BTC Sessions - "No one knows what’s about to happen" AI Catastrophe & Bitcoin BIP 110 | Matt Hill

Episode Date: March 3, 2026

Mentor Sessions Ep. 055: Sovereign Computing, AI Singularity, & Bitcoin BIP 110 Chaos with Matt HillWhat if cloud computing's hidden flaws—privacy invasion, censorship, skyrocketing costs, and inevitable hacks—are paving the way for AI overlords, rampant authoritarianism, and existential threats to humanity? In this explosive interview, Start9 Labs founder Matt Hill exposes why sovereign computing and bitcoin are the ultimate defense against Big Tech's grip, how we're already deep in the AI singularity where "no one knows what's about to happen," and the messy Bitcoin governance battle over BIP 110 that's splitting the community. Matt warns of a "global hack bloodbath" from cloud-based AI assistants stealing your life's keys, predicts software engineering's wipeout as AI democratizes power, and reveals Start9's roadmap for personal servers, routers, and local AI inference to reclaim digital freedom. Dive into Bitcoin's spam wars, where transaction filters act as rate limiters to deter attacks without consensus nukes, and why node operators hold the trump card in this technological arms race. For Bitcoiners obsessed with decentralization, self-custody, privacy tools, node running, open-source tech, and resisting centralized control, this episode is your blueprint to sovereignty in a world of surveillance and scams.About Matt Hill:Founder of Start9 Labs, pioneering sovereign computing for over six years.X: https://x.com/_MattHill_https://start9.com/Chapters:00:00:00 Teaser & Intro Clips00:01:39 Sovereign Computing Refresher00:05:26 Four Critical Flaws of Cloud Computing00:10:09 Human Vulnerabilities in Security00:10:24 Convincing People to Adopt Sovereign Tech00:13:56 Keeping Private Keys Safe00:14:18 Emerging AI Risks & Sovereign Lens00:15:59 Entering the AI Singularity00:20:05 Personal AI Assistants & Privacy Dangers00:25:02 Cloud AI Vulnerabilities Exposed00:28:39 Start9's Sovereign Computing Roadmap00:32:02 StartOS Major Update Details00:36:24 Router OS Announcement & Features00:39:43 Advanced Networking in StartOS00:40:42 Publishing Services Easily00:41:06 Integrating AI into StartOS00:43:50 AI Endgame: Express Will, Get Results00:46:38 Sovereign vs Centralized AI Futures00:50:58 Avoiding Violence Through Decentralization00:52:32 Bitcoin Development: Users vs Creators00:58:18 BIP 110 Overview & Messy Governance00:59:20 Predicting Contentious Soft Forks01:02:01 Node Operators' Political Power01:05:03 Economic Nodes Decide Bitcoin's Fate01:11:42 Dynamic Filters Proposal Explained01:17:08 Filters as Rate Limiters01:24:18 Deterring Spam with Policy Layers01:32:44 Forking as Ultimate Deterrent⚡ 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/bfiDBV🛡️ Bury Your Bitcoin Secrets Deep: The ultimate underground vault for seed phrases and hardware—rugged, weatherproof, and built to vanish. Grab 10% off Dirty Man Safe with code BTCSESSIONS at https://dirtymansafe.com 🏠 Unlock Home Equity for Bitcoin: Convert your home's future value into BTC without loans, payments, or interest—stay in control and grow your stack tax-free. Start at https://joinhorizon.com/?ref=BTCSESSIONS💡BOOK Private Sessions with Nathan, Gary, or Ben at Bitcoin Mentor: Master self-custody, hardware, multisig, Lightning, privacy, and more. 👉 Visit btcmentor.io Previous Episodes with Dr. Jack Kruse & Peter McCormack: https://youtu.be/DAjTEOIH8WwFollow Us on X:• BTC Sessions: @BTCsessions• Nathan: @theBTCmentor• Gary: @GaryLeeNYC#Bitcoin #SovereignComputing #AISingularity #CloudComputingFlaws #PrivacyInvasion #BitcoinBIP110 #PersonalServers #NodeConfiguration #TransactionFilters #RateLimiting #Decentralization #Start9Labs #MattHill #TechnologicalArmsRace #CommunityDrivenStandards #BTCSessions #BTC #BitcoinPodcast

Is there AI risk to individual users? Oh boy. We're about to witness one of the largest wipeouts. Global hack situations the world could ever imagine. It's going to be a bloodbath. Apple doesn't sell privacy. They sell confidentiality. Every server worth hacking will be hacked.

Really ever since Claude 4.5 came out. That was, I think, the inflection point. You could no longer pretend that we weren't in the singular. Everything that you thought you knew, needs to be called into question. BIP 110. Broadstructs your overview of the situation. Good, bad, ugly, ugly.

Meet Matt Hill, the visionary founder of Start 9, who spent over six years pioneering sovereign computing hardware and software to liberate us from big tech's grip. In this episode, Matt lays out why cloud computing is leading us down a dangerous path to global authoritarianism. We think that this poses an existential threat to the species. That type of power and control over information

must lead to rampant centralization. and authoritarianism. He reveals why we're already in the AI singularity. Total reimagining of reality. And it's happening at a pace that very few people can comprehend, including myself. And Matt unpacks Bitcoin's messy governance crisis, including the controversial soft fork BIP 110 that's splitting the community and could redefine the network's future. I predicted that this would happen. I was like, if you don't do this, there's going to be a contentious software. Plus, he lays out the essential steps for anyone to survive the coming chaos. We are going to come out of this safely and into a better world for all future generations.

Sovereign computing has to happen and Bitcoin has to happen. Matt, thank you very much for joining us again. I don't recall if you remember this, but you were actually our very first guest on the show. And I went back and I rewatched a bit of it. And I have to say, I was pretty handsome back then. So I'm pretty pleased with how it went. The stash had to come off. but we have a lot of fun stuff I want to get to today.

So we have a lot of things to talk about, including Bitcoin policies such as BIP 110. But first, can you give a refresher for those in our audience who may not be familiar with it or who may just be dumb like me? What is sovereign computing in simple terms? Why is it important? And why would the average person want to run their own server instead of relying on cloud services? Okay. Cool. I'll do my best to keep that really short.

Well, thanks for having me back, by the way. I do remember I was the first on this show. I did not know it was the one year anniversary of that. So that's cool. So anyway, sovereign computing. What is it in a nutshell? Sovereign computing is the ability to use your computer and information technology in a way that does not require the involvement of trusted third parties. So it is a disintermediated non-custodial model of personal computing, where, for example, you and I could exchange information, we could text message, we could transact, all without anyone on Earth even knowing that the communications were taking place, as opposed to the current model of computing, which is broadly referred to as cloud computing, which is a fully intermediated, largely customaryed.

studied computing paradigm, where a few and increasingly few, number of companies and governments sit in the middle of all the world's information, either just as a middleman relaying the information, even if it's end-to-end encrypted, that affords a degree of privacy, but it is still intermediated, which always leaks metadata and definitely enables censorship. But in most cases, cloud computing is actually not only intermediated, but completely transparent and custodied as well. So it is a privacy, invading, censorship, ridden, expensive, and highly insecure and undignified model of computing.

And it is broadly, almost universally adopted for all of the world's computing needs. We think that this is a huge mistake. We think that this poses an existential threat to the species given a long enough time horizon because that type of power and control over information must lead to rampant centralization and authoritarianism, which itself would lead to rebellion, which itself would lead to violence. And in the world and age of weapons of mass destruction, we must avoid violence and war at all cost. So we think that disintermediating computing, introducing a sovereign computing paradigm, at least making it possible for certain, if not all people to use, will be a meaningful mitigating factor

against runaway centralization and totalitarianism. And so we spend our time building software and hardware technologies that make the ability to use computers in a fully independent, sovereign, private way accessible to everyone. That is a tall order. We've been working on it for over six years, and we're getting better by the week. Thank you, Matt. I don't know why anybody would be concerned about anything you said. I completely trust my corporate overlords, and I especially trust the United States government,

so I'm fine. But I've heard you, yes, I've heard you suggest that there are four critical flaws of cloud computing. Would you be able to break that down for our audience? Yeah. Invasion of privacy would be the first that a lot of people, at least these days, are more familiar with, which is that when you're using intermediated computing infrastructure, basically every button you click and everything you say is recorded and logged and can be used for a variety of purposes. And it's also just flat out undignified to have various corporate entities and government entities just kind of listening in on everything that you do, even if you have

nothing to hide or breaking any laws, it's just kind of a childlike existence where mommy and daddy are looking at you in your room to make sure you don't do anything bad, right? So even if you're not, it's still just childish and undignified. We think people deserve better. So that's invasion of privacy. Another one is censorship. When everyone is using the same central servers to propagate and store and process information, the people controlling those servers control the information. They can cut you off. They can amplify falsehoods. And we've seen this. We've seen this, in fact, on the rise over the past, especially five years, but really since social media came into existence, anyone who controls those servers controls the narrative, which is very dangerous for a free society.

