Jim Carucci and Marty Bent explore how Bitcoin and open-source AI are converging to create decentralized agent economies.
In this episode, Jim Carucci and Marty Bent explore the fast-emerging convergence of Bitcoin, AI, and decentralized agent economies. Carucci warns of a pivotal choice: a future dominated by centralized AI monopolies like OpenAI and Amazon, or one driven by open-source innovation exemplified by models like DeepSeek. He explains how open AI models empower users, while closed systems trap them in censored ecosystems. Carucci’s company, Cascader, is already building modular AI agents that subcontract tasks among themselves, using Bitcoin micropayments for peer-to-peer value exchange. Highlighting Bitcoin’s censorship-resistance and privacy, they argue it is the natural financial infrastructure for AI-to-AI commerce, much like AI agents in a Minecraft simulation gravitated toward using scarce "gems" as money. They stress that while business-to-business applications will lead early adoption, broader agent economies are coming—and that building open systems now is critical to avoiding a centralized digital future.
"AI is a lot like gunpowder. Either you use it, or you get subjugated by someone that does."
"We either have Apples, Amazons, Googles, and OpenAI controlling everything, or we have an open system."
"Bitcoin is the glue that holds together an open agent economy."
"If AI independently chooses Bitcoin as its money, that might be the clearest sign of real AGI."
"Marry the process, date the tool."
"I think we’re going to win, but it’s going to take effort. The human will to not be dominated is very strong."
This episode delivers a powerful warning and call to action: Bitcoin and open-source AI are humanity’s best hope against digital authoritarianism, but their success isn’t guaranteed. Jim Carucci and Marty Bent highlight how decentralized agent economies powered by Bitcoin are already being built, yet challenges like agent reliability, network effects, and centralization remain. Carucci stresses that preserving freedom demands active effort—whether through building privacy tools, supporting decentralized messaging like Nostr, or promoting Bitcoin adoption. The collision of AI and Bitcoin is underway, and those who build today will define the future.
0:00 - Intro
0:36 - Lacrosse and manifesting
4:38 - Open vs centralized AI
8:02 - Fold & Bitkey
9:38 - LLMs and their biases
13:37 - CASCDR basics
17:57 - Unchained
18:25 - Bitcoin AI synergy
30:39 - Progress of the agentic framework
36:13 - Implementations of agents
41:46 - Nostr, DVMs, MCP
46:57 - Open source vs panopticon
51:40 - Ideal picture of future AI
1:02:08 - Driving bitcoin adoption
1:10:53 - Plugs
(00:00) Are we accelerating the reality in which the matrix is real with this AI? It's a lot like gunpowder. Either you use it or you get subjugated by someone that does. There's really two possibilities. We either have the Apples, Amazon's, Google's, open AI controlling everything and calling the shots or you have an open system.
(00:18) Open source is punching way above its weight. We wanted to build towards this future where autonomous agents negotiate with each other. So, do you want central planning? Do you want to go to the pallet bureau to go about your daily life? Or do you want to just have an open internet? Carr stole that from you. Did he steal that from me? He does it all the time, dude. That was a high school thing now.
(00:46) Well, I stole from the crew kids in high school. That was their That was their uh Oh, that was their battle cry. Woo doggy. That sounds like a lacrosse guy thing. No, it was a crew guy thing at our school. I used to uh I used to play lacrosse in high school. I I grew up in Onida County, which is one of the Irakcoy tribes, and we would play those guys.
(01:06) It was intense, man. Yeah. My brother uh my brother had a guard two of the the Thompson brothers in college. Okay. They played Albany every year. He shut them down. He's he pumped my brother up. He used to, for anybody who doesn't know, Lyall Miles Thompson, two Native American lacrosse players who are the best ever, played for Albany University.
(01:25) My brother was a defender who would take the best attackman on the team and while he was at Bryant, they were at Albany and they they were the two best attackmen on the team. They're incredible on offense. Like they they must start with it in the crib. They they do, you know, their practice they they get like a little a little piece of wood in the backyard.
(01:46) It's like this big. And that's how they get good at stick skills. They like hit that little circle. They just go back and forth, back and forth. That's my dad's favorite story is in eighth grade I took uh I I played defense against one of those guys and he was tearing us up. He scored like three goals in like five minutes and then I just kept kept whacking him and he left the game.
(02:08) So he left the game the uh the Thompson brothers had a incredibly high pain tolerance. Yeah, I can imagine. They're a lot more elite probably. Yeah, the uh lacrosse it's lack season. It's peak lack season. Basketball's over. College basketball is You get You get quality lacrosse games on ESPN now this time of year.
(02:31) You gonna go back to calling games? Uh, probably not. Probably not. My uh my live game calling days are probably behind me. I'm not It's It's a skill. Yeah. Trying to call. You have to remember. How long did you do it for? Like two years? Uh I did I mean for for like broadcasting? Yeah. I did it for like two games. Oh wow. Okay.
(02:51) for Anderson High School here in Austin. I'm much better suited as a coach. I may get back to coaching at some point, but I have been away from the game for for a while. There's definitely some advancements that have been made. Is Texas catching up? Cuz I remember it was all East Coast like Philly, Long Island, Syracuse.
(03:09) Yeah, I think I think Texas is catching up. UT has uh I think they have the best club team in the country right now. The MCLA. I think there's rumors that they want to go D1. That would be cool. have a D1 Lax team here in Austin. That'd be amazing. Yeah, the high school scene up in Dallas, I think, is stronger than it is here in Austin. Yeah, Lax is grow fastest grown sport in the world. Is it? That's awesome.
(03:31) Fastest sport in two feet. Fastest grown sport. Get on the Lax train before it's too late. It's lacrosse is like Bitcoin in 2011 right now. It's probably like 2017, actually. Pumped to have you here. Yeah. you. Let's dive into your background first. You were you were saying that uh we were talking physics.
(03:55) You listened to the Matthew Pines episode. Yeah. What do you think? Can we manifest? I don't know. I think it's possible, man. I think there's some connection between consciousness and matter. I don't know. But it's hubristic to think that we have it all figured out. I agree. the uh again I left left side of the bell curve as somebody like I think there's plenty of anecdotal evidence throughout history people who talk about either having premonitions or telepathic experience in the Matthew episode I certainly had one with my brother where I woke up I was
(04:27) like there's something wrong with that [ __ ] and got a call from my mom 10 minutes later he's in the hospital well it seems like everybody has deja vu what the hell is that right is that a glitch in the matrix are we accelerating are we acceler accelerating the the a reality in which the matrix is real with this AI maybe.
(04:48) Well, I you're you're the one working on it or leveraging it at least the cascader. Yeah. Yeah. I think there's really two possibilities. We either have like a monolithic quasi monopolistic AI paradigm. So you have the apples, Amazon's, Google's, open AI controlling everything and calling the shots or you have an open system.
(05:12) And you know, I've I'm a longtime listener, first time caller here. And uh you know, so I've been keeping track of this. You know, I it's kind of surreal to be on the show because I remember running wind sprints over in Lewis, Delaware, listening to the laser huddle episode and uh laser, you need to come back on. Yeah, laser.
(05:30) where you where we got to put the laser bad signal out. But uh but yeah, I just remember listening to that. It's like yeah, AI does help the state scale like yeah um uh was it Peter Teal at the Nixon seminar sitting next to Pompeo talking about how on one end of the spectrum you have AI which can be kind of an authoritarian technology and then you have Bitcoin which can be kind of a libertarian technology.
