← Cheeky Pint
Cheeky Pint
Anthropic CEO Dario Amodei on designing AGI-pilled products, model economics, and 19th-century vitalism
01:02:51
Loading summary
Transcript178 lines
- [00:00]
Lex Fridman
I'm excited to finally learn what is it like to start a company with your sibling.
- [00:04]
Dario Amodei
I don't know why you're asking me that question.
- [00:06]
Lex Fridman
Because, you know, it's like, you know, the models want to learn. The models want to be extraordinarily successful in the market.
- [00:11]
Dario Amodei
Yes. Right. In addition to having this learning impulse, the models have this, like, capitalistic impulse. Sometimes people think of the API business and they say, oh, it's not very sticky, or it's going to be commodity business.
- [00:22]
Lex Fridman
I love API business.
- [00:22]
Dario Amodei
No, no, exactly, exactly. I think we're going to be in a world where the models will make mistakes much less often than humans, but they'll be stranger mistakes.
- [00:33]
Lex Fridman
So we need to invent slurring for a lens and that's incorrectly poured. Oh, wow. Dario is CEO of Anthropic, one of today's frontier AI labs. He's gone from being an AI researcher just a few years ago to now running one of the world's fastest growing businesses. Cheers.
- [00:48]
Dario Amodei
Cheers.
- [00:51]
Lex Fridman
So I'm excited to talk a bit about the Anthropic business. You studied physics and computational neuroscience.
- [00:59]
Dario Amodei
Yes.
- [01:00]
Lex Fridman
You then worked at Baidu, then Google Brain, then OpenAI, and then started Anthropic.
- [01:07]
Dario Amodei
Yes.
- [01:07]
Lex Fridman
And we'll get into the Anthropic business. But I'm excited to finally learn what is it like to start a company with your sibling.
- [01:14]
Dario Amodei
I could ask the same question of you, but it's almost like there's two things you need to do when you're running a company. You need to, you know, operationally execute and, you know, you need to have a good strategy and kind of see the most important thing or the thing that no one else sees. And so my job is the second and Daniela's job is the first. And we're both good at the things that we do. And so, you know, I think it's allowed us each to spend most of our time on the thing that we're best at.
- [01:43]
Lex Fridman
Presumably there's something about the trust side of things as well, where co founder teams in general, in tech as an AI as well. I mean, there are unstable pairings in just having someone where you have a long running and deep trust.
- [01:54]
Dario Amodei
Yeah, yeah. Where you have just total and complete trust. I mean, I think even beyond that, you know, Anthropic has seven co founders. When we founded it, basically the advice from pretty much everyone was like, seven co founders is a disaster. The company will fall apart before you know it. Everyone will be fighting with each other. There was even more negativity on My decision to give everyone the same amount of equity. But what we found, and I think it was because, you know, you know, obviously me and Danielle are siblings, but then all seven of us, you know, some of us knew each other for a long time or had history of work, not just knowing each other, but working together in the past. And I think that really allowed us, you know, to always be on the same page. And I think especially as the company grows, the idea that you have seven people who really carry the values of the company and project them to a wide set of people, it allows you to scale the company to a much larger size while kind of holding on to the values and the unity that we have.
- [02:52]
Lex Fridman
So I want to ask about the anthropic business because again, it's an incredible story where it was reported recently that you'd blown through $4 billion in ARR. And so it's a lot of discussion, correctly, about the technology that you're developing. But also this is just one of the fastest growing businesses in history. And so I want to talk a bit about the AI market and maybe the place to start is what is everyone doing with AI? There's coding, there's customer service work, but where does all this revenue come from?
- [03:24]
Dario Amodei
Yeah, there's a wide range of things and it's kind of changed over time. I would say definitely the application that has grown the fastest, although it's not very far from the only application we have, a wide range of them is definitely coding. And my theory on why it's grown so fast, other than that we focused on coding, the models are good at coding. It's actually really a statement about kind of society, which is that if we look at today's AI models, I think in every area there's a huge overhang in terms of what they could do compared to how they're actually being deployed today. Because there's some friction. People at large enterprises are not familiar with the technology. I look at what a bank does or what an insurance company does, and there's huge potential even if the model stopped getting better, right? Even if we stop building products on top of the model, there's still huge billion dollar potential in an individual enterprise. And often the CEOs of companies that I talk to understand that perfectly well. But if the company is a 10,000 or 100,000 person company, companies that size, they're set up to operationally do a certain thing, a certain way, and it takes time to change them. But in code, the people who write code are very socially and technically adjacent to the folks who develop AI models. And so the diffusion is very f. They're also the kind of people who are early adopters, who are used to new technology. And so I think the big growth in code, I would say the biggest cause of that is just that the people doing it and the startups devoted to it are fast adopters who understand the technology super well. But it's by no means limited to code at all. If you look at there are a bunch of companies that do things like tool use. There are, as you mentioned, customer service. We work closely with companies like Intercom. We're starting to see some things on the biology side. So we're working both with pharmaceutical and healthcare companies, and we're working on the side of kind of basic scientific research. So, you know, we work with companies like Benchling, for example, but, you know, we also work with some of the very large pharma companies. There was something done a while back where we worked with Novo Nordisk to write clinical study reports. So clinical study reports are like, you've done a clinical trial and then you kind of write up the results. And it's like, these are the adverse events, these are the statistics. And the clinical study report normally takes like nine weeks. Well, Claude could do it in like five minutes. And then it took a human a few days to check it. And so you can really see the opportunity for acceleration. And as the models get better, you know, they'll reach into the deep research as well. So I guess a way to summarize it would be to say that kind of code is out in the lead. But we see a long tail of quite a lot of other stuff, including some very, very significant use cases. I think code is maybe an early indicator, like a premonition of what's going to happen everywhere else. It's the same exponential. It's just faster. It's just happening faster.
- [06:34]
Lex Fridman
Right. So there are many places where there's significant AI uplift, but engineers are used to adopting. You think about hacker news and people arguing over who are the best tools people are passionate about.
- [06:42]
Dario Amodei
And, you know, like two hours after we released Claude code, there's some person out there who's like, you know, who's like, you know, tried 10,000 different things with it and plugged it into all the frameworks. And, you know, Twitter forms one opinion after two hours and then like revises an opinion in two hours. And, you know, you think of the speed of that as compared to the speed that like a pharmaceutical company can use it, can use it in research. Right. Or that traditional retail company. And, you know, we want to bring everything. Some of the biggest benefits in the world are touching the physical economy and we want to get there, but it just intrinsically does not happen at the same speed.
- [07:18]
Lex Fridman
How do you decide which verticals to do yourself versus which to allow? Platform like, you have CLAUDE code, and obviously there's also platform companies like Windsurf and Cursor and everyone like that. You launch cloud for financial services. Presumably there are other verticals where you say, well, we're not building a tool there.