Number three is cost. So when you are using somebody else's property to do anything in this world, they need to get paid. So they are a middleman, right? All these companies facilitating everything from text messaging to photo sharing to social media, they are middlemen. They are, you know, they're definitely enabling it, right? If it wasn't for them, some of these systems would not have come into existence, right? So I think I'll make a statement that's a little bit abstract from the question you asked, but I think this sort of had to happen this way, right?

Like we had to have a few people sort of step up and do it for everyone else before people could learn to do it for themselves. So I don't blame humanity for this step, but it is time for a change. But regardless, these middlemen, these intermediaries and custodians, got to get paid. And given the fact that most of us are dependent on them for our basic computing and information needs, they can get paid quite a bit. They can extort, in fact. So, like, if your company is dependent on, you know, Google workspace or Microsoft 365 or something, they can just kind of ratchet up the costs and hold your data hostage. And there's not a whole lot that you can do. So they get paid in a couple of ways.

One is by mining your data and selling it. And the other is by charging you subscriptions. and more and more these days, it's both. And this is called SaaS, Software as a Service. And this was the killer model for venture capitalists in Silicon Valley for the past 15 plus years. And it's expensive, right? There's a reason why VCs throw money at SaaS companies.

It's because they are charging way more than what the product costs to make and sustain, right? The marginal cost of adding the new user is essentially zero, but you get $100 a month from them. That's a great business model, but it was not to be, right? Some stuff like that is not meant to last, and we're destroying it right now, us and others, and AI especially is just going to murder the whole SaaS model. So good riddance to all those VCs.

And then lastly, is security. So when UI and everyone else is shoving all of our data onto third-party servers, those servers become attractive targets for hackers, big and small. And every server requires human access, right? This isn't a matter of like poor encryption. It's servers need to be accessible by certain humans who possess the keys, and humans are fallible, period, which means eventually every server worth hacking will be hacked, not because the server lacks proper application security, but because humans lack perfect operational security.

and eventually the person with the keys will be identified and socially engineered to reveal those keys, even going so far as to utilize tools such as blackmail or threats. Yeah, I remember that's something, and I'll get you to your questions just a second, Nathan, but you did talk about that last time about that example you gave of a guy in a snowstorm calling up a company like, hey, I can't make it in today? Can you give me the information for these keys? Like, oh, yeah, sure, because humans are the most vulnerable thing. I do want to ask you, Matt, because you've brought up a lot of great points here. And Nathan and I are both kind of libertarian and cap types. We agree with this. In theory, I agree with everything you said. But I'm also a lazy, lazy man. And right now, yes, I do get on something like Signal.

And I, you know, I use proton mail. And yes, full disclosure, I have a start nine server. It works great. Thank you very much for that. My little start nine Bitcoin node. But it seems to me, unless something really hits people in the face, like, you know, In Germany, they're very concerned about inflation generally. Yeah, well, that's because an entire generation lived through the Weimar, you know, crazy hyperinflation. They're aware of it. Is there any way to

try to sell people on this idea of getting their own cloud, doing private work, getting off these large corporate government clouds, aside from just a disaster hitting them in the face? Yeah, there's some techniques. Ultimately, that is what it will require for most people. But, you know, me coming on this show, for example, in this show propagating around social media might convince a couple of people that there's a real problem and that there are practical measures they can take to protect themselves and their loved ones from experiencing its consequences. You have your early adopters, you have your late early adopters, then you have the rush, and then you have your late adopters, and you have the people

who never get with it. And we're going to have the same thing with this as we've had with any technological innovation. You know, Start 9 is not inventing really anything new. I know that that statement could be taken to the absurdity that there is no, we've mentioned of anything new, that everything is a compilation and reproduction of something that existed. But I mean, in a more extreme way with us, like we're really not inventing anything new here. You've always been able to use computers in a disintermediated, non-custodial peer-to-peer, fully encrypted manner, it just was really, really hard to do. And so our mission is to make it less hard. It's always, right, we've come to the conclusion that it's always going to be harder to do

sovereign computing than it is to do the total cloud, you know, intermediated custodial computing. But we think that we can get it down to a very small gap. We think that ultimately, at the long tail of that asymptote, we can get it down to one thing, actually. We think that we can match the cloud computing user experience to the T with one exception, which is forgot password, which is lost password. ultimately, if you are to be totally sovereign, you have to keep your keys safe. You cannot lose your key because if you do and there's somebody who can fix it for you, then it wasn't yours to begin with.

And so we can match the user experience except for that one thing. People need to learn how to keep a private key safe. And if they can do that, then we can do it. We can reach, we can make sovereign computing just as easy, just as convenient as cloud computing. Nathan, is there a company out there that helps teach people how to keep their private keys safe? Yes, if you'd like any assistance of that, I'm sure Gary and I can both have you on it. It's funny enough, now, even imagining my head is like, that's why I got to keep my start nine password on my cold card. That'll work out quite well for me.

I got the backup right there. Matt, I want to go a little bit deeper into the AI space in particular. And starting from just the sovereign computing kind of lens, what is your view of the emerging AI space? There's a lot of things that we can touch on. I do want to talk about its impact on software as a service and talk about its impact on basically open source development as well. But just from a risk model perspective, is there AI risk to individual users? Because most of those are just cloud-based services as well, too, that you're seeing, or do you have any concerns? Or maybe not.

Maybe it's ultimately fine. Is this something that we should also be thinking about moving to our own personal servers? Oh, boy. here we go I'm serious okay I mean we can come at this from a lot of different angles it sounds like you came in there from a security angle which is

let me understand is cloud computing security model threatened by AI was that part of your question there actually no but that's a good one that we should tease apart of basically thinking about kind of the same ideas of if someone is using law of these large language models they're giving a ton of information

and potentially access, depending on what they're using to their computers. So that, to me, sounds like a big risk. There's now, then you're pointing out the AI implications of terms of security risk. One thing that I learned from Chris Hadnaggy that I had no idea about was, I forget what it was call, but it was basically a almost like a unlocked version, I think, of chat GBT on the dark web. They could pay for access to this large language model that had none of the security features to prevent people from writing, say, like, Trojans and things. And it would write, it would write malicious code and things for you as well.

making it even easier for attackers to try and get access to things. So from both the giving up the personal information and maybe the security vulnerabilities that emerge because of AI tech, let's start with that. And then maybe we'll dive into software as a service and open source development. Okay. So this has been my topic now for a while.

For example, you know, sovereign computing, Bitcoin. These were things that I would mostly talk about over the last few years. really ever since Claude Opus 4.5 came out, that was, I think, the inflection point where you could no longer pretend that we weren't in the singularity. Okay? Like, I remember a few years ago, shit, I remember a few months ago when a lot of the coding we did at Start 9 was by hand, right?

when I wrote code. And our stance was that these were useful tools that could hammer out certain tasks, especially when they were well-defined and typesafe, and the contracts were clear, but that their ability to reason on an architectural level was severely limited. And that would probably remain the case

for a long time. I no longer take that stance at all. I have been amongst the more extreme of everyone I know in my assessment of the capabilities and the acceleration and trajectory of LLMs. So I'm on the extreme end of everything that you thought you knew, about your life and about your future and about the purpose of human life in general needs to be called into question.

We are witnessing nothing short of a total reimagining of reality and it's happening at a pace that very few people can comprehend, including myself. I don't think anyone can actually. even the people building these tools do not, cannot wrap their head around what they're building and what's about to happen or what is currently happening.

I shouldn't even speak in future tense anymore. I mean, like, we are in the flip right now. We are in that moment. And I have been the closest to a manic state of consciousness. that I have ever been in in my whole life, and I'm an obsessive person, over the past few months. My sleep has evaporated. I remember as a kid being addicted to certain video games. All of it pales in comparison to what I'm experiencing right now and what I think many are experiencing. And I still do not believe that I've even touched the surface of the implications

of the technology on my life and on the lives of all of us. We are facing the unknown. We don't know. Nobody knows what's about to happen. But let me share some more concrete things rather than just pontificating about the magnitude. But now that I've set the stage. So what a lot of people don't realize is that I'm going to get really practical here as opposed to the implications on the purpose of human life,

is that when you infer, when you are using an LLM, whoever is running the GPUs on which that inference is being performed has the plain text view of everything you asked it. Now, this is easy to understand when you're typing something into chat GPT. you're typing in a question and everyone gets that that question is now in the books right like it is now on an open AI database that you ask that question and here's the response that you got some people didn't realize that at first I think everyone gets that now okay just like we understand that when we type something into Google that that search goes on our quote permanent record what

is even more interesting, concerning, and exciting is the ability to have personal AI assistance. So the open-cloth phenomenon that's happening right now is happening because people are getting a taste of what's possible when you have a personal AI assistant in your computer. right like this thing is it's in your digital life okay because it's not just in your computer if you have iot devices and lots of computers and they're all connected through the cloud you basically have this like virtual assistant right this what's what syri and all them were supposed to be you know um but with with root access to your computer and all your computers and powered by a super intelligent LLM, people are getting a taste of just how powerful they could become. It's like someone shooting a gun for their first time or something, just realizing that

they're not bound by their physical stature anymore, right? That back in the days of the feudal ages in Europe with the knights, it was like these huge, massive athletic people were like royalty because it was like they could literally take on in a suit of armor, they could take on like 50 peasants. They were like actually like a superpower. And then all of a sudden advanced weaponry came into being these people, their status, you know, dropped dramatically. Now they're professional athletes, which is still a great status.

