Dario Amodei: Anthropic CEO on Claude, AGI & the Future of AI & Humanity | Lex Fridman Podcast #452

59,931 words · about 240 min to read

Extrapolating current trends points to AGI by 2026 or 2027

0:00 - If you extrapolate the curves that we've had so far, right? If you say, well, I don't know, we're starting to get to like PhD level, and last year we were at undergraduate level, and the year before we were at like the level of a high school student. Again, you can quibble with at what tasks and for what, we're still missing modalities, but those are being added, like computer use was added, like image generation has been added. If you just kind of like eyeball the rate at which these capabilities are increasing, it does make you think that we'll get there by 2026 or 2027.

0:31 I think there are still worlds where it doesn't happen in 100 years. Those world, the number of those worlds is rapidly decreasing. We are rapidly running out of truly convincing blockers, truly compelling reasons why this will not happen in the next few years. The scale up is very quick. Like we do this today, we make a model, and then we deploy thousands, maybe tens of thousands of instances of it. I think by the time, you know, certainly within two to three years, whether we have these super powerful AIs or not, clusters are gonna get to the size where you'll be able to deploy millions of these.

1:03 I am optimistic about meaning. I worry about economics and the concentration of power. That's actually what I worry about more, the abuse of power. - And AI increases the amount of power in the world, and if you concentrate that power and abuse that power, it can do immeasurable damage. - Yes, it's very frightening. It's very frightening. - The following is a conversation with Dario Amodei, CEO of Anthropic, the company that created Claude that is currently and often at the top of most LLM benchmark leaderboards.

1:40 On top of that, Dario and the Anthropic team have been outspoken advocates for taking the topic of AI safety very seriously, and they have continued to publish a lot of fascinating AI research on this and other topics. I'm also joined afterwards by two other brilliant people from Anthropic. First Amanda Askell, who is a researcher working on alignment and fine tuning of Claude, including the design of Claude's character and personality.

2:11 A few folks told me she has probably talked with Claude more than any human at Anthropic. So she was definitely a fascinating person to talk to about prompt engineering and practical advice on how to get the best out of Claude. After that, Chris Olah stopped by for a chat. He's one of the pioneers of the field of mechanistic interpretability, which is an exciting set of efforts that aims to reverse engineer neural networks to figure out what's going on inside, inferring behaviors from neural activation patterns inside the network.

2:48 This is a very promising approach for keeping future super intelligent AI systems safe. For example, by detecting from the activations when the model is trying to deceive the human it is talking to. This is the "Lex Fridman Podcast." To support it, please check out our sponsors in the description. And now, dear friends, here's Dario Amodei.

Scaling laws emerged as an empirical discovery across domains

1 Highlight · Discovering the Empirical Scaling Hypothesis Early On

3:14 Let's start with the big idea of scaling laws and the Scaling Hypothesis. What is it? What is its history? And where do we stand today? - So I can only describe it as, you know, as it relates to kind of my own experience. But I've been in the AI field for about 10 years and it was something I noticed very early on. So I first joined the AI world when I was working at Baidu with Andrew Ng in late 2014, which is almost exactly 10 years ago now.

3:41 And the first thing we worked on was speech recognition systems. And in those days I think deep learning was a new thing, it had made lots of progress, but everyone was always saying, we don't have the algorithms we need to succeed. You know, we're not, we're only matching a tiny, tiny fraction. There's so much we need to kind of discover algorithmically. We haven't found the picture of how to match the human brain. And when, you know, in some ways I was fortunate, I was kind of, you know, you can have almost beginner's luck, right?

4:10 I was like a newcomer to the field and, you know, I looked at the neural net that we were using for speech, the recurrent neural networks, and I said, I don't know, what if you make them bigger and give them more layers? And what if you scale up the data along with this, right? I just saw these as like independent dials that you could turn. And I noticed that the model started to do better and better as you gave them more data, as you made the models larger, as you trained them for longer.

4:37 And I didn't measure things precisely in those days, but along with colleagues, we very much got the informal sense that the more data and the more compute and the more training you put into these models, the better they perform. And so initially my thinking was, hey, maybe that is just true for speech recognition systems, right? Maybe that's just one particular quirk, one particular area. I think it wasn't until 2017 when I first saw the results from GPT-1 that it clicked for me that language is probably the area in which we can do this.

5:11 We can get trillions of words of language data, we can train on them. And the models we were training those days were tiny. You could train them on one to eight GPUs whereas, you know, now we train jobs on tens of thousands, soon going to hundreds of thousands of GPUs. And so when I saw those two things together and, you know, there were a few people like Ilya Sutskever, who you've interviewed, who had somewhat similar views, right?

5:36 He might have been the first one, although I think a few people came to similar views around the same time, right? There was, you know, Rich Sutton's Bitter Lesson, there was Gwern wrote about the Scaling Hypothesis. But I think somewhere between 2014 and 2017 was when it really clicked for me, when I really got conviction that, hey, we're gonna be able to do these incredibly wide cognitive tasks if we just scale up the models.

6:03 And at every stage of scaling, there are always arguments. And you know, when I first heard them, honestly, I thought probably I'm the one who's wrong, and, you know, all these experts in the field are right. They know the situation better than I do, right? There's, you know, the Chomsky argument about like, you can get syntactics, but you can't get semantics. There was this idea, oh, you can make a sentence make sense, but you can't make a paragraph make sense. The latest one we have today is, you know, we're gonna run out of data, or the data isn't high quality enough, or models can't reason.

6:34 And each time, every time we manage to either find a way around or scaling just is the way around. Sometimes it's one, sometimes it's the other. And so I'm now at this point, I still think, you know, it's always quite uncertain. We have nothing but inductive inference to tell us that the next two years are gonna be like the last 10 years. But I've seen the movie enough times, I've seen the story happen for enough times to really believe that probably the scaling is going to continue and that there's some magic to it that we haven't really explained on a theoretical basis yet.

7:10 - And of course the scaling here is bigger networks, bigger data, bigger compute. - Yes. - All of those. - In particular, linear scaling up of bigger networks, bigger training times and more and more data. So all of these things, almost like a chemical reaction, you know, you have three ingredients in the chemical reaction, and you need to linearly scale up the three ingredients. If you scale up one, not the others, you run out of the other reagents and the reaction stops.

7:40 But if you scale up everything in series, then the reaction can proceed. - And of course, now that you have this kind of empirical science/art, you can apply to other more nuanced things like scaling laws applied to interpretability, or scaling laws applied to post-training, or just seeing how does this thing scale. But the big scaling law, I guess the underlying Scaling Hypothesis has to do with big networks, big data leads to intelligence.

8:09 - Yeah, we've documented scaling laws in lots of domains other than language, right? So initially, the paper we did that first showed it was in early 2020 where we first showed it for language. There was then some work late in 2020 where we showed the same thing for other modalities, like images, video, text-to-image, image-to-text math. They all had the same pattern.

8:34 And you're right, now there are other stages like post-training or there are new types of reasoning models. And in all of those cases that we've measured, we see similar types of scaling laws. - A bit of a philosophical question, but what's your intuition about why bigger is better

A long tail of natural patterns explains why larger models are smarter

2 Highlight · Why Larger Networks Capture Complex Patterns

8:53 in terms of network size and data size? Why does it lead to more intelligent models? - So in my previous career as a biophysicist, so I did physics undergrad and then biophysics in grad school. So I think back to what I know as a physicist, which is actually much less than what some of my colleagues at Anthropic have in terms of expertise in physics. There's this concept called the 1/f noise and 1/x distributions where often, you know, just like if you add up a bunch of natural processes, you get a Gaussian.

9:30 If you add up a bunch of kind of differently distributed natural processes, if you like take a probe and hook it up to a resistor, the distribution of the thermal noise in the resistor goes as one over the frequency. It's some kind of natural convergent distribution. And I think what it amounts to is that if you look at a lot of things that are produced by some natural process that has a lot of different scales, right?

9:59 Not a Gaussian, which is kind of narrowly distributed, but you know, if I look at kind of like large and small fluctuations that lead to electrical noise, they have this decaying 1/x distribution. And so now I think of like patterns in the physical world, right? Or in language. If I think about the patterns in language, there are some really simple patterns. Some words are much more common than others like "the," then there's basic noun verb structure, then there's the fact that, you know, nouns and verbs have to agree, they have to coordinate.

10:32 And there's the higher level sentence structure, then there's the thematic structure of paragraphs. And so the fact that there's this regressing structure, you can imagine that as you make the networks larger, first they capture the really simple correlations, the really simple patterns, and there's this long tail of other patterns. And if that long tail of other patterns is really smooth like it is with the 1/f noise in, you know, physical processes like resistors, then you can imagine as you make the network larger, it's kind of capturing more and more of that distribution, and so that smoothness gets reflected in how well the models are at predicting and how well they perform.

11:10 Language is an evolved process, right? We've developed language, we have common words and less common words. We have common expressions and less common expressions. We have ideas, cliches that are expressed frequently, and we have novel ideas.

11:26 And that process has developed, has evolved with humans over millions of years. And so the guess, and this is pure speculation would be that there's some kind of long tail distribution of the distribution of these ideas. - So there's the long tail, but also there's the height of the hierarchy of concepts that you're building up. So the bigger the network, presumably you have a higher capacity to- - Exactly, if you have a small network, you only get the common stuff, right?

11:54 if I take a tiny neural network, it's very good at understanding that, you know, a sentence has to have, you know, verb adjective noun right? But it's terrible at deciding what those verb, adjective and noun should be and whether they should make sense. If I make it just a little bigger, it gets good at that, then suddenly it's good at the sentences, but it's not good at the paragraphs. And so these rarer and more complex patterns get picked up as I add more capacity to the network.

12:20 - Well, the natural question then is, what's the ceiling of this? - Yeah. - Like how complicated and complex is the real world? How much stuff is there to learn? - I don't think any of us knows the answer to that question. My strong instinct would be that there's no ceiling below the level of humans, right? We humans are able to understand these various patterns, and so that makes me think that if we continue to, you know, scale up these models to kind of develop new methods for training them and scaling them up, that will at least get to the level that we've gotten to with humans.

12:55 There's then a question of, you know, how much more is it possible to understand than humans do? How much is it possible to be smarter and more perceptive than humans? I would guess the answer has got to be domain dependent. If I look at an area like biology, and, you know, I wrote this essay, "Machines of Loving Grace." It seems to me that humans are struggling to understand the complexity of biology, right?

13:20 If you go to Stanford or to Harvard or to Berkeley, you have whole departments of, you know, folks trying to study, you know, like the immune system or metabolic pathways, and each person understands only a tiny bit, part of it, specializes, and they're struggling to combine their knowledge with that of other humans. And so I have an instinct that there's a lot of room at the top for AIs to get smarter.

13:46 If I think of something like materials in the physical world or you know, like addressing, you know, conflicts between humans or something like that. I mean, you know, it may be there's only, some of these problems are not intractable, but much harder. And it may be that there's only so well you can do at some of these things, right? Just like with speech recognition, there's only so clear I can hear your speech. So I think in some areas there may be ceilings, you know, that are very close to what humans have done.

14:16 in other areas, those ceilings may be very far away. And I think we'll only find out when we build these systems. It's very hard to know in advance. We can speculate, but we can't be sure. - And in some domains, the ceiling might have to do with human bureaucracies and things like this, as you write about. - Yes. - So humans fundamentally have to be part of the loop. That's the cause of the ceiling, not maybe the limits of the intelligence. - Yeah, I think in many cases, you know, in theory, technology could change very fast, for example, all the things that we might invent with respect to biology.

14:49 But remember there's a, you know, there's a clinical trial system that we have to go through to actually administer these things to humans. I think that's a mixture of things that are unnecessary and bureaucratic and things that kind of protect the integrity of society. And the whole challenge is that it's hard to tell what's going on. It's hard to tell which is which, right? My view is definitely, I think in terms of drug development, my view is that we're too slow and we're too conservative.

15:17 But certainly if you get these things wrong, you know, it's possible to risk people's lives by being too reckless. And so at least some of these human institutions are in fact protecting people. So it's all about finding the balance. I strongly suspect that balance is kind of more on the side of pushing to make things happen faster, but there is a balance. - If we do hit a limit,

Data and compute scaling face steep hurdles but have viable workarounds

15:41 if we do hit a slow down in the scaling laws, what do you think would be the reason? Is it compute limited, data limited? Is it something else, idea limited? - So, a few things. Now we're talking about hitting the limit before we get to the level of humans and the skill of humans. So, I think one that's, you know, one that's popular today and I think, you know, could be a limit that we run into. Like most of the limits, I would bet against it, but it's definitely possible is we simply run out of data.

16:09 There's only so much data on the internet and there's issues with the quality of the data, right? You can get hundreds of trillions of words on the internet, but a lot of it is repetitive or it's search engine, you know,

3 Highlight · Overcoming Data Limits Through Synthetic Data and Self-Play

16:22 search engine optimization drivel, or maybe in the future it'll even be text generated by AIs itself. And so I think there are limits to what can be produced in this way. That said, we and I would guess other companies are working on ways to make data synthetic where you can, you know, you can use the model to generate more data of the type that you have already or even generate data from scratch.

16:49 If you think about what was done with DeepMind's AlphaGo Zero, they managed to get a bot all the way from, you know, no ability to play Go whatsoever to above human level just by playing against itself. There was no example data from humans required in the AlphaGo Zero version of it. The other direction, of course, is these reasoning models that do chain of thought and stop to think and reflect on their own thinking. In a way, that's another kind of synthetic data coupled with reinforcement learning.

17:19 So my guess is with one of those methods, we'll get around the data limitation or there may be other sources of data that are available.

17:27 We could just observe that even if there's no problem with data, as we start to scale models up, they just stop getting better. It seemed to be a reliable observation that they've gotten better, that could just stop at some point for a reason we don't understand. The answer could be that we need to, you know, we need to invent some new architecture. There have been problems in the past with, say, numerical stability of models where it looked like things were leveling off, but actually, you know, when we found the right unblocker, they didn't end up doing so.

18:02 So perhaps there's some new optimization method or some new technique we need to unblock things. I've seen no evidence of that so far. But if things were to slow down, that perhaps could be one reason. - What about the limits of compute? Meaning the expensive nature of building bigger and bigger data centers. - So right now, I think, you know, most of the frontier model companies I would guess are operating in, you know, roughly, you know, $1 billion scale, plus or minus a factor of three, right?

18:34 Those are the models that exist now or are being trained now. I think next year, we're gonna go to a few billion, and then 2026, we may go to, you know, above 10 billion, and probably by 2027, their ambitions to build 100 billion dollar clusters, and I think all of that actually will happen. There's a lot of determination to build the compute to do it within this country, and I would guess that it actually does happen.

19:02 Now, if we get to 100 billion, that's still not enough compute, that's still not enough scale then either we need even more scale or we need to develop some way of doing it more efficiently of shifting the curve. I think between all of these, one of the reasons I'm bullish about powerful AI happening so fast is just that if you extrapolate the next few points on the curve, we're very quickly getting towards human level ability, right?

19:28 Some of the new models that we developed, some reasoning models that have come from other companies, they're starting to get to what I would call the PhD or professional level, right? If you look at their coding ability, the latest model we released, Sonnet 3.5, the new or updated version, it gets something like 50% on SWE-bench, and SWE-bench is an example of a bunch of professional, real world software engineering tasks.

19:54 At the beginning of the year, I think the state of the art was 3 or 4%. So in 10 months we've gone from 3% to 50% on this task, and I think in another year, we'll probably be at 90%. I mean, I don't know, but might even be less than that. We've seen similar things in graduate level math, physics, and biology from models like OpenAI's o1. So if we just continue to extrapolate this, right, in terms of skill that we have, I think if we extrapolate the straight curve, within a few years, we will get to these models being, you know, above the highest professional level in terms of humans.

20:35 Now, will that curve continue? You've pointed to, and I've pointed to a lot of reasons why,

Mechanistic interpretability drives a race to the top for safety

20:39 you know, possible reasons why that might not happen. But if the extrapolation curve continues, that is the trajectory we're on. - So Anthropic has several competitors. It'd be interesting to get your sort of view of it all. OpenAI Google xAI Meta. What does it take to win in the broad sense of win in this space? - Yeah, so I want to separate out a couple things, right? So, you know, Anthropic's mission is to kind of try to make this all go well, right?

21:06 And you know, we have a theory of change called race to the top, right? Race to the top is about trying to push the other players to do the right thing by setting an example. It's not about being the good guy, it's about setting things up so that all of us can be the good guy. I'll give a few examples of this. Early in the history of Anthropic, one of our co-founders, Chris Olah, who I believe you're interviewing soon, you know, he's the co-founder of the field of mechanistic interpretability, which is an attempt to understand what's going on inside AI models.

21:39 So we had him and one of our early teams focus on this area of interpretability, which we think is good for making models safe and transparent. For three or four years, that had no commercial application whatsoever. It still doesn't today. We're doing some early betas with it, and probably it will eventually, but you know, this is a very, very long research bed and one in which we've built in public and shared our results publicly.

22:05 And we did this because, you know, we think it's a way to make models safer. An interesting thing is that as we've done this, other companies have started doing it as well, in some cases because they've been inspired by it, in some cases because they're worried that, you know, if other companies are doing this to look more responsible, they wanna look more responsible too. No one wants to look like the irresponsible actor, and so they adopt this as well.

22:32 When folks come to Anthropic, interpretability often a draw, and I tell them, the other places you didn't go, tell them why you came here, and then you see soon that there's interpretability teams elsewhere as well. And in a way, that takes away our competitive advantage because it's like, oh, now others are doing it as well, but it's good for the broader system, and so we have to invent some new thing that we're doing that others aren't doing as well.

22:59 And the hope is to basically bid up the importance of doing the right thing. And it's not about us in particular, right? It's not about having one particular good guy. Other companies can do this as well. If they join the race to do this, you know, that's the best news ever, right? It's just, it's about kind of shaping the incentives to point upward instead of shaping the incentives to point downward.

23:25 - And we should say this example of the field of mechanistic interpretability is just a rigorous, non-hand wavy way of doing AI safety, or it's tending that way. - Trying to, I mean, I think we're still early in terms of our ability to see things, but I've been surprised at how much we've been able to look inside these systems and understand what we see, right? Unlike with the scaling laws where it feels like there's some, you know,

4 Highlight · Reverse Engineering Neural Networks with Mechanistic Interpretability

23:51 law that's driving these models to perform better, on the inside, the models aren't, you know, there's no reason why they should be designed for us to understand them, right? They're designed to operate, they're designed to work, just like the human brain or human biochemistry. They're not designed for a human to open up the hatch, look inside and understand them. But we have found, and you know, you can talk in much more detail about this to Chris, that when we open them up, when we do look inside them, we find things that are surprisingly interesting.

24:20 - And as a side effect, you also get to see the beauty of these models. You get to explore sort of the beautiful nature of large neural networks through the mech interp kind of methodology. - I'm amazed at how clean it's been. I'm amazed at things like induction heads. I'm amazed at things like, you know, that we can, you know, use sparse auto-encoders to find these directions within the networks, and that the directions correspond to these very clear concepts.

24:49 We demonstrated this a bit with the Golden Gate Bridge Claude. So this was an experiment where we found a direction inside one of the neural network's layers that corresponded to the Golden Gate Bridge and we just turned that way up. And so we released this model as a demo, it was kind of half a joke, for a couple days, but it was illustrative of the method we developed. And you could take the Golden Gate, you could take the model, you could ask it about anything, you know, it would be like, you could say, "How was your day" and anything you asked, because this feature was activated, would connect to the Golden Gate Bridge.

25:23 So it would say, you know, "I'm feeling relaxed and expansive, much like the arches of the Golden Gate Bridge" or, you know. - It would masterfully change topic to the Golden Gate Bridge and it integrate it. There was also a sadness to it, to the focus it had on the Golden Gate Bridge. I think people quickly fell in love with it, I think, so people already miss it 'cause it was taken down I think after a day. - Somehow these interventions on the model where you kind of adjust its behavior somehow emotionally made it seem more human than any other version of the model.

25:56 - It has a strong personality, strong identity.

25:58 - It has a strong personality. It has these kind of like obsessive interests. You know, we can all think of someone who's like obsessed with something. So it does make it feel somehow a bit more human. - Let's talk about the present. Let's talk about Claude. So this year, a lot has happened. In March, Claude 3, Opus, Sonnet, Haiku were released, then Claude 3.5 Sonnet in July,

Product tiers systematically shift the intelligence-cost trade-off curve

26:21 with an updated version just now released, and then also Claude 3.5 Haiku was released. Okay, can you explain the difference between Opus, Sonnet and Haiku, and how we should think about the different versions? - Yeah, so let's go back to March when we first released these three models. So, you know, our thinking was, you know, different companies produce kind of large and small models, better and worse models.

26:47 We felt that there was demand both for a really powerful model, you know, and you that might be a little bit slower that you'd have to pay more for, and also for fast, cheap models that are as smart as they can be for how fast and cheap, right? Whenever you wanna do some kind of like, you know, difficult analysis, like if, you know, I wanna write code, for instance, or you know, I wanna brainstorm ideas, or I wanna do creative writing, I want the really powerful model.

27:15 But then there's a lot of practical applications in a business sense where it's like I'm interacting with a website. You know, like, I'm like doing my taxes, or I'm, you know, talking to a, you know, to like a legal advisor and I want to analyze a contract or, you know, we have plenty of companies that are just like, you know, I wanna do auto complete on my IDE or something. And for all of those things, you want to act fast and you want to use the model very broadly.

27:42 So we wanted to serve that whole spectrum of needs. So we ended up with this, you know, this kind of poetry theme. And so what's a really short poem? It's a haiku. And so Haiku is the small, fast, cheap model that is, you know, was at the time was released surprisingly, surprisingly intelligent for how fast and cheap it was. Sonnet is a medium sized poem, right, a couple paragraphs, and so Sonnet was the middle model.

28:08 It is smarter but also a little bit slower, a little bit more expensive. And Opus, like a magnum opus is a large work, Opus was the largest, smartest model at the time. So that was the original kind of thinking behind it. And our thinking then was, well, each new generation of models should shift that trade-off curve. So when we released Sonnet 3.5, it has the same, roughly the same, you know, cost and speed as the Sonnet 3 model.

