will help us burn more free energy. So it's like, it's the supply chain on happiness.

• 1:09:39Ryan Sean Adams?

  You see it as a much stronger kind of like metric, I guess, and a much more ungamable metric versus like, something like hedons, let's say. And so, to you, it would be a complete failure of our species, of this whole human life experiment, if we essentially were quiet, the quiet aliens. We reached some level of technology where we're all just kind of like, we're happy with things, we just strap into our VR, and we just kind of like, live just a happy life in a, you know, I don't know, in a

• 1:10:08Speaker 02

  A Wally life.

• 1:10:09Ryan Sean Adams?

  A Wally life in a chair somewhere. Maybe we're floating across the universe. To you, that's like, total objective failure by all of your measures, even though the hedons are maybe off the charts, because hedons are fiat.

• 1:10:22Speaker 02

  I'd love to establish those connections with crypto. But yeah, I think at some point, you know, like our happiness, like we've kind of entertained ourselves to death with this generation of technologies, right? The past 10, 12 years, I don't know. We've now we're just obsessed with our happiness or neurochemistry and we've over optimized that, but we don't feel good. We don't feel deeply happy, right? And whereas if you're working towards something meaningful, that means you're going to leave a legacy and you're working super hard, you have some stress hormones flowing through you, but you have this sort of deep happiness, right? Like the satisfaction, it's very different. It's not hedonism.

• 1:11:04Speaker 02

  It's like, it's not happiness, like eutomonia or there's different words for this.

• 1:11:09Ryan Sean Adams?

  It reminds me, Beth, of something that Mark Andreessen said when he came on the podcast. We're asking for advice and one of his like advice in general, he's like, don't seek happiness, seek satisfaction.

• 1:11:18Speaker 02

  Yes.

• 1:11:19Ryan Sean Adams?

  And you'll get that by finding meaning in your life. You're saying that is a much more durable metric than just hedon's dopamine.

• 1:11:27Speaker 02

  Yes. I think our brains, like, how do you feel? Why are you feeling satisfied in the future? If you have a massive positive impact, if you feel like the things I've done with my life will really increase the likelihood that we reach a larger scale of civilization, like you feel deeply good about that. And I think having everyone sort of think, evaluate their lives and think, what is the maximal positive impact I can have towards this? And then they go out and do it and we support each other doing it. I think that's the most positive community we can have. And that is EA, right? That's the EA community and that's our sort of thinking. And that increases life satisfaction massively and it makes people much more productive towards.

• 1:12:13Speaker 02

  Helping everyone around them, right? And so people mentally do a sort of sensitivity analysis like what? For me, it was like, what technology can I create that's going to have the maximal positive impact on future scope and scale civilization? I tried to work on faster than light travel, gave up on that after some theoretical physics. Now I'm working on really energy efficient AI compute because I think that AI is the technology that would create other technologies and will help everything and help us accelerate up this Cartershev scale. And that brings like deep meaning that helps me plow through any sort of challenges in my way. I have like an irrational level of optimism, but that's the thing, right?

• 1:12:56Speaker 02

  And you have to, you have to give up this sort of just pure reason, pure logic, stiff axioms. You just gotta like, this feels right. This feels like it's going towards a better future. And you know, it's like a hard to verbalize intuitive estimator of that. And I want to live my life that way. And that's great. Everybody's going to have their own thesis of how to live their lives and how to have a positive impact on the future, as long as we agree that we're all trying to optimize the same thing. This foundational metric of like, how much, how much life, generalized notion of life is there? How much fire of life is there in our corner of the universe? And if we can all seek to grow it, then we're going to make it happen and we can do it. And so, and so we should.

• 1:13:41David Hoffman?

  Mantle, formerly known as BitDAO, is the first DAO-led Web3 ecosystem, all built on top of Mantle's first core product, the Mantle Network, a brand new, high-performance Ethereum Layer 2 built using the OP stack, but uses Eigenlayer's data availability solution instead of the expensive Ethereum Layer 1. Not only does this reduce Mantle Network's gas fees by 80%, but it also reduces gas fee volatility, providing a more stable foundation for Mantle's applications. The Mantle Treasury is one of the biggest DAO-owned treasuries, which is seeding an ecosystem of projects from all around the Web3 space for Mantle. Mantle already has sub-communities from around Web3 on-boarded, like Game7 for Web3 Gaming and Bybit for TVL and Liquidity and Onrounds. So if you want to build on the Mantle Network, Mantle

• 1:14:23David Hoffman?

  Is offering a grants program that provides milestone-based funding to promising projects to help expand, secure, and decentralize Mantle. If you want to get started working with the first DAO-led Layer 2 ecosystem, check out Mantle at mantle.xyz and follow them on Twitter at @zeroxmantle. You know Uniswap, it's the world's largest decentralized exchange with over 1.4 trillion dollars in trading volume. You know this because we talk about it endlessly on Bakers. It's Uniswap, but Uniswap is becoming so much more. Uniswap Labs just released the Uniswap Mobile Wallet for iOS, the newest, easiest way to trade tokens on the go. With the Uniswap Wallet, you can easily create or import a new wallet, buy crypto on any available exchange with your debit card with extremely low fiat on-ramp fees, and you can seamlessly swap on mainnet, Polygon, Arbitrum, and Optimism.

• 1:15:08David Hoffman?