- [07:35]
Dario Amodei
Yeah, we have things like CLAUDE for Enterprise, which is not a vertical, but like a general play to go with enterprise. I think the way we like to think about it, I think we think of ourselves as a platform company first. So the analogy here would be maybe clouds or something. If you think of a really large platform business of the size we're trying to get to in hopefully a small number of years, there are a number of reasons why you would also want to have things that are first party and that some. Some verticals end up being more first party heavy. One is when you want to have direct exposure to the users, the end user gives you some sense of how exactly are they using it, what are they most looking for? If you're a pure platform and you don't have that direct connection, you can be disadvantaged in various.
- [08:25]
Lex Fridman
It's hard to build the best products.
- [08:26]
Dario Amodei
Yeah, it's hard to build the best products. It may even be hard to know where the model really, really needs to go. Right. People say things like coding, but like, there are many models that seem to be good at coding, but they aren't good in the way that's actually relevant. Right. We've actually managed to make Claud good in a way that's relevant to what people actually use. So I think that's one reason. Another reason goes back to the large enterprises where building on an API sometimes is more challenging for a more traditional company to do that, and you need to give them something that's a little bit easier to use, either a kit to help them build things or you need to give them an app. Enterprises have also liked Claude code, and we're gradually developing CLAUDE for Enterprise into what we call like a virtual coworker.
- [09:14]
Lex Fridman
But I find it hard to picture anthropic developing CLAUDE for oil and gas exploration. And, you know, why is that?
- [09:20]
Dario Amodei
Yeah, why is it that you find it hard to imagine or why maybe.
- [09:25]
Lex Fridman
In fact it's the next launch.
- [09:26]
Dario Amodei
But yeah, we're not currently working on CLAUDE for Oil and gas exploration. I would draw a distinction between things we just, you know, things we just like, don't allow. Right? Things that are like, illegal or, you know, things like that. And there are a number of use cases that it's like, okay, you know, we're a platform. People are going to do a bunch of things. We're just not passionate about it. But, like, we're not passionate about it. You know, we're not going to go out and make this happen before the other use cases. So I think there is a component of that where, you know, probably we work on things like science and biomedical out of proportion to its immediate profitability.
- [10:03]
Lex Fridman
Because you guys think it's worthwhile.
- [10:03]
Dario Amodei
Because. Because we think it's worthwhile. We feel the same way about things in the developing world. One, I'll give you that's controversial. People think about it the opposite way. So, you know, the work we do on defense and intelligence, people are often like, oh, these guys are selling out. I think about it the opposite way, right. So, you know, there was this contract with a ceiling of 200 million with, you know, the DOD and intelligence community. People are like, oh man, you know, anthropics, anthropic selling out. It's exactly the opposite. Getting another 200 million from some coding startup would take like an order of magnitude less effort than like getting that contract.
- [10:39]
Lex Fridman
Defense is very worthwhile.
- [10:40]
Dario Amodei
We're doing it because we want to defend democracies and we do it within bounds. There are some things where, you know, we're concerned about. I'm deeply concerned about abuse of government authority. On the domestic side, you know, we think more on the kind of outward directed side. But that's an example of the things we prioritize are things that we think are good, not necessarily things that feel good or that people will. That we think kind of external buzz will be positive. We actually have conviction around some things and we do them regardless.
- [11:11]
Lex Fridman
You reference the kind of business you want to build. What are your aspirations for the anthropic business in say, 3 to 5 years time?
- [11:17]
Dario Amodei
AI is strange in a number of ways. I think one of the ways it's strange is that because it's an exponential, we have a hard time calibrating exactly how big the business will be. So we had the following experience. So in 2023, I'd never raised money from institutional investors before, and so our revenue was zero at the beginning of 2023 because we had not released a product. So I was putting together something and I'm like, oh, I Think we can probably get $100 million of revenue in the first year. And this caused some investors to say, this is crazy. This has never happened in the history of capitalism. You've lost all credibility. Goodbye. Goodbye. And then we actually did it. And so, you know, then the next year I was like, oh, well, I think we can go from 100 million to a billion. And actually that was having done it the first time, people were like, it was a little bit less dismissed as crazy, but, you know, still often dismissed as crazy. And then we did it again. And, you know, this year we're halfway through the year. You know, we're, as you mentioned, well past 4 billion in revenue. So, you know, in kind of logarithmic space to add another order of magnitude. So there's a bunch of different futures. There's one where once things get to a certain size, the curve slows down. But there's a provocative world where the exponential continues. And in two or three years, these are the biggest businesses in the world. And I think one of the fundamental experiences and uncertainties of working at or running something like anthropic is you kind of don't know, you make this exponential projection. And it sounds crazy, it might be crazy, but also it might not be crazy because that trend line has followed before. And I've said much the same thing in the context of training AI models, in the context of the cognitive capabilities of AI models on the technological side. But now we're seeing the same kind of continuous lines on the business side.
- [13:14]
Lex Fridman
So what's the analogy to scaling laws here where you scale up the relevant inputs for model quality in parallel and you get a much better model performance. Is there something where you put better models in and I know the right organization?
- [13:29]
Dario Amodei
Yeah, there's something like there's some curve where you spend 5x or 10x more to train a model, or you have 5x or 10x more data or whatever the scaling laws say. And there's some transfer curve for revenue where I spend 10 times more on the model. And the model goes from being a smart undergrad to a smart PhD student. And then I go to a pharmaceutical company and I'm like, well, how much more is that worth? Often they end up saying, that's worth more than 10x, that's worth about 10x. Where these kind of power law distributions occur in a bunch of contexts. Going on the technical side, when you train the model, there's a longer and longer tail of correlations that you're capturing as you train the model, correlations in the structure of language, in the world, in patterns. And that correlation is what's thought to lead to the scaling laws because there's this kind of logarithmic distribution. And then as you think of the model getting more and more capable in terms of cognitive tasks, there must be, or we're seeing empirically so far as if you think of the like uses of the model in the economy. Right. You know, if I think of, you know, the way that companies are organized. Right. There's a kind of power law. There's a power law structure of like, of like the org charts, the org charts of companies. And it almost feels like you're climbing that power law distribution of value. And then I guess the way I think about product and go to market is that the model wants to be on that exponential of revenue and product and go to market. They're kind of a way to clean the window and let the light shine through. Right. It's a way to kind of open the aperture and let the exponential happen.
- [15:19]
Lex Fridman
It's like the models want to learn. The models want to be extraordinarily successful in the market.
- [15:23]
Dario Amodei
Yes, right. In addition to having this learning impulse, the models have this capitalistic impulse that they want to embody unless they're given a bad product or bad sales to.
- [15:33]
Lex Fridman
Go with it because they're really useful. That intelligence is really useful to people. So it kind of gets pulled out of you.
- [15:38]
Dario Amodei
Yes, yes, yes. That is a way to think about it.