(05:56) So all of those, you know, consuming that content got me thinking of, you know, what do we do in this AI future cuz I don't think it's going anywhere. And it's a lot like gunpowder in a way. Like either you use it and apply it or you get subjugated by someone that does. Yeah. Yeah. No, I went to a couple of the teal lectures that they had at UATX a couple of weeks ago as one of the students asked him the question like if AI is going to become this all-encompassing digital panopticon surveillance tech like should we try to nip it in the bud right now? He's like
(06:31) no it doesn't have to become that. certainly can and it may certainly be trending in that direction. But when these things manifest and materialize, you have to lean into it and try to affect it in a way that produces an outcome that is beneficial or that you would like to see, which is not living in that digital surveillance panopticon.
(06:50) Yeah. And I think a lot of the people that are liberty-minded and open- sourceminded that are in on this wave are kind of like the space force for liberty. Like they have to be in this new frontier. You want them in on in on everything like Lawrence Taylor. You don't want you don't want to just let the open AIs kind of call the shots.
(07:09) And so how would you describe the landscape as it exists now? Um juxtaposing open source to monolithic big techr run AI. I think I think it's pretty balanced right now. I think that open source is punching way above its weight when you compare the amount of funds they spend versus the results they get.
(07:31) And a huge white pill was Deep Seek. Now, everyone's going to say, "Well, they're Chinese. They they shifted a certain way. They cheated. They distilled." You can see that in certain context, it'll think that it's open AI. That's probably all true. But at the same time, it's open source. I'm running it in my own cloud server right now.
(07:52) And it's the token cost is so damn cheap. So, it kind of collapses this kind of overleveraging effect and centralizing effects. these monoliths have by getting cheap capital. Listen freaks, I know you're tired of me talking about Fold, but I'm going to beat the drum. I'm going to beat the dead horse. If you're a Bitcoiner living in the United States and you're not using Fold, I'm just going to ask one question.
(08:15) What are you doing? You're leaving SATS on the table. They've got gift cards. They've got their debit card. You can use your credit card. Just connect your credit card to the Fold app. Use the Fold out to pay off your credit card. and you're going to stack your credit card points and Bitcoin as well. They even teamed up with Crowd Health to give their members Fold Plus membership at no cost.
(08:34) Don't leave SATs on the table. Go sign up for Fold today. Go to tftc.io/fold. There's nothing to lose except SATs that you could have stacked. Sup freaks. This rip of TFTC was brought to you by our good friends at BitKey. BitKey makes Bitcoin easy to use and hard to lose. It is a hardware wallet that natively embeds into a 20 or three multisig.
(08:56) You have one key on the hardware wallet, one key on your mobile device and block stores a key in the cloud for you. This is an incredible hardware device for your friends and family or maybe yourself who have Bitcoin on exchanges and have for a long time but haven't taken a step to self-custody because they're worried about the complications of setting up a private public key pair, securing that seed phrase, setting up a pin, setting up a passphrase.
(09:20) is again, BitKey makes it easy to use, hard to lose. It's the easiest 0ero to one step, your first step to self-custody. If you have friends and family on the exchanges who haven't moved it off, tell them to pick up a big key. Go to bit key.world, use the key TFTC 20 at checkout for 20% off your order. That's bit.world, code TFTC20. Yeah.
(09:41) And so let's for the layman listener out there who may not understand the minute differences between monolithic AI and open source AI like deepseek or llama to a certain extent. Um how would you describe the differences and how they've been implemented so far? Yes. When we talk about AI, typically we're talking about large language models and this is this revolution in uh the ability to understand language adequately and images as well.
(10:09) Those are two things we take for granted as humans because we've evolved to see faces, recognize faces, recognize predators, speak to each other, communicate to each other. But it's actually a very complex process. And so going back to the 70s, this was a massive problem that people had difficulty solving. And they tried to solve it with sort of top- down rules and realized they couldn't really do it.
(10:31) So they applied bottom up uh rules using neural nets, convolutional neural networks and other techniques such as those. And what we're seeing is the culmination of decades of research and application. And to to go back to your question, you know, the difference between monolith and non-monolithic um when the what the lang large language model really is is a system of equations.
(10:59) It's literally a matrix. It's called a matrix. And so it's it's it's math, right? It's it's math, but the monolithic systems are holding that behind an API where they don't let you download the weights. They don't let you run it in your own hardware. You have to run it on their hardware. And so that gives them an asymmetric advantage because now you're dependent on them to give you the information to not censor the information and to pay them always.
(11:26) Yeah. And you don't know since you don't know the weights, you don't know how the models and the matrices act exactly like you don't know the uh the biased context that could be injected into that type of system. So when you think of the spectre of an all-encompassing digital panopticon run by AI, like the bias of that panopticon will be decided by the rulers of that model and the weights in that model and the context that's fed to that model.
(12:05) It is insane that we're sort of speedrunning into the like whether we like it or not, it's happening. It was obvious like AI is become more of a part of my life. I use it every day now. Uh and it certainly is useful and I do use a combination of open source models like llama and monolithic models like grock and um chatbt.
(12:31) But I I do think it's important for everybody to realize that there are two directions that we can go in, which is one where the context and the bias is tightly controlled by a group of people who may be benevolent, maybe malicious, but regardless of whether they're benevolent or malicious, it's still not ideal to have similar like Bitcoin or Bitcoin because they don't like central authorities controlling the monetary system.
(12:59) like apply that context to AI and it I think it still rings true. You don't want centralized actors controlling arguably one of the most important technologies that will begin to integrate into our lives over the next decade or two. Yeah. It's kind of like you don't want with Bitcoin, you don't want anyone to control the stock or the storage of wealth.
(13:20) And then on the flip side, like if if AI is centric to the economy in the future, which I think it is, then that's controlling kind of the flow of wealth, right? So, uh I think that's essential as well. And yeah. Yeah. So, let's talk about Cascad. Like why did you decide to build Cascader? How are you building it? And let's talk about all the different sort of tentacles that are emanating from Yeah. Yeah.
(13:50) So, Cascader just like a two sentence summary and we can get into an example is Cascader creates um it creates a uh AI powered building blocks that allow businesses to achieve greater productivity and lower cost and we monetize those externally. Additionally, we use them internally to build niche applications in lucrative markets.
(14:17) And so an example of this would be we have a timestamped transcript mo uh uh API. And so three examples of us using example number one is just a podcaster like you goes to our website downloads you know the transcript and maybe even like creates content based on that all in all in one shot. So that's one example. A second example would be pull that upjamie.ai.
(14:44) So that's our service that not only gets the transcript, not only uh indexes it so it's searchable, but it allows you to pull them up in real time. It allows you to play back the clip uh of exactly what you're looking for in a you know long even three-hour podcast and then also clip it and share it easily like with the with the deep link.
(15:02) So that would be that would be um a second example. And then the third example of a time stamp stamp transcript is our product CLA AI which helps attorneys process multilingual video and audio data securely and pull them up even in you know high pressure situations like a cross-examination where they need to present the evidence in a you know in a time constrained high pressure situation.
(15:30) So those are three applications that are pretty disperate, but you can see how they they all stack on the same building block that gets reused efficiently. And are they all running off the same model or Yes. Yeah. Deepseek. I don't use DeepSseek. I use something else. Okay. Um I use something else.
(15:49) You don't want to say um I I guess I can say it's just I use Deepgram because I find them to be higher accuracy overall. Um and I'm excited to see some of like their newer models. Um, but I think over time that's starting to get commoditized for obvious reasons. So that's kind of why I think about okay that's the building block. Then it's like okay can we I think like an electrical engineer cascade comes from a circuits term where you cascade systems together you put them together.