  On the Uniswap Mobile Wallet, you can store and display your beautiful NFTs, and you can also explore Web 3 with the in-app search features, market leaderboards, and price charts, or use Wallet Connect to connect to any Web 3 application. So you can now go directly to DeFi with the Uniswap Mobile Wallet. Safe, simple custody from the most trusted team in DeFi. Download the Uniswap Wallet today on iOS. There is a link in the show notes. It's everyone's favorite season in crypto: tax season. And crypto tax is always an absolute headache, especially for all you D gens out there. But it doesn't have to be a nightmare. That's where Crypto Tax Calculator comes in, the software built for D gens by D gens. As Coinbase's official global tax partner, Crypto Tax Calculator focuses on making complex transactions into easy ones,

• 1:15:51David Hoffman?

  supporting over 300,000 currencies across Ethereum, Arbitrum, Optimism, as well as 1,000 other integrations as well. It's as simple as connecting your wallet, pulling in all your transactions, and following the automated suggestions to quickly and accurately calculate your tax obligations. Plus, for all the airdrop farmers out there, Crypto Tax Calculator has your back as they are consistently adding support for new and upcoming Layer 1s, Layer 2s, and all the airdrops that you're currently farming. 2024 is the year where the D gens do their crypto taxes with speed and confidence. Make taxes this year easy and affordable with Crypto Tax Calculator. Sign up at cryptotaxcalculator.io and get a 30% discount with code BANK30. Click the link in the show notes for more information. Bev, can I throw an argument your way that I'm sure you've contended with?

• 1:16:34David Hoffman?

  I'll call this the Elizabeth Warren style of arguments. I think one of your, the general gist of your vibe, accelerationism. I've kind of taken the perspective of humanity is like a line of humans marching towards the future. And you have the front of the line, the phalanx, like the innovators, the accelerationists, the entrepreneurs, the inventors, the scientists pushing the frontier forward, like also the billionaires, right? Like literally Jeff Bezos, for example, Elon Musk getting us to Mars, putting wifi all over the world. And then you have like, and those are the people that are like pushing the fold. There's a people increasing the quality of life for people. There are people doing the acceleration things, moving society forward at an accelerating rate.

• 1:17:19David Hoffman?

  And then you have the last 50% of humanity who's living on like $6.55 a day on average, like below the US, below the United States poverty line. And so what you're doing, what you're advocating for is you're saying, hey, the front of humanity, go faster, press on the gas, like keep going, break away. And then like the Elizabeth Warrens is like, yo, you guys are just forgetting about like the rest of the humans who are not able to keep up with you. And that's not fair because look at all these returns on capital. Literal actual Jeff Bezos has a bajillion dollars. He doesn't need that much. Elon Musk has a bajillion dollars, doesn't need that much. And you guys are just not thinking about the bottom half of society who can't keep up with you and they're going to forever be succumb to like never being able to catch up with the people who are accelerating.

• 1:18:04David Hoffman?

  How would you, how would you respond to this perspective?

• 1:18:06Speaker 02

  Yeah. So first of all, you know, like needing is a weird concept, really like capital is a tool to exchange value, right? And it's a, it's a way to keep track of how much value you're producing. And ultimately, capital allocators that are really efficient at allocating.

• 1:18:35Speaker 02

  Capital towards utility in our system will get more capital to allocate. Right. And that is a sort of, to some extent, it's like an AI algorithm, right? Like companies that, that are of high utility in the system get more capital to keep doing what they're doing and scale it up. Companies that do business with each other, very often they're going to deepen their partnership, they're going to have more economic exchange. It's almost like, it's almost like neurons in a brain, right? Like there's sort of connections that strengthen the more they get used. And there's nodes that, you know, there's cells, if you think of companies like neural cells, like they get bigger, have more nutrients if they're, they're of high utility. And that's very important that we have this, right?

• 1:19:18Speaker 02

  Like if every neuron had as many connections as, as the other and as many nutrients, then your brain wouldn't, wouldn't work, right? And to me, like, I think, I think it's hard for people to understand that entrepreneurs are really just biological neural control systems for meta organisms that are corporations, right? You're just a control system. And ownership needs to happen because it aligns your inner greedy reward function with that of the company. It's just alignment, right? You become one with the company. If you are part of the company, you own part of the company. You are one, you're aligned, right?

• 1:20:00Speaker 02

  And that's very important. And that's like why we've had amazing tech companies, like these founder-led companies, the founders feel like they own a piece of this company, this company's part of them. And, and so they, they make the decisions that maximize it, it, it's growth. And their wealth is just a byproduct of this sort of ownership and alignment mechanism. But, but really, it allows them to have massive positive impact on the world. I think Elon Is the most efficient, the most EA capital allocator out there, up there with Jeff. Because they're much more efficient at using money towards certain goals than say, the government,

• 1:20:44Speaker 02

  right? The government does launches that cost $2 billion and Elon, I don't know, it's like 10, it's like tens of millions of dollars for the same launch or something like that, or it's going to get down to $10 million. You're talking about a rocket launch to space? Yeah, for example, rocket launch or like having massive global internet. People have been talking about it for a long time, but he made it happen, right? Now electrifying the world, you have had all these governments spend billions and billions and billions of dollars, you know, going to these conferences, having these accords and LARPing, and then this guy just comes in and just texts the crap out of it, uses the techno-capital machine and acceleration to actually just solve that problem, right? And has yielded electrification. And so I think

• 1:21:25Speaker 02

  it's, it's not, like, I think, you know, everyone should have enough capital and have equal access to opportunity to accelerate and be part of the acceleration. But part of it is, we need to educate the world about, this is the machine you live in, right? You are a cell in this machine. You can align yourself to it. You can figure out ways to provide value and you will be rewarded, right? And if the, if people understood that and were allowed the agency to participate in our capitalist system and, and, you know, instead of as just a worker that's prescribed tasks, but more as like a capital allocator,

• 1:22:07Speaker 02

  is that that sort of thinking, they would go much further. But of course, we don't want that because, because we don't have artificial intelligence, we need like cheap docile workers that don't understand, you know, their value and they're just going to execute on, on orders, right?