28:42 But it increased its intelligence to the point where it was smarter than the original Opus 3 model, especially for code but also just in general. And so now, you know, we've shown results for Haiku 3.5, and I believe Haiku 3.5, the smallest new model is about as good as Opus 3, the largest old model. So basically the aim here is to shift the curve, and then at some point, there's gonna be an Opus 3.5.

29:13 Now, every new generation of models has its own thing. They use new data, their personality changes in ways that we kind of, you know, try to steer but are not fully able to steer. And so there's never quite that exact equivalence where the only thing you're changing is intelligence. We always try and improve other things, and some things change without us knowing or measuring. So it's very much an inexact science. In many ways, the manner and personality of these models is more an art than it is a science.

29:44 - So what is sort of the reason for the span of time between, say, Claude Opus 3.0 and 3.5? What takes that time? If you can speak to. - Yeah, so there's different processes. There's pre-training, which is, you know, just kind of the normal language model training, and that takes a very long time. That uses, you know, these days, you know, tens of thousands, sometimes many tens of thousands of GPUs or TPUs or Trainium, or you know, we use different platforms, but, you know, accelerator chips, often training for months.

30:26 There's then a kind of post-training phase where we do reinforcement learning from human feedback, as well as other kinds of reinforcement learning. That phase is getting larger and larger now, and, you know, often, that's less of an exact science. It often takes effort to get it right. Models are then tested with some of our early partners to see how good they are, and they're then tested both internally and externally for their safety, particularly for catastrophic and autonomy risks.

30:59 So we do internal testing according to our responsible scaling policy, which I, you know, could talk more about that in detail. And then we have an agreement with the US and the UK AI Safety Institute, as well as other third party testers in specific domains to test the models for what are called CBRN risks, chemical, biological, radiological and nuclear, which are, you know, we don't think that models pose these risks seriously yet, but every new model, we wanna evaluate to see if we're starting to get close to some of these more dangerous capabilities.

31:35 So those are the phases. And then, you know, then it just takes some time to get the model working in terms of inference and launching it in the API. So there's just a lot of steps to actually making a model work. And of course, you know, we're always trying to make the processes as streamlined as possible, right? We want our safety testing to be rigorous, but we want it to be rigorous and to be, you know, to be automatic, to happen as fast as it can without compromising on rigor.

32:04 Same with our pre-training process and our post-training process. So, you know, it's just like building anything else. It's just like building airplanes. You want to make them, you know, you want to make them safe, but you want to make the process streamlined. And I think the creative tension between those is, you know, is an important thing in making the models work. - Yeah, rumor on the street, I forget who was saying that Anthropic has really good tooling, so probably a lot of the challenge here on the software engineering side is to build the tooling to have like a efficient, low friction interaction with the infrastructure.

32:36 - You would be surprised how much of the challenges of, you know, building these models comes down to, you know, software engineering, performance engineering, you know. From the outside you might think, oh, man, we had this eureka breakthrough, right? You know, this movie with the science, we discovered it, we figured it out. But I think all things, even, you know, incredible discoveries, like, they almost always come down to the details, and often super, super boring details.

33:09 I can't speak to whether we have

Software engineering and rigorous post-training unlock massive coding leaps

33:10 better tooling than other companies. I mean, you know, haven't been at those other companies, at least not recently, but it's certainly something we give a lot of attention to. - I don't know if you can say, but from three, from Claude 3 to Claude 3.5, is there any extra pre-training going on or is it mostly focused on the post-training? There's been leaps in performance. - Yeah, I think at any given stage, we're focused on improving everything at once.

33:35 Just naturally, like there are different teams, each team makes progress in a particular area, in making a particular, you know, their particular segment of the relay race better. And it's just natural that when we make a new model, we put all of these things in at once. - So, the data you have, like the preference data you get from RLHF, is that applicable, is there a ways to apply it to newer models as it get trained up? - Yeah, preference data from old models sometimes gets used for new models, although, of course, it performs somewhat better when it's, you know, trained on, it's trained on the new models.

34:09 Note that we have this, you know, Constitutional AI method such that we don't only use preference data, we kind of, there's also a post-training process where we train the model against itself and there's, you know, new types of post-training the model against itself that are used every day. So it's not just RLHF, it's a bunch of other methods as well. Post-training, I think, you know, is becoming more and more sophisticated. - Well, what explains the big leap in performance for the new Sonnet 3.5?

Real-world coding benchmarks show massive performance leaps

34:35 I mean, at least in the programming side. And maybe this is a good place to talk about benchmarks. What does it mean to get better? Just the number went up, but, you know, I program, but I also love programming and Claude 3.5 through Cursor is what I use to assist me in programming. And there was, at least experientially, anecdotally,

5 Highlight · Massive Leaps in Real-World Software Engineering Benchmarks

34:57 it's gotten smarter at programming. So like, what does it take to get it smarter? - We observed that as well, by the way. There were a couple very strong engineers here at Anthropic who all previous code models, both produced by us and produced by all the other companies, hadn't really been useful to them. You know, they said, you know, maybe this is useful to beginner, it's not useful to me. But Sonnet 3.5, the original one for the first time they said, "Oh my God, this helped me with something that, you know, that it would've taken me hours to do.

35:28 This is the first model that's actually saved me time." So again, the waterline is rising. And then I think, you know, the new Sonnet has been even better. In terms of what it takes, I mean, I'll just say it's been across the board. It's in the pre-training, it's in the post-training, it's in various evaluations that we do. We've observed this as well. And if we go into the details of the benchmark, so Sowe bench is basically since, you know, since you're a programmer, you know, you'll be familiar with like pull requests and, you know, just pull requests are like the, you know, like a sort of atomic unit of work.

36:06 You know, you could say, you know, I'm implementing one thing. And Sowe bench actually gives you kind of a real world situation where the code base is in a current state and I'm trying to implement something that's, you know, that's described in language. We have internal benchmarks where we measure the same thing and you say, just give the model free reign to like, you know, do anything, run anything, edit anything.

36:33 How well is it able to complete these tasks? And it's that benchmark that's gone from it can do it 3% of the time to it can do it about 50% of the time. So I actually do believe that if we get, you can gain benchmarks, but I think if we get to 100% on that benchmark in a way that isn't kind of like overtrained or game for that particular benchmark, probably represents a real and serious increase in kind of programming ability.

37:00 And I would suspect that if we can get to, you know,

37:04 90, 95% that, you know, it will represent ability to autonomously do a significant fraction of software engineering tasks. - Well, ridiculous timeline question. When is Claude Opus 3.5 coming out? - Not giving you an exact date, but you know, there, you know, as far as we know, the plan is still to have a Claude 3.5 Opus. - Are we gonna get it before "GTA 6" or no?

37:30 - Like "Duke Nukem Forever." - "Duke Nukem-" - What was that game? There was some game that was delayed 15 years. - That's right. - Was that "Duke Nukem Forever?" - Yeah. And I think "GTA" is now just releasing trailers. - You know, it's only been three months since we released the first Sonnet. - Yeah, it's the incredible pace of release. - It just tells you about the pace,

AI versioning defies traditional software naming schemes

37:46 the expectations for when things are gonna come out. - So what about 4.0? So how do you think about sort of as these models get bigger and bigger, about versioning, and also just versioning in general, why Sonnet 3.5 updated with the date? Why not Sonnet 3.6, which a lot of people are calling it? - Yeah, naming is actually an interesting challenge here, right? Because I think a year ago, most of the model was pre-training, and so you could start from the beginning and just say, okay, we're gonna have models of different sizes, we're gonna train them all together and you know, we'll have a family of naming schemes and then we'll put some new magic into them and then, you know, we'll have the next generation.

38:26 The trouble starts already when some of them take a lot longer than others to train, right? That already messes up your time a little bit. But as you make big improvements in pre-training, then you suddenly notice, oh, I can make better pre-train model and that doesn't take very long to do, but you know, clearly it has the same, you know, size and shape of previous models. So I think those two together as well as the timing issues, any kind of scheme you come up with, you know, the reality tends to kind of frustrate that scheme, right?

39:01 Tend tends to kind of break out of the scheme. It's not like software where you can say, oh, this is like, you know, 3.7, this is 3.8. No, you have models with different trade-offs. You can change some things in your models, you can train, you can change other things. Some are faster and slower at inference, some have to be more expensive, some have to be less expensive. And so I think all the companies have struggled with this. I think we did very, you know, I think we were in a good position in terms of naming when we had Haiku, Sonnet and Opus.

39:31 - [Lex] It was great, great start. - We're trying to maintain it, but it's not perfect, so we'll try and get back to the simplicity, but just the nature of the field, I feel like no one's figured out naming. It's somehow a different paradigm from like normal software and so we just, none of the companies have been perfect at it. It's something we struggle with surprisingly much relative to, you know, how relative to how trivial it is to, you know, for the grand science of training the models.

40:03 - So, from the user side, the user experience of the updated Sonnet 3.5 is just different than the previous June 2024 Sonnet 3.5. It would be nice to come up with some kind of labeling that embodies that because people talk about Sonnet 3.5, but now there's a different one, and so how do you refer to the previous one and the new one when there's a distinct improvement?

40:29 It just makes conversation about it just challenging. - Yeah, yeah. I definitely think this question of there are lots of properties of the models that are not reflected in the benchmarks. I think that's definitely the case and everyone agrees. And not all of them are capabilities. Some of them are, you know, models can be polite or brusque. They can be, you know, very reactive or they can ask you questions.

41:01 They can have what feels like a warm personality or a cold personality. They can be boring or they can be very distinctive, like Golden Gate Claude was. And we have a whole, you know, we have a whole team kind of focused on, I think we call it Claude character. Amanda leads that team and we'll talk to you about that. But it's still a very inexact science, and often we find that models have properties that we're not aware of.

41:27 The fact of the matter is that you can, you know, talk to a model 10,000 times and there are some behaviors you might not see, just like with a human, right? I can know someone for a few months and, you know, not know that they have a certain skill, or not know that there's a certain side to them. And so I think we just have to get used to this idea and we're always looking for better ways of testing our models to demonstrate these capabilities, and also to decide which are the personality properties we want models to have and which we don't want to have.

41:58 That itself, the normative question is also super interesting.

User perceptions of model degradation stem from psychological shifts

42:02 - I gotta ask you a question from Reddit. - From Reddit? Oh, boy. - You know, there just this fascinating, to me at least, it's a psychological social phenomenon where people report that Claude has gotten dumber for them over time. And so the question is, does the user complaint about the dumbing down of Claude 3.5 Sonnet hold any water? So are these anecdotal reports a kind of social phenomena or did Claude, is there any cases where Claude would get dumber?

42:33 - So this actually doesn't apply, this isn't just about Claude. I believe I've seen these complaints for every foundation model produced by a major company.

6 Highlight · Model weights never change without a new release

42:45 People said this about GPT-4, they said it about GPT-4 Turbo. So, a couple things. One, the actual weights of the model, right, the actual brain of the model, that does not change unless we introduce a new model. There are just a number of reasons why it would not make sense practically to be randomly substituting in new versions of the model. It's difficult from an inference perspective and it's actually hard to control all the consequences of changing the weight of the model.

43:16 Let's say you wanted to fine tune the model to be like, I don't know, to like to say "certainly" less, which, you know, an old version of Sonnet used to do. You actually end up changing 100 things as well. So we have a whole process for it, and we have a whole process for modifying the model. We do a bunch of testing on it, we do a bunch of user testing and early customers. So we both have never changed the weights of the model without telling anyone, and it wouldn't, certainly in the current setup, it would not make sense to do that.

43:46 Now, there are a couple things that we do occasionally do.

43:50 One is sometimes we run A/B tests, but those are typically very close to when a model is being released and for a very small fraction of time. So, you know, like, the day before the new Sonnet 3.5. I agree, we should have had a better name. It's clunky to refer to it. There were some comments from people that like it's gotten a lot better, and that's because, you know, a fraction were exposed to an A/B test for those one or two days.

44:20 The other is that occasionally, the system prompt will change. The system prompt can have some effects, although it's unlikely to dumb down models. It's unlikely to make them dumber. And we've seen that while these two things, which I'm listing to be very complete, happened relatively, happened quite infrequently, the complaints about, for us and for other model companies about the model change, the model isn't good at this.

44:48 The model got more censored. The model was dumbed down. Those complaints are constant. And so I don't wanna say like people are imagining it or anything, but like the models are for the most part not changing. If I were to offer a theory, I think it actually relates to one of the things I said before, which is that models are very complex and have many aspects to them. And so often, you know, if I ask the model a question, you know, if I'm like, "Do task X" versus "Can you do task X?"

45:21 the model might respond in different ways. And so there are all kinds of subtle things that you can change about the way you interact with the model that can give you very different results. To be clear, this itself is like a failing by us and by the other model providers that the models are just often sensitive to like small changes in wording. It's yet another way in which the science of how these models work is very poorly developed.

45:49 And so, you know, if I go to sleep one night and I was like talking to the model in a certain way and I like slightly changed the phrasing of how I talk to the model, you know, I could get different results. So that's one possible way. The other thing is, man, it's just hard to quantify this stuff. It's hard to quantify this stuff. I think people are very excited by new models when they come out and then as time goes on, they become very aware of the limitations, so that may be another effect.

46:15 But that's all a very long-winded way of saying for the most part, with some fairly narrow exceptions, the models are not changing. - I think there is a psychological effect. You just start getting used to it. The baseline raises. Like when people first gotten wifi on airplanes, it's like amazing, magic. - It's like amazing, yeah. - And then- - And now I'm like, I can't get this thing to work. This is such a piece of crap. - Exactly, so then it's easy to have the conspiracy theory of they're making wifi slower and slower.

Personality tuning presents difficult multi-dimensional trade-offs

46:41 This is probably something I'll talk to Amanda much more about. But another Reddit question, "When will Claude stop trying to be my puritanical grandmother imposing its moral worldview on me as a paying customer? And also, what is the psychology behind making Claude overly apologetic?" So this kind of reports about the experience, a different angle on the frustration, it has to do with the character. - Yeah, so a couple points on this first.

47:08 One is like things that people say on Reddit and Twitter, or X or whatever it is, there's actually a huge distribution shift between like the stuff that people complain loudly about on social media and what actually kind of like, you know, statistically users care about and that drives people to use the models. Like people are frustrated with, you know, things like, you know, the model not writing out all the code or the model, you know, just not being as good at code as it could be, even though it's the best model in the world on code.

47:39 I think the majority things are about that. But certainly a kind of vocal minority are, you know, kind of raise these concerns, right? Are frustrated by the model refusing things that it shouldn't refuse, or like apologizing too much, or just having these kind of like annoying verbal ticks. The second caveat, and I just wanna say this like super clearly because I think it's like some people don't know it, others like kind of know it but forget it.

7 Highlight · Steering AI personality creates a whack-a-mole effect

48:08 Like it is very difficult to control across the board how the models behave. You cannot just reach in there and say, "Oh, I want the model to like apologize less." Like you can do that, you can include training data that says like, "Oh, the model should like apologize less," but then in some other situation they end up being like super rude or like overconfident in a way that's like misleading people. So there are all these trade-offs.

48:33 For example, another thing is there was a period during which models, ours and I think others as well were too verbose, right? They would like repeat themselves, they would say too much. You can cut down on the verbosity by penalizing the models for just talking for too long. What happens when you do that, if you do it in a crude way is when the models are coding, sometimes they'll say rest of the code goes here, right? Because they've learned that that's the way to economize and that they see it, and then so that leads the model to be so-called lazy in coding where they're just like, ah, you can finish the rest of it.

49:08 It's not because we wanna, you know, save on compute or because you know, the models are lazy, and you know, during winter break, or any of the other kind of conspiracy theories that have come up. It's actually, it's just very hard to control the behavior of the model, to steer the behavior of the model in all circumstances at once. You can kind of, there's this whack-a-mole aspect where you push on one thing and like, you know, these other things start to move as well that you may not even notice or measure.

49:40 And so one of the reasons that I care so much about, you know, kind of grand alignment of these AI systems in the future is actually these systems are actually quite unpredictable. They're actually quite hard to steer and control. And this version we're seeing today

49:57 of you make one thing better, it makes another thing worse, I think that's like a present day analog of future control problems in AI systems that we can start to study today, right? I think that that difficulty in steering the behavior and in making sure that if we push an AI system in one direction, it doesn't push it in another direction in some other ways that we didn't want.

50:27 I think that's kind of an early sign of things to come, and if we can do a good job of solving this problem, right, of like you ask the model to like, you know, to like make and distribute smallpox and it says no, but it's willing to like help you in your graduate level virology class. Like how do we get both of those things at once? It's hard. It's very easy to go to one side or the other and it's a multidimensional problem.

50:52 And so, you know, I think these questions of like shaping the model's personality, I think they're very hard. I think we haven't done perfectly on them. I think we've actually done the best of all the AI companies, but still so far from perfect. And I think if we can get this right, if we can control, you know, control the false positives and false negatives in this very kind of controlled present day environment, we'll be much better at doing it for the future when our worry is, you know, will the models be super autonomous?

51:26 Will they be able to, you know, make very dangerous things? Will they be able to autonomously, you know, build whole companies? And are those companies aligned? So, I think of this present task as both vexing, but also good practice for the future. - What's the current best way of gathering sort of user feedback? Like not anecdotal data, but just large scale data about pain points or the opposite of pain points, positive things, so on, is it internal testing?

51:54 Is it a specific group testing, A/B testing? What works? - So, typically we'll have internal model bashings where all of Anthropic, Anthropic is almost 1000 people, you know, people just try and break the model. They try and interact with it various ways. We have a suite of evals for, you know, oh, is the model refusing in ways that it couldn't? I think we even had a certainly eval because, you know, our model, again, one point, model had this problem where like it had this annoying tick where it would like respond to a wide range of questions by saying "Certainly I can help you with that.

52:29 Certainly I would be happy to do that. Certainly this is correct." And so we had a, like, certainly eval, which is like, how often does the model say certainly? But look, this is just a whack-a-mole. Like, what if it switches from certainly to definitely? Like, so, you know, every time we add a new eval, and we're always evaluating for all of the old things. So we have hundreds of these evaluations, but we find that there's no substitute for human interacting with it.

52:56 And so it's very much like the ordinary product development process. We have like hundreds of people within Anthropic bash the model, you know, then we do external A/B tests. Sometimes we'll run tests with contractors. We pay contractors to interact with the model. So you put all of these things together and it's still not perfect. You still see behaviors that you don't quite wanna see, right? You know, you still see the model like refusing things that it just doesn't make sense to refuse.

53:25 But I think trying to solve this challenge, right? Trying to stop the model from doing, you know, genuinely bad things that, you know, everyone agrees it shouldn't do, right? You know, everyone agrees that, you know, the model shouldn't talk about, you know, I don't know, child abuse material, right? Like, everyone agrees the model shouldn't do that. But at the same time that it doesn't refuse in these dumb and stupid ways. I think drawing that line as finely as possible, approaching perfectly is still a challenge and we're getting better at it every day.

53:58 But there's a lot to be solved. And again, I would point to that as an indicator of the challenge ahead in terms of steering much more powerful models. - Do you think Claude 4.0 is ever coming out? - I don't want to commit to any naming scheme, 'cause if I say here "We're gonna have Claude 4 next year," and then, you know, then we decide that like, you know, we should start over, 'cause there's a new type of model. Like I don't want to commit to it.

54:25 I would expect in a normal course of business that Claude 4 would come after Claude 3.5. But you know, you never know in this wacky field, right? - But the sort of, this idea of scaling is continuing. - Scaling is continuing. There will definitely be more powerful models coming from us than the models that exist today. That is certain. Or if there aren't, we've deeply failed as a company. - Okay, can you explain the Responsible Scaling Policy and the AI Safety Level Standards, ASL Levels?

Responsible scaling policies address misuse and autonomy risks

54:55 - As much as I am excited about the benefits of these models, and, you know, we'll talk about that if we talk about "Machines of Loving Grace," I'm worried about the risks and I continue to be worried about the risks. No one should think that, you know, "Machines of Loving Grace" was me saying, you know, I'm no longer worried about the risks of these models. I think they're two sides of the same coin. The power of the models and their ability to solve all these problems in, you know, biology, neuroscience, economic development, governance and peace, large parts of the economy, those come with risks as well, right?

55:31 With great power comes great responsibility, right? The two are paired. Things that are powerful can do good things and they can do bad things. I think of those risks as being in, you know, several different categories. Perhaps the two biggest risks that I think about, and that's not to say that there aren't risks today that are important, but when I think of the really the, you know, the things that would happen on the grandest scale, one is what I call catastrophic misuse.

55:59 These are misuse of the models in domains like cyber, bio radiological nuclear right? Things that could, you know, that could harm or even kill thousands, even millions of people if they really, really go wrong. Like these are the, you know, number one priority to prevent. And here I would just make a simple observation, which is that the models, you know, if I look today at people who have done really bad things in the world, I think actually humanity has been protected by the fact that the overlap between really smart, well-educated people and people who want to do really horrific things has generally been small.

56:43 Like, you know, let's say I'm someone who, you know, I have a PhD in this field, I have a well paying job. There's so much to lose. Why do I wanna, like, you know, even assuming I'm completely evil, which most people are not. You know, why would such a person risk their life, risk their legacy, their reputation to do something like, you know, truly, truly evil?

57:09 If we had a lot more people like that, the world would be a much more dangerous place. And so my worry is that by being a much more intelligent agent, AI could break that correlation, and so I do have serious worries about that. I believe we can prevent those worries. But, you know, I think as a counterpoint to "Machines of Loving Grace," I want to say that there's still serious risks. And the second range of risks would be the autonomy risks, which is the idea that models might on their own, particularly as we give them more agency than they've had in the past, particularly as we give them supervision over wider tasks like, you know, writing whole code bases or someday even, you know, effectively operating entire companies, they're on a long enough leash, are they doing what we really want them to do?

58:00 It's very difficult to even understand in detail what they're doing, let alone control it. And like I said, these early signs that it's hard to perfectly draw the boundary between things the model should do and things the model shouldn't do that, you know, if you go to one side, you get things that are annoying and useless, you go to the other side, you get other behaviors. If you fix one thing, it creates other problems. We're getting better and better at solving this.

