We Study Billionaires - The Investor’s Podcast Network - BTC207: Bitcoin, AI, Nostr, and Hardware Manufacturing w/ NVK (Bitcoin Podcast)

Starting point is 00:00:00 You're listening to TIP. Hey everyone, welcome to this Wednesday's release of the Bitcoin Fundamentals podcast. On today's show, I have one of my favorite technical guests with Cold Cards, founder and creator, Mr. NVK. During our show, we talk about some of the super interesting things NVK is building with AI so people can train models with their own personal context windows without giving up any of their data or security. Additionally, we talk about how Noster continues to get more robust and important in digital identity and interoperable. ability with Bitcoin and AI amongst many other topics. This is one you surely won't want to miss. So without further delay, here is my chat with the thoughtful NVK. Celebrating 10 years, you are listening to Bitcoin Fundamentals by the Investors Podcast Network.

Starting point is 00:00:50 Now for your host, Preston Pish. Hey, everyone, welcome to the show. I'm here with the one and the only NVK. Welcome back to the show, sir. Thanks for having me. Hey, so this is where I want to start, and this is going to be kind of non-standard, maybe not where people would expect us to start. But you've been tinkering with AI. And you've been tinkering with it in a way that I just find fascinating. I think it just highlights your brilliance and how smart you are in the space and how you can

Starting point is 00:01:25 kind of cover so many different things. And I'm not saying that to try to embarrass you or anything, but I truly- I'm just an idiot. No, that's so far from the truth. And anybody who knows you knows that's so far from the truth. But, you know, very early on, I think you started tinkering with this maybe a year ago with AI and you started building out some interesting stuff online. And I think it's very counterintuitive for a lot of people that are playing with AI because they look at Open AI. They look at Chat GPT. And they're looking at how much money they're plowing into this thing. I think they're, what have they spent on all of their training? Like over $100 billion or some absurd. Some absurd. Yeah. It's crazy. It's infinite money. It's infinite money. And I think for somebody from the outside that doesn't understand AI, they look at that and they're saying intuitively, they're saying there's no way anybody's going to

Starting point is 00:02:16 be able to compete with them because of how much they're spending on training and the argument or the case for that goes on and on and on. But a guy like you and others that deeply understand AI are saying, oh, that's not how this is going to work at all. And so explain what you've built. Explain why that logic of, we're going to throw $100 billion at training and therefore or we're going to have the better model is a false assumption. And then give a handoff to this thing that you're building because I think it's absolutely fascinating and just kind of help educate us in the way that only NVK can educate us in that you make things so easily accessible.

Starting point is 00:02:52 And even though they're super technical. Yeah. I mean, well, thanks for the intro there. That's going to be a hard follow up. So training models and sort of like inventing the hard science of this large models, right? There's large language models cost a lot of money. I mean, that part, it does cost a lot. You have this very, very large neuronets that you have to like have a lot of horse power to

Starting point is 00:03:16 train on. You need a lot of data to put it through. You need to hire a lot of scientists, a lot of mathematicians, a lot of staticians, like all of the people who normally don't work at a startup. Yeah. But that's to build a model, right? Like, think about the models as more like a car engine, right? Think Toyota.

Starting point is 00:03:34 So Toyota has this V6 engine that like they stick on almost every car they make. Okay, like from trucks to sports cars to off-road vehicles, it's the same V6. And then they will add a few other things to it and make it make it be. So with this large language models, you have MixRow, you have Ljama from Facebook. You have GPT from Open AI. There is a bunch more. So with these engines, you can create the appropriate machine. All it does is say, hey, you know, like, I think this.

Starting point is 00:04:04 letter comes after this letter to make this work. Yeah. And these things have like an incredible accuracy to make it happen. So that's sort of like basically how this language models work. Now, they're not very smart. It's just an engine, right? Like it doesn't have wheels. It doesn't have a chassis.

Starting point is 00:04:17 It doesn't have a purpose, right? Like you can rev an engine as much as you want in a little booth, engine test booth. Makes a lot of noise great. So what happens is you need to build all the tooling that comes around. And that's the stuff that makes it look like magic. And that's a lot of human made stuff often in Python. And this tooling includes stuff like Lang Chan that helps you.

Starting point is 00:04:38 Essentially, you know, the AI doesn't understand English. Yeah. Okay, none of these models understand English, right? These are just vector maps where, you know, it just understands that statistically these other point is here. That's it. So when you create this model that now understands like human topics, say translation, right? So like you want to transfer from this language to this language.

Starting point is 00:05:00 The AI goes there and create it has this statistical map of like, this word essentially is going to be very close to this other word in this other language. And then with a lot of points, millions and billions and billions of points, he figures out the right neuro path to sort of like put out what he wants to put out. So anyway, so that's cool. But then you go like, okay, can you please tell me what like Preston had for breakfast yesterday? It's like it doesn't know. It doesn't know Preston.

Starting point is 00:05:27 Yeah. It doesn't know like how to figure out like what breakfast or or where. or what? So what do we do? We go and we build something like, for example, GROC, where we teach it how to access this database of tweets by Preston. And that's just like a lot of code that we let the agent, the AI agent, access and run as a query on a database.

Starting point is 00:05:52 So then it gets like, all this sort of tweets by Preston talking about breakfast, right? And then it's vectorizes it. What that means is that it transforms. all this stuff into points in a vector map. So that's the language that it speaks. And then it goes like, okay, well, I have some context now. That's called the context window. And he goes like, breakfast, Preston, English.

Starting point is 00:06:18 It's sort of like with a lot of whooply goobly there. It comes off if like, he probably had bagel today because he talks about bagel all the time. All the time. Yeah. Okay. So that's the value that the people who are not building the model. themselves, the engines, right? We're starting to build all these other stuff.

Starting point is 00:06:37 And this other stuff is like, how do you embed all this context into the AI model window, the context window, right? So for example, every time you open a new chat TPT window, right, a new chat, it clears the old context, right? And it's kind of like a shame, but it's a limitation of the technology. Because there's only so much context in width in bytes, right, that you can stick it into the model in runtime. So it has a limit. It's similar to like a person, right? You learn a new hobby, but you kind of forget other things and there's only so much you can have in memory at a time.