- [15:44]
Lex Fridman
What is the terminal market structure here? Like, is there a few large scaled players or do we kind of keep seeing new upstarts for kind of specific spaces?
- [15:56]
Dario Amodei
It's very hard, you know, it's hard to tell for sure. And I think there was, you know, quite a lot of uncertainty two or three years ago, but I think we might be relatively close to the finals head of players, if not necessarily the final market structure or the roles of the players. I would say there's probably somewhere between three and six players, depending on how you count. And those are the players that are capable of building the players that are capable of building frontier models and have enough capital to plausibly bootstrap themselves.
- [16:32]
Lex Fridman
I would love to understand how the model business works where you invest a bunch of money upfront in training and then you have this fast ish depreciating asset though maybe with kind of a long tail of usefulness and hopefully kind of you pay that back. Thus far, I think the image people have from the outside world is ever larger amounts of capex and you know.
- [16:54]
Dario Amodei
How to get kind of burned. There's kind of like two different ways you could describe what's happening in the model business right now. So let's say in 2023, you train a model that costs $100 million and then you deploy it in 2024 and it makes $200 million of revenue. Meanwhile, because of the scaling laws, in 2024, you also train a model that costs a billion dollars. And then in 2025, you get $2 billion of revenue from that $1 billion and you spend $10 billion to train the model. So if you look in a conventional way at the profit and loss of the company, you've lost $100 million the first year, you've lost $800 million the second year, and you've lost $8 billion in the third year. So it looks like it's getting worse and worse. If you consider each model to be a company. The model that was trained in 2023 was profitable. You paid 100 million and then it made 200 million of revenue. There's some cost to inference with the model, but let's just assume in this cartoonish cartoon example that even if you add those two up, you're kind of in a good state. So if every model was a company, the model is actually, in this example is actually profitable. What's going on is that at the same time as you're reaping the benefits from one company, you're founding another company that's like much more expensive and requires much more upfront R and D investment. And so the way that it's going to shake out is this will keep going up until the numbers go very large, the models can't get larger, and then it'll be a large, very profitable business. Or at some point the models will stop getting better. Right? The march to AGI will be halted for some reason, and then perhaps it'll be some overhang. So there'll be a one time, oh man, we spent a lot of money, we didn't get anything for it. And then the business returns to, you know, whatever scale, whatever scale it was at. Maybe another way to describe it is, you know, the usual pattern of venture backed investment, which is that things cost a lot. And then, and then you start making it is is kind of happening over and over again in this field within, within the same, within the same companies. And so we're on the exponential now. At some point we'll reach equilibrium. The only relevant questions are at how large a scale do we reach equilibrium? And is there ever an Overshoot.
- [19:23]
Lex Fridman
Right, right. And yeah, you referenced the cloud companies as a point of comparison. But I don't know, there's something about the cloud companies where it feels like their data center capex is more continuous, they're just always doing new data centers, whereas there's something about how discrete these generations are that maybe it's like, you know, the way the engine manufacturers, they keep coming up with new technologies like.
- [19:43]
Dario Amodei
It'S like the F16 or something or it might a little bit like drug development, like you know, kind of an R and D heavy thing.
- [19:49]
Lex Fridman
Yes. And when do you actually go to the effort of training?
- [19:52]
Dario Amodei
Yeah, yeah. You know it's almost like a drug company where it's like you develop one drug and then like, you know, if that works you develop 10 drugs and if that were you develop 100 drugs. You know the drug development market does not work like that numerically, but it is as if it did.
- [20:06]
Lex Fridman
Right. So we can look at each of these models as individual programs and look at their individual P and ls and you're saying that the payback math on those, at least in the models we've seen to date in the industry is not actually that challenging. Where like I think most like in, you know, when you're acquiring a customer, if you have a nine month payback on acquiring a customer, you'll do that all day long, that's like very easy to underwrite. And you're saying the paybacks are kind of 9 months, 12 months. So like they're very easy to do.
- [20:30]
Dario Amodei
I don't want to make any specific claims but qualitatively if you look at the business this way, model by model, it looks very viable.
- [20:40]
Lex Fridman
Yes. Because the ever growing Capex is masking the underlying quality of the model businesses.
- [20:47]
Dario Amodei
Yes.
- [20:48]
Lex Fridman
In 2023 everyone is talking about the data wall. Is this how we solved our way out of the data wall?
- [20:53]
Dario Amodei
Yeah. So I don't know, people talk about things in public and you know, sometimes they're, you know, rumors or suppositions or whatever. I wouldn't even necessarily assume that there's a data wall. One thing I will say is that the idea of using RL has been around for a while. If we go all the way back to when Google DeepMind won beat the world go champion with AlphaGo, it was RL first and then we built these language models and now we're kind of uniting the two together by putting RL on top of the language models. That's all chain of thought or reasoning is, it's just A fancy way of saying RL, where the RL environment is that the model writes a bunch of things and then gives an answer. There's nothing more to it than that. It just kind of has a fancy name. And so I think of these as kind of the two key ways of learning. Right. I think of, like, base LLM training as learning by imitating, and RL as learning by trial and error. I think those are the two styles of learning. Right. You know, if I'm like a child, there's two ways for me to learn. You know, I look at my parents and I'm like, oh, they do something and I try and learn what they do, or I can just kind of like, experiment with the world and learn things. And it's very clear in developmental psychology that people use both. And so we're now seeing that recapitulated in the language models. And so we have a stage where we do the imitative learning, and we have a stage where we learn by trial and error. So it seems very natural to me.
- [22:20]
Lex Fridman
The other thing that's obviously notable to I think, people not in the AI industry looking at it is all of the talent wars and the fact that your IP walks out the door each evening. And you referenced in a recent interview you gave $100 million secrets that were a few lines of code. And obviously I think you were talking about that in a national security context, but you can also think about it in a talent context. And so how does one like, in the pharma industry, they protect their secrets of patents in Wall street, where also they have $100 million secrets that are just a very simple idea. Renaissance Technologies, the hedge fund, just very successfully locks up its employees. How do you make keeping a commercial lead work in kind of the current AI environment?
- [23:08]
Dario Amodei
Yeah. So one thing I will say is that there are some things that are like that, but I think more and more, as the field matures, it starts to be more about know how and ability to build complex kind of objects.
- [23:21]
Lex Fridman
Yes.
- [23:22]
Dario Amodei
Right. So, you know, some of the ideas we work with are simple. But I would say the simple ideas, the ones that are like, oh, yeah, twiddle this element of the transformer or something, those tend to be independently discovered or anyone knows them before too long. But there are things like, oh, man, this thing is actually really hard to implement from an engineering sense, and we have it implemented, or this thing is just kind of a pain to do, or there's a know how to doing it.
- [23:49]
Lex Fridman
Yes.