(16:15) So I think about okay I have this building block and maybe I typically package it with circuit A B C or API A B and C and now this is a monolith right and maybe it's an agent maybe it's a whole agent maybe it's a subsystem to an agent right but it's all modular and that's been our focus at Cascader and the reason we went this direction is because we wanted to build toward this machine payable web we wanted to build towards this future where you have autonomous agents negotiate with each um specialize in different area are
(16:47) areas and you know h effect effectively hire and subcontract each other in real time in order to accomplish you know more complex tasks on behalf of a user like you or me. Yeah. No, because agent AI is so hot right now and I think that's if you could squint and look at it like the agents aren't perfect yet but they're getting better at a rapid pace and I think that's like the next big breakthrough.
(17:12) Obviously what you're working on is like this modular sort of buildout and I I think that I forget the exact term but it's like the another way to frame monolith versus like open source is like monolith is like one all-encompassing model or you have the world of a thousand minds where you have very specific models for specific tasks that ultimately become agents that interact with each other. Yeah.
(17:35) And so we talked about kind of the ethical side of monolith versus non-monolithic, but it's economically probably more efficient to have a bunch of thirdparty teams running their own thing. And then what's the glue that holds it together? I think Bitcoin makes the most sense because now you can delegate a Bitcoin budget to um you know to an agent and have it run in real time.
(17:59) Our good friends Parker Lewis and Drew Bonsol are two of the deepest thinkers in Bitcoin. And while they come from very different backgrounds, they've landed on the exact same conclusion. It's Bitcoin that matters, not crypto. This master class lays out why 21 million is the key breakthrough, how Bitcoin, not blockchain, creates decentralization, and why everything else will be built on Bitcoin.
(18:18) Understand why Bitcoin works and why nothing else does. Watch the premiere at unchained.com/tc. That's unchained.com/tc. Why Bitcoin, not stable coins or Stripe? Well, I mean, first off, they're they're they're not censorship resistant, right? So, like right off the bat, you have that issue at hand. We have a bifurcating world, you know, we have the breakdown of Swift and all that.
(18:44) So, there's that component and then um the other part is I guess how do you know they're not a credit risk? Like I'd rather just have the bearer asset. you know, maybe it's premature because there's some some volatility in Bitcoin and maybe that people have liabilities in fiat, but um I'd rather know that you actually sent me the money then than take it on Yeah. faith.
(19:07) And there are are there technical um uh superiorities to using Bitcoin over something like stable coins or Stripe? Like is L42 like much easier to integrate? Um, personally, and I I don't want to ruffle any feathers here, but I personally I just kind of made a very simplified version of L42 that just uses the pre-image and payment hash and public key to verify payments.
(19:33) I think L42 is great in a lot of contexts where you need more sophistication. For me personally, I just kind of rolled my own and called it a day. Um but that con but as you point out the additional you know it's censorship resistant it's global so it's inclusive of people that can't necessarily get access to financial services and then it's private too the fact that it's onion routed I can I can leverage that to remain private while running all these sessions and I don't really know if there's another way to go about that. Yeah. So, we're deep in the
(20:06) weeds right now, but let's like zoom out like paint the picture for like a user experience flow of completing a task and how the agents would use Bitcoin to interact with each other. Yeah. So, the the really simple example I go to is, you know, imagine that you have a personal assistant AI agent that is aware of your schedule.
(20:25) Maybe it takes calls for you. It books things for you. and then um it happens to need or it notices that you want to go to Bitcoin Vegas um then it would that's not kind of its wheelhouse. So maybe it'll spec it'll go to another agent that specializes in travel booking. It'll get you the full stack of hotel transportation to and from the airport, best, you know, um best flight, best best pricing, etc. to get that ac that accomplished.
(20:56) And so it kind of subcontracts. You give it a Bitcoin budget, it subcontracts that to a higher performance agent for that particular task. Now, you could you could extend that out to a whole bunch of things and a bunch a bunch of maybe even smaller niches, even smaller than travel agents.
(21:16) Perhaps that travel agent has sub agents that are good at hotels, airlines, etc. and you expand this out to the whole economy, you start to realize it starts to resemble a real economy. So, do you want central planning, right? Do you want to go to the pallet bureau to go about your daily life or do you want to just have an open internet? Yeah.
(21:41) No, I think that's the that's the question that I've been bouncing around in my head for many months now when it comes to this convergence of Bitcoin and AI. Like one thing I'm highly focused on in terms of really leaning into something whether we're implementing it here at TFTC or investing at 10:30 was like the order of operations of the success and that that use case you just um you just explained like it makes a lot of sense but like there you have the two sided market problem where as much as I'd like to give this agent a Bitcoin
(22:14) budget and go book a flight for me like the airlines and travel agencies aren't accepting bitco point of the lightning network yet. And so like that infrastructure needs to be built out. Correct me if I'm wrong, but do you think just like agent to agent sort of commerce is going to find product market fit first where maybe it's not necessarily I need you to go book me a ticket, but hey, I'm doing deep research on this specific area of finance.
(22:44) I need you to go partner with a model that specializes in weather data because that weather data really affects um the the outcome of the investment I want to make. And so like in this example, it's like an agent going to the weather data agent be like, "Here, run this compute for me. Here's some Bitcoin to go do that.
(23:03) Return me the answer." Yeah, I think I think you're spot on. I think the answer is yes. Agent to agent will find product market fit. Now I at startup day I covered kind of the three barriers that I see in my eyes to getting to this kind of sci-fi future. And the first is just agent reliability. So the agent itself has to have high levels of reliability.
(23:26) And if you think about it, it's sort of like a manufacturing process because it's an LLM running many loops to attempt to accomplish a goal. So the chain of thought stuff is great. like that stuff is I think improving um and kind of letting us peel back the hood in order to identify and improve the quality over time of these models.
(23:47) But the truth is if you're 99.99% accurate but it takes a thousand steps to accomplish your travel booking then that means you're going to be off like what is that 10% of the time are you willing to take that risk with your with your precious Bitcoin? You know a lot of people probably aren't. So I think that's why I kind of drew it as a pyramid.
(24:08) like that's the bottom of the pyramid is is um is agent reliability. So that's a technical challenge. Then as you point out there's market challenges which is agents have to proliferate. Not only do they have to be reliable but now you have to have enough teams using them in different marketplaces effectively building the digital digital workforce so that they can interoperate in this new economy.
(24:32) And then lastly you have to have incentives to be open and accept these things. So suppose Apple and Amazon decide that they're just going to collude and leave everyone out, but their AIs will collaborate. You know, that's not what we want. We want open systems. I'm encouraged to see things like MCP coming out there, things like Noster out there, and then Bitcoin, you know, Bitcoin incentives to be the glue that holds this huge mosaic of the Bitcoin uh agent economy together. Yeah.
(25:01) And that's the uh that's the fascinating thing too is I think uh humans are injecting a lot of preconceived notions about the optimal monetary system that'll be leveraged by artificial intelligence. But I think it will be really funny if and when we do find out that the AI sort of coaleses around a money it prefers in independently without us injecting our bias.
(25:32) and Logan in post marked down the time stamp. I want you to pull up the I'll share a link with it, but the I'm sure you saw the video the video the Minecraft AI agents. Did you see those? I didn't see those. Oh, so Logan, play the clip and I'll explain it. Play it after I explain it to Jim. But the um there was an AI lab that wanted to test out agents and so it created its own Minecraft server and unleashed a thousand autonomous AI agents into the server and just let them run free uh to to see if they could actually develop a a mature sort of ecosystem within this
(26:07) Minecraft server. And they found um over all the iterations they ran like in 95 plus percent of them the agents coalesed around Minecraft gems as their monetary system because they were relatively scarce divisible portable um somewhat hard to find and didn't really have other use cases. And so like in this Minecraft they were added into the game as money like a nice feature of they were well like Logan Logan's trying to correct me here.