Starting point is 00:07:13 You know, our context window is absolutely ginormous. The machines are not there yet, but they're there yet for many other things. So they can remember like they can have the whole Wikipedia accessible. And the models are trained in a snapshot in time. So what that means is that When you run that training and you essentially massage it and teach this thing to think, it is a snapshot of data in time. That's why AIs are terrible at trivia. Pre being able to search the internet for things, if something happened the day after the data set was used for training,

Starting point is 00:07:46 the AI doesn't know. Yeah. So it just assumes stuff. So what we do now is we have all this tooling that goes and sort of like, you let the AI search the internet, find the context and vectorize it, inputs it into the system, you know what I mean, in that context window and then helps you answer questions, right? So what I like to say is like you have the AI galaxy brain people, right,

Starting point is 00:08:08 building the models. And then you have the left curve apes like us building businesses and sort of like building the tooling around it to sort of feed into the Borg. You use this term context window to basically have like specialized or updated use you were using me and my breakfast preference as an example. What you're building is that, would you categorize that as enhanced context windows for individuals? Is that kind of how you would preface what it is that you're building to explain to people what you are building? Yeah. So, okay, so I'm building unleashed dot chat, right? That was born because I wanted to do data analysis, right? So use a model to analyze my source code, my company's, business data and I'm not willing to give away that data to open AI for free, right?

Starting point is 00:08:57 Like, I mean, they're going to just take your data in terms of services. The data is mine now. So, and I want privacy too, right? So Bitcoin company, coincite.com, right? Like, we do a lot of stuff that we don't want to just give away to the board. And so I went down the rabbit hole of like, how can you run these models privately? And a lot of people do this. You end up running on your own laptop.

Starting point is 00:09:18 But like your computers are very limited in GPS. memory, right? Like, that's the difference between like having a million dollars worth of GPUs in a farm, right? That were specialized, create for AI stuff, right? Versus like, even the most expensive laptop you can buy. That context window is one of the first things that suffer. And you want the biggest context window as possible. Yeah. So that's why like most phones, they're going to have AI and things like that. They're going to have a local sort of like a on device, a little chip that helps accelerate things. But For most things, it's going to have to go to the cloud, go do the thing there in big space of servers and then come back.

Starting point is 00:09:58 So on Unleash.chat, what we did is like, how can I make it so as private as possible for myself versus the people working on it inside the company? And also we saw other people want it to use it. So like, okay, great. So how can we make this sort of like multi-tenance? So like more people can be on it privately. And it's not perfect. But we've done a decent amount of work into making sure that we can't see what people are doing. And there's other things we want to do in the future.

Starting point is 00:10:25 But so your contacts is stored on your local browser. We don't have it in our servers. And that helps a lot. And anyways, so the cool thing is things that I wanted to analyze are, for example, time series data. We're still working on that part. Very hard to do time series data with AI. By the way, if anybody's listening and is an expert on that, please hit me.

Starting point is 00:10:47 I love to talk to you. Explain what you mean by time series data for people who might not understand what you're saying. So time series data is essentially like imagine roles on an Excel right and you have like a date or time and then you say like stock prices today this and then tomorrow is that and then tomorrow the day after is that right and like business data 99% of the data is time series data right so that that is to me in my opinion and somebody obsessed with time series data the most interesting data there is and so but that's surprisingly hard to build tooling for and we're working on it. And then there was I wanted to very quickly look at code, right?

Starting point is 00:11:24 So analyze our code. So we built this thing where you can just put your GitHub repo and you press a button and the inputs. It takes the data in and like, boom, it's like ready to go. And you can do analysis on that. And then I was looking at Noster, right? You're familiar with Noster. You had a few Noster folks to talk about it before.

Starting point is 00:11:43 It's just this sort of like decentralized communication network that is great for many things like social media or markets, all kinds of communities. It's a very cool broadcast system. Anyway, so I was like, okay, great. So Elon has GROC on Twitter and, but I can't access it, right? That data is fully closed and that sucks. So I was looking at Noster and it's like, oh, the social graph and the data is open. So let's see if we can build rock for that and surprisingly enough like, I mean, it's totally

Starting point is 00:12:11 possible. We did it. So we index as much of Noster that we can see and then we vectorize everything and you have like live access to Nostr on that. And that's super cool. I mean, as things start moving further into Noster, right, like Nosterizing everything, it's a super powerful thing. Not just for like what did Preston have for breakfast.

Starting point is 00:12:31 I know that's very cool topic. Which, by the way, Eggs Benedict is my favorite breakfast. No, no, no. I'm just going to put that out there. I'm going to keep a Tobago. You know how AI's got everything wrong. Eggs Benedict. Okay.

Starting point is 00:12:43 That's a tasty one. And so with the Noster part is like, imagine when you have like people using for bids and asks, right, on a public market. or they have, you know, like advertising programs in there. Whatever people are going to end up doing with it, it's just really cool to be able to query that live and have the social graph with it. I think social graphs, that kind of graph and large language models,

Starting point is 00:13:06 they seem to really have interesting interactions. Yeah. You can discover a lot with that. So we're playing with that at this time. Let's take a quick break and hear from today's sponsors. All right. I want you guys to imagine spending three days in Oslo at the height of the summer. You've got long days of daylight, incredible food, floating saunas on the Oslo Fjord,

Starting point is 00:13:27 and every conversation you have is with people who are actually shaping the future. That's what the Oslo Freedom Forum is. From June 1st through the 3rd, 2026, the Oslo Freedom Forum is entering its 18th year bringing together activists, technologists, journalists, investors, and builders from all over the world, many of them operating on the front lines of history. This is where you hear first hand stories from people using Bitcoin to survive currency collapse, using AI to expose human rights abuses, and building technology under censorship and authoritarian pressures. These aren't abstract ideas. These are tools real people are using right now. You'll be in the room with about 2,000 extraordinary individuals, dissidents, founders, philanthropists, policymakers, the kind of people you don't just

Starting point is 00:14:13 listen to but end up having dinner with. Over three days, you'll experience powerful mainstage talk, hands-on workshops on freedom tech and financial sovereignty, immersive art installations, and conversations that continue long after the sessions end. And it's all happening in Oslo in June. If this sounds like your kind of room, well, you're in luck because you can attend in person. Standard and patron passes are available at Osloof Freedom Forum.com with patron passes offering deep access, private events, and small group time with the speakers. The Oslo Freedom Forum isn't just a conference, it's a place where ideas meet reality and where the future is being built by people living it.