- [23:51]
Dario Amodei
And those tend to be more collective things that, you know, that are more difficult to leak. And so I think those things are substantially more defensible. That said, you know, there's still leakage and we still don't want it to happen again. Both for commercial competitive reasons and for national security reasons, both are problems. And so a few things we do. One is, you know, we tend to compartmentalize information. So you know, if you talk to any intelligence agency, that's, that's how they operate. You're only told what you need to know. And I think everyone within, within anthropic.
- [24:23]
Lex Fridman
But that's probably quite different to a normal Silicon Valley culture where, you know, everything's just flying around the company.
- [24:27]
Dario Amodei
Yes, we actually do that at the same time as we have a very open culture. I say things to the company that, you know, maybe other person would put it in kind of PR speak or. But when there is a secret, then I think that actually leads to people trusting that it's something that you actually need to know. And then finally having better retention rates and losing less people is one of the most important things here. So we have the highest retention rate of all the AI companies. I think the differences are even starker because everyone has, you know, kind of a non regretted attrition rate that's maybe constant. So if you subtract that off, then the difference is even larger. Sometimes when people leave, they come back.
- [25:11]
Lex Fridman
I saw that recently.
- [25:12]
Dario Amodei
If you look at, you know, you can see publicly the list of people who went to the meta superintelligence lab. Even if you normalize for our size.
- [25:19]
Lex Fridman
Yes.
- [25:20]
Dario Amodei
It's not, you know, and then many, many, you know, many turned them down.
- [25:25]
Lex Fridman
So in the crazy $100 million comp wars that everyone's been talking about, you guys have not had to time of that.
- [25:30]
Dario Amodei
I think relative to other companies, we've, we may even have been relatively advantaged. It's like a mixture of true belief in the mission and belief in the upside of the equity. Like, you know, I think, you know, we anthropic has developed a reputation for doing what it says it will do for in some cases making less promises, but keeping those promises that we make.
- [25:52]
Lex Fridman
Yes.
- [25:53]
Dario Amodei
And you know, being very clear on what we stand for and you know, being consistent over the years and standing for it that creates a unity around the company. And I think it's a good guard.
- [26:03]
Lex Fridman
Against cynicism when you're talking about the upside of the equity. When you're pitching investors or maybe candidates, how do you pitch the anthropic business? Like we're building a Very large business. That's a good start. What else goes into it?
- [26:14]
Dario Amodei
So often I'll talk about the platform and the importance of the models. For some reason, sometimes people think of the API business and they say, oh, it's not very sticky, or it's going to be commodity business.
- [26:26]
Lex Fridman
I love API.
- [26:26]
Dario Amodei
No, no, exactly, exactly. You know, and they're, they're, they're even bigger ones than, than, than both of ours. I would point to the clouds again.
- [26:33]
Lex Fridman
Yes.
- [26:34]
Dario Amodei
You know, those are, those are $100 billion API businesses. And, you know, when the cost of capital is high and there are only a few players. And relative to cloud, the thing we make is much more differentiated. Right. Like, these models have different personalities. They're like talking to different people. You know, a joke, A joke I often make is like, you know, if I'm sitting in a room with like 10 people, does that mean I've been commoditized?
- [26:57]
Lex Fridman
Yes.
- [26:58]
Dario Amodei
You know, there's like nine other people in the room who have a similar brain to me. They're about the same height. They like, you know, so, so who needs me? But, you know, we all know that human labor doesn't work that way. And so, so I feel the same way about this. So, you know, I think the API business is, is, is a great business and, and then, but, you know, we want to go broader than that. You know, the way I think about it is other players, such as OpenAI and existing incumbents such as Google, are very focused on the consumer side. The idea of providing AI to businesses is something that we are trying to get better and better at.
- [27:32]
Lex Fridman
Yes.
- [27:32]
Dario Amodei
And I think we're out to an early lead in that. I'm not sure because I don't know for sure what the revenues of the other players are, but I think we probably at this point have the plurality of the API market, most likely. And AI for business market, perhaps.
- [27:49]
Lex Fridman
Yeah. No, it's funny when you talk about the kind of the commodity argument where we obviously grew up facing this as a skeptical argument. And I remember finding it so striking when AWS finally had to break out their numbers in 2015. Remember, they used to be wrapped up.
- [28:05]
Dario Amodei
In Amazon's numbers and they broke them up.
- [28:07]
Lex Fridman
And people had been talking about, pundits had been saying, oh, cloud is a commodity. It's uninteresting. And then they broke it out. And it was one of the greatest businesses of all time. And there's something where a business can have competitors and it can have buyers who care about price, but that's very different from being a commodity. And as you say, all the products work differently.
- [28:27]
Dario Amodei
No, no, no, no, no. Exactly. I mean we're like one of the biggest customers of the clouds, right. And we use more than one of them. And like, you know, I can tell you the clouds are much less differentiated than the AI models. Right.
- [28:39]
Lex Fridman
For sure. Because it feels like one. The behavior is non deterministic which, that's the by design trying to make it hard. But that just naturally means that, oh, we get the customer service answers we prefer with this model versus that model. And I don't know why.
- [28:50]
Dario Amodei
Exactly. You don't know why. You know, it's a little like baking a cake. Right. It's like, you know, you put in the ingredients, it just, it kind of comes out a certain way. And like, you know, one chef makes it this way and the other chef makes it this way and you're. If you're like make it exactly like that chef makes it.
- [29:04]
Lex Fridman
You can't.
- [29:05]
Dario Amodei
Right? You just can't.
- [29:06]
Lex Fridman
And presumably it's striking to me none of the AI products are that personalized right now, but it feels like personalization will be a huge deal.
- [29:14]
Dario Amodei
Will be a huge deal and will.
- [29:16]
Lex Fridman
Be a big source of stickiness for the. Because you won't want to switch products. And I don't know exactly what that looks like, but given the amount of. For both the consumer and the business use case.
- [29:25]
Dario Amodei
Absolutely, absolutely. I think we've just started to scratch the surface in terms of models that are customized in various ways for working with a particular business or a particular person within the business. So I think we're just seeing the beginning of the API business. But I don't think AI for business is just about API with things like Claude code. You know, we're selling that to not just individual developers, but enterprises as well.
- [29:51]
Lex Fridman
Yes.
- [29:52]
Dario Amodei
And they find it something useful. Claude, for enterprise that is selling to a lot of enterprises. I actually see it and you see this with some of the clouds where they have a bunch of different services. Right. Some of them are apps, some of it is the underlying cloud itself. And what they are is, you know, the way that AWS or GCP or Azure will present themselves. And the way that we are starting to present ourselves is hey, we want to be your one stop shop for AI or for cloud and you can buy all of these things and you can talk to us about which to use for what. And so I think that starts to create the outlines of a more durable business.
- [30:31]
Lex Fridman
If you think about a typical Fortune 500 company, how you know, they're probably playing with AI for customer service, or engineers maybe have AI powered coding tools. How AI adopted are they compared to how much they should be?