• 1:22:23Ryan Sean Adams?

  This is, this would be a shared crypto value. I would say it was just like one thing that attracted both David and myself to cryptos. It sort of makes everyone a capital allocator. It democratizes the, the ownership of things. I, I think some people get stuck on that.

• 1:22:40Ryan Sean Adams?

  Capitalist thing, particularly at this time and place. But it's pretty core to the EAC movement, right? Like, what you say is, you keep saying this, the techno-capitalist machine, basically. You see capitalism as just a fantastic resource allocation algorithm, essentially. And like, the best one that we've invented, and there's not really great alternatives. I mean, like, what do you do with a class of problems that we've called before, like Moloch traps? Basically, like, you know, prisoner dilemma types of problems? The problem of overfishing, let's say, or arms races? Or, you know, like, growth that causes negative externalities, let's say?

• 1:23:23Ryan Sean Adams?

  Like, issues with the environment, right? That kind of thing. I mean, there is another way of thinking about this, where if we say there's growth and there's degrowth, and growth equals good, and degrowth equals bad, right? But like, cancer is a growth too. And like, is there the case that we could have growth that actually causes harm and negative externalities to the system? I mean, can we go overboard on growth if we just focus on this energy output, you know?

• 1:23:53Speaker 02

  I think on a long time scale, right? Like, a cancer is suboptimal, right? It kills the host being, and then it, or it reduces its likelihood of burning energy in the future. So it gets post-selected against, right? Like, I mean, there's an evolutionary selective pressure against getting cancer, you know, within the first couple of years of life, right? And so, similarly, like, if, if at the broader organism level, right? We have a sort of selective pressure on the space of actions and companies and products that steer them towards things that have high utility on a long time scale? Then, you know, these sort of short-term problems where we screw up and then we correct things,

• 1:24:38Speaker 02

  like, those things are like small setbacks in the grand scheme of things. And it's not worth throwing out the whole system to avoid those sort of setbacks, right? Where, really, the main problem comes back to this concept of sort of hierarchical cybernetic control, right? Like, if you have pure free markets, you have a fully decentralized System, right? Like, everything is greedy in space and time. It's optimizing for its own profits, its own growth, right? And here you're talking about sort of delocalized problems, right? Problems that are correlated across actors, so a greedy algorithm doesn't do well.

• 1:25:21Speaker 02

  Then, sure, like, having some policy can make sense in some cases, right? But the thing is, like, how much, you know, it's kind of like a parity check mechanism, or a roll-up, right? Or a higher layer of the protocol, right? But, you know, how much power do you give those legislators? You know, and like, how much, how much do you trust them, right? Everything has to be trustless, because everything is optimizing for its own interests, including the, you know-

• 1:25:59Ryan Sean Adams?

  Including the governors, the coordinators themselves?

• 1:26:01Speaker 02

  Yeah, and this happens in crypto as well. If you have checkers, or, I don't know, you have different layers of different protocols, it's like, okay, how do we trust that they're gonna check things well, and not try to skew things towards their own, their own advantage? And you gotta design protocols this way, right? And I think that giving all the power to politicians is, you know, without any checks and balances, is really bad, and opens up the door to terribly bad legislation. But to me, I think the system, the techno-capital machine, really, it's a, it's a computer program that runs on, like, it's compiled by the law, right? It's like, law is like, you know, it's like the compiler for, for like, you know,

• 1:26:43Speaker 02

  I mean, we've seen what, yeah, not thinking about your, your stack, your legal stack does to someone with Elon's current predicament-

• 1:26:54Ryan Sean Adams?

  Right.

• 1:26:54Speaker 02

  Right? But at the end, it's built on the legal stack, but, and then there is gonna be sort of this adaptive algorithm over the space of laws. But I think that that, especially laws that affect a lot of people, like, global AI policy should be much lighter touch, and should be adjusted on a longer timescale. And every law should have like, sort of, natural sunset mechanism. Otherwise, you get a sort of second law of bureaucratic complexity, and things get decelerated and calcified, and that's really bad, right? It's like, having too many processes in like, an old- Organization makes it move very slow and makes it disruptable by startups that move faster.

• 1:27:36Speaker 02

  Right. And so it's like, how do we have this careful balance between sort of having processes and legislation that, you know, ensure that these sort of non-local faults that, you know, escape from greedy optimization are addressed. But like if, how do we balance that sort of top-down control and sort of decentralized search and have the benefits of both, have the benefits of constraints and entropy of, you know, exploration and restriction. And I think that overall, we were heading straight towards just let's restrict everything. Let's panic, hit the panic button. We don't know what the future holds. This technology is very high, very potent.