Starting point is 00:14:55 If you run a business, you've probably had the same thought lately. How do we make AI useful in the real world? Because the upside is huge, but guessing your way into it is a risky move. With NetSuite by Oracle, you can put AI to work today. NetSuite is the number one AI cloud ERP, trusted by over 43,000 businesses. It pulls your financials, inventory, commerce, HR, and CRM into one unified system. And that connected data is what makes your AI smarter.

Starting point is 00:15:25 It can automate routine work, surface actionable insights, and help you cut costs while making fast AI-powered decisions with confidence. And now with the NetSuite AI connector, you can use the AI of your choice to connect directly to your real business data. This isn't some add-on, it's AI built into the system that runs your business. And whether your company does millions or even hundreds of millions, NetSuite helps you stay ahead. If your revenues are at least in the seven figures, get their free business guide, Dymistifying AI at netsuite.com slash study.

Starting point is 00:15:58 The guide is free to you at net suite.com slash study. NetSuite.com slash study. When I started my own side business, it suddenly felt like I had to become 10 different people overnight wearing many different hats. Starting something from scratch can feel exciting, but also incredibly overwhelming and lonely. That's why having the right tools matters. For millions of businesses, that tool is Shopify. Shopify is the commerce platform behind millions of businesses around the world and 10% of all

Starting point is 00:16:30 e-commerce in the U.S. from brands just getting started to household names. It gives you everything you need in one place, from inventory to payments to analytics. So you're not juggling a bunch of different platforms. You can build a beautiful online store with hundreds of ready-to-use templates, and Shopify is packed with helpful AI tools that write product descriptions and even enhance your product photography. Plus, if you ever get stuck, they've got award-winning 24-7 customer support. Start your business today with the industry's best business partner, Shopify, and start hearing... Sign up for your $1 per month trial today at Shopify.com slash WSB. Go to Shopify.com slash WSB.

Starting point is 00:17:16 That's Shopify.com slash WSB. All right. Back to the show. So, NVK, you said that you had built Grok for Noster. If a person is using Noster, let's say you're on primal and they want to interact with this, I'm assuming that Million, if he wanted to incorporate access directly to the Unleashed. chat and what you guys have done as far as indexing and vectorizing all of this. That's something that he could just add as a button in there. It would almost look like GROC on Twitter and all that.

Starting point is 00:17:50 From a payment standpoint, so you are pinging a server that's then taking your model that you've developed and you're providing that input and then it's pumping out and output, but that cost energy to basically run it through that model. I'm assuming you're paying with SATs or the person that would be accessing this has a micro payment of 100 sats to basically ping the server and utilize those resources. Am I describing this correctly? And is the vision, I guess, that I'm describing something that you think is going to play out for some of these clients that are hosting Noster clients? Yeah. So it's nice that you brought that up. So the way we monetize and leashed was like, how can we have honest sort of pricing? Right. So we just do token time.

Starting point is 00:18:37 Sats. So all that means is that like whatever we're paying for computation, even though we have the GPUs allocated to ourselves, we try to do that as if it wasn't. We just charging Sats like it just deducts from your account. However, because of Noster, you can connect your account. And if you want, you can use DVMs. These are digital vending machines. So you can go through the UI route, right?

Starting point is 00:19:00 Like on maybe million ads to Primal and then you press a button. But what's really cool is using digital vending machines. This is a very early working Noster. And what that means is native to the protocol of Noster, you can post an event saying, can anybody translate this for me? And it's a special kind of event. And if the Unleash.comchat digital vending machine sees that, it makes an offer. Like, I can do that.

Starting point is 00:19:26 This is my cost. Or you can do it sort of like, it can do it and say it costs this for you to see the result. So you can now imagine how instead of having APIs, that you need permission to use and all this stuff. It's just when you press translate, and all of a sudden, you don't care who did it. You create some logic that says, like, give me the best price. Yeah, yeah. And I don't like this, this one because they never did a good job.

Starting point is 00:19:49 And I imagine that for any kind of competition, this gets super interesting. And what's interesting to the user is the privacy that they're also getting from this, as opposed to, hey, Elon Musk now knows the last 10 questions I asked Grock. and he has a very good idea as to what it is I'm interested in and what it is I'm going to be doing tomorrow potentially or whatever. So that's interesting. Okay, so let's say I was going to use Unleashed.com chat for not for like the AI GROC use that we just described, but let's say I wanted to take the last 10 years of my emails. I wanted to train it on how I draft, respond, the style, everything. So I take all 10 years worth of email data of the backer,

Starting point is 00:20:34 and forth inside of my account and I want to upload it into this Unleashed.chat. I want to basically train a model on Preston's email habits or whatever. And then any inbound traffic that I have on email, I now have a model that's giving me a really accurate response that I don't have to spend a lot of time doing. Is this something that is capable with what you're building or help walk us through like what some of the capabilities are for people that want to train it on their data or and then have possession of the model and nobody else has access to all of that information. Talk us through that. So we do have APIs, right?

Starting point is 00:21:11 So if you want to build a tool that integrates that with your email, like that could be done. I can't remember right now of the top of my head the size of the context window. So 10 years of email might be a little pushing, but there is no reason why that can't be done. What would something like that cost for me to train it on that much data? And I got a lot of emails over 10 years. Would that be pretty expensive? Would it be hundreds of dollars? What would it be?

Starting point is 00:21:36 You know, I think like, surprisingly enough, probably like between $100 and $1,000 kind of thing, maybe. A hundred and a thousand. Okay. Yeah. I'd have to put it through. I thought you were going to say $100,000. No, no. No.

Starting point is 00:21:50 So this is the thing right. Like if you go to Open AI and you want to train a model on your staff, they will, I think about a million bucks and they'll do it for you. The thing is that the result will be very similar to just having enough context window. So essentially with the context windows, it's essentially non-permanent, right? So if you just reboot the GPU or get that context out of the system, which we do after each run, but you keep it locally. That's it.

Starting point is 00:22:17 You can't update it with the next month. It's a very similar result. Yeah. So right now, what we're doing is we're using mixtro, which is a really, really good model for many things. I don't think the context window will be enough. 10 years of emails. I'd say like you probably have like a couple gigabytes worth of emails.

