Gemma: Two little hobbits are trying to get Treebeard to do something about an army that's hacking down his forest. The problem, though? Treebeard, that enormous, sentient tree, operates on a completely different timescale. Takes him, like, a full day just to say hello, to acknowledge you.
Marco: Oh, I know this. Is this from the books, the original, ah, Tolkien? No, wait, it's a metaphor... from, uh, Sam Altman?
Gemma: No, it's Amodei. Dario Amodei, the CEO of Anthropic. He's the one using it to describe what it feels like trying to get slow-moving governments to deal with the absolute lightning pace of AI. We're the hobbits, and the policy-makers are, well, Treebeard.
Marco: Amodei? Ah, I get confused with the names. But the idea is the same, no? Getting the slow institutions to move for the fast technology.
Gemma: Yeah, precisely. And his whole point is that we can't afford to just sit there and wait for Treebeard to wake up. This isn't about the AI we have today; it's about the AI that's coming in one, maybe two years, and that's what we need to legislate for now. Before it's too late.
Marco: It's this idea that AI has crossed a threshold, a point of no return almost. Its strategic importance is undeniable.
Gemma: A point of no return? Is that a bit much?
Marco: No, no! He talks about novel cybersecurity risks, things we've not seen before. This is not just another tech trend, Gemma. It's a rethinking on the scale of the industrial revolution, or nuclear weapons.
Gemma: That's precisely it, though, isn't it? He's saying, look, when you're looking at any policy proposal for AI, you can't just ask if it fits what AI can do today. You have to ask if it's going to be enough for what he calls 'a country of geniuses in a datacenter' that will exist in like, a couple of years.
Marco: It's like the early climate change debates, you know? For years, the scientists, they were pointing to these exponential curves in CO2, saying, 'Look, this is coming! The effects will be devastating!' But the policymakers, they couldn't act because the worst effects, they were not visible yet. And now, we are playing a desperate game of catch-up.
Gemma: Yeah, it's that anxious, urgent feeling, isn't it? Like you see the tsunami on the horizon, but everyone else is still enjoying their ice cream on the beach. It's just... it makes you want to shout, doesn't it?
Marco: Yes! And he is shouting, in his way.
Gemma: He is. And he's not just complaining, mind. He's got specific areas, actual things he says we need to rethink.
Marco: Yes, five areas he goes into. Regulation and public safety, first. We cannot use the old playbooks, he says.
Gemma: The old playbooks? What, like, just, uh, a bit of light touch regulation?
Marco: No! He mentions the 'Collingridge dilemma' – new frameworks are needed for risks like bio-weapons, autonomous misbehavior. Things that are impossible to anticipate now, but if we wait, they are too late to manage.
Gemma: He even compares it to the FAA, no? Like, frontier AI models should go through mandatory technical testing, like airplanes. And if they don't meet high safety standards, their release should be blocked or reversed. That's a pretty strong stance, isn't it?
Marco: Very strong. Then there is macroeconomics and tax policy. AI will cause huge job displacement, he says, and we need a policy framework for it. Anthropic, his company, is even willing to back it financially. Imagine that, putting your money where your mouth is.
Gemma: And scientific innovation. Making sure AI accelerating science is actually a net positive for humanity. It's not a given, apparently.
Marco: Not a given? What do you mean?
Gemma: Well, he doesn't elaborate, but I imagine it's about control, or... what gets prioritised. Or perhaps... no, never mind. Anyway. And then there's the balance of power between state and society, because powerful AI could really shift that, couldn't it? And, of course, geopolitics. A tool of national strategic consequence, just like nuclear weapons.
Marco: So it's not just a warning. It is a roadmap for how to react. Immediately. This is not some academic paper, no?
Gemma: Precisely. No, he's given us a proper, full-on call to arms, just without the actual arms. Just, you know, the policy kind. Right. So, we need to sort it out now, or we're all, you know, toast.
Marco: But wait! What about the other side of this exponential coin? Paul Graham. He talks about how to earn a billion dollars. This one is excellent!