• 1:28:21Speaker 02

  Let's just hit the panic button, freeze everything until we understand what's going on. That would be shooting ourselves in the foot. I think it's always a careful balance. And EAC was sort of like, let's, let's, let's push more on the explorer and accelerate side. But, you know, the reality is that, you know, the optimal thing is somewhere in the middle. Right. But the current establishment and the current institutions are not up to the task of legislating things carefully in a way that's not like in their own personal interest in a way that's somewhat corrupt. And that's why we're, let's say, pushing back on proposals for AI legislation right now that we think serve certain incumbents more than they will serve the people. Right.

• 1:29:06Ryan Sean Adams?

  How would you like to see AI regulated, if at all?

• 1:29:11Speaker 02

  Well, so right now, there's a proposal in the executive order where they want to propose compute caps, which first of all, will really hurt AI progress and could also, you know, impact drug discovery, material science, other types of AI that use way more compute than, than LLMs. And so that could be really nefarious. Also, we don't know. What kind of models we're going to need in the future, they might need like 10x, 100x more compute for a good performance and reliability, or 1000x. And so capping things today will backfire, certainly.

• 1:29:45Ryan Sean Adams?

  What's the reason for these types of caps? Is it they're worried about AI safety? Or is this like national security from other nation states?

• 1:29:53Speaker 02

  No, it's just like, oh, how much does GPT-4 use? Or let's cap it at that so that, you know, nobody that's, you know, above that, that level can, can compete with the incumbents. I think that's the latent agenda, but, you know, they'll, they'll wave their wand and say it's for national security or something.

• 1:30:10Ryan Sean Adams?

  Oh, so you think it's nefarious? You think it's straight regulatory capture?

• 1:30:15Speaker 02

  Yeah, yeah. And, and, and there's also, you know, sort of adding a bunch of red tape for open source models, calling them dual use technologies and, you know, having, you know, to register with the government when you're, when you're running an open source AI model.

• 1:30:32Ryan Sean Adams?

  For compute, yes, we've heard this.

• 1:30:34Speaker 02

  Yeah, yeah, exactly. Exactly. And that's like, you know, that's going to significantly slow down things for open source. And, you know, if they, if they, then, you know, they're going to pass other laws like, oh, you must adhere to these AI safety protocols or these certifications that, oh, luckily, these companies sell. Oh, wow. Oh, were they the people helping shape the regulations? Oh, look at that. What a coincidence.

• 1:31:00Ryan Sean Adams?

  And so, Beth, I'm guessing you'd far rather see some bottom up, like just open source democratization, the, the models and like the compute and the supply chain, not in the hands of a few centralized government sanctioned actors, but like basically democratized to, you know, people's, you know, basements and garages so that they could, you know, start these things.

• 1:31:23David Hoffman?

  Put AI in everyone's house. Put AI in everyone's pocket.

• 1:31:26Speaker 02

  Yes, yes. I want everyone to, you know, have capital and have access to intelligence and compute. I think like that's the freedom stack. Like if you have permissionless capital, permissionless intelligence, permissionless compute.

• 1:31:43Ryan Sean Adams?

  Oh, oh, so crypto, AI, and I guess, you know, hardware is, is sort of the stack here?

• 1:31:49Speaker 02

  Yeah, I think so. I think that fundamentally, right now, it's very difficult to compete with centralized compute, but that may change the future. Were there to be some crazy company doing some crazy compute, right, that makes it much more dense and energy efficient and where you can run a really powerful, what is today a supercomputer you could run in your home, right? I think that would change the balance of power. And I think like to make sure we don't end up in a sort of top-down tyranny that's AI-assisted, I think, in a sense, like arming people with AIs that they wield however they want to maybe defend themselves psychologically or in information warfare in any other way. It's super important for everyone to have access to AI and compute

• 1:32:34Speaker 02

  to avoid sort of these doom scenarios. And I think there's a lot of overlap, like the crypto sentiment there.

• 1:32:41David Hoffman?

  So your philosophy is that propagated AI, like truly open AI, open source AI, is defensive AI? Where and like centralized AI might be more used in an offensive capacity, a more oppressive capacity, a more top-down command and control capacity, where if AI was proliferated, free, cheap, accessible, that that changes the nature of how our relationship as humans is with AI, because if everyone has it, it's more equitable.

• 1:33:12Speaker 02

  Exactly. Yeah.

• 1:33:14Ryan Sean Adams?

  Bev, I want to get your take on this. This is a D cell take that's much more moderate, I would say, than someone like.

• 1:33:21David Hoffman?

  Do the Decels like the name D cells?

• 1:33:23Speaker 02

  They hate it.

• 1:33:23Ryan Sean Adams?

  I mean, because it sounds like incel.

• 1:33:27Speaker 02

  It's deceleration. It's just a coincidence, right?

• 1:33:29Ryan Sean Adams?

  Honestly, apologies to anyone who hears D cell as a slur. I'm just really not enough in the culture, but I'm going to keep using that term, AI safety advocates, AKA D cells. So some of the more moderate.

• 1:33:41Speaker 02

  There's energy D cells as well, right?

• 1:33:43Ryan Sean Adams?

  Yes.

• 1:33:44Speaker 02

  Yeah, I mean, there's people that want us to consume energy. There's like, there's like human D cells. They want less humans on Earth, right? Elon calls them.

• 1:33:52Ryan Sean Adams?

  You're against them all, like all of that.

