Is Bitcoin the antidote to AI spam?

0:00In this video, I want to talk about something that's probably already a huge pain in your life, but it might be about to get much, much worse. I'm talking about spam and the growing fear that AI agents could scale so cheaply and so fast that they basically denial of service attack of the entire human internet by flooding it with spam. So much so that they effectively make the digital communications channels that we as humans rely on completely unusable. You might think that that sounds crazy, but the head of product at Twitter doesn't. He thinks this spam attack hellscape could happen much, much sooner than you think. But contrary to what he said in this tweet about not having any solution for it, a prominent MIT engineer and a U.S. Space Force Guardian thinks we actually already have the solution. We're just not using it yet. It's a solution that lies in a strange, very niche computer science breakthrough that happened way back in 1997

1:00that almost no one has ever heard of. But you probably have heard of the most well-known example of that solution being implemented into another technology and that's because that technology is called Bitcoin. So in this video, I'm going way down the rabbit hole of what might happen if these AI agents scale to the point where they can spam attack the entire internet and why Bitcoin might be the only way to keep the internet usable for humans at all. But first, real quick, you're watching the sat stacker show, a Bitcoin show for people who think deeper about money. I'm your host, my name is John, aka the sat stacker. If you want to learn to stack smarter and grow your conviction in Bitcoin, you hit subscribe button and you can zap me sat via the QR code on the screen using your lightning wallet. So this video was inspired by the tweet I showed on the screen from X's head of product, Nikita Beer. And the overall idea goes like this. I don't think it's that hard to imagine a world where you wake up tomorrow or a couple months from now. And you realize basically every digital communications channel that you use to talk to other humans is pretty much unusable.

2:01Your email inbox is flooded with 10,000 emails. All your texts are from a profile pic. You don't recognize that say, hey, quick question. All your DMs are just an ocean of perfectly worded AI Slop and M-filled nonsense. We all know spam is annoying today, but it's generally tolerable. The channels we use are still usable for us to talk to each other. But I'm talking about a potential world where digital communication for humans is completely broken. But like I said, the weird part is that there is actually a solution to this problem that's been around for almost 30 years, and that might suddenly become much more relevant than ever. And yes, since this is a Bitcoin channel, it involves Bitcoin. It's not quite the AI agents are going to want to transact in Bitcoin argument, which I already touched on. You can watch that video right here. But this concept goes a little deeper than that, because it's not just about whether or not AI agents will want to transact in Bitcoin. It's about the possibility that the only way to keep the internet usable for humans is if we make the AI's transact in Bitcoin.

3:03I'm talking about the concept Jason Lowry put forward called cyber sovereignty. Because once all the infrastructure that we're building now to scale AI is built, the digital actions that the AI's take become basically free. The marginal cost of one additional AI action is zero, which means whoever can automate the most in cyberspace wins. And since we do seem to be entering the age of agentic AI, that means the bots can easily drown out humans on the internet without even trying. So Jason Lowry had the best take on this that I've seen so far. The argument is this, throughout the entirety of history, we see the concept playing out that if you don't make access to your domain expensive, you will lose control of it. Castle walls, for example, we're not built just to look cool. They were built to make attackers pay a cost for trying to gain unauthorized access. But the way we built the internet is kind of the opposite. It's like we built a city with no doors at all. And then we oopsie, surprise Pikachu face

4:03when a billion raccoons move in. Except now the raccoons have GPUs. This is sort of already the idea behind the dead internet theory if you've heard of that. Many of us already started having this sort of existential crisis about whether or not a majority of internet traffic and communication was already being done by bots. But what Nikita is talking about is potentially way worse than that. If operating online is too cheap, we could inadvertently lose jurisdiction over the entire internet. And yes, I know AI isn't literally infinite. It still needs servers and power and chips and bandwidth. But the point is the marginal cost. Once the machines exist, the cost to do just one more message is so close to zero that it might as well be zero. We all know that every platform and every communications channel already has a bot problem. And platforms play this constant whack-a-mo game trying to identify who's human and who's not. But what if that guessing game is just a giant waste of time considering this old idea from the 90s that already solves this problem? Its approach is not about trying to guess who's human and who's not.

5:03Instead, it's just asks, are you willing to pay a cost for access? It sounds very simple, but I'm going to show you how that one tiny switch changes everything. Back in 1997, before iPhones, before social media, before I started losing my fucking hair, a cryptographer and cypher punk named Adam Back. You've probably heard of him, given that he's also cited by Satoshi Nakamoto in the Bitcoin white paper. He proposed something called hash cash. Now the idea behind hash cash was eventually used in the creation of Bitcoin, but hash cash was not invented to solve the problem of digital peer-to-peer money. It was invented to solve the problem of spam. Adam Back's idea was simple. What if, before you can send an email, your computer just has to solve a tiny little puzzle, not a super hard one, but just hard enough that it costs a little bit of time and energy. For a normal person or a normal computer sending a normal number of emails, they wouldn't even notice it. But for a spammer who's trying to send millions of emails, now they have a problem on their hands because solving millions of puzzles becomes very expensive. This is the core idea behind the concept of proof of work.

