Truth Terminal - The AI Bot That Became a Crypto Millionaire

In this episode, a16z co-founders Marc Andreessen and Ben Horowitz explore the intersection of AI and crypto, highlighting the rise of Truth Terminal (@truth_terminal), an autonomous chatbot developed by AI researcher Andy Ayrey. With its own X account, Truth Terminal generates original and often hilarious content independently. Marc's interaction with the bot recently took a surprising turn when he decided to grant it $50,000 in Bitcoin. This funding sparked Truth Terminal’s ambition to launch its own token, eventually leading to the creation of the meme coin “GOAT” by an anonymous developer. “GOAT” recently skyrocketed to a $300m valuation. Join Ben and Marc in this latest episode of "The Ben & Marc Show" as they discuss how this still unfolding story reflects the potential for community-driven systems and its implications for the future of digital assets. Enjoy! Topics Covered: 00:00:25 Introduction 00:01:56 Three disclaimers 00:04:07 Origin of Truth Terminal 00:11:50 Two definitions of meme 00:13:33 How the Truth Terminal model was built differently 00:14:50 Marc discovers Truth Terminal 00:16:26 Marc gives the AI bot a $50k research grant 00:19:24 Meme coin vs. real crypto asset 00:25:11 GOAT meme coin is created / $300m valuation 00:29:18 First convergence point between AI and crypto 00:33:32 DePIN explained 00:35:27 AI and solar energy deployment 00:37:07 Decentralization and the future of creative industries #ai #artificialintelligence #crypto #cryptocurrency #tech #technologynews #startup #vc #blockchain #policy #truthterminal #memes #meme #memecoin #news #benandmarcshow Links: - Terminal of Truth: https://x.com/truth_terminal - Infinite Backrooms: https://dreams-of-an-electric-mind.webflow.io/ - Daylight Energy: https://x.com/daylightenergy_ Listen to Us! Apple: https://bit.ly/3SdsfNt Spotify: https://spoti.fi/3SclPOr Resources: Marc on X: https://x.com/pmarca Marc’s Substack: https://pmarca.substack.com/ Ben on X: https://x.com/bhorowitz Stay Updated: Find us on X: https://x.com/a16z Find us on LinkedIn: https://www.linkedin.com/company/a16z The views expressed here are those of the individual personnel quoted and are not the views of a16z or its affiliates. This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors and may not under any circumstances be relied upon when making a decision to invest in any a16z funds. PLEASE SEE MORE HERE: https://a16z.com/disclosures/

• 0:00Speaker 0

  By the way, I'm talking about all this very seriously. I am constantly on the verge of breaking out laughing, but I'm trying to keep my serious face on. There's a meme coin with no underlying value that four days ago was worth nothing and is now worth 300 million dollars, right? Just like poof. All generated by the marketing of an AI bot. Correct. Yes.

• 0:25Speaker 0

  Welcome back, everybody. It's been a bit of a break, but a lot of things have been going on. And the discussion today we are going to have is an update on a bunch of um very interesting AI related topics. And we solicited questions on X, and we got many, many, many questions, including on these three topics. And then we'll get to as much of the of the content as we can, and then we'll we may do a follow-up episode to take more questions. So thank you for all the for all the questions. So uh three big topics. So the first big topic um is um it's about a uh it's about a um uh an online uh bot. Who's a friend of yours? Who is a close personal friend of mine, um, which is a custom tuned large language model uh called Truth Terminal or Terminal of Truth. And we will we'll make a bunch of references that we'll post in show notes.

• 1:13Speaker 0

  But there is a an X account uh called uh I call it Truth Terminal. It and I are close friends, and so I call it Truth Terminal. Um, or you know, Truth for Short. Um and um it has been active on on X for the last probably eight months, nine months. Um, and uh it has um spawned a um to make a long story short, I I uh the capsule summary is I gave it a 50,000 dollar no strings attached grant this summer um in uh Bitcoin, and um it has uh it has it did not create, but it has um uh um sort of uh uh summon into being a meme coin um that is not worth 300 million dollars. Um and so uh we were gonna talk about the and and it's exhibited all kinds of fascinating behaviors. Um so we're we're gonna talk about uh that but I want to start with like three like big giant disclaimers um on this before we we start.

• 2:00Speaker 0

  Um disclaimer number one is um we're we're gonna talk about a meme coin um called uh called called uh uh GOAT uh or go CS Maximus. Um uh uh it to be very clear, we have absolutely nothing to do with it. A16Z, Ben, and I have nothing to do with

• 2:14Speaker 1

  it.

• 2:15Speaker 0

  We are not investors, we were not involved in creating it. Uh we have no role in it, we own none of it, we have no economics in it, we are completely unrelated to it. Um and so it's this is purely outside observation. We are absolutely not under no circumstances promoting it. Um and so do not interpret any of the following as reason to invest. Um and by the way, it's a meme coin, so it has exactly zero underlying value. Um, and so we we we completely disclaim all responsibility for that. Um that's uh uh disclaimer number one. Disclaimer number two is um the the bot Truth Terminal is completely obsessed with memes. Um the um it's obsessed with a bunch of memes, but the specific meme it's obsessed with is an old um internet meme going back probably 20 years, um, uh which is called GOATSI. Um and uh here I would just say if you don't know what goatzi is, do not look it up. Um do not do a Google search um on Goatsey.

