Dario Amodei

The way I think about it actually is, well, so I think in the early stages, the AIs are going to be like grad students. You're going to give them a project. You're going to say, you know, I'm the experienced biologist. I've set up the lab. The biology professor or even the grad students themselves will say, Here's what you can do with an AI system. I'd like to study this.

And the AI system, it has all the tools. It can look up all the literature to decide what to do. It can look at all the equipment. It can go to a website and say, hey, I'm going to go to Thermo Fisher or whatever the lab equipment company is to the dominant lab equipment company is today. And my time was Thermo Fisher. I'm going to order this new equipment to do this.

I'm going to run my experiments. I'm going to write up a report about my experiments. I'm going to inspect the images for contamination. I'm going to decide what the next experiment is. I'm going to write some code and run a statistical analysis.

All the things a grad student would do, there will be a computer with an AI that the professor talks to every once in a while and it says, this is what you're going to do today. The AI system comes to it with questions. When it's necessary to run the lab equipment, it may be limited in some ways. It may have to hire a human lab assistant to do the experiment and explain how to do it.

Or it could, you know, it could use advances in lab automation that are gradually being developed over, have been developed over the last decade or so and will continue to be developed. And so it'll look like there's a human professor and a thousand AI grad students.

And, you know, if you go to one of these Nobel Prize winning biologists or so, you'll say, okay, well, you know, you had like 50 grad students. Well, now you have a thousand and they're smarter than you are, by the way.

Um, uh, then I think at some point it'll flip around where the, you know, the AI systems will, you know, will, will be the PIs will be the leaders and, and, and, you know, they'll be, they'll be ordering humans or other AI systems around. So I think that's how it'll work on the research side. And they would be the inventors of a CRISPR type technology.

Um, and then I think, you know, as I say in the essay, we'll want to turn, turn, probably turning loose is the wrong, the wrong term, but we'll want to, we'll want to harness the AI systems, uh, to improve the clinical trial system as well. There's some amount of this that's regulatory. That's a matter of societal decisions and that'll be harder, but yeah.

Can we get better at predicting the results of clinical trials? Can we get better at statistical design so that clinical trials that used to require 5,000 people and therefore needed $100 million in a year to enroll them, now they need 500 people in two months to enroll them. That's where we should start.

Uh, and, and, you know, can we increase the success rate of clinical trials by doing things in animal trials that we used to do in clinical trials and doing things in simulations that we used to do in animal trials? Again, we won't be able to simulate it all. AI is not God. Um, uh, but, but, you know, can we, can we shift the curve substantially and radically?

So I don't know, that would be my picture.

Yeah, yeah, yeah. Can we just... One step at a time, and can that add up to a lot of steps? Even though we still need clinical trials, even though we still need laws, even though the FDA and other organizations will still not be perfect, can we just move everything in a positive direction? And when you add up all those positive directions, do you get...

Everything that was going to happen from here to 2100 instead happens from 2027 to 2032 or something.

I think that's going to be one of the areas that changes fastest for two reasons. One, programming is a skill that's very close to the actual building of the AI. So the farther a skill is from the people who are building the AI, the longer it's going to take to get disrupted by the AI. I truly believe that AI will disrupt agriculture.

Maybe it already has in some ways, but that's just very distant from the folks who are building AI. And so I think it's going to take longer. But programming is the bread and butter of a large fraction of the employees who work at Anthropic and at the other companies. And so it's going to happen fast. The other reason it's going to happen fast is with programming, you close the loop.

Both when you're training the model and when you're applying the model, the idea that the model can write the code means that the model can then run the code and then see the results and interpret it back. And so it really has an ability, unlike hardware, unlike biology, which we just discussed, the model has an ability to close the loop.

Um, and, and so I think those two things are going to lead to the model getting good at programming very fast. As I saw on, you know, typical real world programming tasks, models have gone from 3% in January of this year to 50% in October of this year. So, you know, we're on that S curve, right? Where it It's going to start slowing down soon because you can only get 200%.

But I would guess that in another 10 months, we'll probably get pretty close. We'll be at least 90%. So again, I would guess, I don't know how long it'll take, but I would guess again, 2026, 2027, Twitter people who crop out these numbers and get rid of the caveats, like, I don't know, I don't like you, go away. Yeah.

I would guess that the kind of task that the vast majority of coders do, AI can probably, if we make the task very narrow, like just write code, AI systems will be able to do that. Now, that said, I think comparative advantage is powerful. We'll find that when AIs can do 80% of a coder's job, including most of it that's literally like write code with a given spec,

we'll find that the remaining parts of the job become more leveraged for humans, right? Humans will, there'll be more about like high-level system design or, you know, looking at the app and like, is it architected well? And the design and UX aspects, and eventually AI will be able to do those as well, right? That's my vision of the, you know, powerful AI system.