- [30:46]
Dario Amodei
Well, certainly much less than there should be.
- [30:47]
Lex Fridman
Yeah. What is it, like 5%, 30%?
- [30:50]
Dario Amodei
So what I would say is there is almost, there is very often conviction at the top. You talk to the CEO, the CEO gets it. You talk to the cto, the CTO gets it. The struggle they have is that they have 100,000. Pick a company that has 100,000 people who their job is to do something else. Their job is to do banking or insurance or drug development. And they've heard about this AI stuff, but you know, they're not like, this is not what they're an expert in. And so the challenge is often we are working with the leadership of the company to get the 100,000 people in the company really familiar with and using the technology. I think again, the code stuff goes the fastest because the developers are the ones who are most adjacent and most watching the trend. Some of the kind of customer service and process stuff is next to go. But you really have the instinct that even with today's models, it could be 100 times bigger than it is. You really get that sense?
- [31:49]
Lex Fridman
Yes. My intuition is sort of this. We will see the patterns of AI adoption from startups because they're unconstrained by existing organizations, so they can kind of do whatever makes sense versus large organizations are somehow calcified because they have all these people whose job it is to do X and need to be consulted and everything like that. So we'll see the new behaviors from the small startups and then large companies, as you say, the CEOs and CTOs are switched on and they're smart. They say, hey, we should be doing that. And they'll kind of port the new ideas from kind of like the adoption we saw of cloud or many of these other tech brands. Is that what you're seeing?
- [32:22]
Dario Amodei
Is that your mission, the new ideas from the small companies or the small companies will become threatening to them and disrupt them. And that will give them the urgency to kind of drive things through and make them happen? Yes, a pattern I've seen that works pretty well, that I actually recommend if you're a large company, is to kind of make a strike team or strike force that's separate from the rest of the company and kind of develops these prototypes. And then basically you can get momentum behind something. And then there's always this hard work of integrating into the rest of the company. But if you have a lot of momentum and you've done the hard work and you've shown the thing works, then it's easier to do that.
- [33:01]
Lex Fridman
Dworkesh. Did you read his recent blog post on his AI timelines?
- [33:05]
Dario Amodei
Oh, on continual learning? Yeah.
- [33:07]
Lex Fridman
Yeah. He talked about how his fundamental issue with many of the AI models for productivity is that they're like the virtual. They're like the super smart virtual coworker who started five minutes ago, but they remain the coworker that started five minutes ago. They don't learn over time. How will we solve that?
- [33:29]
Dario Amodei
Yes. So the pattern that I've seen in AI on the kind of research and technical side is that what we've seen over and over again is that there's what looks like a wall. It looks like AI models can't do this. It was like AI models can't reason. And recently there's this. AI models can't make new discoveries. A few years ago it was like, AI models can't write globally coherent texts, which of course now they obviously can. You go back a few more years and it was like this Chomsky thing of like, they can get syntactics right, but they can't get semantics right. And every one of those has been blown through.
- [34:08]
Lex Fridman
So what have we blown throughout the new discoveries?
- [34:10]
Dario Amodei
This is a thing that people have said recently. Actually, my view on this, like many of the other things on, on new discoveries, is that it's not really a binary. Right.
- [34:21]
Lex Fridman
They don't get to have their name in the paper.
- [34:23]
Dario Amodei
Yeah, they don't get to have their name in the paper. But like, what is a new discovery? Like, what is genius? I, you know, I remember, I think I remember this developmental psychology book, but they were saying something like, you know, we kind of lionize genius, but like, you know, let's say that like a table's wobbly.
- [34:38]
Lex Fridman
Yes.
- [34:39]
Dario Amodei
And you know, I'm like, oh, I, I take the coaster and I put it under the table and it's not wobbly anymore. Like, that's an idea in a way. That's like a new discovery. Like, you know, even if I've never seen someone do that before, you know, that's like a, that's like a new discovery. And you know the difference between that and the Nobel Prize winning discovery? It's, it's a matter of degree, not, not a fundamentally different matter. And so I would say that the AI models make discoveries all the time. Right. I've had family members where the, they had a medical problem and the AI model Diagno, you know, Claude diagnosed their medical problem when doctors missed it. Like, that's not a big news, but like that. That's a new discovery. Like, you know, and you could say, oh, they're just pattern matching to things that. That happened before, but that new discoveries are like that. You think of writers who have written novels or something that are totally new, and you're like, well, what are your influences? And, you know, they're remixing together and adding a new element, so. So it's all more continuous. And that was the thing I was gonna say about continual learning. I think this idea that it isn't present is. I would say it's present a little bit. Yes. We're gonna find a way to get more of it. So, for instance, the models learn within the context. You talk to them and they absorb the context. Eventually the context is gonna be 100 million tokens. And maybe we'll train the model in such a way that it is specialized for learning over the context. You could even, during the context, update the model's weights. So there are lots of ideas that are very close to the ideas we have now that could perhaps do this. I think people are very attached to the idea that they want to believe there's some fundamental wall. There's something different, something that can't be done. It kind of reminds me of.
- [36:22]
Lex Fridman
Do you think it's a coping mechanism deep down?
- [36:23]
Dario Amodei
Yeah. You know what it reminds me of? So, you know, the 19th century notion of vitalism, you know, this was the idea that the. The human body and organisms that are alive are made of a fundamentally different material than inanimate matter, which, of course, we know scientifically now is not true, but it's something people very much want to believe. And your common sense seems to suggest it. Like, I'm not very much like a table. I've made of very different materials than metal or glass or whatever, but when we actually go down to the fundamental units, of course we're all made of the same thing.
- [37:01]
Lex Fridman
Well, you think people now have this kind of modern concept of vitalism in whatever the fundamental humanity is, and they're saying, oh, you know, models can't do that.
- [37:09]
Dario Amodei
I think there's some. I think there's some tendency to. I think there's some tendency to believe it. And I think, as with vitalism, the way around it is to, you know, is to recognize that a mind is a mind, no matter what it's made of. The notion of the dignity or the specialness of cognition or sentience, it's not that it isn't special, it's that it can be made out of anything.
- [37:28]
Lex Fridman
You referenced the medical use case, which I think is a very cool use case, obviously one because of all the people who have fixed medical issues as a result. But another one is you talked in your Machines of Love and Grace post, which I really enjoyed. I thought it was very well done about the marginal returns to intelligence. What are places where intelligence is the limiting factor? And my read of the popular medical use case is obviously it's kind of a charismatic use case, but also for most normal people they have like some kind of medical issue, low level or serious or something like that. And actually society is very just intelligence limited there. Not that you don't have access maybe to a smart doctor, hopefully you do, but they give you very limited time. You know, they think for 10 seconds about your problem.
- [38:09]
Dario Amodei
Yeah, yeah, yeah, exactly.