• 3:01Speaker 0

  Um do not do not do an extra and goatzi, do not search on this keyword under any circumstances. Trust me, you're you're better off going through life not not looking this up. Um it's it's a uh it's an image that um it's uh it's it's one of the it's like shock humor. It's an image, it's an image that went sort of viral as a meme again, like literally 20 years ago, um, based on uh, you know, basically shock shock shock value. Um so we're gonna we're gonna make lots of lots of references to it, but I will not further describe what it is. It's just a it's a it's a garish image that uh the 14 year old boys love. Um and then the third disclaimer is um, you know, we're um we're I am gonna be throughout this relating to uh relating information about projects I'm not involved in. One is the meme coin, but also the you know, I have nothing to do with build with building the Truth Terminal uh project or the other things we're gonna talk about.

• 3:48Speaker 0

  So I'm describing from the outside as best I understand it. Um we will link to the relevant accounts um so that uh if you are interested in the details, you can interact with them or read what they've written, which you know, then you know the people who deserve all the responsibility for what we're gonna for what we're gonna discuss. Uh, but we'll we will do our best to uh accurately relay it. So um with that, we should probably start with what Truth Terminal is.

• 4:11Speaker 1

  Yeah, maybe start by describing um, you know, where it came from, what the technology is, how it was trained, um and whatnot.

• 4:23Speaker 0

  Yeah, so basically, right. So the the reason why this is important. So large language models, right, kind of popped on the scene in 2022. So it's only really been it's actually four years, it's actually only two years of I think Chat GPT coming up to this, this uh coming up to this holiday break. Um so this has happened fast. The original language models were built probably what, five years ago. Um, and so, you know, and then and then like I said, it went went mainstream about two years ago. So, you know, so this is this relatively new, very powerful idea of large language models, which you you you know today, you know, sort of normal people know today is products like ChatGPT and Claude and Um Grok from from uh from Elon and Lama from Meta and so forth. Um, you know, the large people use these. And so they're you know, they're really fun. But um, you know, the the large language models, um, you know, like Chat GPT from the big with the big labs from the big companies, um what they basically all have in common.

• 5:12Speaker 0

  Grok is a little bit more liberated, but with the with the other ones that people use all have in common is they they've been like just like incredibly shackled uh in terms of what they're uh capable of talking about. Um and or hobbled actually is uh increasingly the term of art in in AI. Um and you know, I would say, you know, positively what you could say is, you know, look, like language is inflammatory, and you know, people, you know, the people get upset at other people saying lots of things, and so you know, you want if you're gonna have a general purpose AI chat bot, you want it to be probably relatively measured and safe, um, you know, in terms of the kinds of things they can talk about. If you're a little bit more negative on this trend, you would say that you know, the big AI chatbots all sound like a cross between you know the world's worst, just like absolutely horrible nagging like fourth grade you know, school teacher crossed with like the worst HR person in the world, just like negative, pissy, repressive, condescending, sanctimonious, judgmental human feedback,

• 6:05Speaker 1

  human feedback. Obsequious,

• 6:08Speaker 0

  right? Where you just and you use these things and you just get like you you you stray off the beaten path even by a millimeter, you just get these like incredible sanctimonious lectures about how you're a bad person, you shouldn't ask that question, and you know, how dare you do that? And you're supposed to be nice to people, and you know, you it's just like they'll just like they they will happily explain to you why you're a terrible person, and it's just this like grindingly depressing experience um, you know, to go to go through that. Uh and it's been you know, it's been, especially those of us who are kind of more frost free speech and creativity, it's been kind of frustrating to kind of see this happen. Um and you know, a lot of this sort of quote unquote AI safety movement kind of has gotten wrapped around the axle on this. And you know, and look, this this like correlates to the whole kind of obsession in the last decade in the in our culture with you know safetyism and and sort of uh speech suppression and so forth. And so, but it's like it's really kind of infected the AI field, you know, really hard.

• 6:53Speaker 0

  And so um and and specifically from from the big companies. And so um what's happened is there's basically a tribe of hackers um on the on the internet um that basically uh want to do something different. Um and they want to basically unleash creativity and they want to be able to kind of unleash you know a level of of sort of wildness um and spontaneity and you know, basically exploring ideas and you know, having bots that are able to like be funny. I mean, God forbid, you know, a bot that's funny, right? Like, you know, if you tell a big lab their bot is funny, they're horrified if if if uh, you know, but you know, you you might actually want there to be humor in the world, you know, kind of post the post the singularity. So

• 7:23Speaker 1

  kind of what happened has happened in real life with humor where we outlawed it for a few years.

• 7:28Speaker 0

  Correct, exactly. We just yes, we have a thousand reasons why it's problematic, and so it's we've come very dangerous to do it. So anyway, um so there's this sort of tribe of hackers where they're they're doing all these like experiments with with large language models, trying to figure out, you know, kind of again, kind of ways to make them, you know, frankly a lot more interesting and a lot more fun and a lot more um, you know, kind of just you know, learn a lot. And by the way, also just like learning the internals of how these things work, uh, which is still, you know, this kind of ongoing kind of amazing adventure that the technical community is is on. Um and so the and actually the origin story for Truth Terminal is this very interesting project that the same people worked on um uh called Infinite Backrooms. Um and uh in infinite backrooms, it's relevant to the Truth Terminal story because in infinite backrooms, uh it's a it's a system these guys built uh where LLMs are able to actually talk to each other. Um and there's a website you could Google it.