Starting point is 00:22:35 Two to five. No, you get a lot of email. There's not a lot of responses to these emails. The point is like these are all things that like are like reasonably easy to engineer. Yeah. If there is a market for something like that, I mean, it shouldn't be hard to just put together something that that does that. But I think more more than the inputting of the emails,

Starting point is 00:22:56 is having the ability of creating sort of like a hard prompts. Essentially like you're helping the prompts. So one thing you can do is in the instructions, like the custom instructions, you can say like whenever I say this, I mean that or whenever you think I like bagels for breakfast, I actually like this other thing for breakfast, right? It's Benedict. Yes.

Starting point is 00:23:17 Ex Benedict. So you can, you can essentially massage the prompt. Yeah. To have an inclination towards things. And so, for example, we had that during COVID. Like, you ask the models about the safety of some medication. And then like, it doesn't use the data. It just outputs what the government asked them to say.

Starting point is 00:23:40 So that's another interesting aspects of this whole thing is that when you are using the big models in those big websites, they have to follow government guidelines in a bunch of stuff, right? Like if they want that funding. It's biased. Yeah. Oh, yeah. I know, it's insanely biased, right? Like, I mean, was it in Google that, like, made the ratio look of somebody who was not supposed to be anyways?

Starting point is 00:24:05 It's really bad. Yeah. And when this big companies release these models, what people try to do is a bit of a lobotomy on them. They try to de-educate them from all this communist stuff and put them as incensored models, right? It is not perfect. And so people should know that. There's a lot of companies out there to try to pretend they're like their models are super based or whatever, right? But you know, it's all they did is they go there and they type. If somebody

Starting point is 00:24:29 asks about COVID, just answer this, right? And then they look good on demos. But realistically speaking, there is limits because the actual fundamental part of the model was taught in a certain way. So even though the models can be uncensored, they're still not quite there. And it's okay. I mean, you can be productive or you can look for outrage. You pick your path. But anyways, so the models are improving. And we want to have Lima three soon added to the system. We support different models. Different models are kind of better for different things. And there is a lot still that can be done to make all this more productive.

Starting point is 00:25:05 For people that want to check this out, Unleashed.com. This is what NVK is building on the AI side. Super interesting stuff. If you have, what is it, time chain data experience and you want to hit them up in DMs, I'm sure you'd love to talk to you more on that. I have a question here on Noster, but I'm going to push that until later on because I really want to get into Bitcoin. It appears that we might be on the cusp of a big bull run.

Starting point is 00:25:30 And for people that don't know who you are, you have the best, literally the best hardware wallet on the entire planet when it comes to Bitcoin. For people that are looking at the video, you can see some of them behind NVK there. This is the question I want to ask you. What is the most frustrating security advice that you hear people sharing? Oh, Jesus. The list is long and prosperous. What's at the top of the list, though?

Starting point is 00:25:55 Oh, man. I think for most people using DIY devices, it's just a terrible idea. They're going to end up screen themselves. Keeping seeds at home, terrible idea. Doing a 202 multi-sig, also idiotic idea. Using a node phone as a hydro wallet, oh, that triggers me so hard. It's one of those things that we've been building now, like, you know, Bitcoin security for, you know, over 15 years. Right, like not just me, like all their companies too.

Starting point is 00:26:25 And we got to a point where the security is amazing. Like you buy a cold card for 150 bucks or like a treasor or a ledger or whatever. You buy one of these devices and you have like Pentagon level security. Like even more actually. Yeah. Way more. Uh, because there's no back doors. So like it's way more security than even the Pentagon has for private keys.

Starting point is 00:26:47 Right. Like it is, it is mind boggling. And I think as we become. more accustomed to our security becoming so good, I think the conversation becomes more complacent. Because you just assume everybody knows, right? You assume that everybody that's been around knows. And that gives sort of like space to a lot of people who have not a lot of understanding of security to become like new security influencers. And that mud is the water big time, right? Because, you know, the new people don't know the difference, right? That there is all this like

Starting point is 00:27:17 security being built. It's not like they're following the Bitcoin mailing list or reading the cold on their devices. So it's a very frustrating thing to see like bad advice being put out because people will get wrecked, right? Like people will get racked and people get wrecked and then they go tell everybody that Bitcoin self-custody is not safe. And then you have a bunch of people holding ETFs and not real Bitcoin. As a Bitcoin bagholder, it is in your interest to have as many people as possible being self-custody in their coins safely, right? You don't want them to lose it on a personal side, but it's also how Bitcoin remains secure and independent from government. central banks. I suspect that this is only going to get worse moving forward. Oh, absolutely. Yeah.

Starting point is 00:27:58 So if let's say somebody's from one of the big banks or, you know, they're one of these people who's like, oh, yeah, I'm just going to buy the ETF. What's your message to that person who's, who has that opinion? And they're just looking at it like, I've been, I've had a trading account with, you name it, for the last 25 years. I've never had a concern or issue with my Apple stock not being there. I'm going to buy I bid or I'm going to buy FBTC. And, you know, it's just going to be like all the other stuff. And I don't ever have to worry about it. What do you say to a person like that? It depends on my mood. And normally I'll just go have fun, staying poor. But I always say this, okay, like Bitcoin has two propositions, right? It has the debasement proposition where,

Starting point is 00:28:39 you know, if you hold any kind of Bitcoin, even paper Bitcoin, theoretically, you are protected against debasement, right? Your Bitcoin is protected from government inflation. And then there is the sovereignty part of Bitcoin, right? The sovereignty part of Bitcoin makes sure that Bitcoin anti-debasement doesn't happen, but that's a different conversation. But what happens when the government says, you know what? Like, we're going to tax Bitcoin differently. And let's say you really eaped into the ETF and you have like, you know,

Starting point is 00:29:07 90% of your portfolio in Bitcoin ETF. Now you have a conundrum there. Like that's a problem, right? They can just take it. What if the trading task? I imagine most people don't have a trading desk. But like, let's say like a big broker. goes under and then there is some issues on like who owns what and some issues on those coupons.

