BTC Sessions - The AI Dangers Bitcoiners Can’t Ignore — And What to Do About It | Odell & Hill

Episode Date: July 14, 2026

Mentor Sessions Ep. 082: Matt Odell and Matt Hill discuss AI model security risks, self-hosting infrastructure, open source vs frontier models, and Bitcoin's role in digital sovereignty in 2026.AI's most powerful models are being quietly locked away by governments — and the cheap API access you're building your business on is a subsidized trap with an expiration date. Matt Odell and Matt Hill join BTC Sessions to unpack what's really happening at the intersection of AI, open source sovereignty, and Bitcoin freedom tech in 2026.You'll learn why the frontier model subsidy era is ending and who gets wrecked when prices normalize, how Start9 runs an entire company on a $200/month plan (and why that won't last), what Odell thinks about the deepfake phishing attack that nearly fooled Matt Hill live on a Zoom call, and why the open source AI arms race may be humanity's last real shot at digital independence. You'll also see how the Start9 router and self-hosted security camera are the trojan horses that bring sovereignty tech to people who've never heard of Bitcoin.⏱️ Timestamps:0:00 - Intro1:32 - AI model landscape and security risks2:15 - Start9 runs on $200 monthly Claude plan3:16 - AI subsidy trap and future pricing shock4:09 - Local inference versus frontier model tradeoffs7:32 - Odell on open source versus proprietary models9:28 - Why Anthropic and OpenAI withheld top models11:23 - Open source software and why control backfires13:36 - China's open source strategy in AI geopolitics18:47 - AI as intelligence of mass destruction risks21:26 - Regulation landscape in Canada UK and Australia27:01 - Operational security and AI-powered phishing threats27:44 - Deepfake Zoom attack that fooled Odell32:40 - Public key cryptography and verifiable identity39:02 - Signal messenger sovereignty case study43:55 - Do Bitcoiners lead digital sovereignty efforts45:05 - Relentless freedom fighters win revolutions55:42 - Drivers of Start9 adoption beyond Bitcoin nodes57:37 - AI as second reason for sovereign servers59:55 - Start9 sold Bitcoin to stockpile RAM1:06:16 - Start9 router for household sovereignty1:09:13 - Building the freedom router from first principles1:14:44 - AI as new interface replacing GUI era1:20:06 - Odell on open source compounding in Bitcoin1:25:56 - Tools to try PBC Maple Hermes LM Studio• Sovereign Sessions — AI, Privacy, and Bitcoin education: http://youtube.com/@SovereignSessions?sub_confirmation=1🔗 Links & Resources:→ Matt Hill: https://x.com/_MattHill_→ Start9 (sovereign server, router, marketplace): https://start9.com→ Matt Odell: https://x.com/ODELLXYZ→ OpenSats: https://opensats.org/💡BOOK Private Sessions with Nathan, Benn and the BTC Mentor Team: Master self-custody, hardware, multisig, Lightning, privacy, and more. 👉 Visit btcmentor.io 📌 Previous Episodes: Prof St Onge & Larry Lepard → https://youtu.be/cS4QXZr5qvc⚡ POWERED by Abundant Mines: Fully managed Bitcoin mining. Learn more at https://qrco.de/bgYKPB🔒 Lockdown your Bitcoin with the BEST gear on the market from Coinkite. Get the 5% Off the COLDCARD visit: https://qrco.de/bfiDBVFollow Us on X:• BTC Sessions: @BTCsessions• Nathan: @theBTCmentor#Bitcoin #BTC #BTCSessions #MattOdell #MattHill #Start9 #SelfHosting #SelfHostedAI #LocalAI #AISovereignty #FreedomTech #DigitalSovereignty #BitcoinPrivacy #OpenSourceAI #SovereignComputing #BitcoinNode #AIDangers #DecentralizedAI #PrivacyTech #SovereignTech #BitcoinAndAI #Cypherpunk #LocalInference #AIDeepfake #Decentralization #AIPrivacy #OpenSource

I'm curious what you read on the AI model landscape is right now. I had a call the other day with Ryan Gentry from Lightning Labs. It was not Ryan Gentry. I'm talking about a video call. This was a targeted fishing attack at me. And it was so good that I made it all the way to the Zoom video call before I realized something was wrong. I have a radar for this shit. There's real danger here.

We're getting away with murder right now. We're spending $200 a month and building what would otherwise cost us 10. employees. It's kind of like a drug dealer setting you up with very cheap drugs in the beginning, and then you get hooked and dependent on it. And a lot of people will probably get wrecked in that transitionary period. The top model from Anthropic is not publicly available, and now the top model from Open AI is not available, which is scary. I think that's a dark route for our country to go in. Freedom money does sweet, all for me if I'm not able to also be private or keep my own data. Bitcoin is the hero of the army. Bitcoin is Achilles. Bitcoin is Hector. Without which, none of it

matters. You do not need a majority to win a revolution. All you need is a relentless, small number of courageous freedom fighters. All right, good morning, gentlemen. Thank you so much for joining me today. Very excited to have this conversation. We got a lot to get into just Bitcoin and AI and privacy and regulation and everything that's going on right now. But I want to start by kind of getting a little bit of the lay of the land. So, Matt, the last time we spoke, you were talking about how incredible Opus 4.6 was from a developer side, from the developer perspective. I'm curious what you read on the AI-I model landscape is right now, both from a software engineering, from a positive, from a constructive lens, as well as possible threat vectors or security risk that

we should be aware of. Well, I don't think anything qualitatively has changed since we last spoke. Obviously, the models have gotten better. The harnesses have gotten better. We at Start 9 have recognized, I mean, order of magnitude productivity increases since, you know, the Opus 4-6, 4-5 drop. We've put an enormous amount of effort into automating all the things, right? You know, dedicated agents for dedicated tasks, interacting with each other. I at this point, you know, manage a fleet of AI. Most of my days spent with, you know, five different cloud code sessions and, and that's it.

I mean, we're building that five times the speed that we could have pre this era. And I'm having a hard time imagining my life without it at this point. I was talking with Aiden about that the other day. We were like, what if, you know, what if for some reason this was just a dream and it disappeared? Like, what would we do? and I maybe 30% joking, 70% truthful was just like

I'd quit. I'm not going back. I'm never going back. And it's a very dangerous kind of headspace to be in in a sense because what it means is that they're undercharging. What it means is that like the current models

and computer being heavily subsidized and absent, you know, ongoing aggressive competition, especially from the open source sector, you know, this will not be a cost savings in the long run, they will fleece us for all we're worth. So it's really, really important to keep the open models pushing and hopeful that one company or, you know, an oligarchy of a few companies backed by government power don't monopolize this thing and gouge the rest of the world for it because it is an essential

technology that everyone needs at this point. Again, it's hard to imagine going back. It's like electricity. We're not going back. We can't go back. I'm curious before I come to Odell here, if you're starting to do any more kind of local inference as well. We talked a lot about how at the time, it was just so powerful that you kind of hold your nose and if we're handing over some information. In the meantime, it's too great of a tool not to. I'm curious if you started push to using some models for simpler tasks and running them locally or what you're kind of thinking from the implications of what you're sharing from the start nine work. Yeah. So we we take a multi, multi-pronged approach, right? You build in the moment and you plan for the future, a combination of

expense and investment. So we are investing a lot into self-hosting, obviously, right? But for example, we still use GitHub as our primary, you know, Git layer. And that is a disaster waiting to happen. And we are aggressively making moves towards self-hosting all of our repositories and using GitHub as a mirror at best. But the reason I mentioned that is because it's like we're seven years in to being a company that promotes and builds self-hosting tools and we're still using GitHub. So what that says is that we are not dogmatic about the process, right? It's not like you need to get where you want to be instantly, right? It's a process. And you identify the things that are most important first, right, the largest threat models and attack vectors first and work your way down to the

lesser important, also eliminating costs first, right? So the things that hurt the most of the things that you fix first, wherever the biggest fire is. And so we're taking the same approach to AI, of course. And so there are obvious, huge outsized benefits and competitive advantages to using the frontier models on the anthropic servers and not using them. I understand if someone wants to take a principled stance on that and just is like, I'm not using them, but they are in a way signing their own death certificate. You've got to compete. You have to stay up. But at the same time, you don't want to become dependent.