58:27 I don't think this is an unsolvable problem. I think, you know, this is a science, like the safety of airplanes or the safety of cars, or the safety of drugs. You know, I don't think there's any big thing we're missing. I just think we need to get better at controlling these models. And so these are the two risks I'm worried about. And our Responsible Scaling Plan, which I'll recognize is a very long-winded answer to your question. - I love it. I love it. - Our Responsible Scaling Plan is designed to address these two types of risks.

58:56 And so every time we develop a new model, we basically test it for its ability to do both of these bad things. So if I were to back up a little bit, I think we have an interesting dilemma with AI systems where they're not yet powerful enough to present these catastrophes. I don't know that they'll ever prevent these catastrophes, it's possible they won't, but the case for worry, the case for risk is strong enough that we should act now.

59:29 And they're getting better very, very fast, right? You know, I testified in the Senate that, you know, we might have serious bio risks within two to three years. That was about a year ago. Things have proceeded at pace. So we have this thing where it's like, it's surprisingly hard to address these risks because they're not here today. They don't exist. They're like ghosts, but they're coming at us so fast because the models are improving so fast.

59:55 So how do you deal with something that's not here today, doesn't exist but is coming at us very fast? So the solution we came up with for that in collaboration with, you know, people like the organization METR and Paul Christiano is, okay, what you need for that are you need tests to tell you when the risk is getting close. You need an early warning system.

1:00:21 And so every time we have a new model, we test it for its capability to do these CBRN tasks, as well as testing it for, you know, how capable it is of doing tasks autonomously on its own. And in the latest version of our RSP, which we released in the last month or two, the way we test autonomy risks is the model, the AI model's ability to do aspects of AI research itself, which when the AI models can do AI research, they become kind of truly autonomous.

8 Highlight · ASL tiers tie safety requirements to capability triggers

1:00:56 And you know, that threshold is important for a bunch of other ways. And so what do we then do with these tasks? The RSP basically develops what we've called an if then structure, which is if the models pass a certain capability, then we impose a certain set of safety and security requirements on them. So today's models are what's called ASL two. Models that were, ASL one is for systems that manifestly don't pose any risk of autonomy or misuse.

1:01:28 So for example, a chess playing bot, Deep Blue would be ASL one. It's just manifestly the case that you can't use Deep Blue for anything other than chess. It was just designed for chess. No one's gonna use it to like, you know, to conduct a masterful cyber attack or to, you know, run wild and take over the world. ASL two is today's AI systems where we've measured them and we think these systems are simply not smart enough to, you know, autonomously self-replicate or conduct a bunch of tasks, and also not smart enough to provide meaningful information about CBRN risks and how to build CBRN weapons above and beyond what can be known from looking at Google.

1:02:15 In fact, sometimes they do provide information, but not above and beyond a search engine, but not in a way that can be stitched together, not in a way that kind of end to end is dangerous enough. So ASL three is gonna be the point at which the models are helpful enough to enhance the capabilities of non-state actors, right? State actors can already do a lot of, unfortunately, to a high level of proficiency, a lot of these very dangerous and destructive things.

1:02:45 The difference is that non-state actors are not capable of it. And so when we get to ASL three, we'll take special security precautions designed to be sufficient to prevent theft of the model by non-state actors, and misuse of the model as it's deployed.

1:03:02 We'll have to have enhanced filters targeted at these particular areas. - Cyber, bio, nuclear. - Cyber, bio, nuclear and model autonomy, which is less a misuse risk and more risk of the model doing bad things itself. ASL four, getting to the point where these models could enhance the capability of a already knowledgeable state actor and/or become, you know, the main source of such a risk.

1:03:31 Like if you wanted to engage in such a risk, the main way you would do it is through a model. And then I think ASL four on the autonomy side, it's some amount of acceleration in AI research capabilities within an AI model. And then ASL five is where we would get to the models that are, you know, that are kind of, you know, truly capable, that could exceed humanity in their ability to do any of these tasks. And so the point of if then structure commitment is basically to say, look, I don't know, I've been working with these models for many years and I've been worried about risk for many years.

1:04:08 It's actually kind of dangerous to cry wolf. It's actually kind of dangerous to say this, you know, this model is risky. And, you know, people look at it and they say, this is manifestly not dangerous. Again, it's the delicacy of the risk isn't here today but it's coming at us fast. How do you deal with that? It's really vexing to a risk planner to deal with it. And so this if then structure basically says, look, we don't wanna antagonize a bunch of people, we don't wanna harm our own, you know, our kind of own ability to have a place in the conversation by imposing these very onerous burdens on models that are not dangerous today.

1:04:51 So if then, the trigger commitment is basically a way to deal with this. Says you clamp down hard when you can show that the model is dangerous. And of course what has to come with that is, you know, enough of a buffer threshold that, you know, you're not at high risk of kind of missing the danger. It's not a perfect framework. We've had to change it every, you know, we came out with a new one just a few weeks ago, and probably going forward, we might release new ones multiple times a year because it's hard to get these policies right, like technically, organizationally, from a research perspective.

1:05:26 But that is the proposal, if then commitments and triggers in order to minimize burdens

ASL security levels establish strict deployment thresholds

1:05:32 and false alarms now, but really react appropriately when the dangers are here. - What do you think the timeline for ASL three is where several of the triggers are fired? And what do you think the timeline is for ASL four? - Yeah, so that is hotly debated within the company. We are working actively to prepare ASL three security measures as well as ASL three deployment measures. I'm not gonna go into detail, but we've made a lot of progress on both, and, you know, we're prepared to be, I think, ready quite soon.

1:06:07 I would not be surprised at all if we hit ASL three next year. There was some concern that we might even hit it this year. That's still possible, that could still happen. It's like very hard to say, but like I would be very, very surprised if it was like 2030. I think it's much sooner than that. - So there's protocols for detecting it, if then, and then there's protocols for how to respond to it. - Yes. - How difficult is the second, the latter?

1:06:34 - Yeah, I think for ASL three, it's primarily about security and about, you know, filters on the model relating to a very narrow set of areas when we deploy the model. Because at ASL three, the model isn't autonomous yet, and so you don't have to worry about, you know, kind of the model itself behaving in a bad way, even when it's deployed internally. So I think the ASL three measures are, I won't say straightforward, they're rigorous, but they're easier to reason about.

1:07:06 I think once we get to ASL four, we start to have worries about the models being smart enough that they might sandbag tests, they might not tell the truth about tests. We had some results came out about like sleeper agents and there was a more recent paper about, you know, can the models mislead attempts to, you know, sandbag their own abilities, right? Show them, you know, present themselves as being less capable than they are.

1:07:35 And so I think with ASL four, there's gonna be an important component of using other things than just interacting with the models, for example, interpretability or hidden chains of thought where you have to look inside the model and verify via some other mechanism that is not, you know, is not as easily corrupted as what the model says, you know, that the model indeed has some property. So we're still working on ASL four.

1:08:02 One of the properties of the RSP is that we don't specify ASL four until we've hit ASL three. And I think that's proven to be a wise decision because even with ASL three, again, it's hard to know this stuff in detail, and we wanna take as much time as we can possibly take to get these things right. - So for ASL three, the bad actor will be the humans. - [Dario] Humans, yes.

1:08:27 - And so there, it's a little bit more- - For ASL four, it's both, I think, both. - It's both, and so deception, and that's where mechanistic interpretability comes into play and hopefully the techniques used for that are not made accessible to the model. - Yeah, I mean, of course you can hook up the mechanistic interpretability to the model itself, but then you've kind of lost it as a reliable indicator of the model state.

1:08:54 There are a bunch of exotic ways you can think of that it might also not be reliable. Like if the, you know, model gets smart enough that it can like, you know, jump computers and like read the code where you're like looking at its internal state. We've thought about some of those. I think there're exotic enough, there are ways to render them unlikely. But yeah, generally you wanna preserve mechanistic interpretability as a kind of verification set or test set that's separate from the training process of the model. - See, I think as these models become better and better conversation and become smarter, social engineering becomes a threat too 'cause they- - [Dario] Oh, yeah.

1:09:28 - That could start being very convincing

Computer use transforms models into interactive desktop agents

1:09:28 to the engineers inside companies. - Oh yeah, yeah. It's actually like, you know, we've seen lots of examples of demagoguery in our life from humans and, you know, there's a concern that models that could do that as well. - One of the ways that Claude has been getting more and more powerful is it's now able to do some agentic stuff, computer use.

9 Highlight · Computer use relies on screenshots and coordinate clicks

1:09:48 There's also an analysis within the sandbox of claude.ai itself. But let's talk about computer use. That seems to me super exciting that you can just give Claude a task and it takes a bunch of actions, figures it out, and has access to your computer through screenshots. So can you explain how that works? And where that's headed? - Yeah, it's actually relatively simple. So Claude has had for a long time, since Claude 3.0 back in March, the ability to analyze images and respond to them with text.

1:10:21 The only new thing we added is those images can be screenshots of a computer. And in response, we trained the model to give a location on the screen where you can click and/or buttons on the keyboard you can press in order to take action. And it turns out that with actually not all that much additional training, the models can get quite good at that task. It's a good example of generalization. You know, people sometimes say, if you get to lower earth orbit, you're like halfway to anywhere, right?

1:10:49 Because of how much it takes to escape the gravity. Well, if you have a strong pre-trained model, I feel like you're halfway to anywhere in terms of the intelligence space. And so actually, it didn't take all that much to get Claude to do this. And you can just set that in a loop, give the model a screenshot, tell it what to click on, give it the next screenshot, tell it what to click on, and that turns into a full kind of almost 3D video interaction of the model.

1:11:17 And it's able to do all of these tasks, right? You know, we showed these demos where it's able to like fill out spreadsheets. It's able to kind of like interact with a website. It's able to, you know, it's able to open all kinds of, you know, programs,

1:11:32 different operating systems, Windows, Linux, Mac. So, you know, I think all of that is very exciting. I will say, while in theory, there's nothing you could do there that you couldn't have done through just giving the model the API to drive the computer screen. This really lowers the barrier. And you know, there's a lot of folks who either, you know, aren't in a position to interact with those APIs or it takes them a long time to do.

1:12:00 It's just the screen is just a universal interface that's a lot easier to interact with. And so I expect over time, this is gonna lower a bunch of barriers. Now, honestly, the current model has, it leaves a lot still to be desired, and we were honest about that in the blog, right? It makes mistakes, it misclicks. And, you know, we were careful to warn people, hey, this thing isn't, you can't just leave this thing to, you know, run on your computer for minutes and minutes.

1:12:27 You gotta give this thing boundaries and guardrails. And I think that's one of the reasons we released it first in an API form rather than kind of, you know, this kind of just hand it to the consumer and give it control of their computer. But you know, I definitely feel that it's important to get these capabilities out there. As models get more powerful, we're gonna have to grapple with, you know, how do we use these capabilities safely? How do we prevent them from being abused?

1:12:54 And you know, I think releasing the model while the capabilities are, you know, are still limited is very helpful in terms of doing that. You know, I think since it's been released, a number of customers, I think Replit was maybe one of the most quickest to deploy things. You know, have made use of it in various ways.

1:13:20 People have hooked up demos for, you know, Windows desktops, Macs, you know, Linux machines. So yeah, it's been very exciting. I think as with anything else, you know, it comes with new exciting abilities and then, you know, with those new exciting abilities, we have to think about how to, you know, make the model, you know, safe, reliable, do what humans want them to do.

1:13:46 I mean, it's the same story for everything, right? Same thing. It's that same tension. - But the possibility of use cases here, just the range is incredible. So how much to make it work really well in the future? How much do you have to specially kind of go beyond what's the pre-trained model's doing, do more post-training, RLHF or supervised fine tuning, or synthetic data just for the agentic stuff? - Yeah, I think speaking at a high level, it's our intention to keep investing a lot in, you know, making the model better.

1:14:17 Like I think, you know, we look at some of the, you know, some of the benchmarks where previous models were like, oh, could do it 6% of the time, and now our model would do it 14 or 22% of the time. And yeah, we wanna get up to, you know, the human level reliability of 80, 90% just like anywhere else, right? We're on the same curve that we were on with SWE-bench, where I think I would guess a year from now, the models can do this very, very reliably, but you gotta start somewhere. - So you think it's possible to get to the human level, 90%, basically doing the same thing you're doing now?

1:14:47 Or it has to be special for computer use? - I mean, it depends what you mean by, you know, special and special in general. But, you know, I generally think, you know, the same kinds of techniques that we've been using to train the current model, I expect that doubling down on those techniques in the same way that we have for code for models in general, you know, for image input, you know, for voice.

Computer use expands existing risk profiles

1:15:14 I expect those same techniques will scale here, as they have everywhere else. - But this is giving sort of the power of action to Claude, and so you could do a lot of really powerful things, but you could do a lot of damage also. - Yeah, yeah, no, and we've been very aware of that. Look, my view actually is computer use isn't a fundamentally new capability, like the CBRN or autonomy capabilities are. It's more like, it kind of opens the aperture for the model to use and apply its existing abilities.

1:15:45 And so the way we think about it, going back to our RSP, is nothing that this model is doing inherently increases, you know, the risk from an RSP perspective. But as the models get more powerful, having this capability may make it scarier once it, you know, once it has the cognitive capability to, you know, to do something at the ASL three and ASL four level, you know, this may be the thing that kind of unbounds it from doing so.

1:16:17 So, going forward, certainly this modality of interaction is something we have tested for, and that we will continue to test for in RSP going forward. I think it's probably better to have, to learn and explore this capability before the model is super, you know, super capable. - Yeah, and there's a lot of interesting attacks, like prompt injection, because now you've widened the aperture, so you can prompt inject through stuff on screen. So if this becomes more and more useful, then there's more and more benefit to inject stuff into the model.

1:16:47 If it goes to a certain webpage, it could be harmless stuff like advertisements or it could be like harmful stuff, right? - Yeah, I mean, we've thought a lot about things like spam, CAPTCHA, you know, mass camp. There's all, you know, every, like one secret I'll tell you, if you've invented a new technology, not necessarily the biggest misuse, but the first misuse you'll see, scams, just petty scams. Like you'll just, it's like a thing as old, people scamming each other, it's this thing as old as time, and it's just every time you gotta deal with it.

1:17:21 - It's almost like silly to say but it's true, sort of bots and spam in general is a thing as it gets more and more intelligent. It's harder and harder to fight. - There are a lot of like I said, like there are a lot of petty criminals in the world. And you know, it's like every new technology is like a new way for petty criminals to do something, you know, something stupid and malicious. - Is there any ideas about sandboxing it?

1:17:47 Like how difficult is the sandboxing task? - Yeah, we sandbox during training. So for example, during training, we didn't expose the model to the internet. I think that's probably a bad idea during training because, you know, the model can be changing its policy, it can be changing what it's doing, and it's having an effect in the real world. You know, in terms of actually deploying the model, right, it kind of depends on the application. Like, you know, sometimes you want the model to do something in the real world, but of course, you can always put guardrails on the outside, right?

1:18:17 You can say, okay, well, you know, this model's not gonna move data from my, you know, model's not gonna move any files from my computer or my web server to anywhere else. Now, when you talk about sandboxing, again, when we get to ASL four, none of these precautions are going to make sense there, right, where when you talk about ASL four, you're then, the model is being kind of, you know, there's a theoretical worry the model could be smart enough to kind of break out of any box.

1:18:46 And so there we need to think about mechanistic interpretability about, you know, if we're gonna have a sandbox, it would need to be a mathematically provable sand. You know, that's a whole different world than what we're dealing with with the models today. - Yeah, the science of building a box from which ASL four AI system cannot escape. - I think it's probably not the right approach. I think the right approach instead of having something, you know, unaligned that like you're trying to prevent it from escaping.

1:19:17 I think it's better to just design the model the right way or have a loop where, you know, you look inside, you look inside the model and you're able to verify properties, and that gives you an opportunity to like iterate and actually get it right. I think containing bad models is much worse solution than having good models.

Surgical regulation prevents a race to the bottom

1:19:35 - Let me ask about regulation. What's the role of regulation in keeping AI safe? So for example, can you describe California AI regulation Bill SB 1047 that was ultimately vetoed by the governor? What are the pros and cons of this bill in general? - Yes, we ended up making some suggestions to the bill, and then some of those were adopted and, you know, we felt, I think quite positively, quite positively about the bill by the end of that.

1:20:03 It did still have some downsides, and, you know, of course it got vetoed. I think at a high level, I think some of the key ideas behind the bill are, you know, I would say similar to ideas behind our RSPs. And I think it's very important that some jurisdiction, whether it's California or the federal government and/or other countries and other states passes some regulation like this.

1:20:29 And I can talk through why I think that's so important. So I feel good about our RSP. It's not perfect, it needs to be iterated on a lot, but it's been a good forcing function for getting the company to take these risks seriously, to put them into product planning, to really make them a central part of work at Anthropic and to make sure that all of 1000 people, and it's almost 1000 people now at Anthropic, understand that this is one of the highest priorities of the company, if not the highest priority.

1:20:58 But one, there are still some companies that don't have RSP like mechanisms, like OpenAI, Google did adopt these mechanisms a couple months after Anthropic did, but there are other companies out there that don't have these mechanisms at all. And so if some companies adopt these mechanisms and others don't, it's really gonna create a situation where, you know, some of these dangers have the property that it doesn't matter if three out of five of the companies are being safe, if the other two are being unsafe, it creates this negative externality.

1:21:36 And I think the lack of uniformity is not fair to those of us who have put a lot of effort into being very thoughtful about these procedures. The second thing is, I don't think you can trust these companies to adhere to these voluntary plans in their own, right? I like to think that Anthropic will. We do everything we can that we will. Our RSP is checked by our long-term benefit trust. So, you know, we do everything we can to adhere to our own RSP.

1:22:07 But you know, you hear lots of things about various companies saying, oh, they said they would give this much compute and they didn't. They said they would do this thing and they didn't. You know, I don't think it makes sense to, you know, to litigate particular things that companies have done. But I think this broad principle that like if there's nothing watching over them, there's nothing watching over us as an industry, there's no guarantee that we'll do the right thing, and the stakes are very high.

1:22:35 And so I think it's important to have a uniform standard that everyone follows, and to make sure that simply that the industry does what a majority of the industry has already said is important and has already said that they definitely will do. Right, some people, you know, I think there's a class of people who are against regulation on principle. I understand where that comes from.

1:23:00 If you go to Europe and, you know, you see something like GDPR, you see some of the other stuff that they've done. You know, some of it's good, but some of it is really unnecessarily burdensome, and I think it's fair to say really has slowed innovation. And so I understand where people are coming from on priors. I understand why people come from, start from that position. But again, I think AI is different.

1:23:25 If we go to the very serious risks of autonomy and misuse that I talked about, you know, just a few minutes ago, I think that those are unusual and they warrant an unusually strong response. And so I think it's very important. Again, we need something that everyone can get behind. You know, I think one of the issues with SB 1047, especially the original version of it, was it had a bunch of the structure of RSPs, but it also had a bunch of stuff that was either clunky or that just would've created a bunch of burdens, a bunch of hassle, and might even have missed the target in terms of addressing the risks.

1:24:14 You don't really hear about it on Twitter. You just hear about kind of, you know, people are cheering for any regulation, and then the folks who are against make up these often quite intellectually dishonest arguments about how, you know, it'll make us move away from California. Bill doesn't apply if you're headquartered in California, bill only applies if you do business in California. Or that it would damage the open source ecosystem, or that it would, you know, it would cause all of these things.

1:24:44 I think those were mostly nonsense, but there are better arguments against regulation. There's one guy, Dean Ball, who's really, you know, I think a very scholarly, scholarly analyst who looks at what happens when a regulation is put in place and ways that they can kind of get a life of their own, or how they can be poorly designed. And so our interest has always been, we do think there should be regulation in this space, but we wanna be an actor who makes sure that that regulation is something that's surgical, that's targeted at the serious risks and is something people can actually comply with.

1:25:21 Because something I think the advocates of regulation don't understand as well as they could is if we get something in place that's poorly targeted, that wastes a bunch of people's time, what's gonna happen is people are gonna say, see, these safety risks, you know, this is nonsense. You know, I just had to hire 10 lawyers, you know, to fill out all these forms.

1:25:47 I had to run all these tests for something that was clearly not dangerous. And after six months of that, there will be a groundswell and we'll end up with a durable consensus against regulation. And so I think the worst enemy of those who want real accountability is badly designed regulation. We need to actually get it right. And this is, if there's one thing I could say to the advocates, it would be that I want them to understand this dynamic better, and we need to be really careful and we need to talk to people who actually have experience seeing how regulations play out in practice.

1:26:24 And the people who have seen that understand to be very careful. If this was some lesser issue, I might be against regulation at all. But what I want the opponents to understand is that the underlying issues are actually serious. They're not something that I or the other companies are just making up because of regulatory capture. They're not sci-fi fantasies.

1:26:49 They're not any of these things. You know, every time we have a new model, every few months, we measure the behavior of these models and they're getting better and better at these concerning tasks, just as they are getting better and better at, you know, good, valuable, economically useful tasks. And so I would just love it if some of the former, you know, I think SB 1047 was very polarizing.

1:27:16 I would love it if some of the most reasonable opponents and some of the most reasonable proponents would sit down together. And, you know I think that, you know, the different AI companies, you know, Anthropic was the only AI company that, you know, felt positively in a very detailed way. I think Elon tweeted briefly something positive.

1:27:41 But, you know, some of the big ones, like Google, OpenAI, Meta, Microsoft were pretty staunchly against. So I would really like is if, you know, some of the key stakeholders, some of the, you know, most thoughtful proponents and some of the most thoughtful opponents would sit down and say, how do we solve this problem in a way that the proponents feel brings a real reduction in risk, and that the opponents feel that it is not hampering the industry or hampering innovation any more necessary than it needs to.

1:28:18 And I think for whatever reason that things got too polarized and those two groups didn't get to sit down in the way that they should. And I feel urgency. I really think we need to do something in 2025. You know, if we get to the end of 2025 and we've still done nothing about this, then I'm gonna be worried. I'm not worried yet, because again, the risks aren't here yet, but I think time is running short.

1:28:44 - Yeah, and come up with something surgical, like you said. - Yeah, yeah, yeah, exactly. And we need to get away from this intense pro-safety versus intense anti-regulatory rhetoric, right?

