5 Ways to Make Money From an AI the Government Fears - Ep. #310

Chris Koerner (0:00)

The most powerful AI ever released to the public was online for only three days. Three, and I had a chance to play around with it. But then, the United States government sent a letter, and the company that built it shut it off worldwide in a matter of hours. Not throttled down, we're talking off off. So, I'm going to tell you what it was, why the government killed it, and the stuff it did in those three days that honestly scared the crap out of the same people that built it. Then, I'm going to hand you five businesses that you can set up right now, while this AI model is still dark. So, the day that this AI model comes back online, which it will, you are going to be primed and ready to make some money on it. At the end, I'm going to show you the one signal that everyone out there is getting wrong. They're reading it backwards. People see the word band and they run away, but I'm going to show you why the word band is the strongest buy signal the government has ever handed a regular person. Let's dive in.

Okay, let's get on the same page. The company is Anthropic. They make Claude. This is the AI that a lot of you already use as DUI. On June 9th, they released a new model called Fable 5 This was kind of a big deal because Fable came off of a tier that they've been holding back. And this tier is a class of AI models that they were calling Mythos. Think of their normal models like Opus, Sonnet, stuff that you're probably familiar with as the strong stuff that everybody gets. But Mythos was like the locked vault in the back. The one they said was too powerful to hand out. So Fable 5 was anthropic, finally cracking that vault open for the public with safety guardrails bolted on. And that's important. I'm going to get back to that. And you're probably non-technical like I am. You've probably never written a line of code. So this matters to you because the thing that Fable can do is the exact thing that every online business has been waiting for. But before I get to that, I want to get to the part that reads like a movie. Here's why they pulled Fable. So three days after launch on June 12th at around 521 in the evening, Anthropic got a letter from the US Commerce Department. This letter ordered them to cut off access to Fable 5 for any foreign national. But listen, that doesn't really work. Anthropic can't just check your passport when you log in. And if they tried, that friction would turn almost everyone away. So to follow this order from the US Commerce Department, they had no choice. They had to shut down the whole thing for everybody. You, me, paying customers, enterprise, doesn't matter. Even their own employees who weren't US citizens, they had to shut it down. But the craziest part about this is that this is the first time in history that the US government used export controls on an AI model itself. Export controls, like we're talking physical goods. This has never happened before. Here's the kind of company that this AI is now keeping. Export controls are the rules that cover stuff like nuclear material, fighter jets, missiles and the guidance systems beside them, or spy satellites, night vision. The whole list of stuff that government decides is too dangerous to let walk out of the country. That's the club that Fable just got drafted into. A thing that you just log into with a password is sitting on the same shelf as weapons. But individuals have never had access to weapons or missiles or nuclear reactors until now. It's called Fable. Couple years back, the government banned the best AI chips, the Nvidia chips that everybody wanted. They banned it from being sold to China. So Nvidia just turned around and built weaker versions on purpose so they could stay legal. Of course, they don't want to stop selling to China. So then they went after the machines that built the chips. There's this Dutch company called ASML that makes the one machine on earth that can actually print the most advanced chips. One machine costs over 100 million bucks and it weighs like 170 tons. And the US government said that machine does not go to China, period. So follow the pattern. First, they control the chips. Then they control the machines that make the chips. And now for the first time ever, they're controlling the model itself, the AI. They skipped the hardware and went straight for the brain. So here's the part that should make like the hair on the back of your neck stand up. This is the second time in history that the government has looked at a piece of software and called it a weapon. The first time set off a fight that changed the internet forever. And how that fight ended is the whole reason you should be paying attention right now. But I'm going to get to that later. And why? Why did they do this? Well, there's the official reason and the unofficial reason. The official reason is because of a security hole. The government said somebody found a way to jailbreak Fable. Remember the early days of the iPhone? You could jailbreak it and have it do things that Apple didn't authorize. Someone was doing that allegedly with Fable. And this was a trick to slip past its safety rules, its guardrails. And they got worried about what a foreign actor could do with it. I mean, think about it. I like to think about things in upside and downside. Basically, the government was like, it's kind of more downside than upside here, at least in its current state. And that's the whole thing is what these mythos class models like Fable are scary, scary good at is finding weaknesses in software. So this is awesome if you're defending a company, but bad if you're a bad guy and you want to break into one. So of course, Anthropic, which by the way, is privately valued at about a trillion dollars, okay? They pushed back hard. They said that, listen, that jailbreak that you heard about was narrow and other public AI models can already do the same thing. So yanking a model that are used by hundreds of millions of people over one alleged jailbreak is a massive overreaction. Now, of course, Anthropic is biased. Of course, they're saying that they know what this means to their bottom line or top line, both. And the timing of this could not have been worse for Anthropic because they're right about to go public. It just happened to SpaceX that went public up 40% in the first two days. We're talking to a $3 trillion company. Anthropics licking their chops, hoping for something similar. Terrible timing for them. And I don't know the full politics behind this. I'm sure there were some conversations that went on behind closed doors that nobody will ever know about. But what I do know is the result. The single best AI tool in the world went dark. It had never happened before. And the window to get ready for it to come back swung wide open. Now we kind of know what we're getting and we can prepare for it. Remember when I said the first time the government called a piece of software a weapon? Here's how that ended. So back in the 90s, there was a program called PGP. And this program let regular people scramble their messages so nobody could read them. The government freaked out. They classified that software as a munition, the same legal bucket as a missile. And they opened a criminal case on the guy who built it. He was facing literal prison time for letting code leave the country. And how did he beat it? He simply printed the code in a book. Because a physical book is protected speech. So they were stuck. The government was stuck. Their hands were tied. The case fell apart and they lost. And that weapon they tried so hard to lock down turned into the little padlock that you see on every website. It's the whole reason you're able to put your credit card into the Internet at all. So this thing that in the 90s they called too dangerous to export became the whole freaking foundation of the entire online economy. But let me tell you why I care this much about this model. And again, this is not sponsored. Anthropic does not pay me, okay? This is just Chris Koerner talking here. Here's the one thing that makes Fable different the most because everything else that we talk about is gonna flow from that. So every AI you've used so far is smart for like a minute. Okay, you ask it, it answers. And if the job is big and the context window is what it's called, gets too big, too onerous for the model to handle, it loses the plot. It starts hallucinating, you're like, whoa, what are you talking about? Why are you talking about this? We're over here, halfway through the chat, and you're gonna be back to babysitting it. But Fable was built to be left alone, right? Like a really good employee, you give it a real job. Let's say it has 10 steps to it and it plans it out. It does the work. It checks its own work, fixes its own mistake, and then it hands you the finished product with a nice little bow on top. And it can run for hours. It did. It can run for days. It can even spin up copies of itself, split the work, and review the copies. And that's the whole game here. The magic is not that it's just smart. Lots of AI is smart. It's all smart. The magic is that you can walk away and it'll just do. It'll just work while you sleep and you can trust the result enough to put your name on it and to sell it. Okay, that's why we're here. To sell it, that has never existed before. So if you've tried AI for real work and just watched it quit on you halfway, you're not crazy. That wall is exactly what Fable just knocked down. Here's some receipts. Here's some examples of how crazy this is. So Stripe, you know, giant payments company, they tested Fable early. They had a code base with 50 million lines of code and they needed to overhaul it. They needed to update it. And something like this would usually take a whole team of engineers two to four months. Fable did it in a freaking day. One day.