You don't want to set yourself up for a future attack vector. So we are building out local infrastructure as well. We have two DGX sparks that we use for company matters and smaller tasks. we have self-designed harnesses. So, you know, we built our own custom harness that we use that allows us to, you know, maybe don't tell them, but, you know, we run our entire company on a Claude Max plan of $200 a month. We don't use the API. We circumvented the restrictions that they put in place, possibly in violation of their terms of service.

To be honest, I don't really know because I haven't read them or handed them to a lawyer. But I know that they are that the way that we are using their max plan is not the way that they want us to be using it. We're basically pretending to be a human, but it's not. And so, yeah, I mean, we're getting away with murder right now. We're spending $200 a month and building, you know, what would otherwise cost us 10 employees. And we know we're not going to get away with that forever. So that custom harness has a fail safe and fallback already built such that not only can we fall back to AP. usage in an emergency if we need to, but we can also fall back to local inference if we need to as

well. But obviously that would come at a huge penalty of productivity. No 100%. Odell, I'm curious, your thoughts on all this. And in particular, even the open source versus the frontier models, both from a usability and kind of user control and also the open source implications of using the tools to build open source tech, if that makes sense. First and foremost, as the largest investor in Start 9, I'm constantly impressed. that match resourcefulness in saving money and running a lien operation. So that's fantastic to hear. I didn't realize they were doing that.

But I do think, I mean, he touched on a key piece here, right? Which is right now specifically with the proprietary frontier models, right? And specifically anthropic and open AI. We live in a world where they're heavily subsidized. And I think what's going to be really interesting is, you know, it's kind of like a drug dealer setting you up with very cheap drugs in the beginning, and then you get hooked and dependent on it. And a lot of people will probably get wrecked in that transitionary period. I think with that comes opportunity. I mean, a lot of people will be able to transition effectively and keep their token costs down and be able to use these things in a relatively free way. And those people will find success. But I think a lot of people are getting hooked kind of on this. this this cheap intelligence vaccine right and so when when you think about it from that perspective what do we want to see i mean first of all we want to see the open source models and the open

weight models specifically become more competitive with the frontier models they have not yet i mean i think there's a lot of hype out there influencers it's it's you get a lot of engagement on an ex if you say you know glm 52 is basically opus quality or fable they even go even further they're like oh, this is like basically Fable 5 and super cheap. That's not the case. It's not true. And so that's one piece. I think the other piece is, you know, Open AI just recently announced that they're coming

out with 5-6, which is their latest model, and that the government requested that they don't release it publicly yet. So that's the second time we've seen this. Now there's the top model from Anthropic is not publicly available, and now the top model from Open AI is not available, which is scary. I think that's a dark route for our country to go in if we want our country to be leading on the AI front. So that's a development that I'm watching closely.

And then I mean, the last note, Odell, I want to just double click on that. Why do you think that is? For me, I wonder, like it almost feels like if we're just creating some sort of hype, like they're so incredible, we can't release this, that movies so scary, people are fainting and having heart attacks. What do you think is the motivation behind keeping these behind a closed wall? I mean, look, it's a little bit too early, I think, to make a definitive call.

With Anthropic first, they kind of did it to themselves. They kept saying, like, this is basically a weapon of mass destruction. And then they also didn't, they weren't on good terms with the U.S. government. So, like, it kind of made sense both from the U.S. government point of view, like, let's fuck these people that are not playing ball with us. And it made sense from the anthropic side, which is like, oh, it's great marketing. We have this model that is so good, that is too dangerous for. you to have as the average person. So I didn't really, I didn't want to put too much mental bandwidth in that and I wanted to see how it developed. But now having it happened

to open AI at the same time, I think you're starting to see a little bit of a trend. And so it's something that I think we should be watching. I think this is a tendency with regulators with all new technology, right? They want to try and keep it in a box. They think like if they can control it, then they'll have more power and more influence. And usually it backfires. I mean, this is why open, open source software and open protocols are so important because nobody can control them. And they're viral in nature. But I mean, to that point, like, the thing that will be really interesting is I'm not sold, you know, I'm not sold on the idea of people buying hardware right now yet. Like, I think it obviously makes sense from like a freedom and sovereignty and privacy

point of view. I don't know if it necessarily makes sense from a cost point of view. And part of that is me coming from the Bitcoin lens is like, we haven't even really hit like the ASIC hardware revolution yet of models. Like I think, and hardware. Like I think a lot of people that aped into GPUs or early ASICs on Bitcoin ended up getting absolutely wrecked from a financial point of view. So I do caution, I think that if people are thinking about this, they should be very clear to themselves. Like, I'm not making this necessarily for a financial investment. I'm making this for a freedom investment. And that, to me, makes sense. But for the average person, especially on small business side, being able to just adapt on the fly to different APIs and not get

locked into like a vertically integrated Claude setup or vertically integrated code setup and being able to switch and adapt, I think is the key differentiations. that will set you up for success. O'Dell, would you say that my purchase of an M5 Macs with a 128 gigs of RAM was a local top then? I don't know. Not yet. It's not a local top yet because Apple just announced

they're raising prices, right? I think people who bought like Mac studios and decked out MacBooks like two months ago are right now sitting on a better deal than if they bought them today. But I do, I think it's important to think in longer than six-month timeline. And I question where we're going to be in 12 to 18 months and whether or not people are going to regret paying $10,000 from Mac Studio because they thought they were setting up their family for generational intelligence when the thing is already outdated.

One other thing I was going to ask, you were mentioning kind of the open versus closed models and kind of where we're at. I wonder if either of you actually could comment on what kind of threw me for a loop with all this stuff was. the, what I consider a counterintuitive tactics of the West leaning into the closed models and China leading into more of the open models. Why do you think that manifested? What do you think were the factors, sir? Speculating on, you know, international politics, information age, wartime strategy. You know, look, if you're lagging, you're going to try to poke at the beast any way that you can. I mean, undermining the West hegemony on frontier models in any way that you can. I mean, open source in a way does ruin business models, right? Certain business models.

And so China might not be trying to win. They might just be trying to make sure that we don't. Maybe, right? So, yeah, I mean, I will follow on a little bit of what Matt was saying as well. you know, we live in the information age, okay? A lot of critical infrastructure and personal security and survivability depend on digital systems. And when you have tools that are capable of causing massive disruption to digital systems, it would make sense from a government, right, the U.S. government's vantage point to treat them like weapons. like, you know, there exists in the world today, killer drones and biological weapons and chemical weapons and nuclear weapons. And we do not have access to those things, right?

In fact, pursuing access to those things is considered criminal or treasonous behavior. And I don't see why they would treat intelligence of mass destruction any differently. Now, I think they should, obviously, right? I think that it is a losing battle to try to contain this. It's not like mining uranium and building nuclear weapons that requires an enormous amount of real world, meet space, activity. But it's going to be a lot easier, I think, to compete or to make available superintelligence to smaller entities, even individuals, and that it can't be contained,

but they're going to try, right? it makes sense that they're going to try. We have, in my opinion, reached what I would perceive to be the sort of like guerrilla war stockpile era of weaponry right now. Right? Like if open models reached a level of efficacy of Opus 4.8 that could be run on reasonably priced commodity hardware, I think we have a fighting chance. You know what I mean? Like the guerrillas never have the same level of things. firepower as the state.

But you can win so long as you have some degree of firepower, right? Now, maybe that goes out the window of true AGI emerges and then it's just like, well, a million opus 4.8s, even in the guerrilla tactics are going to be useless. But I don't think we're going to get there. I think that like so long as we can get the open models to be run cheaply and easily with the same level of intelligence and power as Opus 4.8, I think we're going to be okay. I personally would prefer a lot more, of course. I played with Fable 5 for played with.

I built an enormous amount of things with Fable 5 in the four days that it was available. And it was an improvement on Opus 48 for sure. And I want it back. Sorry, I'd say, really, to that, it was that much of a step up even just with the Fable there. Yeah. I mean, look, there's a lot of, there's a lot of fog out there. Like, again, with the open models and people comparing, you know, GPT 5.5 to Opus 4Aid and Fable 5 and, you know, deep seek.

And I'm telling you, there's a degree of wine connoisseurie going on here. Like, there are certain people who can really tell the difference. And other people are like, hey, they both taste fine. Okay. And so I'm not sitting here saying that I am the authority or expert or connoisseur here, but I'm saying I probably have a better feel for quality and efficacy than many others do, just having a, you know, long engineering background. And again, I'm not talking about anything besides.

besides coding here, just coding, okay? Like, the ability to build digital products. Opus 48 is unparalleled, in my opinion. And Fable was a step function improvement on Opus 48th. So, I mean, they're just the leaders. Anthropic is just winning. The other piece that makes it really difficult is like the benchmarks don't tell the whole picture. Like, it's more, it's more nuanced than that.