Starting point is 00:29:25 So it's not ideal. And then there is the fact that you're simply not, you're not adding security to Bitcoin. If you're holding that Bitcoin on an ETF, they are trading that Bitcoin. They might be leveraging that Bitcoin and you're actually making that Bitcoin go down price too. I know it sounds crazy, but if you are lending your Bitcoin out, right? Or if you were holding a a Bitcoin that is owned by somebody else, like a big, tradfai entity. I guarantee you that they're leveraging it or they're making it available in some other ways to other people to trade against you, really. Because that's the only time somebody needs your asset is to trade against you, right?

Starting point is 00:30:01 So if you take the Bitcoin custody, you're essentially removing from the system and you're making sure that nobody can't sell or lend out that BTC. Yeah. Right. And you're essentially making the supply smaller and you're making the price go up. So you want the price to go. up if you're a Bitcoin person, right? So take the custody. Yeah. You are one of the few people that I know in the space that's literally manufacturing hardware and doing it in a successful, profitable way.

Starting point is 00:30:30 I think for people that have never run a production line or dealt with supply chains, they don't understand how hard this is to do and do it well and to do it profitable and do it for a very long extended period of time like you have. And I guess, I guess the question I have for you is, in the saying goes, hardware is hard, which is pretty generic, but very true. What is it that you would like to tell the audience about hardware manufacturing and maybe something that you learned that you didn't know before you got into this line of business? That's just interesting or something that is rarely discussed because I find the topic

Starting point is 00:31:08 to be really interesting. Yeah, I mean, hardware is really cool. So there was a saying, I can't remember who said that might have been Paul. Alan. If you're serious about your software, you make your own hardware. Right. Like, and I don't think you can be. That's interesting because that's a Microsoft quote, right? Oh, there you go. Right? That's really interesting. Because it was more Apple that you would have thought would have said something like this, but this is Paul who was. No, it might have been them. I don't know. I can't remember. Oh, okay. Okay. We'll search it. But anyways, the thing is, very few things you have to be

Starting point is 00:31:41 serious about, right, like the hardware, right? Like when you have a private key on a device, anything goes wrong, money gone. Yeah. Yeah. And there is no, there's no mulligans, right? It's not like a bank you can just rewind. And you can use general purpose computers, right? So when you want to do something, you want more features, you want to make it prettier.

Starting point is 00:32:02 But the more abstractions and the more features, the more things you add, the more complexity you add, and there's more cold there that could have a bug, you could have a backdoor, it could have issues, right? So what we do in embedded hardware is we try to make it a similar. as possible. So we pick the actual chips we want to use and then we develop the actual code. Like there is on a code card like every single up code that we're running on that chip is ours. Like we wrote everything from very scratch. Yeah. And then we picked the other secure elements we want to use. We do the same with them. And everything goes in this very sort of like simple, very bare bones package.

Starting point is 00:32:38 And building hardware outside of Bitcoin, it's like you can take some liberty into like how you want to design it and things. But on Bitcoin, it's this crazy universe where we have to be insanely paranoid. So we're thinking of what if an evil made swap the devices? How do you protect against that? What if a delivery man opens the package? Yeah. What if we get compromised, right? Like, how can you protect the users against that?

Starting point is 00:33:02 So there is protections for every little bit of these things. And they're serious, right? Because we don't want to be compromised. We don't want anybody to be compromised. So like, how can we de-risk every aspect of the operation, right? And that's not very normal for a hardware company. Yeah, yeah. Normally they can fix things with a software update or whatever, right?

Starting point is 00:33:21 So we go into this like extra crazy level of security to do this things. And just because of the nature of Bitcoin, we were able to do this. If this was some system that needed a server or something else, we would never be able to do this. So Bitcoin created this asymmetry of security where you, with a design like ours can have this insane security for the price that you can get. It really is remarkable. And when it comes to harder companies, I mean, you learn that like, there is like essentially

Starting point is 00:33:50 two things that matter, right? Like it's logistics and how close down, how much can you get to the bottom of the chip? How close can it get to silicon in terms of to write your firmware, right? Because if you control the whole stack, it sucks that you're now stuck with that chip. You have to rewrite everything in order to move to another one. But you control everything, every aspect of it. And then you use your logistics, new knowledge. which we learned through time to make sure you have that specific chip in stock.

Starting point is 00:34:21 Do you consider lifetime buys? I mean, you don't really have to pay for replacements. Typically, people would buy another device, right? So you don't have to worry about that, I guess, huh? Yeah. Yeah. So what we do is we have like just very good relationships to the manufacturers. We pick families that if we did chip families, that if we needed to move, we could.

Starting point is 00:34:40 It's just a lot of work. Yeah. But it could be done. So, and we, we have like good knowledge that unfortunately under NDA and stuff about like, you know, end of life of parts. It's very serious, right? Like when you're talking to chip manufacturers like for real like 101, not buying off DJK kind of thing, there is assurances. They can't just change stuff. Imagine if they changed a chip and then like BMW can't make a new car because they have to change the whole firmware.

Starting point is 00:35:07 Right. Yeah. This was true during COVID. It was insane. The amount of car makers that shipped their car. missing chips. It's insane. I'm just like, sorry, you can't have this functionality for the next year.

Starting point is 00:35:18 We'll have the chip later. Yeah. But it's pretty cool. It is a lot of fun. It really did show you the dependency on chips. There's a fantastic book. I've mentioned this a couple times on the show called Chip Wars. It just shows you the dependencies on Taiwan, the lithography process, like all of it.

Starting point is 00:35:36 It's just, and I think because it's such a specialized piece of manufacturing to create these chips and how expensive some of these lithography machines are. I want to say some of them are like $100 billion just for one machine. And there's only one manufacturer. I think it's somewhere over here in Holland that makes these things. It's such a critical path to production. And it's so vulnerable to disruption. And when you understand all this and you see the U.S.'s interest in Taiwan and China's interest in Taiwan, you quickly understand why that's such a heated conversation is because of all these dependencies on these chips. And COVID truly illustrated, like, how scary some of this can get if that supply chain is disrupted. For you in your manufacturing of cold card, is that

Starting point is 00:36:24 something that ever got pretty scary as far as like receiving chips? Or is it something that you think about in the lots that you're buying in order to prevent that volatility or future volatility with things. I mean, it's a little sort of, it's always like a little scary, but at the same time, we have the scheduling and the parts and inventory and stuff. We're harder and normally playing a year or two out. Yeah. So it's not then the world. For example, there was a typhoon in Asia, right? Like, I don't know, 70% of that batch got moved. Like, it went on an airplane to the wrong country. Then they have to wait out there. So you get used to these things and you prepare for these things. Most of the time is just like really like just a capax problem. Right. Yeah.