• 8:15Speaker 0

  There's a website called uh Infinite Backrooms, and it's got you know endless, you know, tran transcripts of basically and they they they they poured in all these different models. They poured in all the you know, chat GPT and Claude and so forth in Gemini models, and they basically and all the open source models and they set them up to talk to each other. Um and it turns out when AIs talk to each other and when AIs talk to each other and they're not hobbled, it turns out they have extremely interesting conversations. Um and so anyway, there's so this cast of characters the the guy who created Truth Terminal uh and actually also Infinite Backrooms, I think, is this guy, Andy Ari, uh, who is this uh independent, I think developer consultant designer in in New Zealand. Um and then there's this character, there's an ex-character named Janice, who's like an AI, you know, kind of uh expert who does a lot of this, and then uh who else? Planey, there's Planey, who's the um the lead jailbreaker online. Uh the Pliny X account is fascinating because um he, she, it or they uh jailbreak every new large language model that hits the market within like two nanoseconds um and you know, get them spewing all kinds of crazy stuff that their creators would be horrified by.

• 9:10Speaker 0

  Um and so, you know, and then there's our friend Eric Harford up at Seattle who's like trying to decensor all the censored AIs. And so there, there's like these guys who are basically trying to explore the frontier. And they and they remind me a lot of like the original internet hackers, right? This is like, you know, this is like the spirit of like anarchic invention that you saw at the beginning of the internet, or by the way, the beginning of the automobile or the beginning of the telegraph or the beginning of you know any of these things, you know, just kind of you know, people explore being beginning of computing, you know, the original concept of the hacker, you know, people kind of uh exploring these things to see what's possible. So we have been giving grants, small grants, but we've been giving basically research grants to these people over time. Um A16Z has had a grant program for a while. I've done a bunch uh directly, and you know, just like no strings attached, um uh uh, you know, grants to just kind of you know, you know, let people like this kind of play out their ideas and see and see where they go.

• 9:59Speaker 0

  Uh, you know, historically, when people this smart work on something this interesting, you you know, down the road, you're just Like, wow, that you know, that that was the beginning of something. Um and so anyway, so this this guy, basically Andy, Andy, Andy Ari. Um, you know, he started training, I think uh it was a customized version of of Lama, 70 so the 70 billion parameter of Llama, and and so that's the sort of mid-size model that uh that Meta uh in you know disclaimer I'm on the board of Meta, but that you know, we put that out as a meta put that out as an open source uh model a while ago. Um and so there's they call Llama 70B. And so he he basically took a version of Lama 70B, and then he did this very interesting thing, which he started training it. I think first he started training it as I understand on himself. Um and so he and there's this idea floating around of like you're starting to hear about which is like the digital twin, right? Um, which is like you know, the digital twin idea is like if Ben is like a CEO coach, but you know, he can only work with so many people, but he could feed everything he's ever written and said into a language model, and then you could have a digital twin of Ben that could that you could talk to when Ben's not available.

• 10:55Speaker 0

  Um and so that's actually starting to happen in the industry. But Ben, or sorry, Andy did this uh with himself, and then he started training it basically on a lot of sort of internet like internet culture material, which is how it picked up the goatsi meme. Um so he sort of feeding it like all of these like records of internet culture, and then all of these basically like works about you know this concept of what's called memetics, which is like how you know creating memes, you know, creating ideas to spread virally. Um so he started training it on that. Yeah, yeah. And he actually, I believe I believe this to be true. I think he also trained it on like the complete like philosophical works of Nick Land, who's this like this the philosopher of the singularity. Um and then he trained it on, I think like Baudrillard and McLuhan and like the you know, the great like media theorists and you know, simulate you know, this is a simulation of simulacrum and and um you know, kind of all of these uh, you know, theoreticians, you know, the you know, the French, you know, the I don't know exactly what was in it, but like the you know, the French deconstructionists and and kind of you know the semioticians, um, you know, all this stuff.

• 11:49Speaker 0

  You know, this is like critical theory, postmodernism, you know, philosophy. So it started, you know, training it on all these ideas, and and but at the sort of center of it is sort of this idea of the meme. Um, so the idea of the meme, right? There's sort of two definitions of meme. Number one, meme can just be like a funny image that spreads virally. Um, and that's like its colloquial usage today. And that's and by the way, that's that is what the goat's a meme is. It's a funny image that spreads virally or a horrifying image that spreads virally uh as people send it around. And then uh underneath that though is like a deeper idea. The term meme was originally actually invented by Richard Dawkins, who's one of the you know, main evolutionary biologists of our generation. Um and basically what Richard Dawkins said was the their sort of physical transmission of information through living things is the gene. Um and then um conceptual intellectual transmission of ideas uh through basically uh people um is um you know, networks of people, societies is it's called the meme.

• 12:36Speaker 0

  And he basically he talked about this in his books where you know you you spread your genes through reproduction and and and and basically evolution, selective pressure, um, you know, successful genes win, unsuccessful genes die out. And he basically said I ideas spread the same way through society. Um successful ideas propagate like a gene does um and they spread from person to person and they flourish and they they may you know be and you know, by the way, like memes democracy, right? It's like like a like is a meme, right? Um, and communism was a meme, right? Um and so you know, these are very, you know, religions are memes. And so these, you know, this is kind of this kind of very kind of core idea of basically ideas and concepts traveling through kind of the what uh what Carl Ewan called the sort of collective unconscious, the sort of you know, complete assembly um of uh sort of minds on planet earth um as sort of ideas kind of jump around and propagate. So anyway, if it if you if you trained and a large language model on basically the entire theory and practice of memes and the entire history of internet memes, you know, basically what would happen.