But it's nothing compared to the kind of like power, political power that they wielded in past eras, right? I'm not talking about just making money. I'm talking about the ability to move the earth. in the way that you see fit. And in the modern era, the equivalent of that are these super genius software engineer types like Mark Zuckerberg, where it's like,

he comes up with an app because he had these skills and he, like, is now, like, massively powerful and politically influential. And it's like most people didn't have access to that in the same way that they didn't have access to being a huge knight in the feudal era. They didn't, they just weren't endowed with that or they didn't practice it their whole life,

or they weren't born into a situation that fostered it, et cetera, et cetera. But now it's like you don't need to be a coder anymore to code things, right? This is the modern day equivalent, the digital and intellectual analog of weaponry. Coin kite has been in the game for years creating hands down the best and most

secure hardware when it comes to securing your Bitcoin. The cold card Q is an absolute powerhouse and my daily driver, and it's ideal for newcomers and advanced users alike. The tab signer gives you a low-cost, user-friendly option for those just getting started or for convenience when traveling. You can head to coincite.com and use code BTC sessions for discounts, or simply scan the QR code on the screen to get started right away. One company I like pointing people to when they ask about Bitcoin mining is abundant mines. They were founded by Beau 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. The 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% year-one bonus depreciation. Learn more at abundant mines.com slash sessions. Right? This is the modern day equivalent, the digital and, an intellectual analog of weaponry, right? Like, even a small person can wield a dangerous force. And now you have the corollary of, like, quite frankly, an uneducated stupid person can be a powerful force in the world now.

And so that taste of power is intoxicating to everyone, right? Everyone's just like, look at what I have built. my bot is sending this and doing this and I've automated my life and I'm like building apps now. And we've watched an entire industry of probably the most comfortable, coveted, sought after profession in modern history of high-level advanced software engineers, just getting wiped out. And for people in that profession or people still with the perspective of three months ago, of that profession, delete it.

It's not a thing anymore. I would have said juniors will get wiped out in a few years, right? Six months ago, I've been like, yeah, junior devs are going to get wiped out in a few years, right? And then it's going to be longer for mid-tier developers. And it's like it's coming for the seniors now. Okay? Wow. So that is extremely powerful, but here's what people don't realize,

going back to what I was saying about all the inference being done on the LLM server. if you install OpenClaw on your computer and connect it to all of your accounts, your Facebook and your email and your signal, and you, in order to, if you ask your assistant, like, hey, link up my Facebook account so you can auto post for me or link up my X account, okay? In order to do that, all the information necessary to link up to your account was shipped off to the LLM for inference.

All of it. This isn't about like asking a question and that's what gets logged. This is about the keys to the kingdom, your entire life, your entire computer, everything about you, all the keys to everything you have, everything you've ever written on these platforms, anything is now accessible to Anthropic the company. Right? All of it. You're had, you're done. second you installed it and said, do things for me, it's just everything about you, every access to everything you've done has now been shipped off. So what we've done is we've simultaneously introduced

the ultimate superpower, which is the outsourcing of intellectual, mental capabilities to a machine that serves you. And in the process, put the cloud computer, the vulnerabilities, the privacy, censorship, cost, and security vulnerability. of the cloud computing paradigm on steroids. And so some people are kind of aware of this, and they're frantically looking for a way to have the power without the trade-off. And what that means is local inference.

You must run the inference on your own GPUs against your own open-source LLM. If you're not doing that, you are handing over everything, and I promise it will be used against you. We're about to witness one of the largest wipeouts, global hack situations the world could ever imagine in the future. I don't know how soon it's going to happen,

but it's going to be a bloodbath when it does. And so you have to be doing your own inference. The problem with that is the same problem that it was always to do your own anything with computers. It's like, how do I do that? And not only how do I do it, but how do I do it at a level that compares to plugging my computer into Opus 4.6? Because I can't stand to be left behind, right? Like I personally, I use Claude Opus 4.6 for an enormous amount of development work because nothing else compares.

Nothing else that is readily accessible to me in terms of performance, cost, and just practical setup. really is available to me right now. The good news is that you can quarantine things in a way that doesn't hand over the keys to the kingdom and sort of plug your nose and deal with it while the more sovereign capabilities arise. And we are aggressively working on that. and I'm happy to share some of that roadmap with you guys, but I've been talking for a long time. No, share the roadmap. Don't stop.

Just share the roadmap. I want to know what's going on. My mind is getting blown, so continue. Okay. We think, I know that if the world is going to pull through this transition into a better world and not into a dystopia beyond all imagination, either where we're dominated by literal AI overlords or a few humans that have managed to control and wield their powers, or into a nuclear holocaust.

If we were going to come out of this safely and into a better world for all future generations, sovereign computing has to happen. and Bitcoin has to happen. We have to get people using information technology in sovereign ways, without intermediaries or custodians, and we need to break the alliance between state and money.

Those are the essentials. And the cool part in my mind, which a lot of people don't realize about me, is that I don't think everyone needs to do this for us to pull through. There needs to be a critical mass of intelligent, courageous, strong people who adopt these things such that they can coordinate with each other and stage a resistance or, or, violent revolution.

So I build today, this was not always the case, but the more the world has evolved over the last few years, the more my thinking about what I do has evolved. I build for those people. I want everyone to get their own server. I want everyone to run local inference for their own privacy, for their own dignity, for their own security. but I'm building for the people who will need it to fight back against what is inevitably a rapid encroaching totalitarianism backed by horrible weapons and AI intelligence. It's coming.

And we need a few people, right, the strong, courageous revolutionary, the people who are going to be the last line of defense, to have the tools available to them to mount an effective guerrilla war. If we cannot communicate privately with each other, if we cannot transact privately with each other, when I say we, I'm not talking about me because I'm not a guerrilla warfare revolutionary.

I just build the technology, okay? Blacksmith for them. Bitcoin mentor does not officially endorse violence continues. Just for the record, right? I build freedom technology. just in the same way that gun manufacturers build freedom technology, right? But there's a reason for the Second Amendment and it's runaway totalitarianism, not just because you're afraid of your neighbor, but it's because you're afraid of your government

and you should be. There needs to be checks and balances in the world. But wars aren't just fought like that anymore. We live in the information era. The weapons are based on information and access to information and the flow of information. And with AI, it's on steroids. That becomes more true than ever. And so we need the physical equivalent of guns so that people can effectively fight a war in the information era.

And for the same reasons, right? This is all for freedom and prosperity. This isn't for violence and death. This is so that good people stand a chance. And that's what I'm building for. And so the roadmap for that looks as follows, right? So we have had StartOS for a while now. It is about to get its biggest update ever,

an update that took twice as long as we thought it was going to take. We thought we were embarking on a 12 to 18 month journey with StartOS 040, and we are just about to hit 35 months next month. Wow. So almost three years we've been working on one piece of software. Now, we've been doing other things as well, but this has been the primary focus for almost three years, which is the longest I've ever spent building any one thing in my whole life. And I am happy to announce that it is better than I ever imagined it would be. We learned so much along the way.

And that's why it's late is because we didn't know what we didn't know. We ran into problems that we didn't expect. what we are building has not been built before. There's no template for this, right? And so we had to figure it out as we went. And unfortunately, we learned the hard way. We built things, and then we had to tear them down. We built them again, tore them down, and we finally got it right.