It's almost like you kind of have to feel it, right? You have to use it and feel it. And a lot of these model providers, particularly the challenge or Chinese, these models are very focused on benchmark maxing. So you see benchmarks come out and you're like, oh, like, this is an amazing improvement. Then you actually use the tools and you realize, oh, the level of competency is completely separate. That's interesting. Odell, I'm curious. Matt mentioned the idea of intelligence of mass destruction. What is the risk that we actually might have here? What is it that we're concerned

about going forward? Because the one thing I was thinking about yesterday is like encryption works. So even from like a security standpoint, are we even really that vulnerable to these models as they improve? If we just have good encryption and good practices already in existence. Yeah, I don't know. Like, I mean, I think there's a lot of fear mongering out there. There always is with new technology, but also some of it is based in reality. I think if anything, there should be some concern over like this temporary transition period as you have these models. out in the wild and their pressure testing stuff.

But particularly with robust open source protocols, open source software and stuff, they'll get hardened as those vulnerabilities are found. I mean, we've seen multiple examples now of you run a frontier model and you basically say, like, check this shit for me. Like, did I make any mistakes? And it finds mistakes that have been sitting in open source software for a decade and nobody realized. And that's just a human condition.

Like these things can parse. information on a level that most humans, that humans just can't, right? They're really, really good at parsing through large data sets and testing a bunch of things. But I think after that, then we should be in a much, much stronger place than the risk is greatly reduced. And then I think ultimately you have to make a conclusion, right? With every great technology, there's a trade-off balance. Does it empower criminals, doesn't empower malicious people? Yes. But does it net net empower the average person and make society flourish? Yes. And you see that with Bitcoin. You see that with encryption. You see that with things like guns, right? Drones. The drone argument is going to be an

interesting corollary here, particularly because a lot of them are going to be running autonomous with local models. But do I think that the average farmer having access to a sophisticated drone system improves their lives and improves the quality of our food system and makes it more robust? Yes. Do I think governments are going to try and regulate the hell out of it because of like assassination risks and and asynchronous attacks situations where like you have no idea who sent a drone swarm at somebody? A hundred percent. But does it net net benefit to society? Yes. Is it going to be hard to stop? Pretty much impossible. And I think the jurisdictions that overly regulate these things are going to put their citizens and their economies at a disadvantage compared to the

jurisdictions that don't. No, fair enough, Ben, I'm curious your thoughts on the regulation landscape right now, because Canada's gotten a hell of a lot worse. The UK and Australia already sucks. You've been playing around with these tools. If they start to clamp down on AI usage and say our neck of the woods, we'll be able to still continue on the open source stuff that's available that's already out there today. Yes, this is what I don't know, because it's not like I've had the experience of living

in a fully draconian hellhole. We came close a few years back. You're pretty damn close. But yeah, like, you know, you know, government shuts it on the internet. It's like, well, okay, how many people have already preloaded mesh network software? Like, it's just how far do they go, I suppose? You know, I'm super interested in striking that balance. I know, you know, Matt, you're talking about how much better.

And Odell as well as saying how much better the frontier models are versus some of the best open models. But I wonder for a person that's trying to future proof. And Matt, you were saying that you guys are trying to set up infrastructure for when shit hits the fan and prices hit reality. is there a kind of balance to be struck right now where somebody can begin to etch out, this is where I go to the frontier models, this is where I go to the local open or cloud access but private through something like Maple or PPQ, like is there a balance to be struck there? And further to that point, is that also?

a way of actually resolving some of the eventual price shock when it's the fan can people more efficiently choose not only what they're using cloud-wise but locally so I know it's a two-prong question to either of you but I'll let you type in there because I don't know the answer we are absolutely already at the stage where you can begin thinking about your usage of AI and agentic systems as being a triaged hierarchical, you know, approach. Like, there is no reason to send your daily, you know, Cron that checks some emails and summarizes what's in there for you to Opus 48. It's like using a Ferrari to drive down the street.

You can do it, but you don't need to. There are open source, smaller, and even fine-tuned models that you could do to, to a accomplish the same thing, basically for free and totally autonomously. But for your most complex, largest coding tasks, you simply cannot build a coherent quality software system using local open models right now. It hasn't caught up yet. I think we will get there. But we've never taken the stance. I've never taken the stance that it's all or nothing. Any degree of growing your own food is good for the restaurant industry and for your own health, right? You don't need to make a commitment to never eating out again just because restaurants, you know, put crap in the food and because

they could poison you and because it's more expensive. There's real benefits to going out to eat. People enjoy it, right? But there's also real benefits to growing your own food and cooking at home. And so, you know, you find the right balance for yourself, but if you're not thinking about or don't know how to grow your own food, you're setting yourself up for extortion, right? Like you're a dependent at that point. You need to independence is not doing everything for yourself. It's the ability to do things for yourself if needed, right? And so that's what we're looking at. And we always keep our eye on this as we say, okay, well, what's the worst that could happen right now?

it's, we get a scrape on our knee, well, then we just keep doing it. But if what's the worst that could happen right now is a gunshot to the head, we're going to pivot really fast and focus our energies there. Again, where is the biggest threat? Where is the attack vector? What is the highest cost? Focus there and work your way down the ladder. I did want to follow up on earlier, just for a second here when we were talking about, you know, intelligence of mass destruction. And you had asked Nathan about, you know, like application security and increasing. and I agree, and Matt's response was largely in that vertical as well. And while there are vulnerabilities out there, I do not believe that those are a meaningful portion of the danger,

right? Even from the state's perspective, like those things will get hardened. Anthropic, I believe the story when they were like, hey, we've created a committee with a few other companies to kind of like get mythos out there and harden their systems before we release it to the public. I believe that they actually did that. And that those systems, are probably getting patched and hardened, and that might be part of the reason that it was temporarily pulled is to allow the industry time to adjust to this new reality. But that's not the important vertical. I think that from the state's perspective, again, I'll put on my status hat here.

There's really, well, one of them is a developer hat, and the second one is a status hat. So I'll do the developer one first, is operational security. Okay. The fishing attempts are going to get really, really good. All right. So to pull off a good fishing attempt today, you need to be a developer. You need to basically spoof a website and know how to set up emails and know how to, like it was hard to do good fishing attempts, which is why most of them suck.

It's because it's some guy who doesn't know what he's doing trying to fish some idiot who doesn't know how to use a computer. But these things are going to get really good, really fast now that those same people have access to frontier models that can build them really quality. you know, hacking, social engineering tools. So application security has never really been the problem. It's operational security. And these things are absolutely weapons when it comes to operational security. You can create bots. You can do fake video. I had a call the other day with Ryan Gentry from Lightning Labs. It was not Ryan Gentry. I'm talking about a video call. This was a targeted

fishing attack at me, and it was so good that I made it all the way to the Zoom video call before I realized something was wrong. And when the Zoom call started to crash on me, and the error message came up as like, hey, your system dependencies are out of date. You can't run Zoom even in the browser. You need to update this low-level system dependency. Here's the command to paste in. I started reading the commands. I was like, I'm being attacked right now. now. And it got all the way to the point. I'm like, I have a radar for this shit. This was so good. They spoofed the Zoom website and did fake AI generated video of Ryan Gentry. I text to Ryan, I was like, dude, this is sophisticated. Somebody's really being you on the internet right now. And so that is more of my concern around the security aspect of full access to these advanced frontier models. Again, I'm still for it. I'm just saying that there's real danger here. But the second one and the bigger one that the state will eventually argue for here is the social danger, right? I feel like this quote right out of Atlas shrugged.

It's like this doesn't pose a physical danger. It poses a social danger, right? Like everyone having access to this, everyone is a lawyer now, everyone is a doctor now. Everyone can do everything. I mean, we're talking about the eradication of society as we know it. And I've talked about this a lot. I mean, we've all talked about this a lot. Like what happens when half of the economy and job market gets wiped out by this?