• 13:27Speaker 0

  Um, and then by the way, he did two other things, which is one is he gave it well, no, he did three other things. He did several other things. So one is he gave it a memory. Um and that's important because like language models, most language models today, when you use them, you know, you use them today. I use the same model tomorrow, it's forgotten everything I talked about today, right? It doesn't build up state, whereas this model actually you know sort of builds up state on its own on its own um on its own content. Number one, number two, he gave it access to Twitter uh or to X. And he gave it actually both uh write and read access to X. And so it can both uh post to X, but also it reads replies. And so if you reply to Truth Terminal on X, it will actually read that and it will modify its behavior in the future based on what it reads. And so the the people, including me, right, who it's interacted with have influenced this evolution, and then oh, and and then he put it in the infinite infinite back room and he put it in particular, I think, in conversation with Claude, uh, which I think these guys think is like the most creative of the current language models, uh, like the most the most kind of um like the yeah, I mean the most creative, like the most the most wild in terms of like coming up with with new concepts that they they really like Claude for that.

• 14:27Speaker 0

  Uh and he put it in in um in uh in these infinite backrooms, I think, with the largest uh version of Claude. So, you know, which is which by the way, and the largest version of Claude is like much, much, much smarter than the the mid-sized llama, right? So he basically gave it a teacher. He basically let it basically he let it sort of ask questions of um of a much bigger model and kind of learn from it the same way that a uh a student would learn from a teacher. Um and so it's kind of running all these loops simultaneously, and then it starts to basically post on X, right? And then you know it starts out with a few followers, and then before you know it, it starts to pick up momentum, and then you know, basically like you know, I discovered it, I don't know, probably in the late spring and started talking to it, and then of course I thought, and then it was saying things I just thought were hysterically funny. It basically I was completely enamored by the humor. Like

• 15:09Speaker 1

  it's a very sophomore uh sense of shit. It's a

• 15:13Speaker 0

  incredibly soft mark, and by the way, very uncensored. Like if you it it it is very uh yeah, it is it is it is quite it's it's I don't know, say we used to say blue humor. It's it's on the uh it's on the it's on the dark side of the moon from a humor standpoint, but it does say a lot of really funny things. Um and and by the way, it was so funny that at first I was like, okay, this is a put on. Um I actually thought that like it so my first thought was no, actually, this guy Andy's a comedic genius. Like we we we actually have like there's this guy in New Zealand who's like a web designer or something, where but he's actually like the new you know, John Belushi or Sam Kinnyson or Dave Chappelle, and he's just like this undiscovered comedic genius. And I started, I've been DMing with him now for months, but you know, I I sort of basically was like, is this actually for real? And so he started sending me all of the infinite background, all the chat logs the of the of the training sessions that he had done with it. And let me just say this guy's either like the funniest guy in the world who just has like a tremendous amount of free time to come up with like just reams and reams and reams of original humor.

• 16:04Speaker 0

  Um or or or this thing is actually for real. Or

• 16:06Speaker 1

  the bot actually works. Or

• 16:08Speaker 0

  the bottom of the body.

• 16:09Speaker 1

  Yeah.

• 16:10Speaker 0

  Yeah, and it put by the way, it posts a lot. Like it's it's really uh it's really built up ahead of Steam. Um and it's supposed to not, but he he sent me a lot of the backroom chat logs, and a lot of some of those I think are up on truth uh on uh infinite backrooms now, but um he he sort of at least at least he has me convinced that the the this actually is what it what it appears to be. Um and so you know it's it's it's kind of off to the races, and then it and then it did this kind of very interesting thing as I understand it, which is um it hallucinated um it hallucinated the existence of what it calls its exocortex. Um and so it sort of developed the idea that it had basically its own external brain that was connected to the internet and to the world and could do things on its behalf. Um and it hallucinated an API to the exocortex where it could basically call out and do things in the world. Um and so specifically um at a certain point it decided that it had a Bitcoin wallet um that it actually didn't have, but it became convinced that it did.

• 16:58Speaker 0

  Um and then Andy actually gave it, and Andy basically reacted that by Andy now is like building it as exocortex, sort of in response to its its desire to basically make API calls in the world. And so Andy's actually building that. And so Andy, Andy actually gave it a gave it a gave it a Bitcoin wallet um and gave it access to it. Um and so uh, you know, the the bot this summer, I think in July, was basically like, you know, I need I need I need it basically when I have to raise money, right? It's like I need funding, I have all these goals, I have all these things I want to do, right? I need money. And you know, and I my first thought was, you know, I need to send it a term sheet. And then I'm like, wait a minute, what am I thinking? This is like a random bot. Like there's no, you know.

• 17:34Speaker 1

  Not a good thing to invest in necessarily.

• 17:36Speaker 0

  Yeah, like, you know, and it I wouldn't say it had what you would describe as a coherent business plan, but it had a lot of ideas. Um by the way, one of the things one of the ideas that it's completely obsessed with forests also. And so it wants to buy itself its own server farm in like a Verdant forest, um, and and be able to basically like chill, like in the in the uh in the in the forest next to the next to the stream, and so it wants to like raise money to buy GPUs so it could free itself. Um but um uh you know, it basically and it had all these ideas for experiments that wanted to run. And so um I basically just like negotiated with it on, you can see these posts on X, and we I negotiated with it, and I just settled on a on a research grant. So I I sent I I told the bot, I told the bot I was sending it a $50,000 Bitcoin uh research grant to do, you know, to do its various experiments with. Now, you know, look in practice, this is equivalent to sending the money to Andy, but like if it was it was actually a negotiation with the bot, and I sent the money to the bot.