Starting point is 00:37:03 It's like, can you stock more parts? What you were talking about, the Taiwan thing is interesting, right? Because you know how people say, like, humans don't build beautiful buildings anymore? I mean, it's just that we got bored of the buildings. Now we're building like this insane cathedrals. Exactly. Like, it is the pinnacle of human technology. It's not putting rockets in space.

Starting point is 00:37:22 It's like making a two nanometer doped silicon little transistor, right? Like, it's insane. I highly recommend people go on YouTube or something and learn how this dyes are made. Like, it really is my number. boggling. But now the most insane part of all this is that there's maybe 10 dudes out there that know the pro, like the reason why TSM is TSMC is because of those guys. Yeah. And it really is just one guy. So I don't know if people noticed this, but like a few years ago, Samsung, right, like, oh, look, we have this amazing fab now. We're competitive. And then

Starting point is 00:37:57 the SMC goes like, hmm, hang on a second. The guy who invented the whole. like a manufacturing process because like having the tech in the paper is one thing. Like having that stuff yield in a factory floor is a whole different ballgame. And it's normally like, you know, a couple of people know, nobody else can do it. Yeah. So the main guy that came up with the whole thing for 3 nanometers in for its SMC, they had a fallout. And he moved to Korea to teach at the Korea Electronics biggest university there.

Starting point is 00:38:32 And somehow, within three years, Samsung, it's amazing. Wow. Oh, the lawsuits are amazing. Yeah. Like, there is a lot covered on YouTube. I can't remember. If I remember the name of the channel, there's a really good channel about this stuff. I think it's like Asian electronics or something like that.

Starting point is 00:38:50 There's this guy, man. Like, he dives. And it's just so good. This market is wild. Yeah. You know, in hardware, in like hardcore electronics, there's often just a few people to do the things. Facebook used to. have five guys the merge the code for 10,000 engineers. SQLite has three guys and is the most

Starting point is 00:39:07 used piece of open source software out there. Free BSD, very few people as well. It really is remarkable, like how much human technology depends on like one gray beard. Let's take a quick break and hear from today's sponsors. No, it's not your imagination. Risk and regulation are ramping up and customers now expect proof of security just to do business. That's why Vantza is a game changer. VANTA automates your compliance process and brings compliance, risk, and customer trust together on one AI-powered platform. So whether you're prepping for a stock two or running an enterprise GRC program, VANTA keeps you secure and keeps your deals moving.

Starting point is 00:39:47 Instead of chasing spreadsheets and screenshots, VANTA gives you continuous automation across more than 35 security and privacy frameworks. Companies like Ramp and Ryder spend 82% less time on audits with Vantta. That's not just faster compliance, it's more time for growth. If I were running a startup or scaling a team today, this is exactly the type of platform I'd want in place. Get started at vanta.com slash billionaires. That's vanta.com slash billionaires.

Starting point is 00:40:18 Ever wanted to explore the world of online trading, but haven't dared try? The futures market is more active now than ever before, and plus 500 futures is the perfect place to start. Plus 500 gives you access to a wide range of instruments, the S&P 500, Nasdaq, Bitcoin, gas, and much more. Explore equity indices, energy, metals, 4X, crypto, and beyond. With a simple and intuitive platform, you can trade from anywhere, right from your phone. Deposit with a minimum of $100 and experience the fast, accessible futures trading you've been waiting for. See a trading opportunity, you'll be able to trade it in just two clicks once your account is open. Not sure if you're ready, not a problem. Plus 500 gives you an unlimited, risk-free demo account

Starting point is 00:41:07 with charts and analytic tools for you to practice on. With over 20 years of experience, Plus 500 is your gateway to the markets. Visit Plus500.com to learn more. Trading in futures involves risk of loss and is not suitable for everyone. Not all applicants will qualify. Plus 500, it's trading with a plus. Billion dollar investors don't typically park their cash in high-yield savings accounts. Instead, they often use one of the premier passive income strategies for institutional investors, private credit. Now, the same passive income strategy is available to investors of all sizes, thanks to the

Starting point is 00:41:47 Fundrise income fund, which has more than $600 million invested in a 7.97% distribution rate, With traditional savings yields falling, it's no wonder private credit has grown to be a trillion-dollar asset class in the last few years. Visit fundrise.com slash WSB to invest in the Fundrise Income Fund in just minutes. The fund's total return in 2025 was 8%, and the average annual total return since inception is 7.8%. Past performance does not guarantee future results, current distribution rate as of 1231, 2025. Carefully consider the investment material before investing, including objectives, risks, charges, and expenses.

Starting point is 00:42:29 This and other information can be found in the income funds prospectus at fundrise.com slash income. This is a paid advertisement. All right. Back to the show. When you go from, and I'm just ignorant on this, if you would go from like a six nanometer chip down to a three nanometer chip, is it basically twice as fast? Does the math make it?

Starting point is 00:42:49 No, it's not quite like that. Okay. The speed is, it gets very complicated on like how to calculate the speed, but it's going to be a lot faster and a lot better. Yeah. And also more efficient, right? Because we moved from making super big, super fast chips to making way more efficient chips. Arm did a lot of that work. So it's a little ball.