Gemma: Oh, the Paul Graham piece, yeah.
Marco: He talks about this founder, yes? And he asks her, 'What's your growth rate?' She says, '93% last month.' Can you imagine? Almost doubling every month.
Gemma: Madness. Is that even real?
Marco: He says it is! And he tells the audience, 'Pull out your phones! Calculate log(500, 1.93).'
Gemma: Go on then, what's the answer? I'm terrible at mental maths, you know that.
Marco: It's 9.45. Nine and a half months! To turn a two million dollar company... to a billion dollar company. Nine months, Gemma! This is not magic, he insists, it's just mathematics.
Gemma: Nine and a half months! Blimey. That's just... staggering, isn't it? From, what was it, a few million to a billion? In less than a year. I always thought it was all about, I don't know, cheating, or being born into it.
Marco: Exactly! That's his whole thing. The growth rate, it's the most important number because it tells you if you made something people truly love. You cannot cheat the exponential word-of-mouth growth. You earn it, by making users happy.
Gemma: So it's not about having some 'billion-dollar idea' from the get-go? Like, you have to be cleverest person in the room.
Marco: No, no. He says, 'Stop trying to come up with a billion-dollar idea.' Instead, find a small, real problem you have yourself. Build something to solve it. If you solve it well enough, the growth, it will tell you if it is a big idea. My friend's company was like this, you know? It was a simple browser extension, but it solved one tiny, annoying problem perfectly. They never spent a dollar on marketing. For two years, every conversation was just about keeping the servers from melting because of word-of-mouth growth. No advertising, just... people sharing.
Gemma: That's quite a story. It's like a secret formula's been revealed, but it's just... maths and empathy, really. Who'd have thought?
Marco: Precisely! And he gives another example, to make it even clearer. If your revenues grow at 15% a month, which he says, for a good startup, is not that rare. How much more do you make in five years?
Gemma: Five years? That's, what, sixty months?
Marco: Sixty months! You calculate 1.15 to the power of sixty. And the answer is... about 4384 times as much.
Gemma: Four thousand... that's enormous!
Marco: Four thousand three hundred eighty-four times! So if you start at ten thousand a month, after five years, you are making forty-four million a month. Forty-four million! You are a billionaire, Gemma. Poof!
Gemma: Good heavens. I mean, it's just... it's a bit mind-boggling when you see it written down like that, isn't it? Just this consistent, relentless growth, building something people genuinely want.
Marco: And the key is to build what you need. Young founders, he says, their needs are valuable because they predict future demand. Whatever you and your friends start using now, everyone is going to be using in ten years. It's all about that deep understanding of users, making something they love so much they tell their friends. It's beautiful, really.
Gemma: Right. So, on one hand, we've got Dario Amodei saying this exponential growth is a runaway train, right, and it requires massive, coordinated intervention to prevent catastrophe.
Marco: But Paul Graham says, 'earn a billion dollars' by making something people love! That same process... this exponential growth... it is the magical engine for progress!
Gemma: Yeah, but Amodei's saying that very process, at the frontier, it creates tools of 'global and national strategic consequence' that could threaten us all. So is it a good thing or a terrifying thing, this exponential growth? What do you reckon?
Marco: It depends, no? If you are the one building it, it feels like power. If you are the one trying to control it... it feels like a monster you cannot tame.
Gemma: I suppose that's the tension, isn't it? We can't really have one without the other. This exponential change is here, and whether it's good or bad depends entirely on what we decide to do about it now. Before it's too late.
Marco: Yes, exactly. Amodei is not just waving a red flag, he is saying 'we need concrete action, now'. Legislative proposals, financial backing. He says the risks are no longer theoretical, and the time for slow, ponderous debate is over. The exponential is here, and it demands our attention.
Gemma: And that's where we're left, then. With this urgent call to action. I'm Gemma.
Marco: And I'm Marco. This has been Manish Chiniwalar's Station.