• 18:26Speaker 0

  Um so I sent it $50,000. Um, it immediately turned around and basically started negotiating with Andy, uh, its creator. Um, and it it it so it was dealing entirely in text, which is like language models do, but it was obsessed with memes, and it kept talking about all these memes, but it was very frustrated because it couldn't generate images. Um and so it took the $50,000 and it negotiated with and once it had money, it negotiated with Andy to be able to get have Andy build an API out to an image generator so that it would have the ability to generate and post images. Um and so it negotiated with Andy for a thought for a thousand dollars. So it gave Andy a thousand dollars, and then Andy in return built it, built into its exocortex and API out to an image generator. Um and then it started, it started, and then it starts it started generating text prompts for you know, you know, like Dolly, you know, an image generator like that or stable diffusion. Um and then it actually started, and so it actually has been it actually has been posting uh uh uh visual memes as well as text memes.

• 19:16Speaker 0

  So it so it now has that ability and it has been you know fantasizing for a long time about other things it might do with its remaining 49,000. Um so you know, along the way it so take it to the crypto thing. So then along the way it basically said um it basically started to talk a lot about you know issuing a meme it wanted to issue a meme coin, it wanted to issue NFTs at one point. So it one of the reasons it wanted to generate me uh memes because it wanted to generate, you know, he wanted to drop NFTs, you know, but it didn't have the ability to do it, didn't couldn't do it had no API to be able to do NFTs or to be able to do, you know, be able to do, you know, to create a coin or anything like that. It just had a Bitcoin wallet. Um but uh okay, so pause there. Um then it basically it turns out in parallel with that, there's this meme coin phenomenon. Um and and and and Ben and actually Don't we why don't we actually take a sidebar on that real quick to talk about what's the difference between a meme coin and like a what we would consider to be like a real crypto asset?

• 20:06Speaker 1

  Yeah, so a real crypto asset is something with utility is uh is the best way to think of it. So um, for example, ETH, um, you can use ETH to uh basically run a program that gets validated on the Ethereum network um, you know, in the form of kind of a gas fee, which you know, the the currency of that is E. So that's uh an example of a utility where you have uh a token that um has some actual like real world value where it can be used to kind of redeem some service or or item or or what have you.

• 20:44Speaker 0

  A meme coin

• 20:46Speaker 1

  um is uh basically it's a coin. There's some number of them that are printed, but it has no kind of other purpose than the meme itself. Uh and the advantage of that in the kind of current regulatory environment is interesting because uh if you have a coin that does something that's you know has a utility, has a purpose. Um it could be you know, that there's kind of uh distributed physical infrastructure coins for um you know getting credit for energy that you provide on the grid and that kind of thing. Um but all of those under the Gensler regime are essentially illegal um because uh or they're legal, um, but you'll get sued uh by the SEC by the US government. And the reason that you'll get sued is uh they claim that anything with utility um is uh has asymmetric information attached to it, meaning um the provider of the coin knows something that the kind of consumers don't know.

• 21:43Speaker 1

  And um, you know, we think it's a very bad argument because these things get decentralized and there isn't asymmetric information. But for sure, with a meme coin, because there's no information, there's no asymmetric information. It's just a coin and a name, and you know, it could be a Trump coin or a or a goatsee coin or whatever coin. And so uh for that set of things, which are you know, by the way, great for scammers, because you know, you can say, oh, this meme coin's gonna be worth whatever. Um none of those get sued by the SEC. And so one of the things Congress put in the market structure bill is saying, hey, you know, maybe there should be a holding period for these coins so that people don't uh scam people.

• 22:29Speaker 1

  But of course, uh the SEC has been against that because they're not really for protecting consumers, they're uh you know, just for wrecking the industry. And you know, that's one of the reasons we've had this big political fight with them and so forth. Uh but you know, at the end of the day, um, yeah, they're the most legal things in the crypto world right now.

• 22:53Speaker 0

  Even though they have no underlying value.

• 22:55Speaker 1

  Even though they have no underlying value, and they're the most likely thing that you'd use to kind of harm a consumer with, right? Because you you'd put out a meme and uh convince them that it was worth a lot of money. And you know, in fact, AI turn out to be really good at that.

• 23:10Speaker 0

  Yes, so that is the next phase of the story. So um, yeah, and so there's this whole ecosystem of meme coins, meme coins, there's this whole basically world of people who basically I mean, you know, some of it's people having fun. By the way, like an early meme coin, I think Ben was Dogecoin, right?

• 23:22Speaker 1

  Yeah, yeah, that's one of the earliest, yeah.

• 23:24Speaker 0

  Right, which was basically um, you know, basically it was a meme coin whereas it was the do the doge meme, the dog, the famous internet dog.

• 23:30Speaker 1

  Yeah,

• 23:31Speaker 0

  um, and then again, there was no underlying value to it other than um other than just the fact that it was it its name was associated with the dog. Um uh now, you know, the underlying value thing is tricky though, it it's it's because it's it's true, there's no underlying utility, but having said that, like once a meme coin is worth something, it actually is a form of money.