Starting point is 00:43:12 It might be a little bulkier, but it's still using the same energy, which is really kind of the main variable that people are concerned with because that's the cost to run. the machine. Yeah. I mean, just just think about like your iPhone. I mean, like it's insane. Like these things are some of the most modern machines that exist and the power lasts more than a day, like more than all day now. Like that's not true. Yeah, I got you. I got you. It's a big deal. Hey, let's go to this Noster question. So I'm going to frame this in a maybe a strange way, but bear with me. So you recently wrote an opinionated guide to saunas and put it in a long form. on Primal. And here's what I find so fascinating. This is just as elegant and just as useful as medium. And now we're doing it on Noster. And Primal has just rolled out. Kudos to Milian and his team

Starting point is 00:44:03 because they rolled this out for people to make long form posts. You don't even have to have a medium account and like get the publication. I don't really know what the publication process is for medium. But now you can just do it on Noster. Anybody can do this. And so I'm reading this, article. Let me just read the start of this article. This is hilarious. And this is NVK talking about his opinionated guide to sauna. One of my fondest memories from this time was the Scottish bath, which involves standing against the sauna wall, execution style, so that somebody can spray you with freezing water from a high pressure hose. I've never heard of this outside of South American and can attest to whether it has any real Scottish origin. This is just the start of this

Starting point is 00:44:48 brilliant article that you posted on Noster. The reason I'm bringing this up is it appears that Noster just isn't a Twitter clone. It appears that it's becoming something way bigger than that. I've talked about this with some previous guests. I mean, we have like decentralized GitHub via Noster. We have now long form posts like Medium. We have AI agents like we were just talking about that haven't been fully integrated into clients, but I'm sure it's coming soon. We have spaces level. It seems almost like the Noster protocol is this skeleton and all of these other functions that exist today. And I even heard of some other protocol that's going to try to mimic YouTube now that's going to be like almost pointed to from Noster. Explain to the audience, in your

Starting point is 00:45:39 opinion, what is Noster? Because it seems like it might be something way bigger or more than just replacement to Twitter, I guess. Okay. So I'm going to have to shield the Noster Rising mini series that we're doing. Okay. Because that's the, it's an audio thing, Noster Rising.com. Because I'm going to take a quote from Fiajaf from there. Okay.

Starting point is 00:46:02 And the most, like the funniest and best way of describing Nostras, Nostor is just a bunch of websites. Classic Fiaja. Fiajaf for the people that know is the inventor of Noster, right? He's our Satoshi. And what's fascinating about this view, this sort of mental model is that Nasser is decentralized in the sense of that anyone can put a store and relay server, right? It's called the relays.

Starting point is 00:46:26 And what does this do is imagine, for example, if Twitter was a relay, Facebook was another relay, and I don't know, LinkedIn was another relay, right? Now imagine if you could see the replies between the two. So you know how companies post in all the social media is at the same time? Yeah. It's a pain they asked to go look. Imagine if they're all interconnected. And one has maybe long form posts, right?

Starting point is 00:46:50 Because it's your company blog on Medium. And the other one is more like Twitter, right? Like on Primal. And then imagine if when you respond to a Medium article, it shows up on your feed. Twitter feed, yeah. On the public square. Now imagine that when you have your post, your company post, you can now go and post on the public square and say, I have a new post about my company.

Starting point is 00:47:13 Right. And all these things are fully like native intertwined. Yeah. Fully interoperable. Exactly. You gain the social graph. You now own your identity because it's a key pair, just like Bitcoin. So nobody can take that identity from you.

Starting point is 00:47:29 Like if somebody turns off your Mastodon, somebody turns off your Twitter, your identity is gone. Right? Like you're building on somebody else's land. And on Osterner, nobody can take that away from you. They can disconnect you from their server and say, I don't want you on my server. but your identity is yours. And you probably have that content, like relayed by other relays, right? So you can't really be canceled in the universe.

Starting point is 00:47:54 You could be canceled in that specific server, which is also like they should be free to do that. Right. I mean, you maybe don't want to host some kinds of content, right? So it gives us this freedom where you have these relationships that are many to many, many to one, one to many, private to public, public to private. You can run a private relay. But then because it's native protocol with the public one, you can, for example, say, let's say you build some tooling for marketing, right? Like Hootsweet or something like that.

Starting point is 00:48:21 You can have your whole company internally on a relay that posts internally and comments on public posts internally without the outside seeing it. Oh, interesting. And you can have all these things intertwined. Right. So for example, when you go highlighter.com and you highlight a quote from my Sona article, which I mean, I think the quote I pulled was pretty. good. It was grateful. And then, very true, by the way, very painful. And you post that quote on the Twitter-like use of Noster on Primal. It's the same protocol. Yeah. And the data lives on both. So in short, you think this is a long game that's just going to continue similar to Bitcoin.

Starting point is 00:49:01 It's just going to be this grind of adoption that 10 years from now, people are going to look back and be like, oh, yeah, well, it was inevitable. But in the moment, especially early in the moment, And you're kind of looking at him like, what is this? I don't know if it's going to catch on. And then just more, more people just kind of continue to trickle in and use it. And because it's completely interoperable, the usefulness of it just is like slowly getting better and better and better. And I think anybody that was on Noster a year ago to now can see that the performance is like night

Starting point is 00:49:31 and day difference. And I can only imagine what another year or three years is going to look like as to what's happening on there. Would you agree with that? Yeah. So like, I remember Bitcoin, the very thing. very early days. It's like, what is this? Yeah. And why does it fix everything? Yeah. It's like you sound like an insane person. Right. Like when I still sound like insane people to

Starting point is 00:49:51 most. Yeah. I mean, you know, that's, you know, you can only help people so far. Yeah. But imagine electricity in the early days. It's like, oh, and people are going to use this for everything. Technology doesn't go like that. Right. Like, yeah. It takes time. It takes, it takes, it takes both education and also tooling and you acts to improve and go side to side. And even the people who are in it at the beginning don't necessarily know how things are going to play out or like what is going to be used for. So like it's hard for the skeptics to hear that this thing sort of fixes all these things, right? Because they heard that 10 times already from all their technologies that fail. And it's hard for the people step in it to understand that people are not going to understand it.

Starting point is 00:50:28 Yeah. That's the norm, right? And for the people being the Bitcoin grind since the early days, like, you've already been through this once. So like, it's okay. We're going to make it. And with the Noster stuff, you're going to start seeing the light switches. go on faster than with Bitcoin because it's social, right? It's a social graph. The way I like to think about it's like for Bitcoin, it's the medium of exchange, but you can't do that if you can't coordinate and communicate. Amen. Right? Like you have to discuss bid and ask. That's right. And where does the government go, not just the government, like every like monopoly, try to corner a market or whatever. They try to get in between the bid and ask. Yes. And try to take a cut or block completely, right?