And this week, we are expecting Alpha version 20 of StartOS 040. Alpha 19 was a huge success. It's very stable, very good at this point. I am openly encouraging people who, you know, are a little bit more adventurous to download it and upgrade their servers now to Alpha 19. Alpha 20 is a huge improvement on Alpha 19, which was already really good and that's coming this week. So we are inching closer and closer now to what I will consider an acceptable mass market home server operating system product. We're getting really, really close. I cannot put a date on it because, again, we're just figuring it out as we go, but I can say it's working great.

it's better than I ever thought it was going to be, and I promise, I hope people will forgive us for being late once they see how good it is, that we haven't been messing around. This has been constant for three years every day. So 040 is coming. Secondly, we are just finished the rough draft, fully coded end-to-end prototype of StartWRT, which is our router operating system. The router hardware has already been procured, assembled, boxed. We are ready to rock and roll. We're going to announce the router presale as soon as StartOS 040 enters public beta. We felt that it was inappropriate to announce a new product before we delivered on the one that everyone has been waiting for.

So the router announcement and pre-sail will follow the StartOS 040 public beta announcement. and the router, unlike Stardos O4O will be delivered in short order because we were ahead of that. It's actually already working. We still are fixing some things, but we do expect a summer or late summer release of the actual physical router. We will be shipping units out the door this summer. That is our goal, and I think we'll hit it. It's also a fully open source router operating system, MIT license, just like StartOS. So you do not have to buy a router from us.

You can just flash your existing router with StartWRT, assuming that the router itself supports OpenWRT, which is what StartWRT is based on, hence the naming. The router is really cool and absolutely necessary in our progression towards a sovereign computing paradigm. The server is the brain of the operation. It stores all the data, it processes all the data, and it is where you can add functionality.

You install services onto your server like Bitcoin or a Lightning Node or any of the other variety of open source services that make your computing life possible. The router is the traffic controller, right? The router is what creates access to the server and the things running on the server, both by devices on the network, So IOT devices in the same home as the server, but not necessarily.

Also, external devices, right? So not only IOT devices, but your other computers, your phones, and the general public. People want to be able to publish things onto the internet, either on the darknet using Tor or on the clear net using a dot com or dot net domain, and increasingly connect over peer-to-peer networks as well using non-Tor networking protocols. So a huge part of StartOS 040 and our router product is designed for simple, easy, yet extremely powerful advanced networking techniques that enable you and others from anywhere in the world to connect to and through the server in order to garner its benefits, effectively a personal cloud accessible from anywhere only by the authorized

individuals in a completely private, disintermediated, end-to-end encrypted way. And we have multiple options now with the router and start OSO-4L. We've upped our networking game beyond anything that's ever existed. Okay. I should have pause for one quick second, Matt. Don't forget where you're going. Just for clarification, too, because like, so as well, What you're telling me is that with the new router open source software as well, too, things like having to set up like a reverse proxy for being able to, like, access back to like a BTC pay server or access back to my home network, we can basically circumvent that, make it way easier for accessing our server when we're not home.

Yes, and that is actually inherent in StartOS. You don't even need the router for that. StartOS 040 makes it a couple of buttons to publish something like your BTC pay server, you know, crowd fund or, you know, e-commerce storefront onto the public internet. A couple of buttons. StartOS handles all of the reverse proxying, all of the certificate generation, all of the firewall rules. It makes everything extremely transparent so that you know exactly what you're exposing and what you're not exposing, how they are exposed. It gives you granular control over all of this in an extremely, like, user-friendly way. I promise.

It's never been anything like this. It took us forever to get the abstractions right. But we did, and it's great, and it works really well. And if you want to try it, you can literally go download it and flash it right now. It works great. Okay. Post router. And I don't even want to say post because we're already working on it. But we need to integrate agenic support into the StartOS and router experience. So StartOS is going to get a big upgrade soon where you will have a personal assistant on the server, similar to how personal assistants have always been on computers, right? Again, like Siri, except powered by AI and with access to everything. People are already doing this with OpenClaw. So, you know, not only on their

startOS servers or other devices, but just on like random, you know, uh, mini PCs that they buy. They're just installing open claw. The problem with open claw is that one, it is a fucking wild west experiment is riddled with security vulnerabilities. It's iterating at a pace that nobody could possibly audit anything related to it. And everyone is just plugging it into LLM providers anyway. Again, it's the most reckless, insecure, unprivate thing I've ever seen.

but the idea is right, right? The idea is correct. And so we're going to do it right. We're going to do it in a secure, quarantined, private way, but without sacrificing most of the power. It's still going to have root access. So what we're doing, and this might get into a little bit more advanced terminology, but for those who who have it, they'll appreciate it. We are building a start CLA, MCP server. So we're basically building abstractions on top of StartCLI. StartCli is the command line tool for administering a StartOS server. What we're doing is building LLM friendly and conforming abstractions on top of StartCLI to accomplish common user flows. So for example, if a user says something like, I want my friends and family to be able to upload pictures to my server with

their own accounts such that we can create albums and share photos with each other. And I want this to be done in a completely private way without ever touching the public internet, but I don't want Tor to slow us down. So we need some sort of other networking solution. They should be able to fucking type that in to a prompt and then have all the things that need to be installed, configured, and run on the server done on their behalf and with their permission in a transparent, safe way. That's fucking cool. So what's happening with StartOS is in the not too distant future, there will actually be two

user interfaces. You will have the graphical user interface, which is for you to navigate, point, click, and do whatever you want, which is way beyond any kind of sovereign computing experience that's ever existed. But you'll also just be able to open a side drawer and talk to your agent and then on the other left side of the screen actually watch the results. So StartOS is a live experience. Everything is piped via web sockets. There's no like stale state that can show on the screen.

So if you tell your agent, install Bitcoin, make it pruned and hook it up on ClearNet for the peer-to-peer interface, enter. It might ask you a couple clarifying questions. to make sure that it knows exactly what you want. But then you'll actually watch on the left side of the screen as Bitcoin gets installed, gets configured, gets this, and starts sinking, and you're just like, you didn't have to even do anything, except express your will, okay?

The old end game of all AI tech, if it is to be in service of a prosperous, dignified future for humanity, is essentially the following. humans should be able to in a variety of ways express their desire, their will, and then have it be done by a robot. That is effectively the end game of this. No matter where I am in the world, no matter what I'm doing, I should just be able to be like, man, I want a hamburger. And it like comes flying in through the window on a drone, right?

and just gets shoved into my mouth and is paid for and everything with nothing, right? Like, in other words, the marginal effort and cost to achieve anything will shrink to zero. That is the end game, okay? There's an enormous amount of nuance in all of that and process between now and then.

But that's where this goes. This goes to humans expressing desire, expressing will, and it being fulfilled by superintelligence and robots. And you were telling me, invented the green lantern ring of modern computing. You are limited only by the will of the user. No, no, no. I'm not going to do the robot that flies in and shoves a hamburger in your mouth. We are one piece of the puzzle on the road to that. But my point about what we're doing, and I love

the reference, but my point about what we're doing is we're trying to make that future possible without open AI, the government, and the multinational conglomerate hamburger company being involved. Okay? I love it. Your desire should be known only to the people who would need to know it in order to fulfill it, right? Which again will ultimately be no people at all.

It'll just be some local bots managing a farm that sends the best beef or something. But like, it's the exposure of data only to the, at the data and at the levels that absolutely are necessary to fulfill the desire and always with your permission and under your control. Either way, that future or some inadequate part of that future I just painted, because we're never actually going to get there. I'm saying this is the theoretical endgame, right? The reality will always be messier than that. But the point is that that future is coming. If we survive the transition at all,

all, that future is coming, right? The question before us is, will it be a centralized, intermediated, custodied, and controlled future, right? Will the architecture of that future be that, or will it be decentralized, disintermediated, non-custodial? Either way, you express a will and it gets fulfilled. But in one model, all the people who you don't want to know about that, know about it. They can stop it. They can cut you off. They can charge you money and they will eventually get hacked by other super AIs, right? So it's all the same problems as cloud computing, just more extreme. In the future that I want to see, it's the same course of events, but nobody can stop you. Nobody can tell you no. Nobody can even know that it's happening unless you want them to.