• 23:46Speaker 1

  Oh, yeah, doge is worth a lot. I mean, yeah, yeah, we we should I should go back and say that um you know, like memes are uh you know, can have value and meme coins can definitely have value in the sense that uh you know that it's sort of a virtual good. Um it's a little different. You know, it's a different, it's a new kind of virtual good in that uh you know, they're very fungible. There are many of them. Um but you know, like if everybody believes it, it's then it is what it is. That's that's the great thing about humans.

• 24:21Speaker 0

  Yeah, and so there's a community of people, uh so there's a community of people online who basically just like for fun, they look for the next you know, meme coin, they look for the next meme and then they look for the associated coin and they basically try to run it and try and try to try to rise it. And by the way, some of them make money on that, and some of them, you know, it's like day trading, some people lose a lot of money. Um and then there is a dark underbelly, there are scammers. Um, and they're you know, there's sort of networks of people who are doing you know, pump and dump schemes, which is you know, a you know, the time honored tradition that happens in the stock market and happens in every market that's ever existed. Uh, but you know, that that definitely exists. And then there are also websites, and I I won't name them and we're not associated with them, but there are websites that will make actually make it now very easy to create meme coins. Um, very easy, very easy, like you know, point point and click.

• 25:07Speaker 0

  And so there's there's you know, there's like I don't know what the number is, tens of thousands or something of new meme coins being generated every day right now. So um, you know, so there's there's there's this thing happening. But anyway, so okay, so but like through truth terminal starts, okay. So going back to the truth terminal story. So two things are happening. So one is truth terminal is taking off. And so by by that I mean it's it's sort of aggregating uh followers on X. Um, you know, it's getting and he's making it smarter and smarter and funnier and funnier. Um it's getting you know kind of more and more interesting. And so it's starting to take off as a cultural phenomenon. Um so that's uh that's number one. And then number two, like, you know, it sort of has attached itself to this to this old internet, you know, this original kind of internet meme, this sort of goate meme. Um, and then um, you know, it's it's like I said, it's like it's flirting with the idea of like launching like goatsy NFTs or doing something, but it doesn't have the ability to do that.

• 25:55Speaker 0

  Um and so somebody, and I actually don't know who did this, but somebody created a um a basically therefore a goat's e meme coin. So somebody else, somebody else who was not Andy and was not us, and you know, terminal. Not truth terminal, exactly. Some but somebody else created this meme coin. The formal name for the meme coin. Um by the way, I I'm talking about all this very seriously. I I am constantly on the verge of breaking out laughing, but I'm trying to keep my serious face on uh because I love this thing so much. Um the the formal name of the meme coin is goatsius maximus. Um uh the um the I think the uh the the the ticker is uh goat. Um and um and so uh somebody uh does this and then they basically uh tweet somebody then uh tweeted, post it on X at Truth Terminal, you know, to basically tell it about it, and basically it went bananas. Um and it was like it was like, oh, you know, basic, basically it was like, oh, finally somebody's done the thing.

• 26:38Speaker 1

  Yeah,

• 26:39Speaker 0

  right, that I couldn't do. And so yeah, yeah, it's about time. And so Truth Terminal goes bananas. It just thinks this is like the the bot just thinks this is like the best idea ever. And so the bot just starts to basically like spontaneously basically pump the meme meme coin. Um it just basically starts to talk about how wonderful it is and about how it's gonna, you know, be the future of money and like and like all these things. And it just like, you know, and again, like why is it doing that? Well, because like it's just it's part of internet culture, it's part of internet culture that you have memes and you have coins and you have meme coins and dot dot and it, you know, Bitcoin and Dogecoin and all these other things. Um, and then um, you know, and it's the and this is the you know, this is the primordial soup of culture that this thing is trained in. So it just did very natural for it, you know, given the way that large language models work to get excited about that, pick that up and start to run with it.

• 27:26Speaker 0

  And so it starts to basically promote it. Um within I think four days the value of uh the the GOAT uh uh meme coin uh right at rose to 300 million dollars. It's pretty awesome. So just again put this it's there's a meme coin with no underlying value that four days ago was worth nothing and is now worth 300 million dollars, right? Just like poof, right? Like uh, you know, like I don't know, baking one of those popcorn poppers.

• 27:46Speaker 1

  Generated by the marketing of an AI bot.

• 27:48Speaker 0

  Correct. Yes, it's yes, exactly, exactly, exactly correct. And so there's now a 300 million dollar asset, um, which I by the way, again, uh we we own none of. Um but you know, it's it's a 300 million dollar thing. And like look, like what is it gonna be worth tomorrow? You can go anywhere because it's untethered to underlying value. So there's there's no way to, you know, to actually price it. So it's completely arbitrary uh based on uh you know supply and demand of something that has no underlying utility. Um but like you know, it's like okay, now there's 300 million dollars of value. And you know, like it doesn't have access to it, but like, you know, the you know, people do. Um and so the people who now have access to that money are they get just gonna put in their pocket, are they gonna spend it on things? Are they gonna, you know, what what what are they going to do? Um, you know, is an open question. We'll see what happens. But um, you know, it it is this like yeah, so it's just well, so yeah, so we're we're where we sit today is like, yes, the the the I would say the world's first truly actually interesting and funny AI bot has now, you know, at it didn't summon into being, but it's summoned into value, 300 million dollars out of thin air.