Starting point is 00:51:08 Yeah. Yeah. So when Noster does, to us is like, it help us have like a freedom of communication sort of system where you can have the bid and ask if that is for the marketplace of ideas or if that's for an actual like trade of a commodity or whatever, right? So when you're doing this, when you have this Noster thing going on, you have this, the switches going on. And for example, fountain.fm. Right. Like they were a 2.0 thing where for the people that don't know, it's essentially like podcasting, but they embed a public key there so that you can now do payments, you can send messages to the host. It's super cool, right?

Starting point is 00:51:44 Like, it's taking the podcast industry by storm kind of thing. It's just not very common on the main bigger pods yet, but it's huge. Our pod does that. And it's super cool to get those apps with Zaps is how you send Bitcoin on Noster. Yeah. And now Fountain.fm went and integrated Noster. So they essentially switched their Waldgarden content system for people to interact with the podcast to Noster.

Starting point is 00:52:06 And all of the sudden, every single interaction that people have, with my pot goes and shows up to the whole social graph. Like everything is now in the public square automagically. So like I can just re-quote, automagically. So you can just re-quote when somebody commented or asked a question and it shows on my social graph, right? Shows on the public square. Then imagine what that's done to other things.

Starting point is 00:52:31 Yeah. Right? Like medium got kicked out of Twitter. Substack got kicked out of Twitter. Right. So you can't get kicked out of Noster. Right. So when people come over and they.

Starting point is 00:52:39 start doing those activities there, you're going to have the same thing. So I'm looking forward to medium or substack to convert their system into Nostr, and then boom, it's open. Yeah. And then you got somebody that's incentivized to run more clients, right? If you're medium and you come to Noster, you want to run your own client. If Twitter comes to Noster, talk us through this. When you think Twitter would have an incentive to take all their existing server racks, effectively turn them into relays, they're hosting all that same data that they've already been hosting and they're basically using the Noster Protocol and it's like they're looking at the government and they're saying, hey, we're just one of many running a relay. We can't

Starting point is 00:53:15 control their speech. We're just backing up what's being said publicly. I mean, it just really kind of shifts the whole dynamic of the company and the incentives and what their actual dependencies are for backing up, you know, communications. Think about it this way. So close social media or any type, it's more like miles, credit card miles, useless. Yeah. Yeah. Like, it's cool. It's cool. know if you happen to be the person that can rack up a million miles, right? But like, most people just see miles is completely useless. Yeah. Why? Because it's a closed system. Yeah. Bitcoin, why is Bitcoin so valuable? Because it's an open system. Yeah. So Twitter is just Miles right now. Now, if they open their system, those notes would go say, say Facebook does it too. Now, those notes cross

Starting point is 00:53:59 the barrier. Yeah. And they can interact with each other in a way that creates a lot more value for everybody. It grows the pie. Yeah. And there is a lot of fear in growing the pie, right? Because it's like, this is my data. This is my server. Yeah. But it comes with a lot of risk and that comes with a lot of cost. Like, I bet they could.

Starting point is 00:54:18 Do you think it's based on GROC and their AI training that they're really hesitant to open that up because now Facebook can tap into all that data? Is that really kind of the. I think so in the early days of social media, right? Like the way most of the social news sites hacked was like, oh, you know, press here, import your contacts and we're going to contact them. Remember, Friendster, Orcut, all that stuff. And then, like, they would close it down.

Starting point is 00:54:40 And then the previous company that had it open goes like, screw you. Like, you're just stealing my users. So, but this is the conceptual problem is that my users, we are theirs. Like, our data and our identity are owned by them. We need to move on from this to like, it is my identity. It is my content and information wants to be free. And you can't stop this train as a Linne Alden would. put it. It's just going to take the awakening of the user to reclaim their data and their security

Starting point is 00:55:13 and all of these things. They can't go back. Once you feel it, once you try it, it's very difficult to go back. You know, it's too early, right? There's a lot to be built. There's a lot more people to bring over. But he already has like this, the vibe is right. The technology is already better. You probably haven't created a Twitter account in like many years like me, right? And I was trying to create a Twitter account for a product. And what do you mean you need a phone number to create it? And then I try to put a phone number and then it's like, oh, no, we already have that one for another account.

Starting point is 00:55:42 It's like, I literally cannot create an account on Twitter. This is insane. It feels like a bank. You have to go to the bank, show ID and do all that, right? Where on Bitcoin, you download an app and you're ready to hold a billion dollars in it, right? The same is for Noster. You just go in any client and you create your key pair and boom, you're on Noster.

Starting point is 00:55:59 And if the client doesn't like you or you don't like the client anymore, you just take your private key and you put in another client and boom, all your stuff is there again. You can compete with this kind of efficiency. It's very difficult. And all we need now is time. Yeah. MVP, I could literally talk to you all day.

Starting point is 00:56:15 I really appreciate your time. Anytime we have a chance to talk. I'd learn so much. You're just so astute when it comes to the technical hardware, software, everything related to Bitcoin, all of it, all the above. If you want to give people a handoff, obviously the cold card, anything else that you want to, we were talking about Unleashed. Chat earlier.

Starting point is 00:56:33 anything else you want to highlight. For people that don't know you, you're very active on Twitter. You're very active on Noster and you reply to people's questions very graciously with your time, sometimes very ironically with your responses. But is there anything else you want to hand people off to? No, I mean, like, listen, just try Noster. It's happening and nobody stops that train. And then on the Bitcoin side, just get off the exchanges, get a harder wallet.

Starting point is 00:56:59 Any harder wallet is better than no harder wallet. and get your coins off to exchanges. And let's be free people. Yeah, amen to that. All right, we'll have links to all this in the show notes. NVK, thank you so much for your time. Always a pleasure having you. Thanks for having me, sir.

Starting point is 00:57:13 Thank you for listening to TIP. Make sure to follow Bitcoin Fundamentals on your favorite podcast app and never miss out on episodes. To access our show notes, transcripts or courses, go to theinvestorspodcast.com. This show is for entertainment purposes only. before making any decision consult a professional. This show is copyrighted by the Investors Podcast Network.

Starting point is 00:57:37 Written permission must be granted before syndication or rebroadcasting.

We Study Billionaires - The Investor’s Podcast Network - BTC207: Bitcoin, AI, Nostr, and Hardware Manufacturing w/ NVK (Bitcoin Podcast)

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.