Long run, I think it's great. We all just get rich, right? It's prosperity and freedom for all. But there's a real transition here where people don't understand this. They can't accept this. They are dependent on the old system, right? If you're like Morpheus from the matrix here, it's like anyone still plugged in is a potential enemy because they will fight with their teeth to beg.

the politicians and status to fight against this technology. The people themselves who would benefit the most from this are going to be its largest opponents. And so that's the other dangers. They're just going to be like, hey, this technology is too disruptive. We have to, gradually transition. We have to shepherd humans through this period. We can't just unleash technology and expect them to fend for themselves. They'll eat each other. This was the thesis, the essential thesis, of Klaus Schwab's fourth Industrial Revolution. That was the basis of it, which is like, hey, these massive technologies, amazing, but man, oh man, people cannot handle this. We have to shepherd them through this.

And that is the purpose of the World Economic Forum, according to that book. That's how they think. When I first got into Bitcoin, I was overwhelmed. The jargon, the security risks, the fear that one mistake could cost everything. I remember staring at my screen and wondering, are my keys safe? Did I do this right? That experience is why I started BTC session. For over a decade, this channel has helped millions of people like you learn how to use and secure Bitcoin.

But I realized something. For many people, videos aren't enough. Everyone learns differently. Some need to ask questions in real time as an expert walks through their setup, their goals, and their threat model. And certain things like advanced cold storage, inheritance planning, privacy, and node or mining setups often can't be fully solved by watching. another tutorial. So I built BTC mentor. I recruited the best Bitcoin educators on the planet to work with you one-on-one. Real experts, real answers, personalized hands-on guidance, tailored to your exact situation. Whether you're brand new or building a complex setup,

we meet you where you're at and walk with you step by step. By the end, you don't just hope your Bitcoin is safe. You know it is. If you're ready for that level of confidence, then head to BTC Mentor.io and book a call with us today. Ben? Yeah, I just want to jump back to that insane attack on you. I'm curious, and maybe I'll toss it to Odell here, your opinion here. Do public-private key pairs become a necessity for individuals, for people to know that they are indeed talking?

to or seeing posts from people that they think it is. Is that a thing that the public eventually has to grapple with and figure out? Look, I don't think, I don't know if there's an easy solution here. I don't know if there's an easy mitigation. There's two different pieces, and it's both on the open source AI access stuff and on this question, which is, okay, yes, public, key, private. key cryptography helps here, like verifiable identities, webs of trust, people actually verifying who they're communicating with and their communications. But what percentage of the population

is going to use that? If you make it too simple so that the friction is less, then it can probably get gamed as well, right? Because it almost requires a level of personal responsibility. And why I say that on the open source AI stuff, too, to your previous question is, is like if the government, if your government says that it's a criminal act to use anything but like Anthropics Cucked model that only tells you certain things, 98% of people just won't break the law, right? It doesn't even matter if they can force it or not. They just won't go around that.

So what do I think is actually going to happen? I think we're going to see this use as an excuse for the spread of KYC, centralized trusted third parties that tell you you're speaking to the correct person. that's where probably the majority of society will go. They will basically trust Elon or Zuckerberg or Dureov of Telegram to be the authority on whether or not someone's real or someone's not and whether or not the information they're getting is truthful. And it will be used as actually another vessel of increased control on people. Interesting, Matt, I'm curious.

Your thoughts on that? I completely agree. I think that's ultimately, this is our lot, right? people I had to learn the very internally kind of devastatingly hard way over the last five years I would say that we can't fix people

all right you can make you can make the systems as secure simple reliable performant as possible but you cannot make people adopt them.

So myself and at Start 9, our plan and our hope here is we build the best possible systems, uncompromising, open source, secure, simple to use, easy to understand systems. We just maximize that, right? We build the best possible systems that we can. We ourselves and also in support of others do the best that we can to educate. as many people as we can on the availability of these systems. We highlight the dangers of centralized information systems, you know, and dangers and costs, right? So invasion of privacy. I talk about these a lot, invasion of privacy, censorship, insecurity of just, you know, aggregating data,

consolidating data onto single servers, and overall economic cost because these things are inherently custodial intermediaries and they need to get paid. And so it's always going to be more expensive. And so we educate on the problems and then we provide solutions. We say, okay, if you are sick of that and you do understand what we're saying or what you yourself have observed, here's an alternative for you. And so it's been kind of interesting to see both results of that, right? Like, in a way, I've been devastated in recognition that the vast majority of the population will just will never do it. They just, whether it's an education issue, a cultural issue, or a human condition, there is just this striving by what it seems to be a huge mass of the

population to be dependence, to be subservient, right? It's almost like, that's what, not everyone wants to be free. It is like a perpetual childlike. It is like a perpetual childlike. existence. They want to be cared for, protected, controlled by somebody that they perceive to be good or powerful. And we can't do anything about that except keep trying. But I've also been impressed with, from a nominal perspective, just the sheer number of people that have arrived at our door and in our community and been like they did get it and they do see it and they want it and they're willing to fight for it. And these are people of all ages, right? This isn't just younger generations.

I mean, we get a wide variety of backgrounds that see this alternative and want to want it for themselves and for their families and for their businesses. And I think, I hope that we have actually reached an inflection point here, a potential multi-generational inflection point where we have a chance actually. like mass adoption, right? I think this thing could snowball, not overnight, but I'm optimistic that in one or two generations, the dominant form of computing on planet Earth could absolutely be open source, self-hosted, fully self-ferifying. Some personal responsibility is unavoidable, but that we could get there. But we cannot get there with the current population. I genuinely believe that you need a generation or two to fundamentally change behavior on a global level, right? Marx was very famous for citing, you know, give me two generations of children and I'll give

you a nation of communists. And I think we're in a similar position now. We honestly, people just don't, like, the internet is new. Computing is new. I mean, there's a sizable portion of the population alive today that grew up without computers at all in any way, shape, or form. And so the way we did it, the way we dove into this new technology as a species was just like naive. It was super naive. And I don't think it's a permanent state of affairs.

I think we will recognize the naivete and the mistakes made and fix them. But there's going to be an enormous amount of pain between now and then. Adele, I'm curious your thoughts. And then, Ben, yeah, I mean, the optimistic take and just a good case study that's actually out there and is not theoretical, I think, is Signal Messenger, right? And so Signal is centralized. They control the servers. But it's a privacy-focused service, right?

You want to have private chats with family and friends and secure, private and security. they're focused while trying to make it user-friendly. Right? And so as a result, they made multiple trade-offs. First, is they require phone numbers. It's a spam mitigation technique, and it's also a network effect bootstrapping technique. But they do use public-key-private key pairs, and most people don't see them. They're behind, they're behind the surface. They're below the surface. Now, public-key, private-key pairs are difficult for people, so they hide it. And then second-in-law, when there's a key transition, they just do it automatically. So if you've used a signal chat,

You might see like your key with BTC sessions has rotated and it gives you a little notification. Now, if you're someone who is is more technically aware, that might be a red flag to you. It's like, oh, someone else has joined the chat now potentially. I got to confirm out of band that it's actually been. But the average person is just going to look right past that. They're going to see the key change happen. And they're going to be, oh, I got a new phone, whatever. I don't really care about it.

That happens all the time. And they'll just proceed with the fishing attack, right? why do I bring this up? Because they had to make all those tradeoffs to hit 100 million users. Now, Telegram, which is none of those things, has 2 billion users, right? WhatsApp has, you know, 4 billion users or something like that. Just to show the scale there of what we're up again. So the optimistic take and both the negative or like the pragmatic realization of who are going to use these tools in the short term is signal. It's like, okay, well, 100 million people is a lot of people. But in the grand scheme of things,

it's a small percentage of the global population. And so I think it's important to build the tools right now for basically the rider dies, you know, less than 5% of the world and then hope that they raise strong families and kids and the next generation picks this up and we start compounding. But in the short to medium term, this is, you know, this stuff is not going to be used by masses, by the mass amount of people. Just not going to happen. Again, I think Les, as Matt was earlier,

mentioning and Odell you bringing up now yeah it's it's very clear that most people don't give a shit even just within Bitcoin like how many people just leave it on in the exchange or you know just just don't care and the answer is a lot and I think that's right like we we go for the lowest hanging fruit right we find the people that are already in line ethos wise and and arm those people with the best tools available to them and make them as accessible and well as as beneficial as possible both those things um my curiosity is is this drive towards better best practices um is this partly maybe going to come from a subset of of the bitcoinsers because i've seen a noticeable shift uh especially this past year maybe it's just bear market vibes people

are getting interested in other things probably a bit of that too uh but But not only with, say, the Human Rights Foundation, there's been a huge push towards opening up their freedom money track to be the Freedom Tech track. And that's inclusive of, of course, Bitcoin, but also online privacy and, you know, sovereign AI and all that kind of stuff. But the other notable thing I found was even just recently being at a conference in Prague, that was also the vibe there.