- [38:11]
Lex Fridman
Test time, compute was actually what we needed there on the medical stuff. But is that kind of your take on this?
- [38:16]
Dario Amodei
That is how I think about it as well. You know, I have talked to Nobel prize winning biologists who say I will only. I mean it sounds a little elitist, but they'll say I'll only go to the top 1% of doctors because the rest of the 99% I can get better advice from an LLM. You know, it really is true. Doctors are busy, they're overworked and just the nature of medical data and medical information, you know, it's a lot of pattern matching. It's a lot of the same things. The level of consistency and the ability to put together many different facts. I think it's something that LLMs are quite good at.
- [38:57]
Lex Fridman
So you talked to this Machines of Love and Grace post about some of the big humanity level areas where we're intelligence limited. But again the personal medical use case is a good example of one where society is intelligence limited. And if you give lots of people much more intelligence on their specific issues, it's very valuable. What are other areas either in the consumer use case or in the business use case where you think we're just very obviously intelligence limited.
- [39:21]
Dario Amodei
Yeah, the places where at least the AI models of today can help the most. The characteristic quality is something is repetitive. But every example is a little different, right?
- [39:32]
Lex Fridman
Yes.
- [39:33]
Dario Amodei
Automation before AI, if you could program exactly how it happened, you could do it. So if you were doing the same thing over and over again. But customer service is like just to take customer service as an example, there's a long tail of stuff, but a lot of it is like you Get a bunch of calls. Each call is different, but each call is basically about one of 10 things. And it's like a different person in a different voice saying basically one of these 10 things in a different way. And that situation where things are repetitive and similar but not the same, and each has its own thing. That's where AI can come in the most. I think.
- [40:11]
Lex Fridman
Dwarkesh had in the same blog post the prediction that you can't yet give an existing AI all of your financial data and forward it all the emails and have it do your taxes. And his prediction for the year in which you can, you know, what is the year where your first tax return is done by just emailing everything to whatever AI you use. His prediction was 20, 28. What do you make of that prediction?
- [40:37]
Dario Amodei
Probably sooner than that.
- [40:38]
Lex Fridman
Okay.
- [40:39]
Dario Amodei
I don't know if it's 26 or 27. Some of that is model, mostly accuracy. I think the model could do that today, but it would make too many mistakes. And so working on ways to have the model check its own work and do less mistakes is one part. There's kind of an interface part of it as well, but I would be surprised if it takes that long.
- [40:58]
Lex Fridman
Okay, 26 or 27. And when you say about mistakes, actually you were running through the list of things that people thought we would never solve in AI. It feels like hallucinations should be on that list. Not they're totally solved, but they've gotten a lot better.
- [41:09]
Dario Amodei
They've gotten a lot better. And I think people have gotten more used to. They kind of know what to trust the model for and what not to trust the model for. The models have also been grounded in citations. I mean, we've done that with CLAUDE AI, We've done that with Enterprise claude. So I think part of the solution is citation. Part of the solution is algorithmically, the models hallucinate less now. And part of the solution is people have adapted and understand the weaknesses of the model. My view on things like hallucinations has always been there's a certain class of critic who points to something where models are weird or worse than what humans do and say, see, they're not like us at all, or they'll never get there. And I kind of get where the instinct comes from, where, like, maybe they're looking to, you know, to see if we've matched the human brain exactly. They're saying, oh, this is so different. It can't be like a human brain. But I basically, I just think it's a fallacy. There's a Notion of kind of general intelligence, but it's made up of a bunch of different things, and you can simply have most of the things and be much worse on some and much better on others. Like, if we look at humans that are.
- [42:14]
Lex Fridman
Have you met humans?
- [42:15]
Dario Amodei
Yeah, yeah. Have you met humans? Right. Like, you know, if you look at, you know, humans who are. Who are, you know, autistic versus humans that are schizophrenic, if you look at the optical illusions that humans face that machines are not fooled by, it's very clear that we have some of these weaknesses, just like, you know, just similar to the model's hallucinations. It's just that they look very different and we're much more used to them because we're surrounded by humans all day.
- [42:41]
Lex Fridman
Yeah. The autonomous vehicle double standard feels like the kind of clearest example of this.
- [42:47]
Dario Amodei
The clearest example of this, yes.
- [42:49]
Lex Fridman
Where people have much higher standards.
- [42:51]
Dario Amodei
People have much higher standards. But I think it's going to be a feature of this technology and it has implications on the business side. I think we're going to be in a world where the models will make mistakes much less often than humans, but there'll be stranger mistakes. And actually that takes some adaptation, because imagine you're an end user. If you work with humans, you get used to it and you have some notion. Right. So if a human makes a mistake 5% of the time, you might have a good understanding of why. You know, like, let's say I'm talking to a customer service agent and they kind of sound incoherent and they're slurring their speech. You know, they've probably had too much of this and they're not doing their job very well. And, you know, that's a bad mistake to happen. But also, if I'm talking to this person, I kind of know what's going on, and I know not to trust what they're saying. Whereas an LLM might make a mistake five times less often. But it's kind of, you know, it's more deceptive. The model sounds just as erudite, just as coherent as it does when it's saying something. That's right. But that's not a, you know, that's an adaptation thing. That's not a fundamental thing. And that's something that, when we talk to our customers, we tell them about that. We tell them they need to get used to that.
- [44:06]
Lex Fridman
So we need to Invent slurring for LLMs is what you.
- [44:08]
Dario Amodei
Right, right, right, exactly.
- [44:13]
Lex Fridman
So you started out as a researcher but now you're the CEO of a company and you're in the business of selling AI. And so what have you had to learn about go to market or dealing with customers?