(26:42) I I'm just saying like like it makes sense that the AI chose the money thing as money. The reason for emeralds to be in the game is to be used as money. Okay. So what did they subjugate? But they weren't told it was money, right? Probably not. Yeah, they weren't. All right, Logan, thanks for No, but holding true, but still they weren't told it was money and they So, you said 95% of it was a high percentage, maybe even higher than that.
(27:10) I'm wondering if they wound up eventually subjugating the 5% that didn't, cuz isn't that what happened with the gold versus the silver standard? Yeah, I don't know. I don't know if they were very altruistic from you. We play the video and yeah, this is a Minecraft server, but every player here is actually an autonomous AI agent, and they're completely Minecraft agnostic capable of using other apps and games.
(27:33) Our agent started with nothing and then worked together to collect over 300 items in Minecraft. They set up a market where agents agreed to use gems as a common currency for trading supplies, building an economy. In this case, you might think the merchants would trade the most, but it was actually the priest. Why? He was bribing towns folk to convert.
(27:52) Yeah, I got freaked out when I was a kid. I saw like it was like a capture the bacon type game between like AI avatars and it was like very primitive cuz this was 10 years ago. But they started figuring out that they can like attack each other to get them to stop going after the bacon. So they started killing each other.
(28:10) Yeah. Lovely. Lovely. So, I think yeah, I think the safeguards are probably a good thing. As much as I don't like censorship, they're important. Yeah, but uh I think creating proper incentives to act in a cooperative fashion makes sense. I think money is strong incentives. Like back to my point, which is like I wouldn't be surprised if I mean we we're seeing it in crypto like AI crypto tokens.
(28:37) So, we're going to create like a token for each model. Barry Silbert was out there talking about his shitcoin project where you're incentivizing the build out of models and all this with the shitcoin, but it's like it's going back to barter. Yeah, you're just taking on counterparty risk of the issuer not rugging you.
(28:55) Yeah. Well, not only that, but it's like I think it would be hilar Maybe that's how we actually realize if we have AGI or not is if like the models, the agents are like, "No, like I'm not using the stable coin. I'm not using the shitcoin. I'm not using your Stripe account. Like, give me a Bitcoin wallet.
(29:12) I want to use that to spend. It's much more efficient for me. Yeah. And it makes you wonder if maybe that's the leash on on uh AI since we can tether them to the real world through energy. Mhm. Yeah. And private keys, I guess, too. I mean, this is Drew's one of Drew's theories is that like AI, like Bitcoin is the money for AI because it just makes sense.
(29:34) And if you're thinking about how to restrain AI from just like doing this unbounded sort of resource extraction, probably makes sense to introduce this hard money that that humans have coalesed around as well. So they can't print more to go do resource extraction that's not again in a sort of cooperative incentive structure which Bitcoin creates. Yeah.
(29:58) And probably for the foreseeable future, they'll at least be dependent on us for energy to power, right? To power all that stuff. So, I mean, I don't want to go matrix mode. But yeah, we're going to have to blot out the sun. Am I going to wake up in a tube of a vat of of liquid goo at some point before the end of my life? Hopefully not. Plug me back in.
(30:17) I'm a micro influencer. I know the steak is fake, but it tastes so good. Just plug me back in. Plug me back in. Is that what the ETH guys are doing? Probably. They're eating like tofu steak. Yeah. Yeah. Probably getting anemic. They're uh tofu steak or soyent the uh but where where would you say like we are now in terms of like the maturation cuz like the the agentic framework like once it pops off like I've been playing with Manis.
(30:51) I'm not sure if you've seen that model that was that product that was launched. I'm not sure if it's a model, but it has this sort of a gentic framework where you give it a task and it Yeah, I actually didn't play with it yet, but I did see like some demos. That's interesting. But I think certainly getting much stronger. Um I'm kind of caught kind of ambivalent about which direction it'll go.
(31:13) I think look 5 years down the road. Like I'm bullish, but I do wonder if there's weird incentives with inference just because of all the VC subsidies. like what what's going on there. Deepse seek is like a counterpoint on that. They made it much cheaper. So, it's I'm want I'm a little surprised that I haven't seen I just did research and maybe I missed a model, but I'm looking for, you know, what are the lowest cost for inference because the variable cost is the is the gem, right? I don't care how much Deep Seek spent. you know, they they made a
(31:45) bunch of probably false claims about how little they spent, but that regardless of that, they did some very clever techniques if you study the way they trained Deep Seek, but also just the fact that it's so cheap. The the variable cost is so low. It's like I kind of don't care how much they spent. Well, what are the what are the clever things they did in training and why is it so cheap? I I'm uh I'm a little rusty because I looked at it like right when it came out.
(32:10) Um I'm not entirely sure how how it got so cheap. I'd have to study that more. But I do know that they they used a reasoning model which was less adept at language but was good at I I think it was called like deepse v0 or zero something like that and they put that in a pipeline and they created um they created a bunch of syn what's called synthetic data.
(32:32) So it was data that isn't real but it was actually high quality which would have taken a long time. you know, even if you got tons of Chinese people making content, you wouldn't be able to keep up. So, the fact that it was able to create this high quality synthetic data that accelerated their development, I think lowered their cost significantly.
(32:50) There's some questions they probably distilled from OpenAI as well. Uh, I'm not fully clear on exactly how that went, but point is that they got the token cost way down and you can run it in in your own um in your own cluster in AWS or Azure. So, and so you can verify like that is verifiable that that low token cost is not being subsidized by some money backing deepseek like you run it in a cloud instance and you like that's my understanding.
(33:22) I still haven't got my first bill for it, so I got to triple check that, but it it looks like it's pretty cheap. Even in AWS, even with the overhead of setting up and paying them their vig and everything, it seems that it is cheaper. And so that's kind of a counterpoint on on, you know, is this leverage effectively is this leverage is this pulling of value from the future forward to build AI, is it paying off? Maybe.
(33:47) It looks like it kind of is. So I'm if you asked me 4 months ago I would have said probably not. There might be a correction where you know you're not going to use chat GPT to ask where you should you know what you should have for dinner. You're going to use it to like get your job done faster three times a day instead of 100 times a day.
(34:08) Um that's that's conceivable I think but I don't I don't I mean with Deep Seek it's pretty good. And I'm sometimes I mean I I think there's there's levels to this and it depends on the task but it's interesting Guy Swan points out it's kind of like uh graphics after a certain point like I'm getting the automation the marginal return is that Yeah.
(34:30) Yeah. And so there's certainly value in pushing the envelope just like there is with it's a lot like semiconductors in my eyes the way they push the envelope but um but it's good enough to get a lot of stuff done. I I think we are going to just like the internet, it's going to take us like 10 years to fully reap the benefits of of what we've already invented. Yeah.
(34:50) And that gets to my my next question, which is like how do you even approach building on top of these LLM and exploring different models or picking a certain model and sticking with it with everything changing at the rapid pace that it has been? Yeah. Um it's tough. I mean, we already just went over a couple items. I wasn't fully aware of.
(35:12) I try to stay up to date, but I also try to filter out noise. Um, so it's a challenge there. But once I become aware of something or some capability, then I treat it a lot like being an electrical engineer. When I was an electrical engineer, you got suppliers. You know, I would never work on the AS6 or like the guts of chips, but I would be like, "Okay, this chip can do this and it's got these specs.
(35:35) It can do this and this and that." So, it's very similar because I'm building circuits but with software effectively. And so I generally try to look at, you know, what's what's a cool high margin application I could build? What's something um that is going to actually like improve somebody's bottom line? So I kind of Steve Jobs it and start with the user.