Like it was still, funny enough, equally busy as it would have been during a bowl market. But a lot of the discussion was in and around widening your horizons, broadening your horizons and learning about a lot of this sovereignty tag. And it's a direct knee-jerk reaction to these draconian measures that we see coming in for, you know, age verification. and censorship and all of this stuff, people see it coming. And they're saying, you know, freedom money does sweet fuck all for me if I'm not able to also, you know, be private or, you know, keep my own data

and all these other things that come with it. So my question is, do Bitcoiners help in spurring the digital sovereignty revolution here? Yeah. Yes. I mean, I've never been so bullish on Start 9. Thank you. Yeah. It's catching, okay?

Like, people are waking up. And while, you know, what Matt said is very true, we're never going to get mass. We're never going to get the whole world here. We're never going to get 4 billion people using this, not in, not in any reasonable amount of generations. But 5% of the world is, what? 300 million people, right? give or take, that's a pretty big market.

You know, that's a, that's a, that's a, from a business perspective, it's a pretty big market. From a, from a human perspective, that's a lot of, that's a lot of free individuals. And you can do a lot with a small number of passionate people on a mission. History has demonstrated this. You do not need a majority to win a revolution. You do not need. All you need is a relentless, small number of courageous freedom fighters. And so that's where we focus.

That is our mission is to build the tools for the freedom fighters, those who see it, those who want it for themselves and for their families and for their world. We were doing our best to make it available to them. And so, yeah, you're right with like the HRF and there's a lot. It's all over the place. There's a big kind of, I don't want to call it a pivot, but an expansion into broader freedom tech thinking. This is, you know, Bitcoin is the hero of the army. Bitcoin is Achilles. Bitcoin is Hector.

Without which none of it matters. The army loses morale. There's nothing left. But it is not the army. It is not the answer. It is the hero of the army. We need Bitcoin to take on the banks. We need Bitcoin to fight Theat.

But we need to come through behind it with ships and tanks and all right, to roll through. and actually take over. And that is the broader freedom tech movement where we have always been focused. And when you ask if Bitcoiners are the spearhead, they are the seedling of this, absolutely. And that's why we started where we started. Start 9 is not a Bitcoin company.

We're not. We never have been. We targeted the Bitcoin market as our sort of niche launching point for very strategic reasons. and because we ourselves are Bitcoiners, and we hold almost every piece of capital we have in Bitcoin, not because we're a Bitcoin Treasury company or because we're a Bitcoin company at all, but because we understand Bitcoin and know that holding the majority of our balance

sheet, if it's not all of it, at times in Bitcoin is the right move as a business. And so, yeah, 100%, man. It's really gratifying and heartening and hopeful to see a lot of what we've been shouting for years start to take a little bit. O'Dell, I'm curious your perspective from with regards to adoption of these tools and Bitcoin, who you think the next cohort of Freedom Fighters is going to be. I found myself thinking about this the other day that I don't, it could be just, you know, bare market vibes are not paying attention to the right areas. But I don't feel like there's a steady stream of new Bitcoiners coming in.

I feel like it's the old crew and we're just expanding in terms of our understanding of the things that we care about, what we're involved in. The idea I was playing around with was all the clampdowns we see on social media. I don't know about you guys, but I remember, like, I figured out how to use Tor and a bunch of shit when I was a kid because I just wanted to pirate stuff and download stuff. And I wonder if it's like, I'm cautiously optimistic. I wonder if the clampdowns we're seeing from governments and regulations on the youth will spur them to learn the tools to just navigate. gate around them, that we might have a little bit of an injection of young cypherpunks, just out of necessity because they're getting blocked from everything. But just overall adoption and where people are coming from, Odell, and curious your thoughts.

Look, I overwhelmingly agree with Matt and co-sign his message. I think, look, on the Bitcoin-specific side, I think what it is, is like we kind of had a lost generation to borrow a term from history in the last cycle, because just, just so many people just got onboarded to like the Orange Thai movement into like Bitcoin stocks and shit. And so I feel like it has felt like there's less new people, like young people. And a perfect example of that, by the way, is Austin BitDev's, which has been a prolific grassroots local community.

And the guys that run it, it's a lot of work, right? And it makes sense that they got burnt out and want to stop Austin BitDev's. But what is disappointing is that there isn't a younger. cohort ready to take over from where they left off. Like, no one should be expected to do this thing for like a decade plus and not just pass the torch to someone else. Like, that's how this compounds. It's easy to think of these things as technology projects, but ultimately it's movement

of individuals, right? And moving to individuals, you need constant new supply. And that's how it compounds over and over again. But to the larger, look, I don't think that means that we're never going to get young people into Bitcoin. There are young people coming into Bitcoin. There are a lot of new excitement and new entrants coming in. I think that will continue.

I think the trend is up into the right of people using Bitcoin as freedom money. And as a percentage, it might be smaller than it has been in the past. Probably is in an overall community. But just the absolute number of people using Bitcoin as Freedom Money has never been higher. And that's just going to keep going up. And then the last piece is to the greater Freedom Tech stuff and Bitcoin is, through my experience, where do you see the greatest progress and adoption and usage of these tools? It's where people need them.

You know, if you, if you try and, if you try and talk Bitcoin to a, specifically Bitcoin as Freedom Money to a financially savvy, well-off American population that has, has good access to traditional financial tools, they don't really have as much of a need for Bitcoin and freedom money. Or they don't realize that they have that need. Maybe that need is in the future. Maybe it's a little bit more abstract, right? But if you talk to someone in the developing world, or if you talk to an activist or something, they immediately realize the need. Their question isn't, why does this exist?

It's how can I use it, right? And I think that translates across the entire Freedom Tech stack. So if you're in the UK and your speech is getting trampled, you're going to figure out ways around it, right? But if you're in America and maybe it's more of a boiling frog and is not outwardly getting trampled, you'll just stay in the existing system. And that is the key benefit we have going for us.

It's almost like an iron sharpens iron thing. Like as the walls close in, the resistance gets stronger. And that is our ultimate bulwark, I think, against, you know, widespread corporate and government surveillance and control. I love it. You brought it up. So I'm going to ask quickly, do you think the orange tie thing is kind of peaked? One company I like pointing people to when they ask about Bitcoin mining is abundant mines. They were founded by Bo and Christine Turner after losing over half a million dollars to broken promises in the mining industry.

and they built their hosting model to remove the usual headaches. With abundant mines, you own your machines and keep 100% of the Bitcoin you mine. There's no revenue share, no hidden skims. Pricing is simple. One flat monthly fee covers power, parts, labor, and repairs. They also guarantee uptime. If a machine goes down, their hash rate redirect system routes hash power from their fleet, so earnings don't just stop.

And every machine is insured at full replacement value. Everything is hosted in the U.S., powered by hydro, and mining equipment may qualify for 100% Your One bonus depreciation. Learn more at Abundantmines.com slash sessions. Do you think the orange tie thing is kind of peaked? If you want to stay sane, you never call peaks. So we'll see. We'll see how this all plays out. I'm going to give you some anecdotal evidence that it's, you know,

not as compelling as it once was from two examples actually the last two events i was at i was doing something uh or presenting something whilst sailor was talking and in the past that meant whatever you were doing was effectively empty um you know pacific bitcoin back in the day i did a panel during that and there was like four people in the room and i said congrats guys. Thanks for being here. I know it takes a lot to skip Sailor's speech and some guy in the back. It's like, oh, fuck, Sailor's on and he ran up the room. But this time, so one of them was in Vegas, myself and Nathan were running workshops. Sailor was speaking. The room was packed. It was lined across the walls. People were cross-legged in front and they were learning, you know, how to actually

use Bitcoin in a private way. The other one was in Prague. Sailor was speaking and I was on a panel called Building for the Apocalypse. And it was, you know, building out your freedom tech. And again, I'm thinking, oh, this will be empty. Again, it was full. And so people are at least a subset, right? It's still a minority. It's not like his talks are not well attended.

But at the same time, people are skipping on those and come into the freedom, the censorship resistance, the privacy tracks, because they're seeing the writing on the wall of what's happening there. So, yeah. Interesting, Matt, your thoughts? I was never interested in it, personally. For me, this has always been single-tracked revolution, right? Like, I'm in this for the revolution.