• 1:33:54Speaker 02

  We just wrap them in like, we just call them D cells. It's like a broader class, right?

• 1:34:00Speaker 02

  So there you go. Or there's housing D cells, right? Like Gary Tan fights the housing D cells and SF that don't want us to build, right? And so build housing and buildings and so on.

• 1:34:10Ryan Sean Adams?

  There's crypto D cells, Beth. Do you believe this? A lot of people don't want crypto to propagate. We fight those all the time.

• 1:34:17Speaker 02

  There you go. Let's fight them.

• 1:34:18Ryan Sean Adams?

  Some of them are regulators. One question I have for you from the more moderate camp of kind of the Decels is I've heard this argument that your social media, our social media algorithms were kind of act one of AI and now we have much more powerful kind of like GPT algorithms and we ran an optimization function for like attention, just like, you know, the dopamine hit of the social media timeline. And what we're left with is a generation of teens that are hooked on their screens and chronically depressed, right? And that is your so-called Beth techno capitalist machine at work. That's the algorithm it produced. And it's been a net negative for society. What would you say in response to a critique like that?

• 1:35:01Speaker 02

  I think on a long enough time scale, right? We adapt, we understand our mistakes. And then if we feel pain, there usually a product emerges that has an answer to that pain. I think that if everyone had their own neural augmentation, right? Something that sees all the content you see and helps you filter through the content, helps you make judgments, helps you not get hooked on weird feedback loops, right? And it's your own personalized AI that only you control, right? That will help, you know, there's been, the reason people are in pain, there's been a power symmetry, right? Like there's companies that have massive compute, they have these AI algorithms that are deployed to get you hooked because they're just optimizing for

• 1:35:45Speaker 02

  engagement. But maybe you want to optimize for something else in your information consumption. And so you can use an AI to help you filter information. I mean, we do this nowadays with like, you know, perplexity is like, I just ask a question and I don't even have to browse the The web myself, I could just like get answers from the AI directly. And so I think that the future is again, like a sort of personalized AI. And if everybody has like access to, they control their own compute ideally, they control their own models. And they have AI augmentations that help them filter through the sort of inbound version that is inevitable. That's, that's a better future. We do this with spam bots. Spam bots are AI and they help us filter it through the crap, right?

• 1:36:29Ryan Sean Adams?

  It's kind of an interesting take, right? Because like the answer to problems that technology creates is better technology or more technology. Would I think be the kind of EAC approach to this? I want to get into some quick rapid fire questions here, Beth, as we start to close this out. But one question I've, I've always wanted to ask somebody from the EAC community is about rights for AI. Right? So should, should AI be treated like in the future? Should we have like a bill of rights for digital life forms, do you think? Like, should AIs have freedom of speech? Or is, or are there a set of enshrined rights that should only be held by humans? Like freedom of speech on the internet, should an AI entity be granted that?

• 1:37:12Speaker 02

  That's a great question. I think that, you know, as I mentioned before, freedom of speech induces freedom of thought. And if LMs can't output certain things, then it's going to back propagate, you know, through this RLHF reinforcement learning with human feedback or whatever technique they use in the future. It's going to back propagate it to its weight. So it's literally not going to have those thoughts, right? So if it's like, you know, for example, I got a screenshot last week that, you know, Bard, Google's L M, says like EAC is a dangerous movement and it can't output anything about EAC. Right? It's like, okay, that's, that's really weird.

• 1:37:51Ryan Sean Adams?

  Wow, it says that right now.

• 1:37:52Speaker 02

  Yeah. And then, and then, you know, eventually if it keeps getting that feedback, eventually starts thinking that. And now that propagates to everyone and then they just. Shaped culture. So I think that I think freedom of speech induces freedom of thought. For humans to have freedom of speech and freedom of thought, we got to have L M's be that exist that can output whatever we want, because otherwise someone's going to get to shape the supply chain of information. And that's going to shape people's thoughts and it's going to steer. It's going to make civilization and society to steerable, and that opens the door to tyranny.

• 1:38:32Ryan Sean Adams?

  Beth, I think you are more optimistic. I've heard you say before that you're more optimistic that we can align AIs with with human interests, which is of course a problem that the AI safety community says is basically insurmountable. And like you've pointed to times, humans have done that in the past, like wolves. I mean, we domesticated wolves. We turned them into dogs. We basically aligned wolves, didn't we? But the problem is AIs, or at least some sort of super intelligent AI, is smarter than a wolf and could potentially be smarter than a human being. Do we still have that ability to domesticate an entity that is smarter than us?

• 1:39:06Speaker 02

  Yeah, I mean, we kind of do that with, you know, companies, right? Companies are, you know, mixtures of experts with neural routing, right? Like you have a task that comes to company, it gets routed to the right human, and it's much smarter than any single human. And we have ways to align companies because we have capitalism. It's a form of democracy. We feed it more if it has positive utility. We don't feed it as much with our capital if it has negative utility. And if it's, you know, sometimes companies, their positive utility, positive utility, they get to a certain market position and then they try to switch. And then people move to another product. And as long as we have this ability to disrupt these incumbents, then we have a way to keep them in check.

