Dario Amodei: uncertainty

Transcript context

…I was re-listening to the interview from three years ago, and one of the ways it aged poorly is that I kept asking questions assuming there was going to be some key fulcrum moment two to three years from now. In fact, being that far out, it just seems like progress continues, AI improves, AI is more diffused, and people will use it for more things. It seems like you’re imagining a world in the future where the countries get together, and “Here’s the rules of the road, here’s the leverage we have, and here’s the leverage you have.” But on the current trajectory, everybody will have more AI. Some of that AI will be used by authoritarian countries. Some of that within the authoritarian countries will be used by private actors versus state actors. It’s not clear who will benefit more. It’s always unpredictable to tell in advance. It seems like the internet privileged authoritarian countries more than you would’ve expected. Maybe AI will be the opposite way around. I want to better understand what you’re imagining here. Just to be precise about it, I think the exponential of the underlying technology will continue as it has before. The models get smarter and smarter, even when they get to a “country of geniuses in a data center.” I think you can continue to make the model smarter. There’s a question of getting diminishing returns on their value in the world. How much does it matter after you’ve already solved human biology? At some point you can do harder, more abstruse math problems, but nothing after that matters. Putting that aside, I do think the exponential will continue, but there will be certain distinguished points on the exponential. Companies, individuals, and countries will reach those points at different times. In “The Adolescence of Technology” I talk about: Is a nuclear deterrent still stable in the world of AI? I don’t know, but that’s an example of one thing we’ve taken for granted. The technology could reach such a level that we can no longer be certain of it. Think of others. There are points where if you reach a certain level, maybe you have offensive cyber dominance, and every computer system is transparent to you after that unless the other side has an equivalent defense. I don’t know what the critical moment is or if there’s a single critical moment. But I think there will be either a critical moment, a small number of critical moments, or some critical window where AI confers some large advantage from the perspective of national security, and one country or coalition has reached it before others. I’m not advocating that they just say, “Okay, we’re in charge now.” That’s not how I think about it. The other side is always catching up. There are extreme actions you’re not willing to take, and it’s not right to take complete control anyway. But at the point that happens, people are going to understand that the world has changed. There’s going to be some negotiation, implicit or explicit, about what the post-AI world order looks like. My interest is in making that negotiation be one in which classical liberal democracy has a strong hand. I want to understand what that better means, because you say in the essay, “Autocracy is simply not a form of government that people can accept in the post-powerful AI age.” That sounds like you’re saying the CCP as an institution cannot exist after we get AGI. That seems like a very strong demand, and it seems to imply a world where the leading lab or the leading country will be able to—and by that language, should get to—determine how the world is governed or what kinds of governments are, and are not, allowed.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.