I always haven't been. And I am just not interested in the use cases, the derivatives, the financial wrappings. It's all just affinity scamming. to be honest, even if it's not scamming. It's using Bitcoin's essence as cypherpunk open, decentralized, permissionless freedom money for alternative purposes, right? And without that being true, right, if Bitcoin is not open, permissionless, decentralized, freedom, fuck you money, then all the derivatives fall apart as well, right?

Affinity scamming against something that is not cool doesn't work. So Bitcoin's brand, its reality as a technology, is what enables these things and makes them inevitable. But I also don't find them interesting, right, even though there are short-term opportunities, you know, to profit or benefit from them. It's just a distraction, in my opinion. It's not worth your effort because you're just as likely to get burned as you are to win. and so why gamble, right? Just focus on the thing that matters, on the signal, and let the noise come and go as it does.

So anyway, that's my take. I've never been interested in sale or any of them. Love it. I have a different question for you, Matt. I'm kind of curious because I feel like at Start 9, you're going to have a very interesting insight into what has people's attention right now. So I imagine still a lot of the customers coming in.

They're coming in for the node first. It's probably still one of the, I would think, the largest drivers for getting your own personal server set up. Is there any other trends, in the sovereign computing space that you can see is kind of driving adoption. And the other thing that I want to add there is we talked earlier about how like Apple's raising prices. I know Nintendo's raising prices.

Compute's getting more expensive, which feels like the almost like the worst possible counterforce against going, trying to stop everybody focusing on cloud compute to get away from cloud compute to get away to self-hosted data. If the hardware's getting more expensive, that kind of throws a bit of a wrench in it. Do you think that trend of rising memory costs, rising compute costs, is going to reverse or is this something like a new paradigm we can't have to deal with going forward? So there's a little bit to unpack there. So first of all, the majority of our sales are to people wanting to run a Bitcoin node, like the vast majority. Very, very few people come to us and

just don't care about Bitcoin, right? They do other things as well, but the main reason they are buying our servers to run a Bitcoin node. And that has always been the case and is something that we know will invert at some point should we succeed. It's not that Bitcoin will go away. It's just that it can't remain as the number one reason why somebody is buying a personal server. And that will gradually shift over a long time as not only the number of offerings grow, but the quality of those offerings and the understanding of the problems with the centralized alternatives. Bitcoiners are just way ahead of the curve here. So it's our job, like I said before, to just make it great.

And we'll be here waiting when you're ready, anyone, right? And then in terms of, what was the second part of your question? I was basically asking if, like, what other aspects of sovereign computing were kind of driving adoption? What's the second thing that's kind of bringing people in? And then additionally, the implications of having rising computer costs. Yeah. So AI is definitely the number two seat.

Really? Okay. Yeah, which is not really viable yet. So that thing is a bit, you know, card ahead of the horse here a little bit. So, you know, everyone wants, like the big hype on the Internet right now is around AI censorship and cucking of the models and, you know, rising costs. And everyone wants to set up these local inference systems and local harnesses and like the open claw, you know, phenomena, I'll call it. But the reality is that the hardware and also the kind of the integrated systems are not there yet. It's really clunky.

You're going to have a bad time. It could be dangerous, right, if you make mistakes. Not just in terms of like security, but in terms of privacy, right? If you don't understand how to structure your file systems and how, you know, context works, you could end up shipping your entire file system off to an anthropic server, right, without even realizing it. And so a lot of dangers. And I'm doing a once a month AI meetup at the Denver Bitcoin space right now where we are doing all the way from 101 to specific workshops on how to use different AI tools and in a fully sovereign way. And by doing those workshops and by seeing the people who are coming there is there's a huge discrepancy between where they think it's at and where it's actually at.

And it's going to take a while, potentially a few years before we really have bulletproof, effective, intelligent local systems. But we are very focused on that, right? We have seen that as the next ship, right? That's the next market in addition to just data and messaging and all the other things that you can do with our technology. We think AI could be the tip of the spear in terms of market expansion. And then one more time, last part, hardware costs, hardware costs.

So, yes, we feel this very keenly. We made a really good move in retrospect back in October, November of last year, where for the first time ever, Start 9 sold some Bitcoin basically at the top in order to buy RAM. And that turned out to be a almost life-saving move, to be honest. What a great investment. You're like the Michael Burry of hardware here. Timed it absolutely perfectly, and it saves us a boatload of money. But we ran out of that stock now, right?

Our sales have been great, which is a good thing, but our purchase orders are not large enough to last more than six months. So we are now subject to the same pricing as everyone else, and we've had to raise our prices as an effect, right? We've raised our prices by 20 or 30 percent over the last couple of months on our servers, which has impacted sales to a degree. They've remained strong, historically speaking, but not what they were in the second half of last year.

And we think a lot of that has to do with pricing. A lot of that has to do with some drama in the Bitcoin space, waning and waxing and waning. I have no idea what you're talking about. What's that? What's that I mean? But strong nonetheless, even despite some oscillation there and the rising costs. So net positive, but yeah, yeah, we see no relief in. site. Our suppliers are not forecasting any relief. SSD prices followed RAM prices. They are now

insanely expensive. The SSDs that we provide in our servers, the 4 terabyte is $600. And so when you look at our price of like a 4 terabyte server one, which is like a really powerful device with 16 or 32 gigs of RAM and a 4 terabyte, and we're still under $1,200. It's like, We're actually not making a lot of money here. You know, like the components are extremely expensive. And we're doing our best to try to keep prices reasonable and by shrinking our own margins and just sort of pushing through this phase. But who knows when this phase is going to be done?

We have no end in sight. And now, just now over the past two months or so and forecasted to continue, we're seeing CPU prices go up. So first it was the RAM. Then they came for the NVMEs. and now they're coming for the CPUs. And none of them are coming back down either. So who knows?

Who knows? We have to weather this storm and hope that there is some degree of bubble taking place in the enterprise AI world, and that when it pops, there's just going to be idle machinery laying around and tons of parts and prices will plummet temporarily. I do not think we're in like a historic bubble. I think we might be in a local. bubble related to AI, but AI is literally the biggest productivity GDP breakthrough in the

history of humanity. And so it's like from a nominal value of like stock market prices and everything like that, we ain't seen nothing yet. I think we're going up, up, up, up. And this touches on something earlier that you guys said about like interest in Bitcoin from like newcomers, like new people entering. You know, a lot of Bitcoin adoption, not Bitcoin usage adoption, Bitcoin investment adoption. What Matt is saying about usage has never been better.

We're doing great. But in terms of people who are buying Bitcoin as a speculative investment or hedge against inflation, I think we're like, I've never seen it less. I've never seen less new people coming into Bitcoin to make money, which is partly because we're now in a bear market and it always happens during the Bulls, but also because there's alternative investments that are crushing Bitcoin right now. Right? Like you look back over the last six years, it's not Bitcoin.

That was not your best bet. There are stocks that are just crushing. And so Americans, people who are in first world, you know, financial situations, Bitcoin just is risky and stagnant and controversial and there's a bad taste there because of all the crap it's associated with, all the affinity scams. So yeah, it's not a good time for Bitcoin adoption on the first world investment front right now. I do think that that will return eventually when some of the...

legacy financial stock market bubble, you know, has a setback. But again, I actually don't think we're in like a real historic. Like there's no outsized malinvestment taking place. There is malinvestment, but it is not outsized. AI is real. Super real. Interesting. Adela, I want to get your take on that.

I also want to highlight that, yeah, apparently I should have invested in Western Digital. That would have been the winning strategy. But I just have Bitcoin in zero stocks. Sandisk. Sand disk. Adel. I mean, two things. First of all, like, I think the rising compute costs cut both ways. For the longest time, big tech strategy has always been you subsidize the hell out of things.

And then you harvest data, right? And the data is where you monetize. Now that compute has real costs, they're going to have more and more difficulty subsidizing. I mean, if you use Anthropic as an example, it's like a $200 max plan is getting you like $6,000 worth of token usage. And that's real cost to Anthropic. We've never seen a tech company try and subsidize to that level, right? You know, $60,000 a year per user and then try and recoup costs afterwards. I think they're going to have a lot of difficulty with that.