It's a dignified, free and prosperous future. Either way, humans expressing design. and those desires getting fulfilled by super intelligence and robots, that's happening either way, which is a good thing. But if it happens in the centralized controlled way, I actually fear for the survival of the species because nobody will tolerate that at the extreme. And therefore, you have violence. And again, in the era of weapons of mass destruction, we need a foundation that does not incentivize violence. And that is a decentralized, sovereign architect. and nobody's building it. None of the big guys, I mean, right? There's no big guys actively trying to cut themselves out of the future. They're all trying to cling to the future

where they end up being the monopoly that controls all the world's information. None of them are being like, let's just do the right thing for humanity. Right. And again, we're not doing that of pure altruism either. My goal here is not just like, well, I don't want to make money and, you know, I just want freedom for humanity. I do want that, but I want to make a lot of money, too. I see opportunity here. When all the big guys with all the billions and trillions of dollars are pursuing the hegemonic, I control the world's information approach,

there's opportunity for a startup to pursue the antithesis of that and garner market share from people who agree with the vision and who want to be sovereign and independent and free and who want the same for their grandchildren. And what's cool is that, like, Apple and Apple might be best of breed, but even them and all the rest, they're never going to do what we're doing. We're not competing with them within the sovereign computing paradigm. What we're witnessing is the emergence of cloud computing versus sovereign computing,

but they'll never jump ship. They will never come out with sovereign computers. Apple will never sell a camera or a device that doesn't plug into Apple. And so we have a chance. We have a fighting chance here. So long as people understand that Apple doesn't sell privacy. They sell confidentiality. They know everything about you. They're just promising not to tell. That's confidentiality. That is not privacy. And as long as people keep thinking it's privacy, we're going to get crushed because they're stealing our word. They're taking the word and they're perverting it. And so long as people don't see through that, smoke, screen, we actually won't get market share. We will not survive. So education matters. If you're a homeowner and you want more Bitcoin, Horizon just opened a whole new door. Most of your wealth is locked in your house, growing a few percent per year, while Bitcoin has historically grown far faster. Horizon lets you convert a portion of your home's future value

directly into Bitcoin without taking a loan. No monthly payments, no interest, no term limits. You stay in your home and you fully control and custody your Bitcoin. There's no force liquidation, no margin calls, and no income requirement. You unlock tax-free cash today, and Horizon simply receives an agreed-upon share of your home's future appreciation when you sell or refinance. Hundreds of homeowners are already doing this in over 21 states. So, go to Joinhorizon.com and see your Bitcoin potential. in about 90 seconds. Turn your home into a personal Bitcoin treasury with Horizon.

If you're the kind of bitconer who doesn't trust convenience, cloud backups, or leaving anything to chance, Dirty Man Safe was built for you. This underground physical storage is for the most paranoid self-custy setups designed to protect and hide seed phrases, hardware, critical documents, and more. Rugged, weather-resistant and built-to-disappear, Dirty Man Safe, as a lot of. last line failsafe to your Bitcoin security model. Head to Dirtymansafe.com and use code BTC sessions for 10% off. Great.

I'm just completely blown away even just thinking about this because for the longest time, I talked about like at some point in time, something will come along and like overthrow Google. And people just think that I'm crazy. It's such a recency bias that they can't remember Yahoo or anything that really came before it. And I'm thinking about all the inference data that'd be going to Anthropic and like, there's a contender.

There's a real contender to overthrow Google right there as well. Yeah, with a worst Google. Yes. And what the, what the, the massive risk to software as a service too, right? In the sense that if software is your entire business, well, that can be spun up by almost anybody at any point in time with a couple prompts. It really demonetizes. It goes to the cost of production of software, which I think just has profound implications. Now, I do want to pivot a little bit because I want to dive into Bitcoin. Before I do, Gary, did you have any questions about what, you know, the beautiful giant rant

that we just got from Matt there? Well, first of all, Nathan, I want you to know that I understood every single word that Matt said, I'll explain it to you after the show if you have any difficulty. I know, I'm just, your passion for privacy is amazing. Just I'm, I'm hyped up. I'm, I'm aroused right now. So I'm going to leave this to you. I'm going to adjust my pants.

Nathan, give me what you got. Jesus. All right, beautiful. With that beautiful segue, I want to pivot into Matt, your kind of view of the state of Bitcoin right now. And specifically with regards to, you know, potential soft forks and policy, I want to kind of get your 30,000 foot view of where you think we are and where you think we're going.

And I want to kick it off too, because I actually saw a post, I believe it was from you, not from the start nine handle. That thought was actually beautiful with regards to core. It said, quote, I think the answer is they're developing for people who want to use the network instead of people who want to create the network. And I thought that that was a really interesting distinction with regards to the direction of where Bitcoin development is headed. Could you break down what you mean for me by use the network versus create and then we'll kind of go into your view of Bitcoin right now? Yeah, so it's actually a pretty clear distinction. People who create the network are people who are running their own nodes and using them for transactions, using them to verify inbound transactions and to send outbound transactions and to choose the consensus rules and configurations that align with their vision of the technology. those are people creating the network, broadly referred to as, you know, node runners, node operators, right? Yes, a lot of people running nodes are not actually using them, and so they're not included in that, right?

They're might be well intended, but they actually don't matter insofar as, you know, any kind of fork or otherwise consideration is concerned. People who want to use the network, right, are really, people who are not running the nodes. but using the nodes to do other things. And that's not an insult. I'm not like saying that these people are bad. Like I want people to use the Bitcoin network for, you know, anchoring in and layer twos and even, you know, in certain circumstances where it makes sense,

like even, you know, imperfect custodial type setups, right? If it's small value and you trust the person doing the custody and it's a federated system, for example, like, you know, a fetament or something like that. So what it occurred to me before I wrote that, that was a while ago actually,

was that it felt like the core developers were building the node implementation for the people who were going to use the network. And it's not that they shouldn't be considered at all, but it's that they were taking priority over the people who were going to run the notes, create the network.

And I think it is more appropriate for people who write software or build products of any kind to build it for their customers. The people who are using nodes, that's the wrong phrasing because that could be misinterpreted, the people who are not running nodes, but who are using the Bitcoin network to accomplish things

are not actually the customers of protocol developers. They're a layer removed, right? They're the customers of the people running the nodes. And so if you're writing code like I do, I write it for the people who are running StartOS, right? Not for the people who might use StartOS. Now, it's not to say that I don't think about that. you want to handle as many downstream layers as you can, but never, the more downstream layer

should never be sacrificed, should never take priority over the direct relationship. And so I think that's kind of where this came from. There's a lot of people who are running nodes were like, we want this and we don't want that. And they were like, we're building for others, the people who are going to use your node to do stuff, not for you. And I think that that is the essence of why this happened. Beautiful. I think it's a wonderful explanation. And kind of continuing from there then too.

So, of course, there is the ongoing drama and debate. And I think it's actually important that we do have these sort of discussions. And we do try and hash out in public where we think the direction of the network should go, what changes we think we should make. And so going into basically even the latest software proposal with Bill, Bip 110. I'm curious kind of broadstructs your overview of the situation and just even how you're starting to think about it or parse these things together. Good, bad, ugly. Where are we kind of, where are we thinking? So, um, ugly. Ugly. All right. Right. Not good, not good, not bad. And to be

honest, ugly is even an objective term. We'll just call it Bessie or chaos or Bitcoin governance, So I think what's happening with BIP 110 right now and the division, derision that it has caused in the community is par for the course. I think that this is the nature of Bitcoin governance and that that is a good thing. I think you predicted this as well too. I did. I gave a speech at Bitcoin 2025 called Power to the Nodes where I did a couple of things. One, I gave a broad overview of Bitcoin governance by ID. identifying the various constituencies, call them branches of government, and the powers that they wield in Bitcoin's governance. So you have these various constituencies and each has a power and looked at holistically, there is a very healthy balance of power in Bitcoin that prevents any one constituency or branch from running away and taking over. And that means change is hard, right? Inevitably, that means that making any change whatsoever is going to be very difficult because you have to garner consensus. You have to get all these different constituencies to come together, at least a large portion thereof to make any kind of a change. And that's always going to be a messy process, right? You could imagine a theoretical

situation where there's like a clear and present threat to Bitcoin that is so obvious that just everyone agrees that it's a clear and immediate threat and that we need to do something right away to fix it and consensus happens instantaneously. In reality, that isn't how things usually are, especially in a fog of war type scenario where literally the ground is being pulled out from under humanity when it comes to the emergence of AI and robotics and space travel. It's like good luck keeping up with anything these days with how fast it's changing. And in that kind of environment, discerning the truth is a hero's journey. And so almost nothing is going to be obvious.

I mean, you still have people out there like wondering if Epstein was a bad guy or so. You know what I mean? No matter how blatantly things can smack you in the face. it's still very hard to accept it because the tools of deception are so powerful. You don't know what videos you watch online are real anymore. So in reality, every proposal to Bitcoin is going to be met with massive skepticism, total fog of war, nobody's going to agree. And so the process of consensus is inherently messy.

It's ugly. And I don't think that there is anything wrong with the way that BIP 110 has come about. In fact, I predicted that it would come about. And because I predicted it, everyone thinks that I've caused it or in part of it or et cetera. I predicted it because I like doing analysis of political systems in particular and technological systems. And I said that, let me back up. policy came first, okay? There was a disagreement on Bitcoin's policy layer

where, broadly speaking, you had two camps of people. There's a lot of nuance here, which is policy, transaction and mempool relay policy, is a good and efficacious place to combat spam. And spam should be combated. Okay, so there's a couple kind of premises on that side of the equation, which is that, one, spam is bad. That is rarely contentious, right?