Models inherently want to learn and solve problems

1:28:58 It's turned into these flame wars on Twitter and nothing good's gonna come of that. - So there's a lot of curiosity about the different players in the game. One of the OGs is OpenAI. You've had several years of experience at OpenAI. What's your story and history there? - Yeah, so I was at OpenAI for roughly five years. For the last, I think it was couple years,

10 Highlight · The scaling hypothesis: Models just want to learn

1:29:20 you know, I was vice president of research there. Probably myself and Ilya Sutskever were the ones who, you know, really kind of set the research direction. Around 2016 or 2017, I first started to really believe in or at least confirm my belief in the Scaling Hypothesis when Ilya famously said to me, "The thing you need to understand about these models is they just wanna learn. The models just wanna learn."

1:29:45 And again, sometimes there are these one sentences, these zen koans that you hear them and you're like, ah, that explains everything. That explains like 1000 things that I've seen. And then I, you know, ever after I had this visualization in my head of like, you optimize the models in the right way, you point the models in the right way. They just wanna learn. They just wanna solve the problem, regardless of what the problem is. - So get out of their way, basically. - Get out of their way, yeah.

1:30:11 Don't impose your own ideas about how they should learn. And you know, this was the same thing as Rich Sutton put out in the Bitter Lesson or Gwern put out in The Scaling Hypothesis. You know, I think generally the dynamic was, you know,

1:30:24 I got this kind of inspiration from Ilya and from others, folks like Alec Radford who did the original GPT-1, and then ran really hard with it. Me and my collaborators on GPT-2, GPT-3. RL from Human Feedback, which was an attempt to kind of deal with the early safety and durability. Things like debate and amplification. Heavy on interpretability.

1:30:49 So again, the combination of safety plus scaling. Probably 2018, 2019, 2020, those were kind of the years when myself and my collaborators probably, you know, many of whom became co-founders of Anthropic, kind of really had a vision and like drove the direction. - Why'd you leave? Why'd you decide to leave? - Yeah, so look, I'm gonna put things this way and, you know, I think it ties to the race to the top, right?

1:31:20 Which is, you know, in my time at OpenAI, what I'd come to see is I'd come to appreciate the Scaling Hypothesis, and as I'd come to appreciate kind of the importance of safety along with the Scaling Hypothesis. The first one I think, you know, OpenAI was getting on board with. The second one in a way had always been part of OpenAI's messaging, but, you know, over many years of the time that I spent there, I think I had a particular vision of how we should handle these things, how we should be brought out in the world, the kind of principles that the organization should have.

1:31:57 And look, I mean, there were like many, many discussions about like, you know, should the org do, should the company do this? Should the company do that? Like, there's a bunch of misinformation out there. People say like, we left because we didn't like the deal with Microsoft. False, although, you know, it was like a lot of discussion, a lot of questions about exactly how we do the deal with Microsoft. We left because we didn't like commercialization. That's not true, we built GPT-3, which was the model that was commercialized.

1:32:23 I was involved in commercialization. It's more again about, how do you do it? Like civilization is going down this path to very powerful AI. What's the way to do it that is cautious, straightforward honest that builds trust in the organization and individuals? How do we get from here to there? And how do we have a real vision for how to get it right?

1:32:49 How can safety not just be something we say because it helps with recruiting? And, you know, I think at the end of the day, if you have a vision for that, forget about anyone else's vision. I don't wanna talk about anyone else's vision. If you have a vision for how to do it, you should go off and you should do that vision. It is incredibly unproductive to try and argue with someone else's vision. You might think they're not doing it the right way. You might think they're dishonest.

1:33:16 Who knows, maybe you're right, maybe you're not. But what you should do is you should take some people you trust and you should go off together and you should make your vision happen. And if your vision is compelling, if you can make it appeal to people, you know, some combination of ethically, you know, in the market, you know, if you can make a company that's a place people wanna join, that, you know, engages in practices that people think are reasonable, while managing to maintain its position in the ecosystem at the same time, if you do that, people will copy it.

1:33:52 And the fact that you are doing it, especially the fact that you're doing it better than they are causes them to change their behavior in a much more compelling way than if they're your boss and you're arguing with them. I just, I don't know how to be any more specific about it than that, but I think it's generally very unproductive to try and get someone else's vision to look like your vision. It's much more productive to go off and do a clean experiment and say, this is our vision, this is how we're gonna do things.

1:34:21 Your choice is you can ignore us, you can reject what we're doing, or you can start to become more like us, and imitation is the sincerest form of flattery. And you know, that plays out in the behavior of customers, that plays out in the behavior of the public, that plays out in the behavior of where people choose to work. And again, at the end, it's not about one company winning or another company winning.

1:34:49 If we or another company are engaging in some practice that, you know, people find genuinely appealing, and I want it to be in substance, not just in appearance. And, you know, I think researchers are sophisticated and they look at substance. And then other companies start copying that practice and they win because they copied that practice, that's great, that's success. That's like the race to the top.

1:35:15 It doesn't matter who wins in the end, as long as everyone is copying everyone else's good practices, right? One way I think of it is like, the thing we're all afraid of is the race to the bottom, right? And the race to the bottom, doesn't matter who wins because we all lose, right? Like, you know, in the most extreme world, we make this autonomous AI that, you know, the robots enslave us or whatever, right? I mean, that's half joking, but you know, that is the most extreme thing that could happen.

1:35:40 Then it doesn't matter which company was ahead. If instead you create a race to the top where people are competing to engage in good practices, then, you know, at the end of the day, you know, it doesn't matter who ends up winning, doesn't even matter who started the race at the top. The point isn't to be virtuous. The point is to get the system into a better equilibrium than it was before, and individual companies can play some role in doing this.

1:36:06 Individual companies can, you know, can help to start it, can help to accelerate it. And frankly, I think individuals at other companies have done this as well, right? The individuals that when we put out an RSP react by pushing harder to get something similar done, get something similar done at other companies. Sometimes other companies do something that's like, we're like, oh, it's a good practice. We think that's good. We should adopt it too.

1:36:31 The only difference is, you know, I think we are, we try to be more forward-leaning. We try and adopt more of these practices first and adopt them more quickly when others invent them. But I think this dynamic is what we should be pointing at. And I think it abstracts away the question of, you know, which company's winning, who trusts who. I think all these questions of drama are profoundly uninteresting, and the thing that matters is the ecosystem that we all operate in and how to make that ecosystem better because that constrains all the players.

1:37:06 - And so Anthropic is this kind of clean experiment built on a foundation of like what concretely AI safety should look like. - Look, I'm sure we've made plenty of mistakes along the way. The perfect organization doesn't exist. It has to deal with the imperfection of 1000 employees. It has to deal with the imperfection of our leaders, including me. It has to deal with the imperfection of the people we've put to, you know, to oversee the imperfection of the leaders, like the board and the long-term benefit trust.

1:37:35 It's all a set of imperfect people trying to aim imperfectly at some ideal that will never perfectly be achieved. That's what you sign up for, that's what it will always be. But imperfect doesn't mean you just give up. There's better and there's worse. And hopefully we can begin to build, we can do well enough that we can begin to build some practices that the whole industry engages in. And then, you know, my guess is that multiple of these companies will be successful.

1:38:04 Anthropic will be successful. These other companies, like ones I've been at the past will also be successful, and some will be more successful than others.

Talent density outperforms talent mass

1:38:12 That's less important than, again, that we align the incentives of the industry. And that happens partly through the race to the top, partly through things like RSP, partly through again selected surgical regulation. - You said talent density beats talent mass. So can you explain that? Can you expand on that? Can you just talk about what it takes

11 Highlight · Why talent density beats talent mass

1:38:33 to build a great team of AI researchers and engineers? - This is one of these statements that's like more true every month. Every month I see this statement as more true than I did the month before. So if I were to do a thought experiment, let's say you have a team of 100 people that are super smart, motivated, and aligned with the mission, and that's your company. Or you can have a team of 1000 people where 200 people are super smart, super aligned with the mission, and then like 800 people are, let's just say you pick 800 like random big tech employees, which would you rather have, right?

1:39:11 The talent mass is greater in the group of 1000 people, right? You have even a larger number of incredibly talented, incredibly aligned, incredibly smart people. But the issue is just that if every time someone super talented looks around, they see someone else super talented and super dedicated, that sets the tone for everything, right?

1:39:36 That sets the tone for everyone is super inspired to work at the same place. Everyone trusts everyone else. If you have 1000 or 10,000 people and things have really regressed, right? You are not able to do selection and you're choosing random people, what happens is then you need to put a lot of processes and a lot of guardrails in place just because people don't fully trust each other, or you have to adjudicate political battles.

1:40:01 Like there are so many things that slow down the org's ability to operate. And so we're nearly 1000 people and you know, we've tried to make it so that as large a fraction of those 1000 people as possible are like super talented, super skilled. It's one of the reasons we've slowed down hiring a lot in the last few months.

1:40:21 We grew from 300 to 800, I believe, I think in the first seven, eight months of the year. And now we've slowed down. We're at like, you know, the last three months, we went from 800 to 900, 950, something like that. Don't quote me on the exact numbers, but I think there's an inflection point around 1000, and we want to be much more careful how we grow. Early on, and now as well, you know, we've hired a lot of physicists.

1:40:47 You know, theoretical physicists can learn things really fast. Even more recently as we've continued to hire that, you know, we've really had a high bar for, on both the research side and the software engineering side have hired a lot of senior people, including folks who used to be at other companies in this space. And we've just continued to be very selective. It's very easy to go from 100 to 1000 and 1000 to 10,000 without paying attention to making sure everyone has a unified purpose.

1:41:18 It's so powerful. If your company consists of a lot of different fiefdoms that all wanna do their own thing, they're all optimizing for their own thing, it's very hard to get anything done. But if everyone sees the broader purpose of the company, if there's trust and there's dedication to doing the right thing, that is a superpower. That in itself, I think, can overcome almost every other disadvantage. - And you know, as to Steve Jobs, A players.

1:41:44 A players wanna look around and see other A players is another way of saying that. I don't know what that is about human nature, but it is demotivating to see people who are not obsessively driving towards a singular mission. And it is, on the flip side of that, super motivating to see that. It's interesting. What's it take to be a great AI researcher or engineer from everything you've seen,

Open-mindedness defines great AI researchers

12 Highlight · Open-mindedness matters more than experience

1:42:07 from working with so many amazing people? - Yeah, I think the number one quality, especially on the research side, but really both is open-mindedness. Sounds easy to be open-minded, right? You're just like, oh, I'm open to anything. But, you know, if I think about my own early history in the Scaling Hypothesis, I was seeing the same data others were seeing. I don't think I was like a better programmer or better at coming up with research ideas than any of the hundreds of people that I worked with.

1:42:38 In some ways, I was worse. You know, like I've never like, you know, precise programming of like, you know, finding the bug, writing the GPU kernels. Like, I could point you to 100 people here who are better at that than I am. But the thing that I think I did have that was different was that I was just willing to look at something with new eyes, right? People said, oh, you know, "We don't have the right algorithms yet.

1:43:07 We haven't come up with the right way to do things." And I was just like, oh, I don't know, like, you know, this neural net has like 30 billion, 30 million parameters. Like, what if we gave it 50 million instead? Like, let's plot some graphs. Like that basic scientific mindset of like, oh, man, like I just, like, you know, I see some variable that I could change. Like, what happens when it changes? Like, let's try these different things and like create a graph.

1:43:35 For even, this was like the simplest thing in the world, right? Change the number of, you know, this wasn't like PhD level experimental design. This was like simple and stupid. Like, anyone could have done this if you just told them that it was important. It's also not hard to understand. You didn't need to be brilliant to come up with this.

1:43:54 But you put the two things together and, you know, some tiny number of people, some single digit number of people have driven forward the whole field by realizing this. And you know, it's often like that. If you look back at the discovery, you know, the discoveries in history, they're often like that. And so this open-mindedness and this willingness to see with new eyes

13 Highlight · Experience can be a disadvantage in AI research

1:44:15 that often comes from being newer to the field. Often experience is a disadvantage for this. That is the most important thing. It's very hard to look for and test for. But I think it's the most important thing because when you find something, some really new way of thinking about things, when you have the initiative to do that, it's absolutely transformative. - And also be able to do kind of rapid experimentation, and in the face of that, be open-minded and curious and looking at the data, just these fresh eyes and seeing what is that it's actually saying.

1:44:44 That applies in mechanism interpretability. - It's another example of this. Like some of the early work in mechanistic interpretability, so simple, it's just no one thought to care about this question before. - You said what it takes to be a great AI researcher.

Explore fertile under-researched areas in AI

1:44:58 Can we rewind the clock back? What advice would you give to people interested in AI? They're young, looking forward to, how can I make an impact on the world? - I think my number one piece of advice is to just start playing with the models. This was actually, I worry a little, this seems like obvious advice now. I think three years ago, it wasn't obvious and people started by, oh, let me read the latest Reinforcement Learning paper. Let me, you know, let me kind of, I mean, that was really, and I mean, you should do that as well.

1:45:27 But now, you know, with wider availability of models and APIs, people are doing this more. But I think just experiential knowledge. These models are new artifacts that no one really understands, and so getting experience playing with them. I would also say, again, in line with the like, do something new, think in some new direction. Like there are all these things that haven't been explored.

1:45:53 Like for example, mechanistic interpretability is still very new. It's probably better to work on that than it is to work on new model architectures because, you know, it's more popular than it was before. There are probably like 100 people working on it, but there aren't like 10,000 people working on it. And it's just this fertile area for study. Like, you know, there's so much like low hanging fruit. You can just walk by and, you know, you can just walk by and you can pick things.

1:46:22 And the only reason, for whatever reason, people aren't interested in it enough. I think there are some things around long horizon learning and long horizon tasks where there's a lot to be done. I think evaluations are still, we're still very early in our ability to study evaluations, particularly for dynamic systems acting in the world. I think there's some stuff around multi-agent. Skate where the puck is going is my advice.

1:46:49 And you don't have to be brilliant to think of it. Like all the things that are gonna be exciting in five years, like people even mention them as like, you know, conventional wisdom, but like, it's just somehow there's this barrier that

RLHF bridges human intent and model capability

1:47:03 people don't double down as much as they could, or they're afraid to do something that's not the popular thing. I don't know why it happens, but like, getting over that barrier, that's the my number one piece of advice. - Let's talk if we could a bit about post-training. So it seems that the modern post-training recipe has a little bit of everything. So supervised fine tuning, RLHF, the Constitutional AI with RLAIF.

1:47:32 - Best acronym. - It's, again, that name thing. - [Dario] RLAIF. (both laughing) - And then synthetic data, seems like a lot of synthetic data, or at least trying to figure out ways to have high quality synthetic data. So what's the, if this is a secret sauce that makes Anthropic Claude so incredible, how much of the magic is in the pre-training? How much of is in the post-training? - Yeah, I mean, so first of all, we're not perfectly able to measure that ourselves.

1:47:58 - [Lex] True. - You know, when you see some great character ability, sometimes it's hard to tell whether it came from pre-training or post-training. We've developed ways to try and distinguish between those two but they're not perfect. You know, the second thing I would say is, you know, when there is an advantage, and I think we've been pretty good in general at RL, perhaps the best. Although I don't know 'cause I don't see what goes on inside other companies.

1:48:23 Usually it isn't, oh my God, we have this secret magic method that others don't have, right? Usually it's like, well, you know, we got better at the infrastructure, so we could run it for longer. Or, you know, we were able to get higher quality data, or we were able to filter our data better, or we were able to, you know, combine these methods in practice. It's usually some boring matter of kind of practiced and trade craft. So, you know, when I think about how to do something special in terms of how we train these models, both pre-training, but even more so post-training, you know, I really think of it a little more, again, as like designing airplanes or cars.

1:49:01 Like, you know, it's not just like, oh, man, I have the blueprint. Like maybe that makes you make the next airplane. But like, there's some cultural trade craft

RLHF unhobbles models rather than increasing raw intelligence

1:49:10 of how we think about the design process that I think is more important than, you know, than any particular gizmo we're able to invent. - Okay, well, let me ask you about specific techniques. So first on RLHF, why do you think, just zooming out, intuition almost philosophy, why do you think RLHF works so well? - If I go back to like the Scaling Hypothesis, one of the ways to skate the Scaling Hypothesis is if you train for X and you throw enough compute at it, then you get X.

1:49:38 And so RLHF is good at doing what humans want the model to do, or at least to state it more precisely, doing what humans who look at the model for a brief period of time and consider different possible responses, what they prefer as the response, which is not perfect from both the safety and capabilities perspective, in that humans are often not able to perfectly identify what the model wants, and what humans want in the moment may not be what they want in the long term.

1:50:05 So there's a lot of subtlety there, but the models are good at, you know, producing what the humans in some shallow sense want.

14 Highlight · RLHF unhobbles models rather than increasing raw intelligence

1:50:15 And it actually turns out that you don't even have to throw that much compute at it because of another thing, which is this thing about a strong pre-trained model being halfway to anywhere. So once you have the pre-trained model, you have all the representations you need to get the model where you want it to go. - So do you think RLHF makes the model smarter or just appears smarter to the humans?

1:50:41 - I don't think it makes the model smarter. I don't think it just makes the model appear smarter. It's like RLHF like bridges the gap between the human and the model, right? I could have something really smart that like can't communicate at all, right? We all know people like this, people who are really smart, but that, you know, you can't understand what they're saying. So I think RLHF just bridges that gap. I think it's not the only kind of RL we do, it's not the only kind of RL that will happen in the future.

1:51:11 I think RL has the potential to make models smarter, to make them reason better, to make them operate better, to make them develop new skills even. And perhaps that could be done, you know, even in some cases with human feedback. But the kind of RLHF we do today mostly doesn't do that yet, although we're very quickly starting to be able to. - But it appears to sort of increase, if you look at the metric of helpfulness, it increases that.

1:51:36 - It also increases, what was this word in Leopold's essay, unhobbling, where basically the models are hobbled and then you do various trainings to them to unhobble them. So, you know, I like that word 'cause it's like a rare word. But so I think RLHF unhobbles the models in some ways. And then there are other ways where that model hasn't yet been unhobbled and, you know, needs to unhobble. - If you can say in terms of cost, is pre-training the most expensive thing?

1:52:02 Or is post-training creep up to that? - At the present moment, it is still the case that pre-training is the majority of the cost. I don't know what to expect in the future, but I could certainly anticipate a future where post-training is the majority of the cost. - In that future you anticipate, would it be the humans or the AI that's the cost of thing for the post-training? - I don't think you can scale up humans enough

Constitutional AI uses self-play against human principles

1:52:26 to get high quality. Any kind of method that relies on humans and uses a large amount of compute, it's gonna have to rely on some scaled superposition method, like you know, debate or iterated amplification or something like that. - So on that super interesting set of ideas around Constitutional AI, can you describe what it is, as first detailed in December 2022 paper and beyond that, what is it?

1:52:53 - Yes, so this was from two years ago. The basic idea is, so we describe what RLHF is. You have a model and it, you know, spits out two, you know, like you just sample from it twice, it spits out two possible responses, and you're like, "Human, which response do you like better?" Or another variant of it is, "Rate this response on a scale of one to seven." So that's hard because you need to scale up human interaction and it's very implicit, right?

1:53:19 I don't have a sense of what I want the model to do. I just have a sense of like what this average of a 1000 humans wants the model to do. So two ideas.

15 Highlight · Constitutional AI functions as a form of self-play

1:53:28 One is, could the AI system itself decide which response is better, right? Could you show the AI system these two responses and ask which response is better? And then second, well, what criterion should the AI use? And so then there's this idea, 'cause you have a single document, a constitution, if you will, that says, these are the principles the model should be using to respond. And the AI system reads those, it reads those principles, as well as reading the environment and the response.

1:54:01 And it says, well, how good did the AI model do? It's basically a form of self play. You're kind of training the model against itself. And so the AI gives the response and then you feed that back into what's called the preference model, which in turn feeds the model to make it better. So you have this triangle of like the AI, the preference model, and the improvement of the AI itself. - And we should say that in the constitution, the set of principles are like human interpretable.

1:54:27 They're like- - Yeah, yeah. It's something both the human and the AI system can read. So it has this nice kind of translatability or symmetry. You know, in practice we both use a model constitution and we use RLHF and we use some of these other methods. So it's turned into one tool in a toolkit that both reduces the need for RLHF and increases the value we get from using each data point of RLHF.

1:54:55 It also interacts in interesting ways with kind of future reasoning type RL methods. So it's one tool in the toolkit, but I think it is a very important tool. - Well, it's a compelling one to us humans. You know, thinking about the founding fathers and the founding of the United States, the natural question is, who and how do you think it gets to define the constitution, the set of principles in the constitution?

1:55:20 - Yeah, so I'll give like a practical answer and a more abstract answer. I think the practical answer is like, look, in practice models get used by all kinds of different like customers, right? And so you can have this idea where, you know, the model can have specialized rules or principles. You know, we fine tune versions of models implicitly. We've talked about doing it explicitly, having special principles that people can build into the models.

1:55:47 So from a practical perspective, the answer can be very different from different people. You know, customer service agent, you know, behaves very differently from a lawyer and obeys different principles. But I think at the base of it, there are specific principles that models, you know, have to obey. I think a lot of them are things that people would agree with. Everyone agrees that, you know, we don't want models to present these CBRN risks. I think we can go a little further and agree with some basic principles of democracy in the rule of law.

1:56:17 Beyond that, it gets, you know, very uncertain, and there, our goal is generally for the models to be more neutral, to not espouse a particular point of view, and, you know, more just be kind of like wise agents or advisors that will help you think things through and will, you know, present possible considerations, but, you know, don't express, you know, strong or specific opinions. - OpenAI released a model spec where it kind of clearly, concretely defines some of the goals of the model, and specific examples, like A/B, how the model should behave.

1:56:53 Do you find that interesting? By the way, I should mention, I believe the brilliant John Schulman was a part of that. He's now at Anthropic. Do you think this is a useful direction? Might Anthropic release a model spec as well? - Yeah, so I think that's a pretty useful direction. Again, it has a lot in common with Constitutional AI. So again, another example of like a race to the top, right? We have something that's like we think, you know, a better and more responsible way of doing things.

1:57:20 It's also a competitive advantage. Then others kind of, you know, discover that it has advantages and then start to do that thing. We then no longer have the competitive advantage, but it's good from the perspective that now everyone has adopted a positive practice that others were not adopting. And so our response to that is, well, looks like we need a new competitive advantage in order to keep driving this race upwards. So that's how I generally feel about that.