And as a result, you do see a lot of movement to self-hosting where even though the self-hosting is more expensive net net compared to where it was. was a year ago. Businesses and individuals can make a financial calculation in their head where they're like, okay, if I lock in my cost now, that's a fixed cost that I do once rather than a constant variable cost. And maybe that financially makes sense to me. I mean, to the AI piece, I've never seen the level of excitement for self-hosting as I did with the early open call phenomenon. People going out paying $10,000 for hardware, right? That would sound ridiculous two years ago when, you know, a start nine was like really trying to pull themselves up from their bootstraps. And now it's it's not so crazy.

Right now all of a sudden a $1,200 server pure seems like a really good deal compared to a Mac studio. When you're doing that calculation in your head. The other piece I would say in terms of self-hosting that I think is really interesting in Matt hasn't touched on at all is the router they're coming out with. And I say that as someone who's beta testing the router, I've had it for at least I think over a month now and I've been using it as much. as my router in my home. I'm using it right now to communicate with you guys. And I'm someone who's been running a Start 9 server, I don't know, for years now, two, three years. The connection or the stickiness I feel with this router is way higher than the server. Like the server I run my note on, you know, self-host some photos, some other apps,

stuff like that. Now I'm playing around with like OpenClaw and Hermes and stuff on it. But the router is like the central nervous system for your entire digital life in your house or in your office or whatnot. And it's something that I think will excite a lot more people than maybe self-hosting a server, at least at first. Like my parents, for instance, are not going to run a start-nine server at home anytime soon. They're obviously already running a router. If they can have a superior router experience that is more secure and open, I think that that is something that I might get them. And I mean, I'd probably help install it for them and said,

but I was always doing that in the past with there are other routers to begin with. So I think that's an interesting piece is the router piece. And finding where people are actually have the need for certain technologies today and kind of meeting them where they're at. You know, maybe in five years people will really see the need for having a self-hosted server. But people today already have a router, or do you want a better router experience? Interesting, Ben, I want to get your thoughts,

but before that, I want to give Matt a chance to respond to, even if I was talking to somebody about why they should have a different router, why they should consider something to start a 9 router, what's the benefit there? First of all, the router market is huge. This isn't even something that needs to really be educated or pitched to reach a sizable number of people around the world. There are a lot of people that buy aftermarket routers

for the purpose of protecting their networks, their home networks, right? privacy and security of making their outbound internet traffic private, of making sure that the home network is secure, making sure that devices connected on the home network are not, you know, grabbing data from other devices and shipping it off to third parties, which is what your smart TV does. There's like a whole culture of like, you know, DIY Home Lab hacker types that are like super geeky about routers. And then there's this whole like derivative market that listens to those people's about which

router, you know, to get and how to use it and all the rest. And the fact of the matter is, is that they all suck. We did not know this at first. We just thought that the ones that we had used sucked. But the more our community asked us about which router they should get, the one that's going to protect their privacy and security and the one that's going to be the easiest to use, and the more we tried to vet routers on their behalf, the more we realized that there was no good recommendation for us to make that checked all the boxes, which is uncompromising power, but with ease of use. There just wasn't a good router out there.

So we set out to build it and we got to be super creative. We got to start from first principles and just be like, let's stop looking at other router UIs and just think about the essence of this thing. What is it? Matt had cited a central nervous system. That's pretty good, right? This is how the signals get passed around. This is how information gets from one device to another.

It protects, it ingests signals from the outside, it sends signals to the outside, It passes signals around from within the body, and you rely on it to do its job well. And so it plays a very important role. And there's a lot of choice to be had there. There's like a lot of preference and choice available to someone and how they configure their router and what it does and how it does it. A lot of these things are not known to people. And if they do know them, they're very difficult to take advantage of. But we were like, all right, let's just start from first principles and say, okay, we want.

maximum power, maximum ease of use, how would we build it? And it was a long process, and we started from the user experience and worked our way back to the lowest level of the operating system and came up with something totally unique. Start WRT, the operating system that we built, which is a fork of open WRT, is, in my opinion, and maybe Matt can attest to this, is he's one of the very few users in the whole world right now, is the best, most powerful, most secure, most private. it easiest to use router operating system in the world, fully open source. And we are going to sell it on open hardware based on Risk 5 architecture. And so not only is it good and usable

open source operating system and user experience, but we're trying to go the full stack, which is kind of unprecedented. It is the only one of its kind that takes it that far. And this is for us a demonstration of what we are and what we plan to be. This is our true sort of coming out party of freedom tech. I think with the router drop and in the not too distant future, the camera, because we're coming out with an open source camera, security camera as well for indoor, outdoor use, that people are going to get it and that these will actually function

as marketing for the server, right? Because somebody will come in and they'll be like, I want a freedom privacy router, right? That is easy to use. That does not compromise. There's no phone home, nothing like that. And in the process, we'll discover that they could self-host their data. They'll discover that they could run their own self-hosted messaging apps

and leave signal and all the rest behind, that there's a pure play available to them. And same thing with the camera, right? camera is just its own market. It's just like, who wants a security camera in their house? Everyone. I'm pretty sure everyone does. And the reason that they don't have them could be a variety of reasons. It could be cost, but probably not because they're super cheap. It could be the effort of setting them up. Probably not because it's, you know, an afternoon. If they really wanted it, they would do it. I think the reason that most people don't have security cameras in their house

is because they know the footage is being piped to a Google server and that they are letting this company into their home, even if they themselves have not fully conceived of and articulated that, I think that they feel it. And so having an alternative, having a security system that is fully self-contained and sovereign in nature is its own massive market. And so a lot of what we're doing here is it's all about the server, right? Like ultimately the server is where the data are processed and stored. But we got to meet people where they're.

they're at. What do they want? And to do that with the same philosophy and the same strategy and gradually over time we build out an ecosystem. There's much bigger answer than you were asking for, but you, yeah. Can I just do a quick little tag in here? How are you thinking about as time passes user-friendliness? Because I know a lot of people that are very interested in that the idea of self-sovereignty and hosting stuff and obviously educating over the past number years people want to do this stuff but they get super intimidated so obviously there's a balance to be made in terms of you know how how the hell do you manage making it super secure and private and all of that stuff but presented in a way that doesn't scare people what have you found that's kind of

most challenging about that? And are there certain solutions that you found that, okay, this has helped or this has hindered our efforts? Yeah, I have a very concrete answer for you there that I think needs to be understood by anyone building a business right now. Okay, there is a fundamental shift taking place that most of the world has not recognized or adopted. It's sort of like the shift to mobile in the late 2000s where, you know, Facebook was this dominant social network thing and then everyone started using phones and it was just an awful experience and it crushed their stock for multiple years until they figured out how to do mobile right. And now mobile is a solved problem. Okay, the world migrated to mobile and it's a solved problem. And there's a similar shift

taking place right now and it's AI. It is agentic support. Okay. So in the past, when we start, started this company, the challenge, the trade-off that normally exists that we were trying to eliminate was the trade-off between a sovereign system, a sovereign, powerful, self-contained, private, secure system, and ease of use. Those things are total opposites. It's why the cloud exists. It's why SaaS exists is to provide ease of use, but take away sovereignty. And sovereign systems were basically impossible to hand roll.

And so we sought to close that gap by coming up with a novel operating system that brought most of the Linux sysadmin command line power into a user-friendly GUI. That was the basic thesis that we set out with. But in the last year or so, as 040 has begun to come of age, as 040 has like really begun to take shape and took much longer than we thought, of course, but again, massive thing we're doing here. This is not some app. It is a true, total new operating system that nobody's ever tried before. But in the last year, as AI started to take shape, we realized that we were building for the previous era. We were building for the era of GUIs, where the graphical

user interface was the essence of adoptability, of the technology. And while we're not discard it, just like Facebook didn't discard the browser or the desktop experience, right? We're still doing the GUI, we also realize that that is not the primary way that people are going to be interacting with their server. They're going to be interacting with their server through a personal assistant. They're going to talk to someone who administers the server on their behalf, and they just want to be able to look and see the results. So the experience, the user experience that we are developing as part of the 040 release is actually sort of agentic first, right? It's still sort of in the work.

and we will release 040 fully. First of all, for anyone watching this, O4O is ready to roll. It's production ready. Upgrade if you haven't upgraded yet. It still has the beta tag because there's some logistics effort that we're still trying to wrap up.

Like the Raspberry Pi 4 builds working well. And we want to make sure that everyone's upgrade goes smoothly and that there's no loss of data. But this is all sort of surrounding logistics. The operating system itself and the service ecosystem behind it is absolutely ready. You should update if you have an hour or two of free time. follow the update guide.