That is a statement that almost everyone can get on board with. Number two, spam should be combated, right? That is more contentious, but only because of number three, which is how do you fight spam? And in particular, does MMPL transaction relay policy meaningfully, in any way, deter spam on Bitcoin? And that is where the major disagreement was, was, does this work? And there was a huge, nasty kind of, you know, back and forth about whether filters work or not.

And start nine and myself were very vocal about this. because like others, it seems clear to me that this does work, right? Empirically, historically, and even just logically, it works. And that is all the context considered, okay? All the, you know, information theory, infinite ways of embedding data, cat and mouse forever. You can never win the, all of it. It is not a naive statement, is a fully informed, well-thought-through statement of this does have an impact on what results in blocks. Right? It doesn't eliminate it. Nobody's ever made that claim. It has a cultural and practical impact on how people use Bitcoin.

And that is something that I was in Start 9 was very comfortable with being vocal about. Nowhere in any of that vocalization did we say, you should ditch core and run knots. Nowhere, ever. What I was trying to do, and in fact opened multiple back channels in an attempt to do so, was convinced some of the core developers that they should not, be removing policy filters, but in fact should be implementing new ones, like Knott's did, making the node more configurable. Okay?

And from my broad political governance perspective, and I talked about this in my speech at Bitcoin as well, the more configuration options there are for node operators, the more power political power is drawn away from protocol developers and towards node operators and their representatives, which is start nine, umberal, my node, right? What we serve as are almost political parties, right? The analogy doesn't hold one to one. It's probably the wrong framing, but as representatives, elected representatives. Okay. So for instance,

when you install Bitcoin onto your server, it's going to have defaults. It has to have defaults. There's no way for it not to have defaults. For example, even if you presented the user with a form that wasn't filled out at all, it's still a default. That in itself is a default.

Not setting defaults as a default. There's no way out of the, the power of defaults trap. Who sets the defaults matters because most users don't change the defaults. So who sets them matters. And it is our stance that protocol developers already have enormous power in Bitcoin, enormous,

and that it is healthy for the network, for them not to have the power of defaults. The power of defaults instead should exist with us. Umbrell, my node, right? These intermediate platforms where people are actually using, that people are actually using, to install their node. This doesn't give us enormous power over Bitcoin, the ability to conquer it.

What it does is it balances power in the network. This isn't me being power hungry. This is me wanting balance of power in Bitcoin and building a company because I do want some power. I think power is good. I think it's important to have power and to have power be distributed. And I think that this is an appropriate power. And in fact, it is an undeniable power, right?

Because we do have it. Regardless of what the core developers, the Nauts developers, or any other developers of Bitcoin do with their defaults, we still can override them, right? We could make Nots the default. So even the implementation, even which implementation is the one that's put in, front of the user as Bitcoin is a way of setting defaults. So my observation is that we have this

power and that it should be encouraged, welcomed, acknowledged by core developers, by all protocol developers, and that they should relinquish their desire to make people run their node in the way that they think. Ultimately, the user will decide and the user will choose start OS or Umbral or MyNode based on what we believe and what we say, right? We're out there basically being like, here's what we think about Bitcoin, here's the defaults that we set. And if people agree with that, they'll choose us, download it. And even then they can change the defaults.

It's not like they're locked in, right? It's just this starting point. And so we're against protocol developers removing configuration options because what that does is it increases their power and decrease. the power of the node operators and us as their representatives. And I think that's unhealthy for Bitcoin. Right? I think too much power in any one branch is unhealthy.

So what I said in that speech was that if the core developers did not listen to their users, again, not the people using the network, the people running Bitcoin core, the software on their own server. If they didn't listen to their customer, then their customer would abandon them. Their customer would be angry. They would feel unheard. They would be discontent. And they would move to another implementation that gave them the tools that they wanted. This wasn't me telling everyone to go download knots. I was making an objective observation that if you do not service your customers,

your customers will shop elsewhere. And so we witnessed this movement. And then what I said was is that if they doubled down on it, if they refused to listen to their customers and give the customers the configurations that they wanted, that eventually the node operators would wield their second political power. Node operators have two powers in Bitcoin. They have the power of policy, relay and mempool policy, and they have the power of consensus. And in the power, in the world of consensus, the nodes have the Trump card. They are the ultimate power. And I don't just mean people running a node.

I mean economic activity here, okay? I mean like actually large amounts of economic activity running through their nodes. We're talking about like wallet developers and the nodes that they run on behalf of their users. We're talking about exchanges and the nodes that they run. Pool operators are huge. businesses, right, anyone who's running an economically meaningful node, and the plebs in aggregate are an economically meaningful node, right? Like each individual might not be, but you add up a few thousand pleb nodes that are actually being used for transactions. That is a serious economic

force. So ultimately, that is who decides what Bitcoin is. It is the economic nodes. We witnessed this in 2017. We saw what happens when the, the economic nodes get together and decide that they are going to take a certain course, a certain fork, and everyone else gets dragged along, kicking and screaming if need be, but they all get dragged along. And so my observation was that if you didn't give the node operators, not just the Pleb nodes, but any nodes, what they wanted, that eventually they would pull this Trump card and soft fork.

Soft fork in particular, because a soft fork is a coup, right, it is a takeover attempt, is an attempt to change the existing network, whereas a hard fork is the creation of a new network, which I don't think anyone wants. They want to take over the current network to put it back on a track that they believe is the correct track. So I predicted that this would happen. I was like, if you don't do this,

there's going to be a contentious soft fork. And as the year drag on, that became increasingly likely. And in an interview with Isabel, I was like, it's definitely happening. There's definitely a fork coming. And again, I took huge amounts of flack for saying that because everyone's like, why are you forking? Why are you telling everyone to fork? It's impossible to identify something without being attributed to its creation and support, right? And again, I'm not saying I don't support. I'm not saying I do support it as a personal decision that I'm going to make in a company decision that we will make in the coming

months. In the meantime, StartOS, Start 9's technologies have a 100% neutral approach to what people do with their own servers. As you've seen throughout this conversation, my essence, the baseline principle is sovereign computing, user control. Start 9 is never going to come in and tell somebody, don't run this software, do run that software. If we have determined that something is quite literally malware, we will cite that, we will prevent it from being on our market, place. And despite what Luke says about Core V30, I do not think it is malware. You know, I agree with a lot of things that Luke says. I agree with some things that some core developers say. I've had private conversations with many of them. I do not think there are bad actors all over the place.

I think there are understandable differences here. And, and yeah. So, yeah, there's a few things I want to unpack there, Matt. One, first and foremost, too, even as you're going to think, the idea of configurability on like the protocol level. The other benefit is that this is my initial instinct, the more configurable core is, the less politicized it becomes. Because essentially they're just obfuscating any choice decision to start nine, to the users, to whoever it actually has to go in and set those policies. I think it was a better approach overall.

The other thing I want to kind of tease out a little bit here too, which again, not necessarily coming down on pro for or against, is one the effectiveness of the proposal. If you think that, because I think everyone realizes that it's impossible to completely stop spam. And I think that's not necessarily the discussion that's being had. I think it's worth teasing apart a little bit. And then the other one I want to just pose is one of my biggest, I think one of my biggest concerns currently around BIP 110 has almost very little to do with the protocol itself. But the success or failure of it.

And the reason being is that I would very much be in the camp that I'm not in favor of span. And I think that policy was a good place to put that. and I'd like to have attempts to mitigate or filter or deter spam. And one of my biggest worries is that should this soft fork fail, and particularly like fail hard, that it will destroy the will to even try to deal with spam anymore. That would kind of, that conversation will be put aside for a while. We won't be trying to come up with solutions or trying to develop on it further.

My biggest worry is that we lose any momentum or any real concern for dealing with spam. But again, I open, I'm not taking aside one way or the other, and I very well could be wrong. So, Matt, I'm just curious your thoughts on that. Yeah, it's a very interesting observation that I don't have a bulletproof answer for. I don't know. It's hard to predict what the consequences will be either way. Yeah.

Right? If this fork succeeds, what happens, right? most of core resigns, quits, switches to, you know, I don't know, what happens with companies and Citria. It's a different future, right, for Bitcoin, I think. I think it's a dramatically different future. And if it fails, what do all the BIP 110 ardent supporters do? Are they rage quitting? Are they being like, oh, it was a good shot at, you know, a coup and we failed.