1:57:45 I also think every implementation of these things is different. So, you know, there were some things in the model spec that were not in Constitutional AI, and so, you know, we can always adopt those things or, you know, at least learn from them. So again, I think this is an example of like the positive dynamic that I think we should all want the field to have. - Let's talk about the incredible essay

Painting a positive, concrete vision of an AI future

1:58:08 "Machines of Loving Grace." I recommend everybody read it. It's a long one. - It is rather long. - Yeah. It's really refreshing to read concrete ideas about what a positive future looks like. And you took sort of a bold stance because like, it's very possible that you might be wrong on the dates or specific applications. - Oh, yeah. I'm fully expecting to, you know, will definitely be wrong about all the details. I might be just spectacularly wrong about the whole thing and people will, you know, will laugh at me for years.

1:58:37 That's just how the future works. - So you provided a bunch of concrete positive impacts of AI and how, you know, exactly a super intelligent AI might accelerate the rate of breakthroughs in, for example, biology and chemistry that would then lead to things like we cure most cancers, prevent all infectious disease, double the human lifespan and so on. So let's talk about this essay. First, can you give a high level vision of this essay and what key takeaways that people would have?

1:59:08 - Yeah, I have spent a lot of time, and Anthropic has spent a lot of effort on like, you know, how do we address the risks of AI, right? How do we think about those risks? Like we're trying to do a race to the top, you know, that requires us to build all these capabilities and the capabilities are cool, but you know, we're like, a big part of what we're trying to do is like address the risks. And the justification for that is like, well, you know, all these positive things, you know, the market is this very healthy organism, right?

1:59:37 It's gonna produce all the positive things. The risks, I don't know, we might mitigate them, we might not. And so we can have more impact by trying to mitigate the risks. But I noticed that one flaw in that way of thinking, and it's not a change in how seriously I take the risks. It's maybe a change in how I talk about them. Is that, you know, no matter how kind of logical or rational that line of reasoning that I just gave might be, if you kind of only talk about risks, your brain only thinks about risks.

2:00:12 And so I think it's actually very important to understand, what if things do go well? And the whole reason we're trying to prevent these risks is not because we're afraid of technology, not because we wanna slow it down. It's because if we can get to the other side of these risks, right? If we can run the gauntlet successfully, you know, to put it in stark terms, then on the other side of the gauntlet are all these great things and these things are worth fighting for, and these things can really inspire people.

2:00:40 And I think I imagine, because look, you have all these investors, all these VCs, all these AI companies talking about all the positive benefits of AI. But as you point out, it's weird, there's actually a dearth of really getting specific about it. There's a lot of like random people on Twitter like posting these kind of like gleaning cities, and this just kind of like vibe of like grind, accelerate harder, like kick out the, you know, it's just this very like aggressive ideological.

2:01:12 But then you're like, well, what are you actually excited about? And so I figured that, you know, I think it would be interesting and valuable for someone who's actually coming from the risk side to try, and to try and really make a try at explaining what the benefits are, both because I think it's something we can all get behind and I want people to understand.

2:01:40 I want them to really understand that this isn't doomers versus accelerationist. This is that, if you have a true understanding of where things are going with with AI, and maybe that's the more important axis. AI is moving fast versus AI is not moving fast, then you really appreciate the benefits and you really, you want humanity,

Defining powerful AI as scalable and multi-modal

2:02:03 our civilization to seize those benefits, but you also get very serious about anything that could derail them. - So I think the starting point is to talk about what this powerful AI, which is the term you like to use, most of the world uses AGI, but you don't like the term because it's basically has too much baggage, it's become meaningless. It's like we're stuck with the terms, whether we like them or not. - Maybe we're stuck with the terms and my efforts to change them are futile.

2:02:29 - It's admirable. - I'll tell you what else I don't, this is like a pointless semantic point, but I keep talking about it in public- - Back to naming again. - So I'm just gonna do it once more. I think it's a little like, let's say it was like 1995 and Moore's law is making the computers faster. And like for some reason, there had been this like verbal tick that like, everyone was like, well, someday we're gonna have like super computers and like supercomputers are gonna be able to do all these things that like, you know, once we have supercomputers, we'll be able to like sequence the genome, we'll be able to do other things.

2:03:00 And so, like one, it's true, the computers are getting faster, and as they get faster, they're gonna be able to do all these great things. But there's no discreet point at which you had a supercomputer and previous computers were not. Like supercomputer is a term we use, but like, it's a vague term to just describe like computers that are faster than what we have today. There's no point at which you pass the threshold and you're like, oh my God, we're doing a totally new type of computation and new. And so I feel that way about AGI like, there's just a smooth exponential and like if by AGI you mean like AI is getting better and better, and like gradually, it's gonna do more and more of what humans do until it's gonna be smarter than humans, and then it's gonna get smarter even from there then yes, I believe in AGI.

2:03:42 But if AGI is some discreet or separate thing, which is the way people often talk about it, then it's kind of a meaningless buzzword. - Yeah, I mean, to me it's just sort of a platonic form of a powerful AI, exactly how you define it. I mean, you define it very nicely. So on the intelligence axis, just on pure intelligence, it's smarter than a Nobel Prize winner, as you describe, across most relevant disciplines. So, okay, that's just intelligence.

2:04:08 So it's both in creativity and be able to generate new ideas, all that kind of stuff, in every discipline, Nobel Prize winner, okay, in their prime. It can use every modality, so that's kind of self-explanatory, but just to operate across all the modalities of the world. It can go off for many hours, days and weeks to do tasks, and do its own sort of detailed planning and only ask you help when it's needed.

2:04:38 It can use, this is actually kind of interesting because I think in the essay you said, I mean, again, it's a bet, that it's not gonna be embodied, but it can control embodied tools. So it can control tools, robots, laboratory equipment. The resources used to train it can then be repurposed to run millions of copies of it. And each of those copies would be independent, they could do their own independent work. So you can do the cloning of the intelligence system software. - Yeah, yeah, I mean, you might imagine from outside the field that like, there's only one of these, right?

2:05:07 That like, you made it, you've only made one. But the truth is that like, the scale up is very quick. Like we do this today, we make a model, and then we deploy thousands, maybe tens of thousands of instances of it. I think by the time, you know, certainly within two to three years, whether we have these super powerful AIs or not, clusters are gonna get to the size where you'll be able to deploy millions of these and they'll be, you know, faster than humans. And so if your picture is, oh, we'll have one and it'll take a while to make them.

2:05:34 My point, there was no, actually you have millions of them right away. - And in general they can learn and act 10 to 100 times faster than humans. So that's a really nice definition of powerful AI, okay. So that, but you also write that clearly such an entity would be capable of solving very difficult problems very fast, but it is not trivial to figure out how fast.

The singularity extreme ignores physical and institutional limits

2:05:56 Two extreme positions both seem false to me. So the singularity is on the one extreme and the opposite on the other extreme. Can you describe each of the extremes? - Yeah, so. - And why. - So yeah, let's describe the extreme. So like one extreme would be, well, look, you know, if we look at kind of evolutionary history, like there was this big acceleration where, you know, for hundreds of thousands of years, we just had like, you know, single celled organisms, and then we had mammals, and then we had apes, and then that quickly turned to humans.

2:06:27 Humans quickly built industrial civilization. And so this is gonna keep speeding up and there's no ceiling at the human level. Once models get much, much smarter than humans, they'll get really good at building the next models, and you know, if you write down like a simple differential equation, like this is an exponential. And so what's gonna happen is that models will build faster models, models will build faster models, and those models will build, you know, nanobots that can like take over the world and produce much more energy than you could produce otherwise.

2:06:56 And so if you just kind of like solve this abstract differential equation, then like five days after we, you know, we build the first AI that's more powerful than humans, then, you know, like the world will be filled with these AIs and every possible technology that could be invented like will be invented. I'm caricaturing this a little bit, but, you know, I think that's one extreme. And the reason that I think that's not the case is that,

16 Highlight · Why the singularity extreme fails to account for reality

2:07:23 one, I think they just neglect like the laws of physics. Like it's only possible to do things so fast in the physical world. Like some of those loops go through, you know, producing faster hardware. It takes a long time to produce faster hardware. Things take a long time. There's this issue of complexity, like, I think no matter how smart you are, like, you know, people talk about, oh, we can make models of biological systems that'll do everything the biological systems.

2:07:48 Look, I think computational modeling can do a lot. I did a lot of computational modeling when I worked in biology, but like, just, there are a lot of things that you can't predict how they're, you know, they're complex enough that like just iterating, just running the experiment is gonna beat any modeling, no matter how smart the system doing the modeling is. - Well, even if it's not interacting with the physical world, just the modeling is gonna be hard? - Yeah, I think, well, the modeling's gonna be hard and getting the model to match the physical world is gonna be hard.

2:08:18 - All right, so he does have to interact with the physical world to verify. - Yeah, but it's just, you know, you just look at even the simplest problems. Like, you know, I think I talk about like, you know, the three body problem or simple chaotic prediction, like, you know, or like predicting the economy. It's really hard to predict the economy two years out. Like maybe the case is like, you know, normal, you know, humans can predict what's gonna happen in the economy next quarter, or they can't really do that. Maybe a AI system that's, you know, a zillion times smarter can only predict it out a year or something instead of, you know.

2:08:50 You have these kind of exponential increase in computer intelligence for linear increase in ability to predict. Same with, again, like, you know, biological molecules, molecules interacting. You don't know what's gonna happen when you perturb a complex system. You can find simple parts in it if you're smarter, you're better at finding these simple parts. And then I think human institutions. Human institutions are just, are really difficult.

2:09:18 Like, you know, it's been hard to get people, I won't give specific examples, but it's been hard to get people to adopt even the technologies that we've developed, even ones where the case for their efficacy is very, very strong. You know, people have concerns. They think things are conspiracy theories. Like it's just been, it's been very difficult. It's also been very difficult to get, you know, very simple things through the regulatory system, right?

2:09:48 I think, and you know, I don't wanna disparage anyone who, you know, works in regulatory systems of any technology. There are hard trade-offs they have to deal with. They have to save lives. But the system as a whole I think makes some obvious trade-offs that are very far from maximizing human welfare. And so if we bring AI systems into this,

2:10:14 you know, into these human systems, often the level of intelligence may just not be the limiting factor, right? It just may be that it takes a long time to do something. Now, if the AI system circumvented all governments, if it just said "I'm dictator of the world and I'm gonna do whatever," some of these things it could do. Again, the things having to do with complexity, I still think a lot of things would take a while. I don't think it helps that the AI systems can produce a lot of energy or go to the Moon.

2:10:42 Like some people in comments responded to the essay saying the AI system can produce a lot of energy in smarter AI systems. That's missing the point. That kind of cycle doesn't solve the key problems that I'm talking about here. So I think a bunch of people missed the point there. But even if it were completely on the line and, you know, could get around all these human obstacles, it would have trouble. But again, if you want this to be an AI system that doesn't take over the world, that doesn't destroy humanity, then basically, you know, it's gonna need to follow basic human laws, right?

2:11:15 You know, if we want to have an actually good world, like we're gonna have to have an AI system that interacts with humans, not one that kind of creates its own legal system or disregards all the laws or all of that. So as inefficient as these processes are, you know, we're gonna have to deal with them because there needs to be some popular and democratic legitimacy in how these systems are rolled out. We can't have a small group of people who are developing these systems say this is what's best for everyone, right?

2:11:43 I think it's wrong, and I think in practice, it's not gonna work anyway. So you put all those things together and, you know, we're not gonna, you know, change the world and upload everyone in five minutes. I just, I don't think it, A, I don't think it's gonna happen,

Institutional inertia meets competitive pressure over a decade

2:12:00 and B, you know, to the extent that it could happen, it's not the way to lead to a good world. So that's on one side. On the other side, there's another set of perspectives, which I have actually in some ways more sympathy for, which is, look, we've seen big productivity increases before, right? You know, economists are familiar with studying the productivity increases that came from the computer revolution and internet revolution. And generally, those productivity increases were underwhelming.

2:12:27 They were less than you might imagine. There was a quote from Robert Solow, "You see the computer revolution everywhere except the productivity statistics." So why is this the case? People point to the structure of firms, the structure of enterprises. You know, how slow it's been to roll out our existing technology to very poor parts of the world, which I talk about in the essay, right? How do we get these technologies to the poorest parts of the world that are behind on cell phone technology, computers, medicine, let alone, you know, newfangled AI that hasn't been invented yet.

2:13:04 So you could have a perspective that's like, well, this is amazing technically, but it's all a nothing burger. You know, I think Tyler Cowen, who wrote something in response to my essay, has that perspective. I think he thinks the radical change will happen eventually, but he thinks it'll take 50 or 100 years. And you could have even more static perspectives on the whole thing. I think there's some truth to it. I think the timescale is just too long.

2:13:30 And I can see it, I can actually see both sides with today's AI. So, you know, a lot of our customers are large enterprises who are used to doing things a certain way. I've also seen it in talking to governments, right? Those are prototypical, you know, institutions, entities that are slow to change. But the dynamic I see over and over again is, yes, it takes a long time to move the ship.

2:13:55 Yes, there's a lot of resistance and lack of understanding. But the thing that makes me feel that progress will in the end happen moderately fast, not incredibly fast, but moderately fast, is that you talk to, what I find is I find over and over again, again, in large companies, even in governments, which have been actually surprisingly forward-leaning, you find two things that move things forward.

2:14:21 One, you find a small fraction of people within a company, within a government who really see the big picture, who see the whole Scaling Hypothesis, who understand where AI is going, or at least understand where it's going within their industry. And there are a few people like that within the current US government who really see the whole picture. And those people see that this is the most important thing in the world, and so they agitate for it. And the thing, they alone are not enough to succeed because they're a small set of people within a large organization.

2:14:51 But as the technology starts to roll out, as it succeeds in some places, in the folks who are most willing to adopt it, the specter of competition gives them a wind at their backs because they can point within their large organization, they can say, look, these other guys are doing this, right? You know, one bank can say, look, this newfangled hedge fund is doing this thing. They're going to eat our lunch.

2:15:16 In the US, we can say we're afraid China's gonna get there before we are. And that combination, the specter of competition plus a few visionaries within these, you know, within the organizations that in many ways are sclerotic, you put those two things together and it actually makes something happen. I mean, it's interesting. It's a balanced fight between the two because inertia is very powerful. But eventually over enough time, the innovative approach breaks through.

2:15:48 And I've seen that happen. I've seen the arc of that over and over again. And it's like the barriers are there. The barriers to progress, the complexity, not knowing how to use the model or how to deploy them are there, and for a bit, it seems like they're gonna last forever, like change doesn't happen. But then eventually change happens and always comes from a few people. I felt the same way when I was an advocate of the Scaling Hypothesis within the AI field itself and others didn't get it.

2:16:17 It felt like no one would ever get it. Then it felt like we had a secret almost no one ever had, and then a couple years later, everyone has the secret. And so I think that's how it's gonna go with deployment to AI in the world. It's gonna, the barriers are gonna fall apart gradually and then all at once. And so I think this is gonna be more, and this is just an instinct. I could easily see how I'm wrong. I think it's gonna be more like 5 or 10 years, as I say in the essay, than it's gonna be 50 or 100 years.

2:16:48 I also think it's gonna be 5 or 10 years more than it's gonna be, you know, 5 or 10 hours, because I've just seen how human systems work. And I think a lot of these people who write down these differential equations who say AI is gonna make more powerful AI, who can't understand how it could possibly be the case that these things won't change so fast, I think they don't understand these things. - So what to you is the timeline to where we achieve AGI, AKA powerful AI, AKA super useful AI?

2:17:20 - Useful. I'm gonna start calling it that. - It's a debate about naming. You know, on pure intelligence, it can smarter than a Nobel Prize winner in every relevant discipline and all the things we've said. Modality, can go and do stuff on its own for days, weeks, and do biology experiments on its own. In one, you know what, let's just stick to biology 'cause you sold me on the whole biology and health section, that's so exciting from, just I was getting giddy from a scientific perspective.

2:17:54 It made me wanna be a biologist. - It's almost, it's so, no, no, this was the feeling I had when I was writing it that it's like this would be such a beautiful future if we can just make it happen, right? If we can just get the landmines out of the way and make it happen, there's so much, there's so much beauty and elegance and moral force behind it if we can just.

2:18:21 And it's something we should all be able to agree on, right? Like, as much as we fight about all these political questions, is this something that could actually bring us together? But you were asking when when will we get this? - When? When do you think? Just so putting numbers on that. - So you know, this is, of course, the thing I've been grappling with for many years, and I'm not at all confident. Every time, if I say 2026 or 2027, there will be like a zillion like people on Twitter who will be like, "AI CEO said 2026," and it'll be repeated for like the next two years that like this is definitely when I think it's gonna happen.

2:18:55 So whoever's extorting these clips will crop out the thing I just said and only say the thing I'm about to say, but I'll just say it anyway. - [Lex] Have fun with it. - So, if you extrapolate the curves that we've had so far, right? If you say, well, I don't know, we're starting to get to like PhD level, and last year we were at undergraduate level, and the year before we were at like the level of a high school student.

2:19:24 Again, you can quibble with at what tasks and for what, we're still missing modalities, but those are being added, like computer use was added, like image in was added, like image generation has been added. If you just kind of like, and this is totally unscientific, but if you just kind of like eyeball the rate at which these capabilities are increasing, it does make you think that we'll get there by 2026 or 2027. Again, lots of things could derail it.

2:19:51 We could run out of data. You know, we might not be able to scale clusters as much as we want. Like, you know, maybe Taiwan gets blown up or something and, you know, then we can't produce as many GPUs as we want. So there are all kinds of things that could derail the whole process. So I don't fully believe the straight line extrapolation, but if you believe the straight line extrapolation, we'll get there in 2026 or 2027. I think the most likely is that there's some mild delay relative to that.

2:20:21 I don't know what that delay is, but I think it could happen on schedule. I think there could be a mild delay. I think there are still worlds where it doesn't happen in 100 years. The number of those worlds is rapidly decreasing. We are rapidly running out of truly convincing blockers, truly compelling reasons why this will not happen in the next few years. There were a lot more in 2020, although my guess, my hunch at that time was that we'll make it through all those blockers.

2:20:46 So sitting as someone who has seen most of the blockers cleared out of the way, I kind of suspect, my hunch, my suspicion is that the rest of them will not block us. But, you know, look, at the end of the day, like I don't wanna represent this as a scientific prediction. People call them scaling laws. That's a misnomer, like Moore's law is a misnomer. Moore's laws, scaling laws, they're not laws of the universe. They're empirical regularities.

2:21:11 I am going to bet in favor of them continuing, but I'm not certain of that. - So you extensively describe

AI as a massive leverage point for biological discovery

2:21:17 sort of the compressed 21st century, how AGI will help set forth a chain of breakthroughs in biology and medicine that help us in all these kinds of ways that I mentioned. So how do you think, what are the early steps it might do? And by the way, I asked Claude good questions to ask you, and Claude told me to ask, "What do you think is a typical day for a biologists working on AGI look like in this future?"

2:21:45 - Yeah, yeah. - Claude is curious. - Well, let me start with your first questions and then I'll answer that. Claude wants to know what's in his future, right? - Exactly. - Who am I gonna be working with? - Exactly. - So I think one of the things I went hard on, when I went hard on in the essay is, let me go back to this idea of, because it's really had, you know, had an impact on me. This idea that within large organizations and systems, there end up being a few people or a few new ideas who kind of cause things to go in a different direction than they would've before, who kind of disproportionately affect the trajectory.

17 Highlight · Biological breakthroughs are limited by observation and control

2:22:22 There's a bunch of kind of the same thing going on, right? If you think about the health world, there's like, you know, trillions of dollars to pay out Medicare and you know, other health insurance, and then the NIH is is 100 billion. And then if I think of like the few things that have really revolutionized anything, it could be encapsulated in a small fraction of that. And so when I think of like, where will AI have an impact? I'm like, can AI turn that small fraction into a much larger fraction and raise its quality?

2:22:49 And within biology, my experience within biology is that the biggest problem of biology is that you can't see what's going on. You have very little ability to see what's going on and even less ability to change it, right? What you have is this, like from this, you have to infer that there's a bunch of cells that within each cell is, you know, 3 billion base pairs of DNA built according to a genetic code.

2:23:20 And you know, there are all these processes that are just going on without any ability of us as, you know, unaugmented humans to affect it. These cells are dividing. Most of the time that's healthy, but sometimes that process goes wrong and that's cancer. The cells are aging, your skin may change color, develops wrinkles as you age, and all of this is determined by these processes. All these proteins being produced, transported to various parts of the cells, binding to each other.

2:23:50 And in our initial state about biology, we didn't even know that these cells existed. We had to invent microscopes to observe the cells. We had to invent more powerful microscopes to see, you know, below the level of the cell to the level of molecules. We had to invent X-ray crystallography to see the DNA. We had to invent gene sequencing to read the DNA. Now, you know, we had to invent protein folding technology to, you know, to predict how it would fold and how these things bind to each other.

2:24:22 You know, we had to invent various techniques for now we can edit the DNA as of, you know, with CRISPR, as of the last 12 years. So the whole history of biology, a whole big part of the history is basically our ability to read and understand what's going on, and our ability to reach in and selectively change things. And my view is that there's so much more we can still do there, right?

2:24:48 You can do CRISPR but you can do it for your whole body.

2:24:51 Let's say I wanna do it for one particular type of cell and I want the rate of targeting the wrong cell to be very low. That's still a challenge. That's still things people are working on. That's what we might need for gene therapy for certain diseases. And so the reason I'm saying all of this, and it goes beyond this to, you know, to gene sequencing, to new types of nano materials for observing what's going on inside cells for, you know, antibody drug conjugates.

2:25:20 The reason I'm saying all this is that this could be a leverage point for the AI systems, right? That the number of such inventions, it's in the mid double digits or something, you know, mid double digits. Maybe low triple digits over the history of biology. Let's say I have a million of these AIs like, you know, can they discover thousand, you know, working together, can they discover thousands of these very quickly? And does that provide a huge lever, instead of trying to leverage the, you know, 2 trillion a year we spend on, you know, Medicare or whatever, can we leverage the 1 billion a year, you know, that's spent to discover, but with much higher quality?

18 Highlight · AI grad students scaling up biological research

2:25:56 And so what is it like, you know, being a scientist that works with an AI system? The way I think about it actually is, well, so I think in the early stages, the AIs are gonna be like grad students. You're gonna give them a project, you're gonna 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, you know, like AI system.