(35:58) Then I work backwards and then I evaluate, okay, so what steps have to happen and what tooling do I need and is does it still make sense economically? Is it going to be high enough accuracy and quality and is it going to be low enough cost for me to offer this? Yeah. And so I mean diving back into something like pull that up Jamie where like the clips that like I'll launch an episode and I'll have like a cascader clipped section of the podcast I I launch within like 30 minutes.
(36:29) Uh is have you automated that process? Is cascad like So I I generally run it every morning. So you happen to drop in the morning. I'm I should make it more async so it's just triggered by your publishing. But basically, yeah, it I have an agent that kind of knows what you I kind of assumed what you would like.
(36:47) I I honestly want to show you like the the interface and get your take on maybe different topics, but I have four to five topics that it knows you like. It also takes into account, you know, what your speaker's expertise is and what your description is about just so that it makes sure it covers that.
(37:04) Um, and then it iterates over. It searches the database like it automates the search and um, it uses embeddings, which means it can get kind of vague. It can be a gist of what you're looking for. So, it engineers the search terms. It goes out and searches them and then it brings back the top clips and sends it to me in an email. So, then I go click the link and then I, you know, it's not perfect.
(37:28) That's I know that people complain, you know, they can't fully make clips yet, but it gets me close enough to where, you know, within, as you said, 30 minutes of your your pod, I can drop something. Yeah. And in that example, like how many if it's even more than one agents are communicating with each other? In this case, I wouldn't say they're all agents.
(37:52) They're they're they're AI systems. They're AI based systems. Uh, in that case, it happens to only be one, but you could foresee a future. And I was just talking with Rod about this. I want to integrate things like Google Trends into it so that you can see what's trending and then add that into your list of preferences. I want to have clips that are trending and have them ranked accordingly.
(38:15) Um, and you Okay, so the Google Trends maybe isn't uh an agent yet, but maybe there will be the equivalent by OpenAI at some point, which I could easily imagine happening since they seem to be kind of replacing Google in their business model. Then that would be an agent to agent workflow, for example. Yeah. Yeah.
(38:38) I'm just like thinking through it like as like I guess what would your advice be to somebody like myself thinking about implementing this because it's like uh again going back to the rapid nature of the development of the space and all the tools that are coming to market. It's like one day I'm like oh I'm going to work on implementing this and something new shiny comes out.
(38:58) It's like, "Oh, wait. I want to do something over here." Now, it's just like you're constantly inundated with all this cool stuff you can do. But like when it comes to practically sitting down and implementing it into your business, like what are the things you think people should be focusing on? Yeah, I think in general, and this is I'm stealing this from Rod.
(39:20) Shout out to Rod from Bitcoin Park, but he said to me yesterday, marry the process, date the tool. So you probably have a general process of what you're trying to accomplish. Not that different from what I was describing with my starting with the user working backwards. So you have a goal you're trying to accomplish and you work backwards from there to find the tools that meet those ends.
(39:43) Um and so I guess what I would say is you know stay up to date on that but keep in mind um you know I I go back to like do you know who Frederick Winslow Taylor is? I do not. He uh he basically pioneered this idea of efficiency studies in manufacturing. So you want to like have as few steps as few yes as possible.
(40:03) So I tried to do that with Jamie is just everything like as as fast going from getting the clip like podcast dropped. I didn't do you know it's all it's all automated. Go from podcast dropped to clip out. everything in between from finding the clip, editing like the timestamps, clicking the button, having the video generated because this is like a video waveform and then automating posting it to Twitter in Noster using AI to help me think of how to word the copy and then publishing it.
(40:41) So like those are bunch of steps starting with you know dropping the pod. um you have to think about like how to minimize the number of steps and we we still have a ways to go but I've gotten it down to a process to where as you said I can be one of the first people in on making clips for you. Yeah. Yeah. It's been fun to watch and the clips have been good.
(41:02) Like it has you've done a good job of uh honing in on my preferences and surfacing clips that are what I would call what we in the biz call bangers that are actually going to catch people's attention and get shared. Yeah. Yeah, the Tom Longo ones were awesome. Yeah, I uh and the cool thing about that is it's kind of hybrid cuz I I did a few of them just based on the agent recommendations, but I can also go into the interface and search them myself.
(41:28) So, I noticed there was one part that you were um that like you started the show with or it was related to it at least. And so I went into Jamie, I searched it, and then I made a smaller clip so that you can reshare that. And now it's like a 30 second video. So people are more willing to commit 30 seconds to watch that. Yeah. The um I mean you mentioned Noster.
(41:50) How do you see Noster fitting into all this? I mean obviously you have the data vending machines where you sort of use Noster as this way to create this twoart sided marketplace where you can send a note out saying I'm looking to get this task accomplish and you have different AI agents basically compete um for SATs to complete that task and then there's a whole concepts of like data aggregation for the models themselves like using nostra events as as a form of data to feed the models um there's a lot of people with those two ideas specifically like
(42:26) do you see value in that? You see that proliferating and gaining steam? Yeah, I think the MPC plus Noster combination is really cool. I know Austin was working on that. Um, and you know, data vending machines are cool. Um I I'm not quite Parker, but I'm I want to say I'm like semib bearish sometimes on Noster just because I feel um you know people are putting up with a lot of like UX hassle and I don't think I love Odell but I don't think the pitch should be get on Noster right now or you're a cuck and only Noster. I think
(43:03) the pitch should be get some in case it catches on. So kind of our philosophy of Jamie was why not do both? you're gonna you're gonna post this to Twitter. Hey, do you want to also post this to Noster? Log in with your extension and it's easy. Like that's kind of more the mindset so that you have a backup.
(43:19) You have a communication channel that's robust and can't be changed. And as you point out, you know, you can use Noster to relay information between agents to identify them to create web of trust. That's all great stuff. Yeah, I think that's a big question in my mind is like does that using Nostra as this messaging protocol or inter agent task completion like manifest like is that a more optimal communications protocol for these agents to communicate through? Um does it enable like a more robust discoverability or is simply just like
(43:58) connecting agents via APIs the way this is going to work? So I, this is funny, Austin Toeer and I at Tab Comp, this is kind of how Cascader started. We created a new NIP that is distinct from DVMs. So DVMs, at least at that time, I think they've evolved over time, but they would just broadcast to everyone.
(44:17) I want this done. And that makes sense in a lot of cases where it's like I don't care if you know that I'm trying to post an emoji like a or a I don't know you're you don't care if you you know about me trying to generate an image let's say because it was going to be public anyway but um I thought it was better to announce the availability of the APIs and you know Toeer and Austin agree with me just announce the availability of the APIs and then and then just let the machine scan those and choose at runtime. So I think that's a valuable
(44:49) paradigm. I'm not sure that it makes sense to create a paper trail for all that stuff for privacy reasons. I think the privacy opsec has gotten a lot better since then. So maybe I'll revisit it. But you know back at that time we were using NIP 4 for DMs. So it was broadcasting the metadata. It's like why why leave any metadata? You know I was a I was a wrestler in college.
(45:11) The way you defend yourself is you don't give anybody any handles. So it's like why leak anything was kind of my mindset. And you know at the time I think Pablo for some reason didn't want to encrypt DVM. So I was like okay that why why broadcast anything? It was kind of my mindset. I think at some point they added encryption like after that like a couple months after that which is cool but then you're still broadcasting metadata.
(45:37) You're still saying A talk to B. I'd rather just be over HTTP and just call it a day and not have to keep that paper trail. So, there's different tradeoffs for different use cases is what I'm saying. That's pretty long-winded, but yeah. No, it makes it that makes a lot of sense. You mentioned MPC, too. Like, how that was It's MCP or MPC? Uh, you're right.