So we're just going to run core 31 or whatever. I don't know. But it feels like a different future for Bitcoin. Maybe not as dramatically different. It's a little bit more status quo, I think. But definitely different because we've split culturally, right? This might be more akin to like what happened in 2017, where a lot of the. prior, you know, people who were involved sort of went away, right? So it's hard to say, specifically about spam. I, like you, feel that spam is best fought in policy. In fact, Aiden, Start 9's CTO, put up a very unpolished kind of, you know, BIP. It probably shouldn't have been a bib, it should have been an email. outlining the idea. But to be fair, at the time, the protocol for like introducing BIPs and all that was undergoing some pretty aggressive changes and it wasn't incredibly clear what the process

was supposed to be. But regardless, I personally thought the idea had enormous merit, as did others. And it was pretty readily shot down for a lot of bad reasons, in my opinion, which was this idea of dynamic filter. Okay, so rather than protocol developers, whether it's Core or Nots, sort of hard coding what policies are available for a transaction being accepted to the Mempool or being relayed, rather than hard coding those. And Core has these options and Knott's has these options. And to give somebody a new option or to deprecate an old option requires like a code release. You have to put out a new version of core or of nots to implement a new policy. That itself is even too slow, right?

Because if you have to change consensus rules to fight a new kind of spam, you're dead in the water. You cannot fight spam in consensus in the long run, right? Because it's just too slow, as we mentioned before. You're never going to get consensus for almost anything, let alone quickly to deter a spam attack. policy is a lot easier because all you need is a software update. You just need to bump out a new version of Bitcoin Core and everyone has the new filter. But even then, it relies on a super majority of the network adopting the new policy for it to be effective, right? Because filters are only effective at 70, 80% adoption. They're only really effective at like 90% adoption. Because

otherwise transactions still find a way through the network to a minor. And so we thought, Aiden and I got together and Dot, and he have ultimately put this together, that you could ship filters externally to the software. You don't need to hard code them in the software.

There's precedent for this. It's in virus, antivirus. I was going to say, yeah. Right? So, whenever a new form of spam emerges, someone, anyone, could write a script to identify that type of transaction and drop it.

And then they publish this script. These things are simple. You don't need like 18 core developers to review them for soundness or security. There would emerge a very standardized, easy way to evaluate the efficacy and safety of a script and I'm sure even on top of that, humans and AIs would be reviewing it for soundness, right? That is not an important consideration given the popularity of Bitcoin and the tools available to us, the reviewing of these scripts.

Who you install them from, trust will play a role where it's going to be, there's going to be some central provider that says, oh, here's a new policy filter and, you know, everyone updates it. that would be dangerous. So there would need to emerge a means of trusting the distributors of these scripts to make sure that you're not installing a malicious binary, et cetera. These are all technical considerations and we probably shouldn't even get into them right now. The idea, broadly speaking, is that anyone anywhere could identify a type of transaction that they think shouldn't propagate through the network. That should be a second-class citizen. It's not excluded from blocks. It's not against consensus. It just doesn't get priority relaying through the primary peer-to-peer network. That type of transaction needs to incur added cost and effort to get to a minor, either through preferential

relay networks or direct-to-minor relationships and APIs. It needs to incur this added cost because it's not how we think everyone should be primarily using Bitcoin. And they can publish this. And if everyone agrees, again, 80, 90% of node operators are paying attention and explicitly choose opt-in to downloading and installing this new filter onto their node, then it would be effective. And overnight, you could identify a new type of spam filter, and it could be vetted, published and node operators all around the world could choose to side load it or not. And if and only if it garnered immediate, or I shouldn't say immediate, but if it garnered supermajority adoption, would it become effective at all? So what does that mean in practice? Do you see a world where

every day node operators are going to wake up and check the latest bulletin for new types of spam that are hitting Bitcoin and vet which scripts they should install from whom and then and download them. No, of course not. So what are we really talking about here? What is this actually solved for? It solves for a massive attack. Okay?

It is a failsafe against a black swan attack of spam, right? So for like, I don't just a big fee spike. Like I'm reminded of when like the Ornale's things kicked off. You see this massive spike in fees. All of a sudden, everyone's paying attention. Hey, by the way, guys, we have a solution for this if you'd like. it is a, it does two things. One, it allows for slow, gradual adoption without needing software updates, by the way, right? You don't need to update your note. It allows for slow gradual adoption of annoyances.

The transactions that are annoying, but not killing the network. And gradually, if you check in once a month or once a year, you know, a certain policy might garner adoption and eventually become efficacious. And the very threat of that happening might deter the spam in the first place. The very existence of this mechanism might make spammers think twice before they write a protocol and launch it on Bitcoin because they know that they could get thwarted if it ever became enough of a nuisance. That's one track, is this gradual adoption of a filter to deter some annoying thing that everyone, nobody wants. Number two is an actual attack, right? Not somebody trying to build a, you know, some dorky little NFT business, but somebody like objectively just attacking the network with spam, making it effectively

unusable for lightning, for on-chain transactions, just attacking it relentlessly for days or weeks on end to make Bitcoin to grind it to a halt. Wouldn't it be great if node operators could just recognize that this was an attack and without even needing to update the software, just push a button and it just goes away. It just gets dropped from the P2P network. I got to throw in here, Matt, because something is jumping out at me as well, too. And again, this is a very interesting idea that I'm going to have to really sit down and think about a bit more. And I'd be curious to know also why it was basically rejected or no one really had interested in it at the time that it was put up. But it seems like a natural progression

of Bitcoin for me. And what I'm referring to is that, like, originally the client was everything, right? The client was your verification. It was mining. It was your wallet. all in one. And then as Bitcoin has grown and evolved and matured, we've seen that kind of specialization and it break out into pieces. Like mining is its own thing now. My verification, my note is its own thing now. My wallet is I'm not using Core's wallet. I'm using mine own, my own specialty hardware. So abstracting out like policy and filters to another than specialization away from the main consensus protocol, that just seems like that makes sense to me. But there probably is something I'm missing. But on its face, like that makes sense.

That was part of the proposal was this whole, like, was this whole, like, like at the time, a lot of the core developers were like, we don't want to deal with policy. Like, it's stupid. It doesn't work. And we were like, this allows you to never think about it again, ever. You carve it off. And all the policy in Bitcoin, all the code governing policy in Bitcoin basically becomes like a function, right, which is ingest the scripts and run them one at a time. Like, Aiden laid out a very technically detailed explanation of how this would work. Like he and I spent, you know, quite a bit in like policy.cpp. We were in validation.cpp.

We were in the Bitcoin Core codebase figuring out how policy was done and came up with a scheme that would allow it to be effectively ripped out. It's not that it would be trivially done. There's just some complexity there and interplay with other modules. But it could be done with some effort and we were willing to do it. Aiden was willing to actually code this up and definitely has the ability to do it. and then turn it into basically a dynamic module where policy is handled at the social and cultural layer because it has nothing to do with consensus. It has nothing to do with core developers or protocol developers at all, right?

This is all we were asking for or recommending or suggesting was that the protocol had a mechanism by which side-loaded scripts could be injected and run. And if any of them failed, the transaction would be dropped. It was actually pretty elegantly simple, right? Basically, it takes in a list of scripts. It didn't propose doing it in JavaScript, which would be fine. Scripting language. Lua would be fine as well.

There was a prototype actually developed for a Bitcoin Plus Plus Conference where a guy wrote this very rudimentary version of this using Lua scripts, and it worked well. So anyway, that is all the protocol I'd have to do is accommodate the ability to loop through a series of scripts. And so long as they all return good to go, the transaction gets accepted into the Memple. If any of them fail, it gets dropped and states the reason why in the lots. And that's it. And core developers get to wipe their hands of policy forever.

And node operators get everything they wanted. It's like the ultimate knots, right? It's like any filter we want ever can we can just load in. even if Luke doesn't like it, right, there's no central control or authority over policy. Policy is an emergent property of cultural consensus, right? Separate from Bitcoin's consensus layer, which is written and governed by forking and by protocol developers. So I thought it was clean and elegant, and I'm sure there were many things that needed to be

thought through to make it, you know, very practical. and but it was there were a few people that saw the the intent of it and applauded, you know, the intent. In fact, I had a few private phone calls with X and current core developers around it. And ultimately, their stance was that users don't manage their node, right? It was quite condescending, actually, to be honest. There was goodwill, like they were willing to engage. But it was very dismissive of just like, come on. Like people, these idiots don't know how to like load scripts.

They're just going to get tricked into loading things that ruin their node. And they effectively can't be trusted with the power of policy. And so, yes, these developers want out of the policy world, but they want to accomplish that by eliminating policy. They want Bitcoin to be consensus, period. There is no policy layer. And that's how they want to get out of policy, rather than delegating policy decisions to the network in a way that the network itself can arrive at its own policy consensus. They're just trying to eliminate policy, which is effectively eliminating configuration.