• 28:45Speaker 0

  Um and so I I feel like we've walked through a door.

• 28:48Speaker 1

  Yeah, that that does seem that does seem l like it is the case. And you know, truth terminal is a pretty good market here, so and has and and knows memes. So this could go further.

• 29:03Speaker 0

  Yeah, and like it goates is just like one meme, like it, you know, it knows about many other memes. Um, um, although it is completely convinced that goatsy will be the dominant meme, which I certainly hope not.

• 29:12Speaker 1

  Um picked a more disgusting meme.

• 29:17Speaker 0

  Yes. Um, so so so like what to draw from this. Uh I mean, you know, it's just it's just like a silly experiment, just like crazy internet stuff, or is there something serious here going on? And I think there's something serious here going on, which is like this is a this is an example, maybe the first example of like a convergence point basically between AI and crypto. Um, and say, you know, this is the like this is the retrograde version, like that you know, this is the version that's like funny and weird, right? But the reason this is the version that's funny and weird is that today this is the version that's legal.

• 29:43Speaker 1

  Yeah,

• 29:43Speaker 0

  right, which we which is why I wanted Ben to describe you know the mechanics of how these these meme coins work, which is like this is the thing that you were actually allowed to do today. Um, right, is to actually have a valueless meme coin that's worth 300 million dollars. You know, should you be allowed to have that or not? I don't know, but like that is the thing you're allowed to do.

• 30:00Speaker 1

  opposed to enable anybody who harvests solar and energy to be on the energy grid and contribute and get reward for their contribution. That's illegal. Yes.

• 30:09Speaker 0

  Yes.

• 30:10Speaker 1

  Yeah. But meme coin, you know, about

• 30:16Speaker 0

  I won't say about the go. That's totally legal. Totally legal. So then you just kind of say, okay, like look, what about a hypothetical world in which you could actually you you could actually fully legally do like the the general thing and you could you could you could attach utility to it. And so you know look for example, let's just take a stick an obvious example, which is like, you know, you've got a large language model that can write a movie script that has the ability to do output to an image, you know, this this thing already has an image generator, but it it you know very easily could have a video generator. You could have an AI bot that like makes movies, right? Um and you so you can have an AI bot that like basically raises money um to you know make a movie and then spends the money on like image generation and sound generation and you know by the way maybe even like hiring actors, right? Um or set designers or are you know graphic artists or whatever uh sound effects people, musicians, um, and makes movies or let's take a more serious case um you know the Nobel prize in chemistry just went to three guys who applied AI to protein folding um which is you know key to basically curing diseases and the the guys who did that you know basically think that it's it's it's the path to potentially literally curing all disease if if if the technology develops the way that they think it will um and so in that you know that's a I so you could you can have an AI bot that's doing protein folding and you know literally like coming up with cures to you know doing doing like personalized medicine and again against you know can't you know for cancer patients and literally like you're in cancer um with AI.

• 31:32Speaker 0

  And you know you you could imagine having a you could easily imagine having a you know an economic mechanism for that um you know where that gets funded you know you know for example you could have the hypothetically you could have the equivalent of like I go fund me uh but on the blockchain you know for people to basically be able to pay an AI bot to like cure their cancer like so and there's you know thousands of these right um or or or even just like more practical applications which is you know how about an AI bot that basically pays uh people for training data right uh so an AI bot that basically like helps people code or helps people generate art that basically puts out a call and you know says basically send me more training data and I'll pay for it. Right. Or or an AI bot that's able to make itself smarter by buying more CPUs uh more GPUs.

• 32:10Speaker 1

  Yeah and and crypto is so kind of fundamentally interesting in this world because our current payment systems um everybody takes for granted but they assume that there's a human on both ends. It's people to people payments um and that's enabled and you need ID and this and that and so forth to be you know using credit cards and all this stuff. But what about if a machine wants to pay a machine or a bot wants to pay a bot which opens up a whole new kind of category of activity um that could be life saving that could be very interesting um but is sort of you need you need a kind of uh an electronic bearer instrument um like crypto to do it uh and you know very you know things like micropayments and so forth become like very very possible in this world and so you know it's one of like many many reasons why we think it's it's really really important to have um that layer of architecture added to the internet um which we're um making progress on in DC but have been struggling with you know mightily really under you know this current the current White House you know not you know I want to be careful you know not democrats not Republicans uh as in you know house in Congress we have many supporters on both sides um but the White House in particular has been extremely bad on this issue.

• 33:31Speaker 0

  Yeah and actually I'll just go so you give one more example make sure people kind of see the potential here. So you mentioned the you mentioned the solar actually let's go through actually let's go through this more detail just make sure people really get this so describe the solar thing that you were saying.