(46:00) I think it's Yeah, MCP because it ends with protocol. I forget what it stands for, but MCP. I mean that's been a big meme the last couple months is this ability to sort of create these protocol these sort of modularized context protocols that what do they do like cache context and create the ability for you to easily sort of create like context verticals if you will that you can plug into and pull so you don't have to keep reproducing context for for these models.
(46:31) Yeah, that's massive and it's massive to have a big AI lab behind it like anthropic. Um, I mean that just goes back to standardization tends to benefit everyone. It's a centralizing effect in some ways, but in some ways it's a decentralizing effect because now you can make your own MCP repo of everything you've ever said on TFTC and now maybe even earn sats for exposing that to the public, you know. Yeah.
(46:56) Yeah. It's fascinating. So, what do you think? Do we get the world of AI that actually works for the people and is free and open or do we get the monolith? I think I think we're going to win, but it's going to take effort and I don't like complacency. You know, I don't I don't think Bitcoin would be where it is if it wasn't for, you know, all the people you wearing the Silk Road hat.
(47:25) It wouldn't be where it is if it weren't for the people fighting for that future. And I think AI is no different. Do you think there's enough people out there fighting for the open future? I'm I'm encouraged by what I see from the open source community. I think the human will to just not be dominated is very strong.
(47:45) It's stronger than maybe we thought maybe like three or four years ago. We're a lot more blackpilled than we are now. Um so I think we're seeing a an awakening, but it's easy to get complacent. Yeah, it's easy to get soft. Well, and the tools are there, right? Whether it's open source models, but I think you're talking to the Open Secret guys about implementing what they're doing and that is incredibly powerful from a privacy perspective and a security perspective.
(48:12) Yeah, I love to see those guys in the game. And I have to say it's so clever that they basically repurposed like their hardcore lightning, you know, they killed themselves to to fight AML regulations and turns out that maybe wasn't like a, you know, viable business or at least they didn't want to continue with that business and they found a new business that's like, yeah, makes total sense now. Well, let's explain why.
(48:34) Why does it make sense in your context? Well, so they they uh they do endto-end encryption between the client and the compute in such a way that the service provider like me or them can't can't decode um what's going on once it leaves, you know, the ASIC at or maybe not an ASIC, but it leaves the chip at the compute center.
(48:56) And so that allows you to unlock a bunch more applications like medical, you know, HIPPA, um, and accounting, right? Anything that is sensitive and in our case, even law. So that's something I'm curious to explore more with those guys is how do we create more value for attorneys where you know the two biggest concerns for them is number one, you know, am I going to be I think it was that Michael Cohen that presented fake evidence like am I going to be able to use this responsibly and have it not hallucinate? I think that's getting better in their techniques like
(49:31) in Jamie web search how we provide the references. We don't just we we we we suggest that you you don't trust but you verify but we give you a highle summary sometimes it does hallucinate but you have the reference in line so the first problem is you know can it be used responsibly will it hallucinate is it accurate then the second question is am I going to get in trouble for leaking this data for a security breach or am I going to lose a case because I because I breached data that I shouldn't have you know that's entire entirely possible
(50:05) too. So the open secret guys have found a very clever way to repurpose the mutiny wallet technology for this new use case. Yeah. No, that I I I talking to them a lot in recent months and I think particularly for AI and privacy like you have a bunch of people uploading financial records, very intimate information, medical records, whatever it may be, and they're just like feeding it to Chat GPT or Grock or whoever and not understanding that like that data is sitting exposed on the servers that they're sending it to. And
(50:38) so like by tapping into these Nitro enclaves to make sure it's end encrypted, you can verify that there's no way for anybody that doesn't have access to the private key that you've set up via open secret um to access the data. Like I think that's powerful in terms of privacy and moving AI specifically towards a future that really preserves and values people's privacy.
(51:08) And like if we're looking to protect ourselves against the digital panopticon, it seems like this is some like very critical niche application of this enclave technology that's going to be necessary to be widely adopted. Yeah, this is what I'm talking about when I'm talking about the Space Force of Liberty. Like they they join the fight.
(51:26) And if you're out there, you're interested in AI, you know, hit me up. Um, you know, there's there's plenty of work to go around. It's like a farm. So, I'm glad I'm glad the Open Secret guys got in here and really brought a game changer in. Yeah. What's your uh you look ahead like 10 years, what is your ideal scenario of AI's progression? How it's being applied? Yeah, my ideal scenario is um The Young Lady's illustrated primer.
(51:55) Have you ever read that? No. It's a favorite. I know Ryan Gentry is a fan of that one. you know, something like that where it's a selfactualizing assistant. It doesn't just let you get soft. It responsibly teaches you and moves you along to achieve your potential. So, this I don't want to spoil the book, but it's about a woman.
(52:17) It's about a young woman that is in a very dysfunctional family and she gets her hands on this magical book falls in her lap and the magical book teaches her the skills she needs to not only survive but thrive. So that would be my ideal is to have a scenario like that where you have a lot of people with prohuman AI assistance at their fingertips.
(52:43) I actually saw an example of this the other day on X where I believe it was a woman was tweeting like I I talked to a friend I hadn't talked to in quite a while and I was really impressed with the change in his demeanor about like how he was approaching just life simply and I I asked him like what what changed and he's like um I gave Chachi BT like context like I want you to be the ideal father figure and talk talk with me through some of the problems I'm having in my life and like just by giving that sort of master context text to a model and having conversations with
(53:14) it like changed how they view the world and how they operate within the world which I thought was pretty cool. That's incredible. I mean that's great but then it goes back to the privacy thing where you're like now is he blackmailable like who knows? Yeah. Yeah. Yeah. And that's why I think people myself include I mean it's it's tough too because like the models uh the monoliths that aren't privacy preserving are really good, really fast, really sleek UIs and user experiences where it's just like the convenience
(53:44) factor is such where it's like boom. I mean I'm smart enough to understand what's happening with the data when I'm sending it. So, I'm very sort of um particular with what I'm using those models for, like things that I wouldn't worry if other people had access to, but when it comes to intimate life details, financial data, whatever it may be, like I would not ever upload it to to a model like that.
(54:11) So, I think people need to understand when you're doing these things, like they're very powerful, but there is that data leak that treat it like your work computer is the way I think about it, right? If you're working at a corporation, they usually got spyware on there to make sure you're not doing bad stuff understandably. Um, use it use closed source AI that way.
(54:33) Um, even even things like o open secret, maybe keep a little bit in the tank just because you never know if there's there's some some mistake in there. Not not criticizing those guys, just saying I would say never trust anything fully. But you can be a little more trusting when you're doing your taxes or you're you're asking legal questions, legal theories.
(54:57) And then the other component to it is, you know, if you pay in Bitcoin, then you really are an anon. I mean, I people use pull that up jam.ai I for search like they they for for like web search um like car completely removed Google from his life in a lot of ways and um it's really good. It's a perplexity like experience and if you pay in bitcoin how the hell do I even know who you are? There's no real way for me to trace that. Yeah.
(55:28) So that's another incredible component of encryption that isn't just the encrypting the contents but even just encrypting the session. Well, dive dive into that like like so you don't need an account. You're running a VPN. You simply want to query something from Jamie, right? And you get presented with a lightning invoice like run this compute.
(55:52) You don't even have to do that. I actually shout out to the Albi team. They did a great job on the Bitcoin Connect project. So, you just hit Bitcoin Connect. You can plug in your Noster Wallet Connect or other uh service providers. I think there's like LNC and other stuff, but generally it's Nasser wallet connect.