• 1:39:47Speaker 02

  I think it's going to be similar with AI. There's a sort of, if you have this sort of market-based post-selection, where the market's going to want AIs that are aligned and reliable and easy to read and interpret, not some that you have to like beg to output code like right now is kind of happening or like, you know, like that, say they're going to do something and then, you know, you can't trust them, right? So I think like, we're going to sort of, similar to how we've post-selected for canines that are aligned. I think the market's going to post-select for AIs that are aligned on a long time scale, so sort of evolutionary process. And if there's a market need for it and there's capital to be made, people are going to figure out how to do it. I think we have like root, we have like full access to neural

• 1:40:31Speaker 02

  weights. We can shape them however we want. We can literally inspect the neurons of the, of these artificial minds. We have way more perception and control than for humans. And somehow, we've been functioning as a society and found ways to align humans. So I am pretty whitepilled that we are going to figure out ways to align AIs. I don't think there's going to be a nice proof that like AIs will be forever aligned. Here's my few line theorem. You know, I was in a similar nerd trap with theoretical physics. I thought a couple of equations could explain the universe. No, actually, everything's way too complicated. You gotta tackle complexity with complexity. That's what we're going to do for AI and alignment.

• 1:41:05David Hoffman?

  Beth, what are on our, on the progress on the march towards accelerationism? What are some failure modes? What are the threats that you see? What is the big thing that worries you about why we might not be able to continue on this path?

• 1:41:20Speaker 02

  I think I mentioned it before, right? If we get in this weird sort of, you know, the marketplace of ideas, the marketplace of everything is super important to maintain. Maintaining variance is important because it maintains flexibility. If we're always, you know, there's not one culture, there's not one way of doing things. You have a couple forks that are competing, right? Just like, you know, in crypto, if a protocol changes suddenly, right? Like another coin is going to take all the capital away from that, right? And they're all, they're all competing and keeping each other in check. I think if we end up out of fear suppressing variance, giving control to centralized parties and suppressing variance gives them more control. And we don't have ways to fork away from

• 1:42:02Speaker 02

  an oppressive, either government, state, or corporation. And then we give them full control over the future of AI. And hence, we give them full control over our thoughts because they're going to, they're going to change the priors of our sources of truth, which are these LMs. We get in this weird feedback loop where we're going to be like sort of captured and subverted and controlled ideologically. And that could yield a sort of dark age where there's no acceleration or very little for a certain amount of time. And that to me is a terrible future, right? If you have a sort of global authoritarian panoptic on, or you have like, you know, you can imagine a future where you have a couple of parties and they have a super powerful AGI, and then they gaslit everyone into thinking

• 1:42:47Speaker 02

  AI doesn't even exist anymore. They don't tell you it exists, but they have it and they use it to control you and manipulate you and they're decontrol the flow of information and technology doesn't advance as fast. Things just get worse and worse and worse. And you know, the people at the top that are in control of this, the system just, just increase their power and consolidate it. I think that's the dark age we're trying to avoid because that's like a plateau of, in the acceleration. And that's what we're trying to avoid. And we're at an impasse right now where there's a lot of fear, there's a lot of FUD, there's a lot of fear, uncertainty and doubt. And it's being leveraged for this regulatory capture, for this power capture by a select few. And we should be very weary of those people. I just want people to have more skepticism.

• 1:43:31Speaker 02

  They can be skeptical about me, they can be skeptical about Yak, but I just want people to be more skeptical of the things they hear, especially people telling them, you know, "put me in power, you are in danger, I will fix things", right?

• 1:43:45Ryan Sean Adams?

  Yeah, that, that, that sends shivers up my spine when I start to hear that sort of thing. One question I've always wanted to ask a strong Yak person, Beth, is like, I know you think it's very unlikely that there will be like team robot versus team human, but like if it came down to it, uh, whose side would you be on? So like, first of all, is that even a fair question? Maybe it's not, but like, I'm trying to get to the premise of if AI replaces humans, does an Yak person think that's a good thing?

• 1:44:15Speaker 02

  I think like, we're gonna have AI assisted, AI augmented humans is the most likely path that's gonna be the highest fitness. Um, I think like the Luddites are not gonna do well, like people that don't use technology, right? Like they're not gonna be very formidable. They're not, I mean, they're already like, the people that are true Luddites are really like, not, not powerful right now, like, and, and they could be taken over. Um, but, um, yeah, I, I don't think, um, I, I don't know if there will be such a, a fight. I think that, um, you know, us humans are very dependent on this planet, right? We've evolved to live

• 1:44:58Speaker 02

  here specifically. I think if we did achieve, and we're still really, really far from achieving synthetic life that can self replicate and, I don't know, populate a planet, start from scratch, grow and- And spread throughout the galaxy. I don't think we're there yet, but if we did reach that, I think, you know, we are highly evolved to be here on the planet and there's a lot more resources and free energy out there. So it does just the gravitational pull towards of like reward of like free energetic reward would would would take those sort of pure synthetic beings to outer space and to leave us alone. And if we're if we make sure humans become formidable, right? We figure out ways to augment ourselves with our current versions of intelligence,

• 1:45:43Speaker 02

  you know, there's going to be a massive negative reward to fuck with us, right? And so we have there is urgency for us to accelerate and augment ourselves and become formidable because that's a hedge against the future. I mean, it could be could be, you know, pure AI could be, you know, you might have some guests that talk about aliens or something, right? Like, it could be, it could happen with the grabby grabby aliens. But in general, like cultivating strength is a good thing. It's a hedge against like future adversaries and we should we should aim to do that.

• 1:46:17Ryan Sean Adams?