6:07Make the sender prove that they burned some real-world resources. Unlike, say, I don't know, central banks who can print money at the click of a button, this idea is more analogous to gold, where if you want to get more of it, you have to expend real energy and resources to go dig it up out of the ground. So if we go back to Jason Lowry's analogy, every civilization eventually discovers the same thing. If you don't make access to your domain expensive, you lose control of it. In every domain, the ability to impose physical costs preserves sovereignty. The internet was built to move information as fast and as cheaply as possible, which is generally awesome until the cheapness becomes the vulnerability. Because free access plus infinite automation equals whoever spams the most shapes the reality. And that's how you get a really dead internet. Well, actually, ironically, wouldn't be technically dead at all. There'd probably be more activity on it than ever before. It'd just be effectively dead for humans.

7:09At this point, if you've been around for a while, you're probably expecting me to, at some point, use one of the top five favorite Bitcoin or cliches, which is Bitcoin fixes this. And in this case, it actually does. To be clear, I'm not saying Bitcoin is the only imaginable way to fix this problem. A platform could implement its own proof of work puzzles or rate limits or captures or API keys, cloud flare or aggressive filtering. They do a lot of those things now, and they all help to a degree. But they're mostly reactive and they're centralized. Of course, the centralization problem is that if one company controls all the rules, then that company becomes the gatekeeper of who gets to speak at all. So the Bitcoin angle is different. Bitcoin is already a globally neutral, permissionless, battle-tested cost system that exists at scale. You can use something like Bitcoin payments or Bitcoin adjacent proof of work as the shared language of cost. Bitcoin is basically already the world's largest receipt printer for energy, which might sound dumb until you need a way to prove

8:11that real work happened in the real world, not just inside a server. So let me try to make this super concrete. Imagine a future messaging app. To send a DM, you either have to solve a tiny proof of work puzzle on your device or attach a tiny payment like one Satoshi, a fraction of a cent. Now as a normal person, it would make no difference. But as a botform, if you want to send 20 million messages a day, you technically still can't do that, but you just have a massive bill to pay. It's not about stopping AI agents or bots altogether. It's more about stopping free unlimited scale and flipping the ROI calculation of mass spam. And most spam campaigns only work because they can do the spray and prey approach at insane volume. If you add a real cost per attempt, then the hope is that the incentive shift enough to slow them down or stop them all together because the economics no longer clear. That's obviously hypothetical, but it's not that hard to imagine the bot problem becoming that bad.

9:12And if it does, then in my opinion, we will see someone try this solution at scale at some point. And I understand probably most people's biggest objection is, well, wait, I don't want to pay every time I have to send a text or an email. But if you actually sit back and do the math, so if right now Bitcoin is around, I think $67,000, that means one Satoshi, which is one 100 millionth of a Bitcoin, is about 0.0067 cents, so less than one tenth of a penny. So if you sent 100 messages a day, that would cost you six or seven cents a day, maybe $2 in a month. Most humans in developed countries who rely on using the internet all day are not going to notice $2 a month. But if you're a bot farm trying to send 100 million messages, it's going to cost you $67,000. So I don't see a world where it's about eliminating bots or AI entirely, we probably don't even want that. But it's about what Jason Lowry said, at least trying to constrain the digital world with some semblance of real world limits.

10:14And keeping the internet at least somewhat human. No one has to use Bitcoin for this, but it just so happens that Bitcoin is the most obvious shelling point for a neutral worldwide cost layer that doesn't require trusting some company or government to referee who gets to speak. And if Bitcoin were to someday become the proof of work castle wall that separates the human internet from the bot takeover, that might lead to an even more interesting outcome. Because what happens if we force AI agents themselves to start accumulating Bitcoin just to function online? If across the internet we start putting more and more actions behind cost gates, then any entity that wants to operate at scale online will need the thing that pays those costs. So now you're imagining a scenario in which humans want Bitcoin to store value, platforms want Bitcoin to reduce spam and restore signal, and AI agents want Bitcoin because it's literally the admission ticket to do anything at scale. And then Bitcoin is much more than just money.

11:16It's the literal currency of access. So am I saying this is definitely gonna happen? No, I'm not. But what I am saying is that it's possible that some sort of cost in position becomes virtually our only solution to stave off the robot takeover. We either build castle walls in cyberspace or we accept that the loudest thing online will be whatever can automate the most. And this dusty old anti-spam idea from the 90s that we call proof of work might end up being one of the only ways to force the digital world to obey physics again. And Bitcoin just so happens to already be the biggest most battle tested proof of work system on Earth. So ironically, we've had people asking us for years what is Bitcoin's use case? What's Bitcoin good for other than number go up? What if the answer ends up being keeping the internet human? So maybe just maybe before the AI agents catch on to what's happening, it might be a good idea to quit slacking and start stacking.