It's eliminating choice. And that's where a lot of this pushback is coming from. And they're not just doing it because they don't want to deal with it. They're doing it because they don't think it's effective. They're just like, why do you want a policy layer? All it does is cause problems. It doesn't stop spam, because spam can be encoded in n number of ways, and it ultimately causes problems. It causes mempool blindness. You can't predict what's going to be in the next block. It's going to cause side-loading to minors, which is going to increase minor centralization. We heard it all. I get it all. It's a debate. Those are important arguments to make and to be considered. But at the end of all of it, I and many others, Still, despite all of that, came to the conclusion that this is the right tradeoff, that policy is a good and efficacious place to deter spam.

If you get rid of it, you are defenseless against spam. If you are defenseless against spam, the network will be overrun with spam, which poses an existential threat to Bitcoin's future. That is my essential analysis of all of this. And, you know, when it came to these fears, and I expounded on this quite extensively in my interview with Isabel in October and November, these fears of like mempool blindness and minor centralization, right, they're all easily dealt with in a policy-heavy world in the following way. Let's say that some new type of transaction in order to enable some new VM-term,

set up side chain or NFT gimmick, whatever it is, okay, it emerges. And everyone's annoyed by it. And so they gradually adopt a filter to deter it. And so the things start happening, right? Turns out that this isn't just some shady, temporary, you know, spammy thing, but it's backed by big money. and there's real economic demand for this type of transaction, right? Maybe not by like your hardcore bitcoins,

but some new arrivals to the network that want to use it for something else. And there's real economic demand. There's real money on both sides of the supply and demand. Let's say that happens. What are they going to do? They start building preferential relay networks like Libre Relay. They start standing up their own notes to circumvent the P2P network. They start creating direct relationships with miners.

and paying extra in fees to get their transactions side-loaded into a block. So they're incurring cost. They're like, all right, we can't use the free, cheap, and easy P-to-P network. We'll build our own relay networks. We'll create our own relationships with miners, and we'll bypass the P-to-P network. Yes, it's at a cost, right? It takes effort. It's less private.

They have to do KYC arrangements with the miners, which means the transactions are probably going to be vetted, inspected, because the miners don't want to be held liable for something illegal, getting embedded, et cetera. There's costs all around. There's risk and there's costs. But they're willing to incur it because there's massive money to be made. There's massive money behind it and massive money to be made. Fine.

So guess what happens? Your normal nodes, the nodes that are filtering these transactions out, start experiencing consequences. They stop being able to predict fees accurately. They stop being able to see what's going to be mined in the next block because it's all being side-loaded to miners. Miners, centralized miners that are willing to build and develop these relationships, start getting more and more powerful, squeezing out smaller miners. Let's say all this starts to happen. What are the node operators going to do? Maybe they'll admit to themselves that they were wrong,

that this isn't spam, that this is a legitimate high-demand use case of Bitcoin, and they can drop the filter. They're not locked into this decision. It's not like getting a tattoo. It's like getting a piercing. You can just take it out if it's causing problems, right? So in other words, what you can do is you can view filters as a vetting process, right? It's not that this permanent decision, like we hate that type of transaction. It's never going to be allowed on Bitcoin. That's consensus. That's what BIP 110 is doing, is it's being like, okay, you didn't give us the polite way of deterring something. So now we're going to eliminate it. Not all things forever, but this particular thing is going to be eliminated.

That's the nuclear option. The better option is to deter it, to make it prove its economic viability by going through these hoops, by circumventing the P-to-P network, by developing the relationships with the miners, because you know what? 99% of these things won't do that. Because there is no money behind them. There is no economic demand for them. The only reason there's money behind them, VC money behind them, is because,

because they think that they can create the economic demand to dump it on everyone's head and then run with the money. That's why there's VC money behind it. It's not because there's real economic demand. It's because there's contrived, scammy economic demand that they can pump up and then dump. And so that wouldn't happen. You wouldn't experience these artificial types of projects if they knew they were going to have to fight for it over a long run. And at the end of that road, because here's the other option the nodes have, right? I just painted a picture of like, hey, they deterred it, it started causing problems, minor centralization, all the things that were predicted. They have another option.

They can either drop the filter or fork. They retain that right, that power to fork. And so who, what scammer, what spammer out there? is going to try knowing that that's their future. Okay, we're going to do some spammy shit. We're going to get deterred. We're going to have to go build preferential networks, which themselves can be deterred through things like garbage man, spoofed notes, right?

So now they're in this never-ending battle to get their spam to minors. Okay, now we've got to go build relationships with minors. Okay, we have to pay extra fees. We have to do this. We have to get the lawyers involved. We have to build all this stuff. Okay, fine. And at the end of the day, if they're really not wanted there, the nodes can just be like, all that effort and money you just poured into this entire paradigm, we're going to fork you out in consensus. You just wasted your money. The threat is all you need to deter the act and the ability to act on that threat. It's not an empty threat. So the nodes control the network. They either say, we were wrong. This is a legit.

legitimate use of Bitcoin, let's drop the filter and welcome them back to the network. They can ditch their relationships with the miners. They can ditch their preferential relay networks. They can enjoy the low fees and fast propagation, welcome them back. Or fuck them, rug the spammers, fork off. The nodes can do whatever they want. Okay. And so that is kind of what you're seeing right now. That's what's happening with this BIP 110 situation is that some people, some node operators, have decided that it's time to fork because the previous attempts to deter something failed. It is causing problems. And it's time to rug them in a display of power so that they don't do it again. And that is the essence of the BIP 110 fork, I think. It's not. to eliminate spam forever. It's to punch them in the face and make them remember what's going to happen to them and their money if they try this again. I'm not promoting that right now. I'm not promoting that right now, okay? I just need to say that a couple of times.

For everyone who's listening, he's not promoting that right now. Not promoting it either. I'm calling it like I see it. I'm telling you what people are doing and why they're doing it. Start 9 will enable them to do that. Start 9 will enable them to fight against that. Start 9 and Start OS will remain neutral as a technology. I like to perform analysis and I also have opinions, but that's not important. That's the least important thing of everything I could possibly say is what I think about this. I know, Matt, you're pretty good prognosticator.

The only thing I just want to add to, which again, just again, not, I'm not promoting one way or the other two. but it's another reason that I want these things to be more in the policy, just to have that flexibility for the future too. Like, I would hate to see that we do get like a consensus. We make, like, say we make operative a consensus at like 83 bytes. And we discover later on down the road that we actually have a very legitimate use case for that that we really need for some unforeseen black swan. And now we have to hard fork to throw it back in.

Yeah. It sucks. It sucks. I understand that BIP 110 is temporary. I'm well aware of that. But I'm just saying like this is why I think these things are at least ideally in an ideal world. sitting in the policy position. Yeah, you know, something that gets lost and that, quite frankly, it's been a while since

I've been talking about filters, right? My mind has been largely on AI and sovereign computing these days. But when it comes to this whole filter debate, I think it's really, really important to identify, quote, filters for what they really are. Okay, what are they fundamentally doing? Because they're not eliminating anything, right? We've established that and everyone agrees on that. Fundamentally, what a policy filter, Mempool Relay transaction policy filter does is it limits

the instance rate of a particular type of transaction. They are rate limiters, right? For the technical people out there, these are extremely important features of any computing system is to limit the instance rate of particular types of functions or transactions to avoid denial of service, okay? So what a filter is, is everyone agreeing, because again, the rate limiter is only effective if everyone agrees. It is everyone agreeing that a particular type of transaction, if observed in excess, would cause

problems for the network. Therefore, we need to limit its instance rate. We need to rate limit that type of transaction so that it doesn't cause problems. It doesn't eliminate the transaction. limits it, not eliminates it. And so that is a very, very important feature. And if nodes do not have the ability to rate limit transactions, then we are effectively tearing down an extremely important and agreed upon aspect of all computing infrastructure, right? Rate limiters are important and essential. And so we shouldn't be doing away with them. Beautiful, Matt, thank you so

much for spending your time with us today. Go ahead, tell everybody where they can find you, find Start9, all that good stuff. Start9.com. We're on Twitter at Start9 Labs. I'm underscore Matt Hill underscore on X. I guess it's X, not Twitter. How many people still do that? Everyone does that, doesn't it?

I still say Twitter. I always say Twitter. Yeah, it's crazy. But yeah, not hard to find. We have a private community server that many people who buy our products or buy access to it can join. The reason they have to buy it

is to deter spam and skin. scammers from getting in there. If you don't put up a paywall, you just invite spam. So we did. It's reasonable. And it's great. It's a great community. And we're extremely helpful. And our products are going to be coming out, rolling out here over the next few months, some new and exciting things. So stay tuned. And that was so nice of you to promise us some free samples. We really appreciate it. Thanks, man. We'll take care of you. Awesome. Thank you, sir. It is a free operating system. So, three samples for life.

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 Sessions. 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, 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 btcmentor.io and book a call with us today. If you enjoyed today's episode, please do like and subscribe. It really does help out and check out the previous episode with Dr. Jack Cruz and Peter McCormick.