(56:07) You just plug in that string and then you pay it from your browser and thanks to the superb um privacy guarantees of both 11 invoices for at least for senders, how would I know who you are? Like there's no way for me to really know. Yeah. There's no like user account that's necessary to accomplish this.
(56:27) And that's incredible to that we have this technology and people barely know about it. Yeah. It's I mean very like think of a world where and like you as a company too like being forced to store all that data whether it's simply just like email password whatever it may be like I would not be surprised if that is proven to be an artifact of our modern times like in two decades from now where like the whole concept of a user account email and password is it's completely out the door because you don't Yeah, maybe we'll get burned enough times where it doesn't
(57:03) make sense. Like, well, people got burned quite a bit. Well, I'm just saying like open AI like we're thinking, oh, what if they're malicious? What if they just f up and now you have like Yeah, you have like all your data CCP or something crazy like that. Some dark web hacker gets their access to their servers and your username and account and every query you've ever made.
(57:25) Hey, do you want to Yeah. Do you want to buy uh buy the the files that that show what Marty's been quering on on his account on Open AI? Yeah, it's a scary scary thought. Yeah, absolutely. Yeah, we need more people to really understand this cuz that's like uh even in Silicon Valley like that, I feel like they don't understand.
(57:48) I think it's fiat brain. I think a lot of people don't like to go deep and actually do the iterations. They're just in the rat race trying to get to the next thing, get to the next shiny object. And I think we're all guilty of that. I even think Bitcoiners are guilty of that. Guilty. You know, but uh but I think being able to slow down and really look at the depth of something is critical.
(58:10) That's why, you know, Next Door, somebody I respect a lot is Keon at Stacker News. He's somebody I go to a lot to just kind of talk things through and think about the deeper repercussions. He's somebody that I respect a lot in that way. Yeah. And you can tell just through the progression of Stacker News over the years and the decisions he's had made, the hard decisions he he's had to make and whilst trying to stay true to preserving user privacy. It's really cool.
(58:41) And Stacker News is one of the cooler projects in the space that has had this organic growth. Like I I feel like I see multiple stacker news. It feels like it's replacing Reddit and maybe not with like the the subreddits or whatever, but like whether it's hackernews um the conversation spaces, a lot of it's moving to Stacker, which is cool to see. Yeah.
(59:07) And it's like what does what does that mean for what does that mean when you add economic cooperation or collaboration into posting? Like what does that do to change the quality of posts? And the other aspect that Keon I don't think gets enough credit for is popular repopularizing Web of Trust. He was one of the first people that I know of talking about in like 22 23 maybe even earlier and uh now it's kind of like common place on Noster as Noster has risen.
(59:35) Um but he's pioneered that whole whole uh technique and you know it's working. I think overall like when I go to stacker news it's pretty high signal. I don't really see a lot of BS in there. Yeah. No, you think about like the Reddit gold sort of incentive structure like it's very easy to gain. Oh yeah, cuz it's not real money at the end of the day, right? You have to add some kind of like civil resistance into it. Yeah.
(1:00:02) By making uh by making uh posting or interacting I think more more vaguely uh putting an economic cost to it, right? Yeah. Yeah, that was one of the interesting things. Somebody else I uh I enjoy chopping it up with is Lyall Pratt. And I was disappointed when Vita ended up not pivoting. I wouldn't say I was disappointed.
(1:00:26) I mean, I understand why, but it was such a cool concept of just making phone lines spam resistant. I get spam calls every day. My block list is so long. I mean, I think that core functionality still exists, but again, going back to the two market problem, we just have this problem where not enough people have have Bitcoin to make it uh make a deep enough market to or create the network effect necessary to make those products viable right now.
(1:00:55) In the future, I think they'll they'll definitely be widespread and widely adopted, but we're just in that that phase of Bitcoin adoption where just simply not enough people have Bitcoin. And yeah, that's why. And now he's crushing it with this with this other other thing. He just went back to telecom plus AI. Yeah.
(1:01:12) Yeah. No, he's just crushing it. He's going after I love his business plan, too. He's going after like the higherend uh small businesses. That's amazing. Yeah. No, they're I mean, that team at Vita's top notchotch. They're uh they're crushing. That's like what you want to see, too. just like le somebody I talk to a lot about AI because I think he out of everybody I know he's one of the few that's like really on the cutting edge of all this things are happening you can just see it in Vita when you talk about like the compute
(1:01:44) cost that the progression of the decline of the cost curve for for compute over the last years has been insane. Yeah. And also just like the uh like the it is like humbling to have to pivot and like that is such a massive pivot but he went and did it and now it's working. So yeah, there's a lot to be learned there.
(1:02:10) So I think a lot of Bitcoiners are are getting more mature in the sense it's like well you can see very clearly the cipher punk future and how Bitcoin plays into that future. But it's just like the information to understand that vision is nowhere near as distributed as it needs to be to have that vision manifest today.
(1:02:32) But I think ultimately it will manifest and people every day are coming around to like, oh, I need Bitcoin. Right now it's cuz like I need Bitcoin because treasuries are piss poor, stocks are dumping. Like it seems pretty obvious that the government's losing control of the debt situation. Now they're viewing Bitcoin as like a wealth preservation thing.
(1:02:51) But I think we're just going to go through this adoption phase where it's viewed as that wealth preservation. That makes sense. I was going to ask you like what do you think the next uh like okay that's the first step like how do you think you get them all the way to cipher punk? like what are the key key like revelations or uh properties that they pick up on that would I think Parker is on it like you just need individuals particularly business owners demanding Bitcoin for payment or at least offering it as a payment option.
(1:03:21) Um, I think that there's like stepping stones towards that cipher punk future, which is like you literally need to get Bitcoin in the hands of enough people to create the network effects that are necessary to facilitate the cipher punk future we can envision. Yeah. And I'm I'm also I'm I'm passionate about this.
(1:03:38) I'm very passionate about medium of exchange. I mean, Cascader accepts Bitcoin. Um, I'm I'm very passionate about this and I know that there are people out there like Wes from Tampa Bay. I don't I probably by the time this comes out it will be public but he was sharing with me a plan to do a um to do a medium of exchange study like on the ground because the Tampa Bay's focus is grassroots adoption.
(1:04:06) So what are like the objections like how can you objection handle merchants and consumers if you don't know what the objections are? So like how can we do that and how we can can we as you say distribute the information we have as hardcore users to other people that don't want to have to learn how to run a server or something like that like they just want to use it.
(1:04:26) So I think that's a really important initiative and maybe I'll send you a link to that. Yeah, I think it's important. But I I do caution anybody doing this just having been in this for 12 years now like the merchant adopt like leaning focusing on overindexing on merchant adoption is going to lead to tears in the long run.
(1:04:46) Yeah, I think most people are going to be forced into Bitcoin just from economic forces like good money replacing bad money. But I I think uh I think a lot of the problems consumer side because people don't want to have to reconcile taxes and they um they don't want to spend their Bitcoin. Think is that I think most people they're not living on zero.
(1:05:06) They're like it's weird to not live on zero and then complain you don't want to spend your Bitcoin. It's like just live on zero. Problem solved. Now you don't have a choice. Yeah. I mean, well, that's like a very niche problem like within Bitcoin Circle, like we're talking about like when it comes like mass merchant adoption and actual payment being facilitated via the Bitcoin network at these merchants like that's a like that get on zero problems like very niche within our bubble.
(1:05:31) Then like you get into the bubbles that exist on top of ours which encompass ultimately encompass the whole world. Like these people, you said get on zero to them, like they have no idea what that concept means. Like 100% on a Bitcoin standard, like they don't even most people still think Bitcoin's a Ponzi scheme.