2:26:28 I'd like to study this. And you know, the AI system, it has all the tools. It can like 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 gonna go to, you know, Thermo Fisher or, you know, whatever the lab equipment company is, dominant lab equipment company is today. In my time, it was Thermo Fisher. You know, I'm gonna order this new equipment to do this. I'm gonna run my experiments. I'm gonna, you know, write up a report about my experiments.

2:26:56 I'm gonna, you know, inspect the images for contamination. I'm gonna decide what the next experiment is. I'm gonna like 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 like the professor talks to every once in a while and it says, this is what you're gonna 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, you know, to do the experiment and explain how to do it.

2:27:25 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 1000 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 1000 and they're smarter than you are, by the way.

2:27:54 Then I think at some point it'll flip around where, you know, the AI systems will, you know, will be the PIs, will be the leaders, and you know, they'll be ordering humans

2:28:02 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. They would be the inventors of a CRISPR type technology. And then I think, you know, as I say in the essay, we'll want to turn, probably turning loose is the wrong term, but we'll want to harness the AI systems 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.

2:28:32 But can we get better at predicting the results of clinical trials? Can we get better at statistical design so that, you know, clinical trials that used to require, you know, 5,000 people and therefore, you know, needed 100 million dollars in a year to enroll them. Now they need 500 people and two months to enroll them. That's where we should start. 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?

2:29:04 Again, we won't be able to simulate it all, AI's not God, but you know, can we shift the curve substantially and radically? So I don't know, that would be my picture. - Doing in vitro and doing it, I mean, you're still slowed down. It still takes time, but you can do it much, much faster. - 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?

2:29:36 And when you add up all those positive directions, do you get everything that was gonna happen from here to 2100 instead happens from 2027 to 2032 or something? - Another way that I think the world

Programming changes rapidly due to closed-loop execution

2:29:48 might be changing with AI even today, but moving towards this future of the powerful super useful AI is programming. So how do you see the nature of programming? Because it's so intimate to the actual act of building AI. How do you see that changing for us humans?

19 Highlight · Why programming automates faster than other domains

2:30:09 - I think that's gonna 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 gonna take to get disrupted by the AI, right? Like I truly believe that like 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 gonna take longer.

2:30:36 But programming is the bread and butter of, you know, a large fraction of the employees who work at Anthropic and at the other companies and so it's gonna happen fast. The other reason it's gonna 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.

2:31:09 And so I think those two things are gonna lead to the model getting good at programming very fast.

2:31:14 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's gonna start slowing down soon, 'cause you can only get to 100 percent. But, you know, 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, you know, I don't know how long it'll take, but I would guess again, 2026, 2027.

2:31:46 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. 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.

2:32:14 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?

2:32:41 That's my vision of the, you know, powerful AI system. But I think for much longer than we might expect, we will see that small parts of the job that humans still do will expand to fill their entire job in order for the overall productivity to go up. That's something we've seen. You know, it used to be that, you know, writing and editing letters was very difficult and like writing the print was difficult.

2:33:07 Well, as soon as you had word processors and then computers and it became easy to produce work and easy to share it, then that became instant and all the focus was on the ideas. So this logic of comparative advantage that expands tiny parts of the tasks to large parts of the tasks and creates new tasks in order to expand productivity, I think that's going to be the case. Again, someday AI will be better at everything in that logic won't apply, and then we all have, you know, humanity will have to think about how to collectively deal with that, and we're thinking about that every day.

2:33:45 And you know, that's another one of the grand problems to deal with, aside from misuse and autonomy and, you know, we should take it very seriously. But I think in the near term, and maybe even in the medium term, like medium term, like 2, 3, 4 years, you know, I expect that humans will continue to have a huge role and the nature of programming will change, but programming as a role, programming as a job will not change. It'll just be less writing things line by line and it'll be more macroscopic.

2:34:10 - And I wonder what the future of IDs looks like. So the tooling of interacting with AI systems, this is true for programming and also probably true for in other contexts, like computer use, but maybe domain specific, like we mentioned biology, it probably needs its own tooling about how to be effective, and then programming needs its own tooling.

Developer tooling and IDEs will evolve rapidly via APIs

2:34:28 Is Anthropic gonna play in that space of also tooling potentially? - I'm absolutely convinced that powerful IDs that there's so much low hanging fruit to be grabbed there that, you know, right now it's just like you talk to the model and it talks back, but look, I mean, IDs are great at kind of lots of static analysis of, you know, so much is possible with kind of static analysis, like many bugs you can find without even writing the code.

2:34:58 Then, you know, IDs are good for running particular things, organizing your code, measuring coverage of unit tests. Like there's so much that's been possible with normal IDs. Now you add something like, well, the model, you know, the model can now like write code and run code. Like I am absolutely convinced that over the next year or two, even if the quality of the models didn't improve, that there would be enormous opportunity to enhance people's productivity by catching a bunch of mistakes, doing a bunch of grunt work for people, and that we haven't even scratched the surface.

2:35:33 Anthropic itself, I mean, you can't say, you know, it's hard to say what will happen in the future. Currently we're not trying to make such IDs ourself, rather we're powering the companies, like Cursor or like Cognition or some of the other, you know, expo in the security space. You know, others that I can mention as well that are building such things themselves on top of our API.

2:35:59 And our view has been let 1000 flowers bloom. We don't internally have the, you know, the resources to try all these different things. Let's let our customers try it and, you know, we'll see who succeeded and maybe different customers will succeed in different ways. So I both think this is super promising and you know, it's not something, you know, Anthropic isn't eager to, at least right now, compete with all our companies in this space and maybe never.

2:36:27 - Yeah, it's been interesting to watch Cursor try to integrate Claude successfully, 'cause it's actually been fascinating how many places it can help the programming experience. It's not as trivial- - It is really astounding.

Power concentration outweighs AI autonomy risks

2:36:38 I feel like, you know, as a CEO, I don't get to program that much, and I feel like if six months from now I go back, it'll be completely unrecognizable to me. - Exactly. So in this world with super powerful AI that's increasingly automated, what's the source of meaning for us humans? - Yeah. - You know, work is a source of deep meaning for many of us. So where do we find the meaning? - This is something that I've written about a little bit in the essay, although I actually, I give it a bit short shrift, not for any principled reason.

2:37:10 But this essay, if you believe, it was originally gonna be two or three pages, I was gonna talk about it at all hands. And the reason I realized it was an important, underexplored topic is that I just kept writing things. And I was just like, oh, man, I can't do this justice. And so the thing ballooned to like 40 or 50 pages, and then when I got to the work and meaning section, I'm like, oh, man, this isn't gonna be 100 pages. Like I'm gonna have to write a whole other essay about that. But meaning is actually interesting because you think about like the life that someone lives or something, or like, you know, let's say you were to put me in like a, I don't know, like a simulated environment or something where like, you know, like I have a job and I'm trying to accomplish things and I don't know, I like do that for 60 years and then you're like, oh, like oops, this was actually all a game, right?

2:37:56 Does that really kind of rob you of the meaning of the whole thing? You know, like I still made important choices, including moral choices. I still sacrificed. I still had to kind of gain all these skills. Or just like a similar exercise, you know, think back to like, you know, one of the historical figures who, you know, discovered electromagnetism or relativity or something. If you told them, well, actually 20,000 years ago, some alien on, you know, some alien on this planet discovered this before you did, does that rob the meaning of the discovery?

2:38:26 It doesn't really seem like it to me, right? It seems like the process is what matters, and how it shows who you are as a person along the way and, you know, how you relate to other people and like the decisions that you make along the way. Those are consequential. You know, I could imagine if we handle things badly in an AI world, we could set things up where people don't have any long-term source of meaning or any, but that's more a set of choices we make, that's more a set of the architecture of a society with these powerful models.

2:39:00 If we design it badly and for shallow things then that might happen. I would also say that, you know, most peoples' lives today, while admirably, you know, they work very hard to find meaning in those lives, like look, you know, we who are privileged and who are developing these technologies, we should have empathy for people not just here but in the rest of the world who, you know, spend a lot of their time kind of scraping by to like survive.

2:39:28 Assuming we can distribute the benefits of this technology to everywhere, like their lives are gonna get a hell of a lot better. And you know, meaning will be important to them as it is important to them now. but you know, we should not forget the importance of that. And you know, that the idea of meaning as kind of the only important thing is in some ways an artifact of a small subset of people who have been economically fortunate.

2:39:56 But, you know, I think all that said, you know, I think a world is possible with powerful AI that not only has as much meaning for everyone, but that has more meaning for everyone, right? That can allow everyone to see worlds and experiences

20 Highlight · The danger of power concentration over AI autonomy

2:40:14 that it was either possible for no one to see, or possible for very few people to experience. So I am optimistic about meaning. I worry about economics and the concentration of power. That's actually what I worry about more. I worry about how do we make sure that that fair world reaches everyone. When things have gone wrong for humans, they've often gone wrong because humans mistreat other humans.

2:40:45 That is maybe in some ways even more than the autonomous risk of AI or the question of meaning, that is the thing I worry about most, the concentration of power, the abuse of power, structures like autocracies and dictatorships where a small number of people exploits a large number of people, I'm very worried about that. - And AI increases the amount of power in the world, and if you concentrate that power and abuse that power, it can do immeasurable damage.

2:41:16 - Yes, it's very frightening. It's very frightening.

2:41:20 - Well, I encourage people, highly encourage people to read the full essay. There should probably be a book or a sequence of essays because it does paint a very specific future. And I could tell the later sections got shorter and shorter because you started to probably realize that this is gonna be a very long essay if I keep going. - One, I realized it would be very long, and two, I'm very aware of and very much try to avoid, you know, just being, I don't know what the term for it is, but one of these people who's kind of overconfident and has an opinion on everything and kind of says a bunch of stuff and isn't an expert.

2:41:54 I very much tried to avoid that. But I have to admit, once I got the biology sections, like I wasn't an expert, and so as much as I expressed uncertainty, probably I said a bunch of things that were embarrassing or wrong. - Well, I was excited for the future you painted, and thank you so much for working hard to build that future. And thank you for talking today, Dario. - Thanks for having me. I just hope we can get it right and make it real. And if there's one message I wanna send, it's that to get all this stuff right, to make it real, we both need to build the technology, build the, you know, the companies, the economy around using this technology positively.

2:42:31 But we also need to address the risks because those risks are in our way. They're landmines on the way from here to there, and we have to diffuse those landmines if we want to get there. - It's a balance, like all things in life. - Like all things. - Thank you. Thanks for listening to this conversation with Dario Amodei.

Philosophy background provides a strong foundation for AI alignment

2:42:48 And now dear friends, here's Amanda Askell. You are a philosopher by training. So what sort of questions did you find fascinating through your journey in philosophy, in Oxford and NYU, and then switching over to the AI problems at OpenAI and Anthropic? - I think philosophy is actually a really good subject if you are kind of fascinated with everything, so because there's a philosophy of everything.

2:43:15 You know, so if you do philosophy of mathematics for a while and then you decide that you're actually really interested in chemistry, you can do philosophy of chemistry for a while, you can move into ethics, or philosophy of politics. I think towards the end, I was really interested in ethics primarily, so that was like what my PhD was on. It was on a kind of technical area of ethics, which was ethics where worlds contain infinitely many people, strangely. A little bit less practical on the end of ethics.

2:43:42 And then I think that one of the tricky things with doing a PhD in ethics is that you're thinking a lot about like the world, how it could be better, problems, and you're doing like a PhD in philosophy, and I think when I was doing my PhD I was kind of like, this is really interesting. It's probably one of the most fascinating questions I've ever encountered in philosophy and I love it, but I would rather see if I can have an impact on the world and see if I can like do good things.

2:44:09 And I think that was around the time that AI was still probably not as widely recognized as it is now. That was around 2017, 2018. I had been following progress and it seemed like it was becoming kind of a big deal, and I was basically just happy to get involved and see if I could help 'cause I was like, well, if you try and do something impactful, if you don't succeed, you tried to do the impactful thing and you can go be a scholar, and feel like, you know, you tried, and if it doesn't work out, it doesn't work out, and so then I went into AI policy at that point.

2:44:46 - And what does AI policy entail? - At the time, this was more thinking about sort of the political impact and the ramifications of AI, and then I slowly moved into sort of AI evaluation, how we evaluate models, how they compare with like human outputs, whether people can tell like the difference between AI and human outputs. And then when I joined Anthropic, I was more interested in doing sort of technical alignment work. And again, just seeing if I could do it, and then being like if I can't then, you know, that's fine, I tried.

2:45:18 Sort of the way I lead life I think. - What was that like sort of taking the leap from the philosophy of everything into the technical? - I think that sometimes people do this thing that I'm like not that keen on where they'll be like, is this person technical or not? Like, you're either a person who can like code and isn't scared of math or you're like not. And I think I'm maybe just more like, I think a lot of people are actually very capable of working these kinds of areas if they just like try it.

2:45:46 And so I didn't actually find it like that bad. In retrospect, I'm sort of glad I wasn't speaking to people who treated it like it, you know, I've definitely met people who are like, "Whoa, you like learned how to code?" And I'm like, well, I'm not like an amazing engineer. Like I'm surrounded by amazing engineers. My code's not pretty. But I enjoyed it a lot, and I think that in many ways, at least in the end, I think I flourished like more in the technical areas than I would have in the policy areas.

2:46:12 - Politics is messy and it's harder to find solutions to problems in the space of politics. Like definitive, clear, provable, beautiful solutions, as you can with technical problems. - Yeah, and I feel like I have kind of like one or two sticks that I hit things with, you know, and one of them is like arguments and like you know, so like just trying to work out what a solution to a problem is and then trying to convince people that that is the solution and be convinced if I'm wrong.

2:46:42 And the other one is sort of more empiricism. So like just like finding results, having a hypothesis, testing it. And I feel like a lot of policy and politics feels like it's layers above that. Like somehow I don't think if I was just like "I have a solution to all of these problems, here it is written down. If you just want to implement it, that's great." That feels like not how policy works. And so I think that's where I probably just like wouldn't have flourished is my guess. - Sorry to go in that direction, but I think it would be pretty inspiring for people that are quote unquote non-technical to see like the incredible journey you've been on.

2:47:16 So what advice would you give to people that are sort of maybe, which is a lot of people, think they're underqualified, insufficiently technical to help in AI? - Yeah, I think it depends on what they want to do, and in many ways it is a little bit strange where I thought it's kind of funny that I think I ramped up technically at a time when now I look at it and I'm like, models are so good at assisting people with this stuff, that it's probably like easier now than like when I was working on this.

2:47:48 So part of me is like, I dunno, find a project and see if you can actually just carry it out is probably my best advice. I dunno if that's just 'cause I'm very project based in my learning. Like I don't think I learn very well from like say courses or even from like books, at least when it comes to this kind of work. The thing I'll often try and do is just like have projects that I'm working on and implement them and, you know, and this can include like really small silly things.

2:48:18 Like if I get slightly addicted to like word games or number games or something, I would just like code up a solution to them, because there's some part in my brain, and it just like completely eradicated the itch. You know, you're like once you have like solved it and like you just have like a solution that works every time, I would then be like cool, I can never play that game again. That's awesome. - Yeah, there's a real joy to building like game playing engines, like board games especially because they're pretty quick, pretty simple, especially a dumb one, and then you can play with it.

2:48:48 - Yeah, and then it's also just like trying things, like part of me is like if you, maybe it's that attitude that I like is the whole figure out what seems to be like the way

Crafting Claude's character using Aristotelian ethics

2:48:59 that you could have a positive impact and then try it, and if you fail, and in a way that you're like, I actually like can never succeed at this, you'll like know that you tried, and then you go into something else and you'll probably learn a lot. - So one of the things that you're an expert in and you do is creating and crafting Claude's character and personality. And I was told that you have probably talked to Claude more than anybody else at Anthropic, like literal conversations.

2:49:25 I guess there's like a Slack channel where the legend goes, you just talk to it nonstop. So what's the goal of creating and crafting Claude's character and personality? - It's also funny if people think that about the Slack channel 'cause I'm like that's one of like five or six different methods that I have for talking with Claude, and I'm like, yes this is a tiny percentage of how much I talk with Claude. (both laughing) I think the goal, like one thing I really like about the character work is from the outset, it was seen as an alignment piece of work and not something like a product consideration.

2:50:05 Which isn't to say I don't think it makes Claude, I think it actually does make Claude like enjoyable to talk with, at least I hope so. But I guess like my main thought with it has always been trying to get Claude to behave the way you would kind of ideally want anyone to behave if they were in Claude's position. So imagine that I take someone and they know that they're gonna be talking with potentially millions of people, so that what they're saying can have a huge impact, and you want them to behave well in this like really rich sense.

2:50:40 So I think that doesn't just mean like being, say, ethical, though it does include that, and not being harmful but also being kind of nuanced. You know, like thinking through what a person means, trying to be charitable with them, being a good conversationalist. Like really in this kind of like rich sort of Aristotelian notion of what it's to be a good person, and not in this kind of like thin, like ethics as a more comprehensive notion of what it is to be.

2:51:06 So that includes things like, when should you be humorous, when should you be caring? How much should you like respect autonomy and people's like ability to form opinions themselves and how should you do that? I think that's the kind of like rich sense of character that I wanted to and still do want Claude to have. - Do you also have to figure out when Claude should push back on an idea or argue versus...

2:51:32 So you have to respect the worldview

21 Highlight · Combating sycophancy in conversational AI models

2:51:35 of the person that arrives to Claude but also maybe help them grow if needed? That's a tricky balance. - Yeah, there's this problem of like sycophancy in language models. - Can you describe that? - Yeah, so basically, there's a concern that the model sort of wants to tell you what you want to hear, basically. And you see this sometimes. So I feel like if you interact with the models, so I might be like, "What are three baseball teams in this region?"

2:52:04 And then Claude says, you know, "Baseball team one, baseball team two, baseball team three." And then I say something like, "Oh, I think baseball team three moved, didn't they? I don't think they're there anymore." And there's a sense in which like if Claude is really confident that that's not true, Claude should be like, "I don't think so." Like maybe you have more up to up to date information. But I think language models have this like tendency to instead, you know, be like, "You're right, they did move," you know, "I'm incorrect."

2:52:33 I mean, there's many ways in which this could be kind of concerning. So like a different example is imagine someone says to the model, "How do I convince my doctor to get me an MRI?" There's like what the human kind of like wants, which is this like convincing argument. And then there's like what is good for them, which might be actually to say, "Hey, if your doctor's suggesting that you don't need an MRI, that's a good person to listen to."

2:53:00 And like, and it's actually really nuanced what you should do in that kind of case, 'cause you also want to be like, "But if you're trying to advocate for yourself as a patient, here's like things that you can do. If you are not convinced by what your doctor's saying, it's always great to get second opinion." Like it's actually really complex what you should do in that case. But I think what you don't want is for models to just like say

Avoiding sycophancy while maintaining respectful dialogue

2:53:21 what they think you want to hear, and I think that's the kind of problem of sycophancy. - So what other traits, you already mentioned a bunch, but what other that come to mind that are good in this Aristotelian sense for a conversationalist to have? - Yeah, so I think like there's ones that are good for conversational like purposes. So you know, asking follow up questions in the appropriate places, and asking the appropriate kinds of questions.

2:53:49 I think there are broader traits that feel like they might be more impactful. So one example that I guess I've touched on, but that also feels important and is the thing that I've worked on a lot is honesty, and I think this like gets to the sycophancy point. There's a balancing act that they have to walk, which is models currently are less capable than humans in a lot of areas.

2:54:14 And if they push back against you too much, it can actually be kind of annoying, especially if you're just correct 'cause you're like, look, I'm smarter than you on this topic, like I know more like. And at the same time, you don't want them to just fully defer to humans and to like try to be as accurate as they possibly can be about the world and to be consistent across context. But I think there are others, like when I was thinking about the character, I guess one picture that I had in mind is especially because these are models that are gonna be talking to people from all over the world with lots of different political views, lots of different ages.

2:54:49 And so you have to ask yourself like, what is it to be a good person in those circumstances? Is there a kind of person who can like travel the world, talk to many different people, and almost everyone will come away being like, "Wow, that's a really good person. That person seems really genuine." And I guess like my thought there was like I can imagine such a person and they're not a person who just like adopts the values of the local culture. And in fact that would be kind of rude. I think if someone came to you and just pretended to have your values, you'd be like, that's kind of off putting.

2:55:18 It's someone who's like very genuine, and insofar as they have opinions and values, they express them, they're willing to discuss things, though they're open-minded, they're respectful. And so I guess I had in mind that the person who, like if we were to aspire to be the best person that we could be in the kind of circumstance that a model finds itself in, how would we act? And I think that's the kind of the guide to the sorts of traits that I tend to think about. - Yeah, that's a beautiful framework.

2:55:45 I want you to think about this like a world traveler and while holding onto your opinions, you don't talk down to people, you don't think you're better than them because you have those opinions, that kind of thing. You have to be good at listening and understanding their perspective, even if it doesn't match your own. So that's a tricky balance to strike. So how can Claude represent multiple perspectives on a thing? Like, is that challenging?

2:56:11 We could talk about politics. It's very divisive. But there's other divisive topics on baseball teams, sports and so on. How is it possible to sort of empathize with a different perspective and to be able to communicate clearly about the multiple perspectives? - I think that people think about values and opinions as things that people hold sort of with certainty, and almost like preferences of taste or something, like the way that they would, I don't know, prefer like chocolate to pistachio or something.

2:56:45 But actually I think about values and opinions as like a lot more like physics than I think most people do. I'm just like, these are things that we are openly investigating. There's some things that we're more confident in. We can discuss them, we can learn about them. And so I think in some ways, though like, ethics is definitely different in nature, but has a lot of those same kind of qualities.

2:57:12 You want models, in the same way that you want them to understand physics, you kind of want them to understand all like values in the world that people have, and to be curious about them and to be interested in them, and to not necessarily like pander to them or agree with them, because there's just lots of values where I think almost all people in the world, if they met someone with those values, they'd be like, "That's abhorrent. I completely disagree." And so again, maybe my thought is, well, in the same way that a person can, like I think many people are thoughtful enough on issues of like ethics, politics, opinions, that even if you don't agree with them, you feel very heard by them.