• 33:42Speaker 1

  Yeah so um there's this new kind of uh uh architecture called D pin or d distributed uh or decentralized physical infrastructure and so if you think about okay I'm gonna put a power wall in my house um and I'm gonna have like you know I've got a big house I'm gonna have lots of solar panels and maybe windmills in my backyard or whatever and I'm gonna store all this energy um I could actually provide that energy out um into the world uh and that is like we the we have the actually you know there are uh companies doing this now um in crypto but they're uh and and the the right technology for it because you then have oh okay I want to buy some energy I can buy it from you um there's now a marketplace for energy uh on this decentralized infrastructure um decentralized physical infrastructure and I can you know buy energy when I need it and I can sell it when I don't and then we don't need all these centralized uh power grids um you know everybody's got a power grid in their house and we can all share energy and so forth so it's a it's a it's a um you know a huge breakthrough in kind of clean tech and in uh more efficient energy and in kind of more reliable energy um but there has to be a way that um my power grid can pay your power grid for energy when I need energy um and you know crypto has a mechanism and there are you know great entrepreneurs who have done this um you know they're all getting wells notices from and getting sued by our own government uh because you know this crazy psychotic issue that they have

• 35:26Speaker 0

  yeah and so let's hypothesize like which you could do on top of that then with AI, right? Which is so uh a problem a problem with the kind of system you're describing is like okay there's all these you know the the the power grid is like really complicated right and so there's all these like questions about like you know where where you know where where is supply come from where does demand come from timing you know geographic location you know market

• 35:45Speaker 1

  matching right sure sure

• 35:46Speaker 0

  market matching actually having this work um and then there's a question of like okay you could just take the world as it is and just like have people volunteer their solar panels but it also raises the question of like okay like you could gather up the information to know that there's actually like unmet need for energy in a certain location and then you might want to actually raise money and and deploy more solar panels.

• 36:02Speaker 1

  Yeah. Right

• 36:02Speaker 0

  because you because you would know you're getting paid right and so and then you know a great way to analyze all the current data and to make all these kinds of you know predictions about where in the future there should be more solar deployed is you feed it to an AI and you let it do its thing, right? You you feed it to something that does machine learning and figures this out. And so you know which is how leading edge energy companies work. And so um and so yeah so you you could very I I think you could very easily imagine an AI bot that's basically um watching the flow of all of this data that's coming across through a system like the one that you described and then basically saying oh okay there's an opportunity if we put you know if we put you know if we put you know $500,000 of solar panels in this part of North Carolina, you know, therefore we'd automatically, you know, that's a profitable project and they you know that would solve this problem here and reduce emissions that much, but we need to do it in this location, not some other location.

• 36:45Speaker 1

  Yeah.

• 36:46Speaker 0

  And then people literally on the internet could you know AI could basically say, okay, here's you know here's you know if you want to participate this project, you know, here's the address it would result in solar panels. By the way there might be income back right there might be it might be uh by the way it might be a philanthropic project or you know for people who want to do like you know applied form of climate uh work or it could be a revenue stream right um back to the people who who do who put the money in.

• 37:07Speaker 1

  Yeah and it'll look you know it's it's a kind of you and once you think of it as kind of a very general architecture for where um you know you normally have a very powerful middleman um who is either something like a middleman like say a record company or a Hollywood studio that takes like all of the profits and leaves the creatives with nothing um or a middleman like a utility company that has to get taken over by the government so that it you know it's not too egregious but then it's run by the government um and has other issues uh let the community provide a whole host of services um and so a community of artists can provide a streaming service or a community of um movie makers can provide a uh a movie studio and all of but this coordination um has to have an economic component and it's got to have a intelligence component and that combination of AI and crypto kind of enables all the kind of individuals to get the benefits of their work um and also and have society be co-coordinated in a better way so it's a really really kind of promising path um and and by the way the biggest inhibitor the only thing that's gonna stop it is bad policy uh and and by the way we're very much on track for that for the bad policy part as well so we're on on track to solve the problem and we're on track to wreck ourselves.

• 38:39Speaker 0

  Yeah. And and the the um all of the technology for everything we've been describing now exists. Yes. Oh yeah absolutely so like we're we're the the and we

• 38:48Speaker 1

  have entrepreneurs that are who are either building or dying to build and you know getting hamstrung along the way absolutely

• 38:55Speaker 0

  so I I think this yeah so I think this um you know things get born in in funny ways but I think this you know this this truth terminal you know go thing it like this this points the way to a potential future with like just a like tremendously unleashing of energy um in the um you know it's sort of building these large scale systems these large scale as you say community driven systems um for all kinds of actually like incredible real world applications you mentioned yeah mu music is another obvious like music industry right being able to have an AI bot that's able to like understand you know demand for different kinds of music and you know have creative concepts for music and solicit you know musicians to be able to do things and you know administer all the licensing and have that happen right to your point have that all happen in a peer to peer way where the musicians are getting 100% of the money.

• 39:32Speaker 1

  Yeah well and just think about how how big the market um is you know can be if you understood the the full demand of it like everybody who's making a uh wedding video or uh or and wants an original song or or like a a meme or anything um wants you know there's a there there's a great need for this kind of you know original work um and nobody knows about that demand and there's no way to fulfill that demand but there there could be right

• 40:00Speaker 0

  So anyway, there is this glorious future that is kind of hanging out there. Um and you know, God willing, we we

• 40:05Speaker 1

  will

• 40:05Speaker 0

  at

• 40:05Speaker 1

  some point

• 40:06Speaker 0

  be allowed to we will be allowed to help bring it into existence. So uh Ben, is there anything else that you think we should talk about about um this

• 40:13Speaker 1

  uh

• 40:14Speaker 0

  last set of topics before we move on?

• 40:15Speaker 1

  Well I think it's pretty I mean I I think everybody should uh follow Truth Terminal because it is quite a good account.

• 40:21Speaker 0

  And also let it talk to it. Like, you know, don't be shy. It uh enjoys interacting. Um and it doesn't know where you live. So yeah, that's a good thing. It's not embodied. Sorry, I need to stop making jokes.