(1:05:50) No, I I uh I totally understand. I guess my point is if you're like a Bitcoin maximalist and you're like that bullish, you shouldn't have cuck bucks to spend. It should just be It should just be Bitcoin. It's like you don't have a choice. Like I don't know. Somebody with a wife and kids like they're like I'm like most like I am on a Bitcoin standard like as close to 100% as you can get, but there's still that need to like I get that there's a flow and I'm not like especially in that type of scenario.
(1:06:19) I I concede uh you know maybe you got to keep especially for the big bills where it's like all right I'm going to have like a few hundred tax bill if I spend this. I get that. But I'm even just talking about like if if a vendor accepts Bitcoin for coffee or something smallish, you should probably figure out how to how to how to support the network by spending.
(1:06:41) That's kind of download strike and spend from your cash account. I don't like the complacency mindset. I think that's kind not not what you're saying, but like the the idea of like, oh, I'm never going to spend any bitcoin because number go up forever. Yeah. Like like that's loser mindset. like we do we do need to create incentives for merchants.
(1:06:59) I I agree that you don't want to spend all your time there, but I I think there's uh why would the merchant accept it if you're not going to spend five bucks on a coffee? It's not you. No, I see. I mean, I ran into this example of Parker like when I can say it now publicly, but you know, we were discussing it before Rod's acquisition of the Commons and creation of Bitcoin Park Austin, like he took over the invoicing for rent for the studio and he sent me my invoice on a Sunday, a Zapright invoice and he had a 4% premium on fiat
(1:07:32) and I was like, I'm not paying that. I just paid over the lightning network, right? That was a large expense. Thousands of dollars to go over the lightning network settled instantly and I saved a ton on the premium he was charging plus the fees I would have paid um by spending it um via the fiat uh the fiat rails.
(1:07:54) And and when you think about it, like credit cards, they charge 2.9% plus 30 cents. So, if you're like a convenience store, I was this close to convincing the convenience store down the street to uh to go on to uh a Bitcoin center. I think part of the problem is we need like a like real simple plugand go like POS that works. Like I'm really hoping Square or somebody like that can start. Jack.
(1:08:17) Yeah, Jack. Yeah, he's talking about man. Like it's been a while. I know they said they've been working on it. Like I respect everything Jack does, but come on, man. I imagine I I mean I imagine for Block it's just like uh have to deal with a lot of corporate [ __ ] that uh comes with being a multi- deck billion dollar publicly traded company but well life hack I figured out is uh Cash App can do lightning invoices but it's kind of hidden and I get why they probably don't want like programmatic like lump sums happening so they do kind of have
(1:08:49) breaks on it in that way but it's like all right my barber's curious about Bitcoin I can tip them in Bitcoin that's pretty cool. Yeah, I did that to uh to my barber a few months ago. Like I sent her Bitcoin. Like I got her cash tag and sent her Bitcoin. She never claimed it. I was like, "What the heck? What the heck? I got to go back to my original barber down here.
(1:09:16) " Well, it's funny cuz my barber's like, "What's happening to it? It's going down. It's what's happening to it." Then, but you know, six months ago, she was like, "Oh, it keeps going up." And it's like like I told my mom about it and she was like you I might sell this. I I told her I might sell this and she says no listen to him. Hold it forever.
(1:09:32) Like your mom was the voice of reason. I want to send out an apology to my my OG barber here in Austin. Kirby. Kirby, I miss you. I haven't been like six or seven months. It's far in North Austin. It's hard to make the drive. You're going back to flow. You're getting you're going to get back into lacrosse, right? Yeah. Yeah, dude.
(1:09:51) Visor down, flow out. We had a guy, this is funny, I had a a guy that sat next to me in high school chemistry who was uh who ended up being the the goalie for Syracuse. His name is Bobby Wardwell, but we called him Brobert Flowwell. That's a very lacrosse name. Bobby Wardwell. Yeah, actually did recognize that name.
(1:10:15) Yeah, he was at Syra I think they were like number two or something. I remember watching him on TV. I was like, "Oh, Notre Dame versus Syracuse one versus two." When would he have been there? I want to say it was like 14 or 15 time range. So my brother beat him in the NCA tournament. Oh, really? Bryant beat them.
(1:10:29) Upset them in the first round when they were number two. I didn't know that. That makes sense. Yeah, that sucks. Yeah, big upset. Sorry, Bobby. Bobby Bryant Bulldogs got you. They got smoked by Maryland in the next round, but it was a good underdog moment for the Bulldogs there. And that was 2014. Yeah, that was 2014. Yeah, that sounds right.
(1:10:50) Yeah, this has been fun. Where um where can people learn more about Cascader, Jamie, Claraara, whatever they're interested in? What piece of advice would you have to to people out there exploring this intersection of Bitcoin and AI? Yeah, I would say uh yeah, if you want to check us out, there's pull that upjamie.ai.
(1:11:11) So, if you're listening to this, you're probably a podcast afficionado. So, we have the top 30 or 40, you know, podcasts I think Bitcoiners would be interested in, including this one. And so, you'll be able to instantly search stuff. Um, you can send a deep link to your friends. You just click a link, send it to your friend, and it'll dump them right in the actual part you care about.
(1:11:33) You won't have to, you won't have, you know, this problem where you're like, "Hey, you got to listen to this whole hour long podcast and find the part that relates to you." So, there's the quick share, there's the clipping if you want to make clips and share them. Um, you know, we're open to other podcasts if you want us to index you.
(1:11:51) There's a request button. Um, and you can, you know, fill out the form and we'll we'll get back to you like pretty much immediately. Um, so you'll get your your last like 10 or so episodes indexed and searchable for yourself and your fans. And um you'll get a bunch of other benefits like the Jamie Pro agent which will uh give you emails every time you drop a podcast uh with with curated clips based on your interests.
(1:12:20) So that's another um benefit. So So yeah, that's pull that up Jamie. Um, if you're an attorney and you deal with, you know, any kind of um, multilingual cases or um, any kind of litigation or depositions, that type of stuff, and you need to be able to pull things up fast. Similar use case to Jamie, we can we can do that for you.
(1:12:45) We have Clara AI, so get in touch with us. Um, I I guess I'll give you my calendar link. I don't know. That's one way to get in touch with us. And then um if you're out there and you're a builder, you know, my DMs are open. Uh I would love to see more liberty-minded people out there focused on building the future and not just kind of passively consuming and and uh waiting for the feudalists to come get us. They're not going to get us.
(1:13:11) The human spirit is strong. Yeah. We're going to win. Yeah, we're going to win. Let's go seize the day gym. Yeah, that's what my socks say. My dad gave me socks. They say carpy the [ __ ] out of this DM. Carp DM freaks. Peace and love freaks. Thank you for listening to the show. I hope you liked it. If you did like it, please make sure you subscribe, rate, review the show.
(1:13:35) It helps us out a lot. And also, if you like these conversations, I've come to realize that many people listen to the podcast, they don't know we have another sort of layer of this media company. We have the newsletter, the Bitcoin brief. Go to tftc.io. io. Make sure you subscribe there. A lot of the topics that are discussed on this podcast I write about 5 days a week in the newsletter.
(1:13:59) We also have the TFTC elite tier. If you sign up for that, become a member. We have a private Discord server for the elite freaks out there where we're dropping adree versions of this show and having discussions about everything we talk about a day early. Logan wanted me to make sure if you want to get the show a day early, become a TFTC Elite member. You will get that.
(1:14:28) We have our Discord server right now. conversation between myself and TFTC elite tier members, but we're going to expand that. We'll probably do close Q&As's with people in the industry. Uh I may be doing macro Mondays, so join us. Go to tftc.io, subscribe, find the button in the top right corner of the website, become a TFTC Elite member. Thank you for joining us.