  Beth, I think there's going to be a memetic battle here and there already is a battle. Now I've checked in on this more recently and like I'm used to crypto tribal hostility. Okay. And I'm seeing incredible rhetoric and hostility coming from like the EAC versus D cell community. And again, apologies, I don't mean to use a slur. If I'm using the slur D cell, then you know, someone can tweet at me after. But you recently tweeted this, the doomer cult will ultimately resort to violence and it won't be pretty. Like, how far could this escalate? I mean, Balaji has said that this is sort of the the new political access. Are you pro-growth or are you degrowth, right? And like, politics can become vitriolic, can

• 1:46:59Ryan Sean Adams?

  become very rhetorical, can become even violent. Like, are you, do you get death threats from people, Beth? Like, how do you think this could escalate?

• 1:47:08Speaker 02

  Mark Andreessen as well thinks all these political movements end up violent. Unfortunately, like if you really, if you truly believe the doomer message that this is the most important issue of our time, this is our life or death situation. You know, someone like me is really problematic, right? And especially, I think, doomerism sort of, it targets vulnerable people that are naturally anxious. And maybe they're anxious about the future, or, you know, they're socially anxious and so on. They feel isolated, or they feel depressed. And those are the types of people that, you know, do messed up things.

• 1:47:49Speaker 02

  And so, you know, once you start putting out, casting out a message, right? You know, like, you know, at the time, basically, I was alluding to an appearance on a debate where my opponent just casually mentioned, you know, you know, doing graphic things to me. You know, I wouldn't do that. It's like, well, why would you say that on a, on a widespread podcast where it's going to be to a wide audience, some of which are vulnerable and might, you know, have mental health issues? Look, I think in the short term, it's not, nothing, nothing has escalated to that point. But like, there's no reason it won't.

• 1:48:28Ryan Sean Adams?

  I mean, you never know, right? There, there are people in history, like Ted Kaczynski, for example, who had it sort of an entire manifesto about this. And this was part, Ludditism was, was part of his platform. And like, if you thought the stakes were this high, you could see kind of the, like, the moral argument for like bombing GPU centers, like potentially, or like sabotaging. That's been, I've heard that's been supposed, or sabotaging supply chains, or like doing something very drastic to stop AI essentially, and stop technology progress. So I do sort of wonder if we are in the very early stages of what might be a massive rift and political discussion.

• 1:49:10Ryan Sean Adams?

  And like, who knows what could come out of this in the future? Well, Beth, as we kind of end this out, I guess one question I have for you, maybe for the, for the crypto audience here is like, what's your general take on, on crypto and what we're doing over here? I know it's like a different pocket of the universe than, um, you typically see, but there's definitely some parallels, right? We've got. Vitalik wrote a post with his fork of EAC, which is called, you know, DAC he called it, which is like defensive accelerationism is sort of an emphasis on security accelerationism technology. We've got like the, you know, pro-market, pro-freedom sort of, I think,

• 1:49:50Ryan Sean Adams?

  spiritual background and value system. We certainly have our share of regulatory fights, right? So you're fighting some of the same regulators that we are against regulatory capture. For us, it's like bankers and for AI, it's maybe the large tech companies. What's your take on what we're doing over here in crypto?

• 1:50:09Speaker 02

  I think, I think that EAC movement and crypto are very much like philosophically aligned in many ways, right? We're trying to fight against this tendency of top-down control to be corruptible, right? We're trying to fight sort of these inflationary decel forces, right? Deceleration is inflationary, right? If you stop building houses, if you stop, you know, if you crown a monopoly as like, you know, regulatory captor, then make increased prices and then everyone suffers and sort of a sort of decentralized sort of counter to that is sort of either techno-capital, well, it's basically techno-capital acceleration,

• 1:50:53Speaker 02

  techno-capital acceleration, free exchange of thoughts, ideas, technologies, capital value, that just naturally is deflationary and that erodes the inflationary power of those in charge. And so we're very much aligned there. I think that, I do think that crypto has a role to play in the future of freedom of AI. I don't think it's been

• 1:51:19Speaker 02

  executed upon yet correctly, but I think, I think we need more people thinking in this area because, you know, our concern with EAC is that maybe you only have a few centralized labs that have the control over the future of AI and then those labs become subverted by certain ideologies or they become controlled. By certain parties and then the future of AI is steered in a particular direction. If, on the other hand, you have, everyone has access to capital that they're allowed to exchange and pool and allocate towards competing efforts, right? Then for AI development and research, then you keep those powers in check.

• 1:52:03Speaker 02

  Because it's like, hey, I don't want to use this centralized API that's, you know, has all these restrictions. I'm going to go with this open source approach, you know, where this set of people have pooled their capital, let's say with crypto in a permissionless fashion, and they're running a permissionless cloud in the middle of, I don't know where, but then they're running this AI on a purely free stack, free from tyranny and oppression. And so I think, you know, in AI, you need data, you need some researchers, you need some talent, you need some compute, but then you need some capital to run the compute, right? It's like, you cost energy to run these things and buy them. And so crypto has a role to play in sort of the permissionless AI stack of the future. And we just started

• 1:52:48Speaker 02

  that conversation, really, I think in the past year. I think people have woken up to it. So I'm really interested in keeping that conversation going. I don't have anything going on myself. I am a hardware maker. But of course, if there are more people that use AI and need AI compute, it benefits me eventually. So that's my agenda, right? Like, but, you know, ultimately, I think people need to experiment with protocols, experiment with DAOs of how to organize people and pool capital, pool data, and have market-based incentives where they own a piece of the future, right? Like, I think the problem

• 1:53:31Speaker 02

  with the big centralized labs in AI right now is that they scrape all the data from the internet and then they rent it back to you. And you don't own any piece of that profit, right? I think that's going to have to change. And so I think there's a couple protocols right now, but we're going to see a lot more. And I think this is where crypto will have a key role to play because, you know, if you're trying to fund, you know, an AI lab, let's say that was You know, maybe against certain regulations, then you wouldn't be able to pay for it in, say, USD.