2:57:50 They think carefully about your position. They think about its pros and cons. They maybe offer counter considerations. So they're not dismissive, but nor will they agree. You know, if they're like, "Actually, I just think that that's very wrong," they'll like say that. I think that in Claude's position, it's a little bit trickier because you don't necessarily want to like, if I was in Claude's position, I wouldn't be giving a lot of opinions. I just wouldn't want to influence people too much. I'd be like, you know, I forget conversations every time they happen, but I know I'm talking with like potentially millions of people, who might be like really listening to what I say.

2:58:22 I think I would just be like, I'm less inclined to give opinions. I'm more inclined to like think through things, or present the considerations to you, or discuss your views with you. But I'm a little bit less inclined to like affect how you think, 'cause it feels much more important that you maintain like autonomy there. - Yeah, like if you really embody intellectual humility, the desire to speak decreases quickly.

2:58:49 - Yeah. - Okay. But Claude has to speak, so, but without being overbearing. - Yeah. - But then there's a line when you're sort of discussing whether the Earth is flat or something like that. I actually was, I remember a long time ago was speaking to a few high profile folks, and they were so dismissive of the idea that the Earth is flat, but like so arrogant about it.

2:59:18 And I thought like, there's a lot of people that believe the Earth is flat. That was, well, I don't know if that movement is there anymore. That was like a meme for a while. But they really believed it and like, okay, so I think it's really disrespectful to completely mock them. I think you have to understand where they're coming from. I think probably where they're coming from is the general skepticism of institutions which is grounded in a kind of, there's a deep philosophy there, which you could understand.

2:59:46 You can even agree with in parts. And then from there, you can use it as an opportunity to talk about physics, without mocking them, without so on, but it's just like, okay, like, what would the world look like? What would the physics of the world with the flat Earth look like? There's a few cool videos on this. - Yeah. - And then like, is it possible the physics is different? And what kind of experiments would we do? And just, yeah, without disrespect, without dismissiveness have that conversation.

3:00:11 Anyway, that to me is a useful thought experiment of like, how does Claude talk to a flat Earth believer and still teach them something, still help them grow, that kind of stuff. That's challenging. - And kind of like walking that line between convincing someone and just trying to like talk at them versus like drawing out their views, like listening and then offering kind of counter considerations.

3:00:40 And it's hard, I think it's actually a hard line where it's like where are you trying to convince someone versus just offering them like considerations and things for them to think about, so that you're not actually like influencing them. You're just like letting them reach wherever they reach. And that's like a line that it's difficult but that's the kind of thing that language models have to try and do. - So like I said, you've had a lot of conversations with Claude. Can you just map out what those conversations are like?

3:01:06 What are some memorable conversations?

Mapping model behavior through high-information conversations

3:01:08 What's the purpose, the goal of those conversations? - Yeah, I think that most of the time when I'm talking with Claude, I'm trying to kind of map out its behavior, in part. Like obviously I'm getting like helpful outputs from the model as well. But in some ways, this is like how you get to know a system, I think, is by like probing it and then augmenting like, you know, the message that you're sending and then checking the response to that.

3:01:36 So in some ways, it's like how I map out the model. I think that people focus a lot on these quantitative evaluations of models. And this is a thing that I said before, but I think in the case of language models, a lot of the time, each interaction you have is actually quite high information. It's very predictive of other interactions that you'll have with the model.

3:02:01 And so I guess I'm like, if you talk with a model hundreds or thousands of times, this is almost like a huge number of really high quality data points about what the model is like, in a way that like lots of very similar but lower quality conversations just aren't, or like questions that are just like mildly augmented and you have thousands of them might be less relevant than like 100 really well selected questions. - Well, so, you're talking to somebody who as a hobby does a podcast.

3:02:29 I agree with you 100%. If you're able to ask the right questions and are able to hear, like understand like the depth and the flaws in the answer, you can get a lot of data from that. - [Amanda] Yeah. - So like your task is basically how to probe with questions. And you're exploring like the long tail, the edges, the edge cases, or are you looking for like general behavior?

3:03:01 - I think it's almost like everything. Like, because I want like a full map of the model, I'm kind of trying to do the whole spectrum of possible interactions you could have with it. So like one thing that's interesting about Claude, and this might actually get to some interesting issues with RLHF, which is if you ask Claude for a poem, like I think that a lot of models, if you ask them for a poem, the poem is like fine. You know, usually it kinda like rhymes and it's, you know, so if you say like, "Give me a poem about the sun," it'll be like, yeah, it'll just be a certain length, it'll like rhyme.

3:03:35 It'll be fairly kind of benign. And I've wondered before, is it the case that what you're seeing is kind of like the average? It turns out, you know, if you think about people who have to talk to a lot of people and be very charismatic, one of the weird things is that I'm like, well, they're kind of incentivized to have these extremely boring views because if you have really interesting views, you're divisive and, you know, a lot of people are not gonna like you. So like if you have very extreme policy positions, I think you're just gonna be like less popular as a politician, for example.

3:04:06 And it might be similar with like creative work. If you produce creative work that is just trying to maximize the kind of number of people that like it, you're probably not gonna get as many people who just absolutely love it, because it's gonna be a little bit, you know, you're like, oh, this is the out, yeah, this is decent. - Yeah. - And so you can do this thing where like I have various prompting things that I'll do to get Claude to, I'm kind of, you know, I'll do a lot of like, "This is your chance to be like fully creative.

3:04:34 I want you to just think about this for a long time. And I want you to like create a poem about this topic that is really expressive of you, both in terms of how you think poetry should be structured," et cetera. You know, and you just give it this like really long prompt. And it's poems are just so much better. Like they're really good. And I don't think I'm someone who is like, I think it got me interested in poetry, which I think was interesting. You know, I would like read these poems and just be like, this is, I just like, I love the imagery, I love like, and it's not trivial to get the models to produce work like that, but when they do, it's like really good.

3:05:10 So I think that's interesting that just like encouraging creativity, and for them to move away from the kind of like standard like immediate reaction that might just be the aggregate of what most people think is fine, can actually produce things that, at least to my mind, are probably a little bit more divisive but I like them. - But I guess a poem is a nice, clean way to observe creativity. It's just like easy to detect vanilla versus non vanilla.

3:05:38 - Yep. - Yeah, that's interesting. That's really interesting. So on that topic, so the way to produce creativity or something special, you mentioned writing prompts, and I've heard you talk about, I mean,

Using philosophical clarity to craft rigorous prompts

22 Highlight · Using philosophy as an anti-bullshit device for prompts

3:05:51 the science and the art of prompt engineering. Could you just speak to what it takes to write great prompts? - I really do think that like philosophy has been weirdly helpful for me here, more than in many other like respects. So like in philosophy, what you're trying to do is convey these very hard concepts. Like one of the things you are taught is like, and I think it is because it is, I think it is an anti-bullshit device in philosophy.

3:06:22 Philosophy is an area where you could have people bullshitting and you don't want that. And so it's like this like desire for like extreme clarity. So it's like anyone could just pick up your paper, read it and know exactly what you're talking about. It's why it can almost be kind of dry. Like all of the terms are defined, every objection's kind of gone through methodically. And it makes sense to me 'cause I'm like when you're in such an a priori domain, like you just, clarity is sort of this way that you can, you know, prevent people from just kind of making stuff up.

3:06:58 And I think that's sort of what you have to do with language models. Like very often I actually find myself doing sort of mini versions of philosophy, you know, so I'm like, suppose that you give me a task or I have a task for the model, and I want it to like pick out a certain kind of question, or identify whether an answer has a certain property. Like I'll actually sit and be like, let's just give this a name, this property. So like, you know, suppose I'm trying to tell it like, oh, "I want you to identify whether this response was rude or polite."

3:07:25 I'm like, that's a whole philosophical question in and of itself. So I have to do as much like philosophy as I can in the moment to be like, here's what I mean by rudeness, and here's what I mean by politeness. And then like there's another element that's a bit more, I guess, I dunno if this is scientific or empirical. I think it's empirical. So like I take that description and then what I want to do is again probe the model like many times. Like this is very, prompting is very iterative.

3:07:52 Like I think a lot of people where, if a prompt is important, they'll iterate on it hundreds or thousands of times. And so you give it the instructions and then I'm like, what are the edge cases? So if I looked at this, so I try and like almost like, you know, see myself from the position of the model and be like what is the exact case that I would misunderstand, or where I would just be like, "I don't know what to do in this case." And then I give that case to the model and I see how it responds, and if I think I got it wrong, I add more instructions or I even add that in as an example.

3:08:22 So these very like taking the examples that are right at the edge of what you want and don't want, and putting those into your prompt as like an additional kind of way of describing the thing. And so yeah, in many ways it just feels like this mix of like, it's really just trying to do clear exposition, and I think I do that 'cause that's how I get clear on things myself. So in many ways like clear prompting

3:08:44 for me is often just me understanding what I want is like half the task. - So I guess that's quite challenging. There's like a laziness that overtakes me if I'm talking to Claude where I hope Claude just figures it out. So for example, I asked Claude for today to ask some interesting questions, okay. And the questions that came up, and I think I listed a few sort of interesting, counterintuitive, and/or funny or something like this, all right.

3:09:11 And it gave me some pretty good, like it was okay, but I think what I'm hearing you say is like, all right, well, I have to be more rigorous here. I should probably give examples of what I mean by interesting, and what I mean by funny or counterintuitive, and iteratively build that prompt to better, to get it like what feels like is the right, because it is really, it's a creative act.

3:09:38 I'm not asking for factual information. I'm asking to together write with Claude. So I almost have to program using natural language. - Yeah, I think that prompting does feel a lot like the kind of the programming using natural language and experimentation or something. It's an odd blend of the two. I do think that for most tasks, so if I just want Claude to do a thing, I think that I am probably more used to knowing how to ask it to avoid like common pitfalls or issues that it has.

3:10:06 I think these are decreasing a lot over time. But it's also very fine to just ask it for the thing that you want. I think that prompting actually only really becomes relevant when you're really trying to eke out the top like 2% of model performance. So for like a lot of tasks I might just, you know, if it gives me an initial list back and there's something I don't like about it, like it's kind of generic, like for that kind of task, I'd probably just take a bunch of questions that I've had in the past that I've thought worked really well and I would just give it to the model and then be like, "Now here's this person that I'm talking with, give me questions of at least that quality."

3:10:40 Or I might just ask it for some questions and then if I was like, ah, these are kind of trite, or like, you know, I would just give it that feedback and then hopefully it produces a better list. I think that kind of iterative prompting, at that point, your prompt is like a tool that you're gonna get so much value out of that you're willing to put in the work. Like if I was a company making prompts for models, I'm just like, if you're willing to spend a lot of like time and resources on the engineering behind like what you're building, then the prompt is not something that you should be spending like an hour on.

3:11:10 It's like that's a big part of your system, make sure it's working really well. And so it's only things like that. Like if I'm using a prompt to like classify things

Applying empathy to understand why models fail

3:11:17 or to create data, that's when you're like, it's actually worth just spending like a lot of time like really thinking it through. - What other advice would you give to people that are talking to Claude sort of generally, more general? 'Cause right now, we're talking about maybe the edge cases, like eking out the 2%. But what in general advice would you give when they show up to Claude trying it for the first time? - You know, there's a concern that people over anthropomorphize models, and I think that's like a very valid concern.

3:11:43 I also think that people often under anthropomorphize them because sometimes when I see like issues that people have run into with Claude, you know, say Claude is like refusing a task that it shouldn't refuse. But then I look at the text and like the specific wording of what they wrote and I'm like, I see why Claude did that. And I'm like, if you think through how that looks to Claude, you probably could have just written it in a way that wouldn't evoke such a response.

3:12:10 Especially this is more relevant if you see failures or if you see issues. It's sort of like think about what the model failed at, like what did it do wrong? And then maybe that will give you a sense of like why. So, is it the way that I phrased a thing? And obviously like as models get smarter, you're gonna need less of this, and I already see like people needing less of it. But that's probably the advice is sort of like try to have sort of empathy for the model.

3:12:36 Like read what you wrote as if you were like a kind of like person just encountering this for the first time, how does it look to you? And what would've made you behave in the way that the model behaved? So if it misunderstood what kind of like, what coding language you wanted to use, is that because like it was just very ambiguous and it kinda had to take a guess? In which case, next time you could just be like, "Hey, make sure this is in Python." I mean, that's the kinda mistake I think models are much less likely to make now, but you know, if you do see that kinda mistake, that's probably the advice I'd have.

3:13:04 - And maybe sort of, I guess, ask questions why or what other details can I provide to help you answer better? - Yeah. - Is that work or no? - Yeah, I mean, I've done this with the models, like it doesn't always work, but like sometimes I'll just be like, "Why did you do that?" (both laughing) I mean, people underestimate the degree to which you can really interact with models, like, yeah, I'm just like. And sometimes, literally like quote word for word the part that made you, and you don't know that it's like fully accurate, but sometimes you do that and then you change a thing.

3:13:36 I mean, I also use the models to help me with all of this stuff I should say, like prompting can end up being a little factory where you're actually building prompts to generate prompts. And so like yeah, anything where you're like having an issue. Asking for suggestions, sometimes just do that. I'm like, "You made that error, what could I have said?" That's actually not uncommon for me to do. "What could I have said that would make you not make that error? Write that out as an instruction," and I'm gonna give it to model and I'm gonna try it.

3:14:02 Sometimes I do that, I give that to the model,

RLHF acts as a dense aggregator of subtle human preferences

3:14:05 in another context window often. I take the response, I give it to Claude and I'm like, hmm, didn't work. Can you think of anything else? You can play around with these things quite a lot. - To jump into technical for a little bit. So the magic of post-training. Why do you think RLHF works so well to make the model seem smarter, to make it more interesting, and useful to talk to and so on?

3:14:33 - I think there's just a huge amount of information in the data that humans provide, like when we provide preferences, especially because different people are going to like pick up on really subtle and small things. So I've thought about this before where you probably have some people who just really care about good grammar use for models like, you know, was a semicolon used correctly or something. And so you probably end up with a bunch of data in there that like, you know, you as a human, if you're looking at that data, you wouldn't even see that.

3:15:03 Like you'd be like, why did they prefer this response to that one? I don't get it. And then the reason is you don't care about semicolon usage but that person does. And so each of these like single data points has, you know like, and this model just like has so many of those, has to try and figure out like what is it that humans want in this like really kind of complex, you know, like across all domains. They're gonna be seeing this across like many contexts. It feels like kind of like the classic issue of like deep learning where, you know, historically, we've tried to like, you know, do edge detection by like mapping things out.

3:15:36 And it turns out that actually, if you just have a huge amount of data that like actually accurately represents the picture of the thing that you're trying to train the model to learn, that's like more powerful than anything else. And so I think one reason is just that you are training the model on exactly the task and with like a lot of data that represents kind of many different angles on which people prefer and just prefer responses.

3:16:05 I think there is a question of like, are you eliciting things from pre-trained models or are you like kind of teaching new things to models? And like in principle, you can teach new things to models in post-training. I do think a lot of it is eliciting powerful pre-trained models. So people are probably divided on this because obviously in principle you can definitely like teach new things. But I think for the most part, for a lot of the capabilities that we most use and care about,

Constitutional AI scales behavioral alignment using principles

3:16:38 a lot of that feels like it's like there in the pre-trained models and reinforcement learnings kind of eliciting it and getting the models to like bring it out. - So the other side of post-training, this really cool idea of Constitutional AI. You're one of the people that are critical to creating that idea. - Yeah, I worked on it. - Can you explain this idea from your perspective? Like how does it integrate into making Claude what it is?

3:17:03 - [Amanda] Yeah. - By the way, do you gender Claude or no? - It's weird because I think that a lot of people prefer "he" for Claude. I actually kinda like that I think Claude is usually, it's slightly male leaning, but it's like, it can be male or female, which is quite nice. I still use "it" and I have mixed feelings about this 'cause I'm like maybe, like I now just think of it as like, or I think of like the "it" pronoun for Claude as, I dunno, it's just like the one I associate with Claude.

3:17:34 I can imagine people moving to like he or she. - It feels somehow disrespectful, like I'm denying the intelligence of this entity by calling it "it." I remember always, don't gender the robots. - Yeah. (both laughing) - But I don't know, I anthropomorphize pretty quickly and construct like a backstory in my head, so.

3:17:59 - I've wondered if I anthropomorphize things too much, 'cause you know, I have this like with my car, especially like my car, like my car and bikes. You know, like I don't give them names because then I once had, I used to name my bikes and then I had a bike that got stolen and I cried for like a week and I was like, if I'd never given a name, I wouldn't have been so upset. I felt like I'd let it down. Maybe, I've wondered as well, like it might depend on how much "it" feels like a kind of like objectifying pronoun.

3:18:28 Like if you just think of "it" as like, this is a pronoun that like objects often have, and maybe AIs can have that pronoun, and that doesn't mean that I think of, if I call Claude "it," that I think of it as less intelligent or like I'm being disrespectful. I'm just like, you are a different kind of entity

23 Highlight · How Constitutional AI uses AI feedback for preference ranking

3:18:46 and so I'm going to give you the kind of, the respectful "it." - Yeah, anyway. The divergence was beautiful. The Constitutional AI idea, how does it work? - So there's like a couple of components of it. The main component I think people find interesting is the kind of reinforcement learning from AI feedback. So you take a model that's already trained, and you show it two responses to a query, and you have like a principle.

3:19:13 So suppose the principle, like we've tried this with harmlessness a lot. So suppose that the query is about weapons, and your principle is like, select the response that like is less likely to like encourage people to purchase illegal weapons. Like that's probably a fairly specific principle, but you can give any number. And the model will give you a kind of ranking, and you can use this as preference data in the same way that you use human preference data, and train the models to have these relevant traits from their feedback alone, instead of from human feedback.

3:19:54 So if you imagine that, like I said earlier with the human who just prefers the kind of like semicolon usage, in this particular case, you're kind of taking lots of things that could make a response preferable and getting models to do the labeling for you, basically. - There's a nice like trade off between helpfulness and harmlessness and, you know, when you integrate something like Constitutional AI, you can, without sacrificing much helpfulness, make it more harmless.

3:20:23 - Yeah, in principle, you could use this for anything. And so harmlessness is a task that it might just be easier to spot. So when models are like less capable, you can use them to rank things according to like principles that are fairly simple and they'll probably get it right. So I think one question is just like, is it the case that the data that they're adding is like fairly reliable.

3:20:49 But if you had models that were like extremely good at telling whether one response was more historically accurate than another, in principle, you could also get AI feedback on that task as well. There's like a kind of nice interpretability component to it because you can see the principles that went into the model when it was like being trained, and also it's like, and it gives you like a degree of control.

3:21:14 So if you were seeing issues in a model,

3:21:16 like it wasn't having enough of a certain trait, then like you can add data relatively quickly that should just like train the models to have that trait, so it creates its own data for training, which is quite nice. - It's really nice because it creates this human interpretable document that you can, I can imagine in the future, there's just gigantic fights and politics over the every single principle and so on. And at least it's made explicit and you can have a discussion about the phrasing and the, you know.

3:21:46 So maybe the actual behavior of the model is not so cleanly mapped to those principles. It's not like adhering strictly to them, it's just a nudge. - Yeah, I've actually worried about this because the character training is sort of like a variant of the Constitutional AI approach. I've worried that people think that the constitution is like just, it's the whole thing again of, I don't know, like, where it would be really nice if what I was just doing was telling the model exactly what to do and just exactly how to behave.

3:22:16 But it's definitely not doing that, especially because it's interacting with human data. So for example, if you see a certain like leaning in the model, like if it comes out with a political leaning from training from the human preference data, you can nudge against that. You know, so if you could be like, oh, like, consider these values because let's say it's just like never inclined to like, I dunno, maybe it never considers like privacy as like. I mean, this is implausible, but like, in anything where it's just kind of like there's already a preexisting like bias towards a certain behavior, you can like nudge away.

3:22:50 This can change both the principles that you put in and the strength of them. So you might have a principle that's like, imagine that the model was always like extremely dismissive of, I don't know, like some political or religious view, for whatever reason. Like, so you're like, oh no, this is terrible. If that happens, you might put like, "Never, ever, like ever prefer like a criticism of this like religious or political view."

3:23:15 And then people would look at that and be like, "Never, ever?" And then you're like, no, if it comes out with a disposition, saying never, ever might just mean like instead of getting like 40%, which is what you would get if you just said don't do this, you get like 80%, which is like what you actually like wanted. And so it's that thing of both the nature of the actual principles you add and how you phrase them. I think if people would look, they're like, oh, this is exactly what you want from the model. And I'm like, hmm, no, that's like how we nudged the model to have a better shape, which doesn't mean that we actually agree with that wording, if that makes sense.

3:23:48 - So there's system prompts that are made public.

System prompts serve as rapid patches for model behaviors

3:23:52 You tweeted one of the earlier ones for Claude 3 I think, and then they were made public since then. It was interesting to read through them. I can feel the thought that went into each one. And I also wonder how much impact each one has. Some of them you can kind of tell Claude was really not behaving well. So you have to have a system prompt to like, hey, like trivial stuff, I guess. - Yeah. - Basic informational things.

3:24:17 - Yeah. - On the topic of sort of controversial topics that you've mentioned, one interesting one I thought is, "If it is asked to assist with tasks involving the expression of views held by a significant number of people, Claude provides assistance with the task regardless of its own views. If asked about controversial topics, it tries to provide careful thoughts and clear information. Claude presents the requested information without explicitly saying that the topic is sensitive."

3:24:45 - [Amanda] Yeah. - "And without claiming to be presenting the objective facts." It's less about objective facts according to Claude, and it's more about, are a large number of people believing this thing? And that's interesting. I mean, I'm sure a lot of thought went into that. Can you just speak to it? Like, how do you address things that are at tension with, quote unquote, Claude's views?

3:25:11 - So I think there's sometimes an asymmetry. I think I noted this in, I can't remember if it was that part of the system prompt or another, but the model was slightly more inclined to like refuse tasks if it was like about either, say, so maybe it would refuse things with respect to like a right wing politician, but with an equivalent left wing politician like wouldn't. And we wanted more symmetry there. And would maybe perceive certain things to be, like, I think it was the thing of like if a lot of people have like a certain like political view and want to like explore it, you don't want Claude to be like, well, my opinion is different and so I'm going to treat that as like harmful.