So while it won't be part of the first time we remove the beta tag, immediately following will be the inclusion in a central way and recommended way that you will actually use your server rather than clicking around and figuring out what a port forward is and how to do all that stuff. You will just open a chat drawer on the side of StartOS and just talk to your agent. And we are fine-tuning small models that will enable this to run not only on the onboard CPU and GPU, but if you have an even reasonably available local GPU or system, then it can get superpowered by that. But it is not so hard. We don't need a massive and

huge intelligent model to utilize the StartOS MCP server. It's just a list of recipes. We can constrain the agent's behavior to such a degree that it doesn't need to be incredibly smart. This is solved through context and MCP servers, basically, which is another, you know, it's API's for AIs. And because it's running on our system and we're designing the agent, we can also make certain guarantees that it is going to be private, that anything that goes out of the system that you're made aware of, that it is secure, like it's fully integrated experience. Now, this is what everyone is trying to do, all the big players, right? Apple finally made Siri intelligent. And like some of the players in the world, the bigger companies, are realizing that the new interface is agentic chat, whether it's vocal or written, and that GUIs are there as a supporting backstop. And so that's how we're going to win this is because I don't know if we could have ever gotten a personal server operating system GUI, no matter how.

how brilliant we were, how good it got. I don't know if we were ever going to get it to a place where grandma could do it. We needed a revolution in interface, right? And that is chat. Grandma can absolutely say, you know, I want my family photos on my server so that Apple can't see them. And it's just going to be like installing NextCloud and installing image. And you're just like your agent is your power now. So that's how we're going to do it.

I like it. Odell, I want to get your response to that, but I kind of want to tag in as well, particularly because you've worked with OpenSats and you're so plugged into the open source world. I want to know what you think about, Matt's take there on this shift to agentic kind of interface versus moving away from GUI,

but the thing that it reminds you of that it feels like a step function improvement. It feels like a 10x user experience improvement that's coming down the pipe here and already kind of here. And someone actually asked me just yesterday. They asked me what I thought the next step, the next tech,

the next thing that happened in Bitcoin is. And I honestly couldn't come up. with one. I was thinking about, like, is there, is there a 10x increase in usability for Bitcoin even still on the table? Or have we kind of scraped everything out at this point in time? I think, look, LLMs in general is a paradigm shift. Like, it changes everything. I 100% cosine mat's point that I think the GUI is, is basically the graphical user interface is just like not the way people will be interacting with tech going forward. The question is more of a timeline and how that exactly looks.

I'm not sold on this idea that that's going to revolutionally change, like, what kind of hardware we interact with. Like, there's all these, you know, AI-focused hardware companies that are trying to, like, completely move us away from cell phones. Like, I think cell phones actually could work really well in an agentic environment, even if the graphic user interface isn't the key mechanism of the way people interact with it. They have cameras, they have speakers, they have microphones. And at the end of the day, you want to see things, right?

Like, you're going to still want to see your photos. You're still going to want to see the output. You know, maybe the outputs to the weather today, right? Like, I don't want necessarily my agent to constantly tell me what the weather is. On the Bitcoin open source side, look, open sets, we use LLM tools internally. We have to be very careful because a lot of the stuff we interact with is very, very important and privacy-wise, like the privacy of our grantees and what they're working on and whatnot, but also a lot of their projects are open source, right? So the actual project data is all out there

and scrappable by any LLM to begin with, and there's a lot of data that we can work with and improve our own internal processes. We have to constantly think about now what we're funding and how we're funding and where it's most useful, right? Like this whole idea that, you know, maybe someone needed like $100,000 to build a certain project or something, and then ultimately at the end of the day right now, the main cost is their token cost, and they can do it in a fraction of the time, then maybe that grant application and grant approval should look different. On the Bitcoin side, honestly, I think there's some really big changes that can happen, not on the protocol level, but like on the app level and how people interact with these

things, that maybe we can't even comprehend right now, because open protocols and open source software can compound really rapidly in this LLM environment. Like there's a bunch of things that we're funding that are effectively like open libraries, right, for people to build other things on top of stuff like BDK, Bitcoin Development Kit, right? It makes it really easy for a, it was originally designed to make it really easy for a developer to build a new Bitcoin wallet, right? Now you just, you know, tell Claude like, yo, use BDK, build this thing for me.

And then you say, oh, maybe you want to integrate Noster, use Nosterdev kit, integrate that for me. oh, you want to use Lightning, use LDK, do it with me. And it's like you're doing like two prompts. But in the background, there was all this work that was done building out these libraries that actually make it so it can happen so effortless and without friction. And then all of a sudden, you know, you have all these different ideas compound on top of each other.

And the end user experience is completely different than anything we could even comprehend or expect. So I've never been more optimistic on open source software and Bitcoin specifically. And once again, I don't pretend to know exactly where we go and how that looks. And honestly, from an openSatch point of view, we've never been, that's not how we're set up as an organization. We're an app-led organization. People go to opensets.org slash apply and they apply for what they think is valuable to build. We don't go out and tell people, build this, build that.

So that was never our directive or our mandate. And it continues to not be the case. So we get applications coming every day, thousands of applications a year. And some of them I'm like, holy shit, I didn't even think about that. That's actually a really interesting idea. Beautiful. Ben, your thoughts? Yeah, I don't know that I have a whole lot to add here. I mean, I'm happy that you guys are doing that and that you're just allowing people to, you know,

you're not trying to influence the direction that people are going in. I think it's a super valuable thing. And hats off to you guys for, how long has Open Sats been going for now? I don't know. It's four years, five years. We've given over $30 million in grants, over 400 grant recipients, 40 plus countries. Our entire treasury is held in Bitcoin. We only send out Bitcoin.

We send out Bitcoin monthly. It's been pretty wild. It's honestly been pretty wild to see. To be honest, we had no experience going into it. kind of just wielded into existence. We wanted to see it exist. Humble through that 30 mil. No, that's amazing.

I think it's a testament to the resolve that you guys have had to make that happen, and especially in just a few years, that's awesome. So I'm sure the downstream effects for decades to come will be very evident. So thank you. I love it. Gentlemen, before I wrap for today, I did want to go around and just get your guys, as input on tech and tools that people should be trying out or using right now. You're all such a wealth of information when it comes to security, privacy, all that fun stuff,

all that sovereign computing. So, Matt, I'll start with you. What things should people be looking at trying right now? Oh, boy. Yeah, I mean, I live in my own bubble. We, they should be trying out all the new services on the Start 9 marketplace. There we go. Because it has doubled.

I mean, pretty much anything you could want to do now with a, cell phone's server is available and working really well on the 04L marketplace. So if you haven't updated yet and you haven't checked it out and seen all the cool new stuff in there, please do. Love it. O'Dell? My biggest one is if you're stuck on Claude Code right now or Codex or something, open up your horizons. I really like, I think ppq.a.i is just a good place to get started.

They accept Bitcoin. Literally every model under the sun is available there from all the open source models to all the proprietary models, except for the ones the governments won't let you have. And it's just a good way to get your feet wet. I think my Bitcoin experience shades everything that I do going forward. And look, you can listen to 10,000 hours of Bitcoin podcast. It doesn't mean you are better at using Bitcoin. You have to actually use Bitcoin. You can listen to 10,000 hours of driving podcasts.

You should get behind a car and actually drive the thing. Same thing with AI. You've got to play with the tools, get your feet wet, and be able to adapt as this transition happens. Love it. Ben? Yeah, I mean, I, diving into a lot of this stuff feels like when I first started in Bitcoin, in that I feel like an idiot. And I'm, I'm learning as I go and then trying to show people what I've learned thus far, which means mistakes will be made along the way.

And, you know, that's okay. And so I would say in general for people that are just kind of trying to get started and figure stuff out on the AI front. I've had been, you know, I'd second the PPQ thing. Maple, really cool. Maybe, you know, Hermes or OpenClaw on your Star 9 server. It'd be cool to tinker around with and start learning with. And yeah, even insofar as just downloading something like LM Studio and just seeing the difference between what you can do on something locally hosted

versus cloud versus the frontier. and maybe split your efforts and start to realize like Matt Hill was saying, you know, you don't need your Google searches to go through opus, right? So yeah, just start to learn, start to tinker, ask questions, and don't be afraid to make the occasional mistake.