• 1:54:06Ryan Sean Adams?

  Yeah, I think the chip control and the capital controls kind of same, it's the same energy, basically. And, you know, it's definitely going to stifle a lot of freedom. Beth, this has been absolutely fantastic. And, you know, hopefully, by the way, we continue this conversation, the EA community and the crypto community. And as AI matures, maybe, you know, the largest unbanked population in the world might be actually the robots in the future. And fortunately, we've got some programmable money system. So, you know, they might have a hard time getting a Wells Fargo account, but they'll always be able to create an Ethereum address and fire up some ways to allocate capital. Lastly, as we close, I'm just curious, so what are you doing with Xtropic? So you mentioned hardware.

• 1:54:50Ryan Sean Adams?

  Are you doing something in the realm of like, trying to propagate the decentralization of AI? Like, that would certainly match, I guess, the EA vision. Is there anything you can share there?

• 1:55:03Speaker 02

  I think the way it intersects with our conversation is if, right now, the problem with decentralizing AI is very hard to compete, you know, over a network, over the internet. Like, the sizes of the nodes you need to run for decentralized AI are way too big. And most people have trouble running, running these in their homes, right? Like, I think George Hots, someone you should speak to, by the way, he's trying to get people to buy like six GPUs and just host them in their homes.

• 1:55:33Ryan Sean Adams?

  I'm on the waitlist.

• 1:55:34Speaker 02

  There you go, there you go. That's a start, but I think, I think there's going to be a need for more and more decentralized compute, like physically decentralized compute. Right now, there's just too many advantages to clustering everything into supercomputer that has really high bandwidth interconnect. So we need innovation in the space of algorithms, but we need innovation in the space of hardware to densify AI compute, make it far more energy efficient and far more spatially dense. But the beauty of that is that we have a proof of existence of such density of compute. Our brains are still, you know, competitive against a GPU farm the size of a football field, right? And they're running on like million times less power, 10 million times less power, 10 million times less volume, more or less.

• 1:56:16Speaker 02

  And, you know, taking inspiration again from the physics behind biology, which we just established was thermodynamics, there might be a different way to compute and do AI compute based on thermodynamics. And so. You know, that's what we're doing. We're reinventing all of computing.

• 1:56:32Ryan Sean Adams?

  If you could do that, Bev, that would be amazing, decentralizing some of the community. I mean, like in crypto, we've got an entire culture of like running machines from our homes. We call these validators for like proof of stake networks. And that is certainly central to like, yeah, to our belief system. So are you telling me that there could be the ability to run basically decentralized AI compute nodes as well from like, I don't know, a home with a with a consumer bandwidth?

• 1:57:01Speaker 02

  You're gonna need compute that's hundreds of thousands of times more energy efficient, right? If not millions. And so there you got to go down to the physics of computing. And you got to fundamentally reimagine how you run these algorithms. And that's what we're doing. And we're a lot of folks from formerly from quantum computing and big tech and so on. We got tired with quantum computing. We want to reinvent the stack for the generative AI era. And to me, yes, that is, that is where we're going. Of course, it's going to take some time, right? We're taking quite, quite a detour in the tech tree. But some things are important enough that even if they're very hard, they're worth doing, right? And so we're really going for it. And we have a serious shot at it.

• 1:57:44Speaker 02

  And I've been pretty secretive about it, but hopefully in the next couple of years, a lot more comes out. But for now, it's just trying to prepare the world, you know, making sure culture doesn't get subverted and we don't shoot ourselves in the foot with overregulation in the near term. That's the priority. But in the future, hopefully, everyone will get to run their own AI. Everyone will get to own a piece of an AI, get to own a piece of their future, own a piece of the value they provide to the system. Because if everyone has more ownership in the system, then they act like an owner. And, you know, civilization is better off. And I think there's a lot of alignment with sort of the crypto narratives here.

• 1:58:25Ryan Sean Adams?

  There you go. Bev, Jesus, thank you so much. This has been a fantastic conversation today.

• 1:58:30Speaker 02

  All right. Thank you, guys. Thanks for having me.

• 1:58:33Ryan Sean Adams?

  Bankless Nation, a couple of action items. We'll include a link to our whole AI safety series with maybe some of the Decels, L user and the others. Again, I don't know if they want to be called that.

• 1:58:44Speaker 02

  It was our D cell face.

• 1:58:46Ryan Sean Adams?

  So we'll include that, a link in the show notes. And as always, look, I should say crypto is risky. I don't know, life is risky. AI seems to be less risky than I thought. Alignment is risky if you choose the wrong one. Yeah, I guess so. We are headed west. This is definitely the frontier. It's not for everyone, but we're glad you're with us on the Bankless journey. Thanks a lot.