- [44:27]
Dario Amodei
Yeah, yeah, yeah. All the rest of the stuff, absolutely. I think my view on this was, you know, I started a company not, you know, not because I was initially excited about, you know, selling things or business or any of that. I'd seen the way that some of the other companies had run and, you know, the magnitude and gravity of what they were trying to build. And, you know, I was just a bit concerned that the people and the motivations were maybe not the best ones. And I knew that there would be a number of players in this space, but it felt like having at least one player that kind of had a strong compass in how we do things could have positive effects on the ecosystem. We would build things in a different way, we would deploy them in a different way, and above all, we would have again, sometimes short list of principles, but we would stick to them as well, as well as we could. So I think that was the initial motive. And of course, I was excited about building the technology. And, you know, I think as that has happened, of course, you know, I and the other co founders have kind of had to learn how to think about, you know, kind of the business and the strategy. I think I've been very naturally interested in the business side of it. Actually, I was surprised at how quickly I became interested in it. And actually the primary reason was that I was curious about all the industries that are customers of us. Right. Somewhat like the clouds and perhaps like your business, the businesses that we serve are, you know, they run across every possible industry. And so, you know, you learn these things about parts of the economy that you've never thought about. And even in areas where nominally, you know, a lot like, you know, I used to be a biologist, so I, in a way, I know a lot about the pharmaceutical business, but, you know, I never thought about it as a, you know, beyond the science. I never thought about the portfolio side of it. I never thought about how clinical trials work and how they could be made cheaper. You know, I never thought about the, you know, defense and intelligence business in any great detail. And so, you know, you run through those and I just, I just find it super interesting to understand what people's problems are and how AI can help with those problems. And so I feel like I took very, very naturally to that. Actually, the product side was one where I was initially more reluctant. I felt like I just had a Natural interest and curiosity in the. In the business side of it. But like building apps, it was somehow initially it was. It was never like, you know, a thing that drew me in, even after I started the. Even after I started the company. But, you know, I think more recently, as I've seen what products have succeeded and what products haven't, I think this idea of how to design products so that they're, you know, what we call AGI pilled. Right. So that the direction of the product is durable and is kind of a bridge to things that are useful in the future. Right. We've all heard this idea of wrapper companies or wrapper products. The idea is, you know, you make Claude n and, you know, someone makes a product that, you know, basically addresses the deficiencies of clawed n, but then you come out with of Claude n plus one and it just kind of eats it. The advice I always give that, I think all the AI, you know, all the folks at the AI companies give is like, you know, don't make that. See the direction of the field and try to make something that's complementary. And I think thinking about how to make products in a new way, in a kind of AGI pilled way, that actually has caught my interest a great deal.
- [47:54]
Lex Fridman
Okay, so glad you brought this up. Doesn't it feel like we have no AI UIs right now? Like we still enter text into text boxes, you know, literally same as terminals from the 1970s. I mean, a bit more rounded corners and everything. Like, we still talk into voice companion modes that are manually triggered, which is the same as pre Transformer Siri. So UIs are just.
- [48:19]
Dario Amodei
Yeah, yeah, there's something not quite right about it. I basically. I agree with you. You know, it reminds me a little bit of, you know, in the early days, days of the Internet. It was like, you know, people. People would make these websites that had, you know, structures that look like they were in the physical world. You know, open the closet and, like, do the horseless carriage. There's some term for this. I forget there's some word for this. I forget. I forget what it is. Skeuomorphism. Yes. It feels like there's some of that going on here. A thing I would say is that as we move more towards agents, we're gonna be in a world where the AI model can do something end to end. Like, we're almost there with Claude can do something end to end and get it right most of the time.
- [49:04]
Lex Fridman
Yes.
- [49:04]
Dario Amodei
And a human's main job is to kind of check. Right. Or check sometimes but interestingly, checking often means getting really into the details of what happened. And so there's some kind of impedance mismatch here that some product or interface is a solution to where you want something that's as slick as possible and just goes off and does something and you don't want to have to pay attention most of the time, but when something's wrong, you might actually need to get quite involved. And I don't feel like any products or interfaces operate on this principle now or handle this problem now.
- [49:42]
Lex Fridman
Yes, yes.
- [49:43]
Dario Amodei
I don't know if that makes sense, but.
- [49:44]
Lex Fridman
No, it does. I agree. I think what you want is your agent to go away and do really good work for, for you and then come back with its work product to let you review, steer it, decide.
- [49:56]
Dario Amodei
But you can't be overwhelmed because it's going to do so many more things than you have time to look at. And if you're always looking at it, it can be slower than if you just did it yourself. And so it actually strikes me as an interface problem.
- [50:09]
Lex Fridman
Yes, yes. The generalization of this is it feels to me one of the most exciting things about AI is we have such an overhang of current capabilities turning them into good products where even if AI progress was frozen right now, we'd have like 10 years of good product.
- [50:27]
Dario Amodei
Oh, oh, I completely agree. And actually the way that products are being built, I think by everyone in the industry, but we've thought about it this way is very different because the progress is continuing. If the progress in models stopped, the way we built products would change instantly. The reason is I don't think we've ever had before a situation in which the technology is changing under you so fast as you're building the product. And so this idea of long term product roadmaps or the usual way of product planning, I've started explicitly again early in Anthropic I was like, I don't know anything about product, I'm a doofus. But now I always try to talk to people when they come in and they say this is not like building products in the non AI space. Right.
- [51:13]
Lex Fridman
They need to be more AGI.
- [51:14]
Dario Amodei
Yeah, you may be the expert at building these, but like the technology is moving under you. So like these ideas about fast iteration, they're even more true than they are normally.
- [51:23]
Lex Fridman
What's a specific example of this?
- [51:25]
Dario Amodei
I think that if you're trying to make like, you know, you're like, we're going to make something and it's going to be ready in six months. I Think that makes even less sense here than it makes, you know, the building in isolation.
- [51:37]
Lex Fridman
So you need to have tighter ship schedules.
- [51:40]
Dario Amodei
You need to have tighter ship schedules. You need to try things. It's very hard to tell, even harder to tell what's going to catch on because a new model may have come out and a new model may suddenly be good at something that makes a product possible. And so much more than anything else, you're trying something that's never been tried. Right. There's a new model, it's only available within the company. So the thing you ought to do is just build something on it, let people internally try it. There's this eternal September vibe to it where it's as if you disc discovered database technology for the first time and you're like, well, what could you build on this? Right? And it's always the first day. Right. That's what is different.
- [52:22]
Lex Fridman
You mentioned database technology and maybe that provides an interesting analogy. And as we think about open source, the first relational databases that were successful in terms of adoption were proprietary. But then the open source guys caught up. How do you keep the gap with the open source?
- [52:41]
Dario Amodei
Yeah. So, you know, open source, I think, has a different meaning in AI models than it has in other areas. Right. And you know, for this reason, some have called it like open weights models. Right. To distinguish. I think the main difference is that if you see the weights of the models and you look in, you can't understand what's actually going on. Right. There's not that kind of composability. I can't read the source code. I can't, I can't.
- [53:05]
Lex Fridman
You can't produce a trivially different model?
- [53:07]
Dario Amodei
Yeah, yeah. You know, I can't produce a trivially different version of it. Now Anthropic is actually working on mechanistic interpretability, which allows you to see inside the models. And so we're actually working on things that would allow some properties, but we're not there yet. We're not anywhere close to there. There are some things you can do. For example, if you have access to the model, you can fine tune the model. We're now through interfaces, kind of allowing people to fine tune the model. So there is a question of how valuable access to the actual model weights is over and above some thick API that lets you do something. There's some question of economics, but note that it costs a significant amount to run the models on the cloud. Someone has to host it, someone has to run fast inference and then you're back to the margin or some portion of the margin.
- [53:54]
Lex Fridman
So you think open weight models are not that useful and fully open source models, there's just a big gap.
- [53:59]
Dario Amodei
I guess what I would say is that the analogy to the, to previous technologies is only partial. Right. It's kind of a different thing that we're still discovering. But I can say from our perspective that when a new model comes out, when a competitor model comes out, we don't really think about whether it's an open weights model or not. We think about whether it's a strong model. So if someone makes a strong model, that's good at the things that we do, like, like that's competition, that's bad for us. Whether it's an open weights model or not, there's not a huge difference between the two.
- [54:43]
Lex Fridman
How is anthropic more AGI pilled than other organizations? So one is faster, like a tighter product release cadence, but maybe more broadly across the organization, not just within product development.
- [54:53]
Dario Amodei
Yeah. So, you know, I mentioned this thing that every couple of weeks I get up in front of the organization and you know, and kind of of describe my vision. And I think one of the purposes of that is to keep people kind of focused on the mission. It's a strange state of the world and I always express uncertainty about it. But I say if I were to bet, I would bet in favor of this, that in one or two or three years, I don't know exactly how long it's gonna be, we'll have what I've described as a country of geniuses in the data center. And this is weird. It's gonna change the economy, it's gonna accelerate the pace of science. It's going to pose global alignment and national security risks. It may pose economic problems. The upside is huge. The potential for disruption is also huge. And I think what I'm trying to fight against is the idea of employees who join and they're like, oh, I worked in this industry, I worked at this kind of company and I'm going to work at an AI company. And maybe a couple years later I'll go to.
- [55:57]
Lex Fridman
This is very categorically different.
- [55:59]
Dario Amodei
This is really different, different thing. And, and I think up and down the organization we want to make sure that, you know, when our, you know, when our finance people, you know, think about financial projections, they, they understand this. Not that there's necess, there's certainly going to be an exponential, but like wild outcomes are possible. Right. When our recruiting thinks they're, they're, you know, they're like, oh yeah, you know, like this crazy cop stuff could happen because. Because it happened. And you know, and when the product people think they make make AGI pilled products, when the policy people interact, they understand the stakes of what may happen. And so I think a big part of my job is keeping the coherence of the organization around this central thesis. Not that everyone has to believe the thesis, right? It's not like there's not a indoctrination and people chanting with robes or anything. But the basic idea that the company is. Is built around this hypothesis that it is possible and perhaps likely that these large changes will happen. And every aspect of the business, as well as the things the company is doing for social benefit, should be constructed around a strong possibility that this may happen.
- [57:11]
Lex Fridman
To put numbers on this, you've talked about the potential for 10% annual economic growth powered by AI. Doesn't that mean that when we talk about AI risk, it's often harms and misuses of AI? Isn't the big AI risk that we slightly misregulated or we slow down progress? And therefore there's just like a lot of human welfare that's missed out on.
- [57:36]
Dario Amodei
I've had the experience where I've had family members die of diseases that were cured a few years after they die. So I kind of truly understand the stakes of not making progress fast enough. I would say that some of the dangers of AI have the potential to significantly destabilize society or threaten humanity or civilization. And so I think we don't want to, you know, we don't want to take idle chances with that level of risk. Now, I'm not at all an advocate of like, stop the technology, pause the, you know, pause the technology. I think for a number of reasons, I think that's just, it's just not possible. Like we have geopolitical adversaries. Like they're not going to not make the technology the amount of money. I mean, if you even propose even the slightest amount of. I have, and I have many trillions of dollars of capital lined up against me, for whom that's not in their interest. So that shows the limits of what is possible and what is not possible. But what I would say is that instead of thinking about slowing it down versus going at the maximum speed, are there ways that we can introduce safety security measures, think about the economy in ways that either don't slow the technology down or only slow it down a little bit. Like if instead of 10% economic growth, we could have 9% economic growth and buy insurance against all of these risks. I Think that's what the trade off actually looks like. And precisely because AI is a technology that has the potential to go so quickly to solve so many problems, I see the greater risk as like, the thing could overheat. Right. And so I basically want. I don't want to stop the reaction. I want to focus it. That's how I think about.
- [59:32]
Lex Fridman
You said if we hit December 2025 and there's no AI law, I'll be really worried. How are you feeling?
- [59:38]
Dario Amodei
There is actually something in California. There's a bill out, SB53. Of course, you know, last year we had the whole SB 1047 thing. You know, we had mixed feelings on SB 1047. There was initial version that, you know, I think was too aggressive. And when I say that, what I mean is the technology is moving fast and it's kind of unhelpful if you're too prescriptive about it, you know, it ends up actually not contributing to safety. And I was worried a little bit. If something like this passes, it's like the tests that were prescribed to run will end up looking stupid. And then, like, all the people in the AI industry will be like, oh, this is what regulation for safety and security looks like. It's really stupid, and they won't take it seriously. They'll kind of, you know, do everything they can to comply in letter and not in spirit. And so as a advocate of thoughtful regulation, I was actually a bit concerned about this. We offered some changes to the bill to a point where kind of we felt good about it, and we tried to make a compromise between kind of industry and the safety advocates. We didn't really succeed, as you saw. But this year, I think we're making a bill that is something more moderate. It's focused particularly on transparency of practices, transparency of safety and security practices, which is something that that anthropic has been very forward about and that I think other companies are starting to do, but not all the companies do it. And there's no way to tell if folks are telling the truth about what they're revealing.
- [61:04]
Lex Fridman
And California regulation is enough because all the companies have nexus here.
- [61:08]
Dario Amodei
Yeah. Yeah. I mean, I think, you know, most of these bills are organized around, you know, doing business in California. And so, you know, it would be difficult to shut off.
- [61:16]
Lex Fridman
People are very AI built here.
- [61:18]
Dario Amodei
People are very AI pilled here. So we'll see what happens. I'm not sure what's going to happen, but we've always had this approach that we kind of are in favor of guardrails including legislative guardrails on the technology. But we recognize the need to be careful. Like, we don't want to kill the golden goose. We just want to stop it from overheating or running off the road.
- [61:46]
Lex Fridman
Yeah, maybe something like modern bank regulation for all people complain is a good example where there's an inherently very risky activity.
- [61:54]
Dario Amodei
Yeah, no, no, the dangers are pretty clear.
- [61:57]
Lex Fridman
I mean, it's like the bank runs are not right, but it all works pretty well in the modern era once we figured out the regulatory environment. Last question. What is your personal AI stack? How do you use AI differently to maybe other people in tech?
- [62:09]
Dario Amodei
Yeah. Interesting. I basically write a lot. Perhaps I have too much pride in my own. In my own writing. I use Claude to generate lots of ideas. You know, I kind of use it as research, but so far I've done the writing myself. Claude is actually maybe closer than the other ones, but it's still not there. Like, I'd be comfortable with it for business emails, but if I'm kind of like writing an essay or something that I want to really get right, it's not quite there yet, but maybe it will be in a year or so. Cool.
- [62:44]
Lex Fridman
Yeah, very cool. Well, Dara, this is awesome.
- [62:46]
Dario Amodei
Yeah, thanks for coming by. Thank you for having me.