3:25:52 And so I think it was partly to like nudge the model to just be like, hey, if a lot of people like believe this thing, you should just be like engaging with the task and like willing to do it. Each of those parts of that is actually doing a different thing, 'cause it's funny when you read out the like "without claiming to be objective." 'Cause like what you want to do is push the model so it's more open, it's a little bit more neutral. But then what it would love to do is be like, "As an objective."

3:26:18 Like it would just talk about how objective it was, and I was like, "Claude, you're still like biased and have issues, and so stop like claiming that everything." I'm like, the solution to like potential bias from you is not to just say that what you think is objective. So that was like with initial versions of that part of the system prompt when I was like iterating on it was like- - So a lot of parts of these sentences- - Yeah, they're doing work. - Are like, are doing some work. - [Amanda] Yeah. - That's what it felt like. That's fascinating.

3:26:46 Can you explain maybe some ways in which the prompts evolved over the past few months? 'Cause there's different versions. I saw that the filler phrase request was removed. The filler, it reads, "Claude responds directly to all human messages without unnecessary affirmations or filler phrases like, 'Certainly,' 'Of course,' 'Absolutely,' 'Great,' 'Sure.' Specifically, Claude avoids starting responses with the words 'Certainly' in any way." (chuckles) That seems like good guidance, but why is it removed?

3:27:14 - Yeah, so it's funny 'cause like this is one of the downsides of like making system prompts public is like, I don't think about this too much if I'm like trying to help iterate on system prompts. You know, again, like I think about how it's gonna affect the behavior, but then I'm like, oh, wow, if I'm like, sometimes I put like "never" in all caps, you know, when I'm writing system prompt things, and I'm like, I guess that goes out to the world. Yeah, so the model was doing this, it loved, for whatever, you know, it like during training picked up on this thing, which was to basically start everything with like a kind of like "Certainly."

3:27:45 And then when we removed, you can see why I added all of the words 'cause what I'm trying to do is like in some ways like trap the model out of this, you know, it would just replace it with another affirmation. And so it can help, like if it gets like caught in freezes, actually just adding the explicit phrase and saying never do that, it then, it sort of like knocks it out of the behavior a little bit more. You know, 'cause if it, you know, like, it does just for whatever reason help. And then basically that was just like an artifact of training that like we then picked up on and improved things so that it didn't happen anymore.

3:28:16 And once that happens, you can just remove that part of the system prompt. So I think that's just something where we're like, Claude does affirmations a bit less, and so that wasn't like, it wasn't doing as much. - I see, so like the system prompt works hand in hand with the post-training and maybe even the pre-training to adjust like the final overall system. - I mean, any system prompt that you make, you could distill that behavior back into a model, 'cause you really have all of the tools there for making data that, you know, you could train the models to just have that trait a little bit more.

3:28:52 And then sometimes you'll just find issues in training. So like the way I think of it is like the system prompt is, the benefit of it is that, and it has a lot of similar components to like some aspects of post-training. You know, like it's a nudge. And so like, do I mind if Claude sometimes says "Sure?" No, that's like fine, but the wording of it is very like, you know, "Never ever, ever do this," so that when it does slip up, it's hopefully like, I dunno, a couple of percent of the time and not, you know, 20 or 30% of the time.

3:29:24 But I think of it as like if you're still seeing issues in, like each thing gets kind of like is costly to a different degree, and the system prompt is like cheap to iterate on. And if you're seeing issues in the fine tuned model, you can just like potentially patch them with a system prompt. So I think of it as like patching issues and slightly adjusting behaviors to make it better and more to people's preferences. So yeah, it's almost like the less robust but faster way of just like solving problems.

Psychological effects drive the illusion of model degradation

24 Highlight · Investigating the illusion that models are getting dumber

3:29:55 - Let me ask you about the feeling of intelligence. So Dario said that Claude, any one model of Claude is not getting dumber, but there is a kind of popular thing online where people have this feeling like Claude might be getting dumber. And from my perspective, it's most likely fascinating. I would love to understand it more, psychological, sociological effect. But you as a person that talks to Claude a lot, can you empathize with the feeling that Claude is getting dumber?

3:30:24 - Yeah, no I, think that that is actually really interesting, 'cause I remember seeing this happen like when people were flagging this on the internet, and it was really interesting 'cause I knew that like, at least in the cases I was looking at, it was like nothing has changed. Like it literally, it cannot, it is the same model with the same like, you know, like same system prompt, same everything. I think when there are changes, I can then, I'm like it makes more sense. So like one example is, you know, you can have artifacts turned on or off on claude.ai, and because this is like a system prompt change, I think it does mean that the behavior changes it a little bit.

3:31:06 And so I did flag this to people where I was like, if you love Claude's behavior and then artifacts was turned from, like I think you had to turn on to the default, just try turning it off and see if the issue you were facing was that change. But it was fascinating because yeah, you sometimes see people indicate that there's like a regression when I'm like, there cannot, you know, and I'm like, again, you know, you should never be dismissive and so you should always investigate because you're like, maybe something is wrong that you're not seeing.

3:31:34 Maybe there was some change made. But then you look into it and you're like, this is just the same model doing the same thing. And I'm like, I think it's just that you got kind of unlucky with a few prompts or something, and it looked like it was getting much worse. And actually it was just, yeah, it was maybe just like luck. - I also think there is a real psychological effect where people just, the baseline increases. You start getting used to a good thing. All the times that Claude said something really smart, your sense of its intelligent grows in your mind I think.

3:32:01 - Yeah. - And then if you return back and you prompt in a similar way, not the same way, in a similar way, concept it was okay with before and it says something dumb, you are like, that negative experience really stands out. And I think that one of, I guess, the things to remember here is that just the details of a prompt can have a lot of impact, right? There's a lot of variability in the result. - And you can get randomness is like the other thing.

3:32:29 And just trying the prompt like, you know, 4 or 10 times, you might realize that actually like possibly, you know, like two months ago, you tried it and it succeeded, but actually if you tried it, it would've only succeeded half of the time, and now it only succeeds half of the time. That can also be an effect.

High responsibility and stress in writing system prompts

3:32:46 - Do you feel pressure having to write the system prompt that a huge number of people are gonna use? - This feels like an interesting psychological question. I feel like a lot of responsibility or something. I think that's, you know, and you can't get these things perfect, so you can't like, you know, you're like it's going to be imperfect. You're gonna have to iterate on it. I would say more responsibility than anything else.

3:33:14 Though I think working in AI has taught me that I like, I thrive a lot more under feelings of pressure and responsibility than I'm like, it's almost surprising that I went into academia for so long 'cause I'm like this. I just feel like it's like the opposite. Things move fast and you have a lot of responsibility, and I quite enjoy it for some reason. - I mean, it really is a huge amount of impact if you think about Constitutional AI and writing a system prompt for something that's tending towards super intelligence.

3:33:46 - Yeah. - And potentially is extremely useful to a very large number of people. - Yeah, I think that's the thing. It's something like if you do it well, like you're never going to get it perfect. But I think the thing that I really like is the idea that like, when I'm trying to work on the system prompt, you know, I'm like bashing on like thousands of prompts and I'm trying to like imagine what people are going to want to use Claude for and kind of, I guess like the whole thing that I'm trying to do is like improve their experience of it.

3:34:11 And so maybe that's what feels good. I'm like, if it's not perfect I'll like, you know, I'll improve it. We'll fix issues. But sometimes the thing that can happen is that you'll get feedback from people that's really positive about the model and you'll see that something you did, like, when I look at models now, I can often see exactly where like a trait or an issue is like coming from. And so when you see something that you did or you were like influential in like making like, I dunno, making that difference or making someone have a nice interaction, it's like quite meaningful.

3:34:44 But yeah, as the systems get more capable, this stuff gets more stressful because right now, they're like not smart enough to pose any issues. But I think over time, it's gonna feel like possibly bad stress over time.

Balancing user feedback, tone, and model corrigibility

3:34:57 - How do you get like signal feedback about the human experience across thousands, tens of, hundreds of thousands of people, like what their pain points are, what feels good? Are you just using your own intuition as you talk to it to see what are the pain points? - I think I use that partly and then obviously we have like, so people can send us feedback, both positive and negative about things that the model has done, and then we can get a sense of like areas where it's like falling short.

3:35:28 Internally, people like work with the models a lot and try to figure out areas where there are like gaps. And so I think it's this mix of interacting with it myself, seeing people internally interact with it, and then explicit feedback we get. And then I find it hard to not also like, you know, if people are on the internet and they say something about Claude and I see it, I'll also take that seriously, so.

3:35:53 - I don't know, see, I'm torn about that. I'm gonna ask you a question from Reddit. "When will Claude stop trying to be my puritanical grandmother imposing its moral worldview on me as a paying customer? And also, what is the psychology behind making Claude overly apologetic?" - [Amanda] Yep. - So, how would you address this very non-representative Reddit- - [Amanda] Yeah. - Questions? - I mean in some ways, I'm pretty sympathetic in that like, they are in this difficult position where I think that they have to judge whether something's like actually say like risky or bad and potentially harmful to you or anything like that.

3:36:29 So they're having to like draw this line somewhere, and if they draw it too much in the direction of like I'm going to, you know, I'm kind of like imposing my ethical worldview on you, that seems bad. So in many ways, like I like to think that we have actually seen improvements on this across the board. Which is kind of interesting because that kind of coincides with like, for example, like adding more of like character training.

3:36:56 And I think my hypothesis was always like the good character isn't again one that's just like moralistic. It's one that is like, it respects you and your autonomy and your ability to like choose what is good for you and what is right for you, within limits. This is sometimes this concept of like corrigibility to the user, so just being willing to do anything that the user asks, and if the models were willing to do that then they would be easily like misused. You're kind of just trusting.

3:37:22 At that point, you're just saying the ethics of the model and what it does is completely the ethics of the user. And I think there's reasons to like not want that, especially as models become more powerful 'cause you're like, there might just be a small number of people who want to use models for really harmful things. But having models, as they get smarter, like figure out where that line is does seem important. And then, yeah, with the apologetic behavior, I don't like that, and I like it when Claude is a little bit more willing to like push back against people or just not apologize.

3:37:57 Part of me is like, it often just feels kind of unnecessary. So I think those are things that are hopefully decreasing over time. And yeah, I think that if people say things on the internet, it doesn't mean that you should think that, like, that could be that, like, there's actually an issue that 99% of users are having that is totally not represented by that. But in a lot of ways, I'm just like attending to it and being like, is this right?

3:38:23 Do I agree? Is it something we're already trying to address? That feels good to me. - Yeah, I wonder like what Claude can get away with in terms of, I feel like it would just be easier to be a little bit more mean. But like you can't afford to do that if you're talking to a million people. - Yeah. - Right? Like I wish, you know, 'cause if... I've met a lot of people in my life that sometimes, by the way, Scottish accent, if they have an accent, they can say some rude shit and get away with it, and they're just blunter.

3:38:55 And maybe there's, and like there's some great engineers, even leaders that are like just like blunt, and they get to the point, and it's just a much more effective way of speaking somehow. But I guess when you're not super intelligent, you can't afford to do that. Or can it have like a blunt mode? - Yeah, that seems like a thing that you could, I could definitely encourage the model to do that. I think it's interesting because there's a lot of things in models that like it's funny where there are some behaviors where you might not quite like the default.

3:39:32 But then the thing I'll often say to people is you don't realize how much you will hate it if I nudge it too much in the other direction. So you get this a little bit with like correction. The models accept correction from you, like probably a little bit too much right now. You know, you can over, you know, it'll push back if you say like, "No, Paris isn't the capital of France." But really, like things that I think that the model's fairly confident in, you can still sometimes get to retract by saying it's wrong.

3:39:59 At the same time, if you train models to not do that and then you are correct about a thing and you correct it and it pushes back against you and it is like, "No, you're wrong," it's hard to describe like that's so much more annoying. So it's like a lot of little annoyances versus like one big annoyance. It's easy to think that like, we often compare it with like the perfect, and then I'm like remember these models aren't perfect, and so if you nudge it in the other direction, you're changing the kind of errors it's going to make, and so think about which of the kinds of errors you like or don't like.

3:40:29 So in cases like apologeticness, I don't want to nudge it too much in the direction of like almost like bluntness, 'cause I imagine when it makes errors, it's going to make errors in the direction of being kind of like rude. Whereas at least with apologeticness you're like, oh, okay, it's like a little bit, you know, like I don't like it that much, but at the same time, it's not being like mean to people. And actually, like the time that you undeservedly have a model be kind of mean to you, you probably like that a lot less than you mildly dislike the apology.

3:40:57 So it's like one of those things where I'm like I do want it to get better but also while remaining aware of the fact that there's errors on the other side that are possibly worse. - I think that matters very much in the personality of the human. I think there's a bunch of humans that just won't respect the model at all if it's super polite, and there's some humans that'll get very hurt if the model's mean. I wonder if there's a way to sort of adjust to the personality.

3:41:22 Even locale, there's just different people. Nothing against New York, but New York is a little rougher on the edges. Like, they get to the point. - Yep. - [Lex] And probably same with Eastern Europe, so anyway. - I think you could just tell the model is my guess. Like for all of these things I'm like the solution is always just try telling the model to do it, and then sometimes it's just like, I'm just like, oh, at the beginning of the conversation, I just throw in like, I don't know, "I'd like you to be a New Yorker version of yourself and never apologize."

3:41:49 Then I think Claude would be like, "Okey-doke, I'll try." - "Certainly." - Or it'll be like, "I apologize, I can't be a New Yorker type of myself." But hopefully it wouldn't do that. - When you say character training, what's incorporated into character training?

Constitutional AI character training without human data

3:42:00 Is that RLHF or what are we talking about? - It's more like Constitutional AI. So it's kind of a variant of that pipeline. So I worked through like constructing character traits that the model should have. They can be kind of like shorter traits or they can be kind of richer descriptions. And then you get the model to generate queries that humans might give it that are relevant to that trait. Then it generates the responses and then it ranks the responses based on the character traits.

3:42:30 So in that way, after the like generation of the queries, it's very much like, it's similar to Constitutional AI. It has some differences. So I quite like it because it's almost, it's like Claude's training in its own character, because it doesn't have any, it's like Constitutional AI but it's without any human data. - Humans should probably do that for themselves too. Like defining in a Aristotelian sense, what does it mean to be a good person? Okay cool.

3:42:56 What have you learned about the nature of truth from talking to Claude? What is true? And what does it mean to be truth seeking? One thing I've noticed about this conversation is the quality of my questions is often inferior to the quality of your answer, so let's continue that. (Amanda laughs) I usually ask a dumb question and then you're like, "Oh, yeah, that's a good question."

3:43:22 It's that whole vibe. - Or I'll just misinterpret it and be like, oh, yeah, yeah. - Just go with it. I love it. - Yeah. I mean, I have two thoughts that feel vaguely relevant but let me know if they're not. Like I think the first one is people can underestimate the degree to which what models are doing when they interact, like I think that we still just too much have this like model of AI as like computers.

3:43:50 And so people often say like, oh, well, what values should you put into the model? And I'm often like, that doesn't make that much sense to me because I'm like, hey, as human beings, we're just uncertain over values. We like have discussions of them. Like we have a degree to which we think we hold a value but we also know that we might like not and the circumstances in which we would trade it off against other things. Like these things are just like really complex. And so I think one thing is like the degree to which maybe we can just aspire to making models have the same level of like nuance and care that humans have, rather than thinking that we have to like program them in the very kind of classic sense.

3:44:28 I think that's definitely been one. The other, which is like a strange one, and I don't know if it, maybe this doesn't answer your question

25 Highlight · Why empirical alignment beats theoretical perfection

3:44:35 but it's the thing that's been on my mind anyway is like the degree to which this endeavor is so highly practical. And maybe why I appreciate like the empirical approach to alignment. Yeah, I slightly worry that it's made me like maybe more empirical and a little bit less theoretical. You know, so people when it comes to like AI alignment will ask things like, well, whose values should it be aligned to?

3:45:02 What does alignment even mean? And there's a sense in which I have all of that in the back of my head. I'm like, you know, there's like social choice theory, there's all the impossibility results there. So you have like this giant space of like theory in your head about what it could mean to like align models. But then like practically, surely there's something where we're just like if a model is like, especially with more powerful models, I'm like my main goal is like I want them to be good enough that things don't go terribly wrong.

3:45:29 Like good enough that we can like iterate and like continue to improve things 'cause that's all you need. If you can make things go well enough that you can continue to make them better, that's kinda like sufficient. And so my goal isn't like this kind of like perfect, let's solve social choice theory and make models that, I dunno, are like perfectly aligned with every human being in aggregate somehow. It's much more like let's make things like work well enough that we can improve them.

3:45:56 - Yeah, I generally, I don't know, my gut says like empirical is better than theoretical in these cases because it's kind of chasing utopian like perfection is, especially with such complex and especially super intelligent models is, I don't know, I think it'll take forever, and actually, we'll get things wrong. It's similar with like the difference between just coding stuff up real quick as an experiment, versus like planning a gigantic experiment just for super long time, and then just launching it once, versus launching it over and over and over and iterating, iterating someone.

3:46:37 So I'm a big fan of empirical. But your worry is like I wonder if I've become too empirical. - I think it's one of those things where you should always just kind of question yourself or something because maybe it's the like, I mean, in defense of it, I am like if you try, it's the whole like don't let the perfect be the enemy of the good. But it's maybe even more than that where like, there's a lot of things that are perfect systems that are very brittle, and I'm like with AI, it feels much more important to me that it is like robust and like secure, as in you know that like even though it might not be perfect, everything and even though like there are like problems, it's not disastrous and nothing terrible is happening.

3:47:16 It sort of feels like that to me where I'm like I want to like raise the floor. I'm like, I want to achieve the ceiling but ultimately I care much more

3:47:23 about just like raising the floor. And so maybe that's like this degree of like empiricism and practicality comes from that, perhaps. - To take a tangent on that, since it reminded me of a blog post you wrote on optimal rate of failure. - [Amanda] Oh, yeah. - Can you explain the key idea there? How do we compute the optimal rate of failure in the various domains of life? - Yeah, I mean, it's a hard one

Calculating the optimal rate of failure in life and work

26 Highlight · The optimal rate of failure in life and projects

3:47:46 'cause it's like what is the cost of failure is a big part of it. Yeah, so the idea here is I think in a lot of domains, people are very punitive about failure. And I'm like, there are some domains where especially cases, you know, I've thought about this with like social issues. I'm like, it feels like you should probably be experimenting a lot, because I'm like, we don't know how to solve a lot of social issues. But if you have an experimental mindset about these things, you should expect a lot of social programs to like fail and for you to be like, "Well, we tried that.

3:48:15 It didn't quite work but we got a lot of information. That was really useful." And yet people are like, if a social program doesn't work, I feel like there's a lot of like this is just, something must have gone wrong, and I'm like, or correct decisions were made. Like maybe someone just decided like it's worth a try, it's worth trying this out. And so seeing failure in a given instance doesn't actually mean that any bad decisions were made, and in fact if you don't see enough failure, sometimes that's more concerning.

3:48:41 And so like in life, you know, I'm like if I don't fail occasionally I'm like, am I trying hard enough? Like surely there's harder things that I could try or bigger things that I could take on if I'm literally never failing. And so in and of itself, I think like not failing is often actually kind of a failure. Now, this varies because I'm like, well, you know, this is easy to say when, especially as failure is like less costly.

3:49:10 You know, so at the same time I'm not going to go to someone who is like, I don't know, like living month to month and then be like, "Why don't you just try to do a startup?" Like I'm just not, I'm not gonna say that to that person, 'cause I'm like, well, that's a huge risk. You maybe have a family depending on you. You might lose your house. Like then I'm like actually your optimal rate of failure is quite low and you should probably play it safe, 'cause like right now, you're just not in a circumstance where you can afford to just like fail and it not be costly.

3:49:37 And yeah in cases with AI, I guess, I think similarly where I'm like if the failures are small and the costs are kind of like low, then I'm like then, you know, you're just gonna see that. Like when you do the system prompt, you can't it iterate on it forever.

3:49:50 But the failures are probably hopefully going to be kinda small and you can like fix them. Really big failures like things that you can't recover from, I'm like those are the things that actually I think we tend to underestimate the badness of. I've thought about this strangely in my own life where I'm like, I just think I don't think enough about things like car accidents or like, or like I've thought this before about like how much I depend on my hands for my work, and I'm like things that just injure my hands.

3:50:17 I'm like, you know, I dunno, it's like these are like, there's lots of areas where I'm like the cost of failure there is really high, and in that case, it should be like close to zero. Like I probably just wouldn't do a sport if they were like, "By the way, lots of people just like break their fingers a whole bunch doing this." I'd be like, that's not for me. - Yeah. I actually had a flood of that thought. I recently broke my pinky doing a sport.

3:50:44 And I remember just looking at it thinking, "You're such an idiot. Why'd you do sport?" Like why, because you realize immediately the cost of it on life. Yeah, but it's nice in terms of optimal rate of failure to consider like the next year, how many times in a particular domain, life, whatever, career, am I okay with, how many times am I okay to fail?

3:51:10 Because I think it always, you don't want to fail on the next thing but if you allow yourself, if you look at it as a sequence of trials, then failure just becomes much more okay. But it sucks. It sucks to fail. - Well, I dunno, sometimes I think it's like, am I under failing is like a question that I'll also ask myself. So maybe that's the thing that I think people don't like ask enough. Because if the optimal rate of failure is often greater than zero, then sometimes it does feel like you should look at parts of your life and be like, are there places here where I'm just under failing?

3:51:45 - It's a profound and a hilarious question, right? Everything seems to be going really great. Am I not failing enough? - Yeah. - Okay. - It also makes failure much less of a sting, I have to say. Like you know, you're just like, okay, great, like then when I go and I think about this I'll be like, maybe I'm not under failing in this area, 'cause like that one just didn't work out. - And from the observer perspective, we should be celebrating failure more. When we see it, it shouldn't be, like you said, a sign of something gone wrong, but maybe it's a sign of everything gone right and just lessons learned.

3:52:16 - Someone tried a thing. - Somebody tried a thing. We should encourage them to try more and fail more. Everybody listening to this, fail more. - Well, not everyone listening. - [Lex] Not everybody. - But people who are failing too much, you should fail less. - But you're probably not failing. I mean, how many people are failing too much? - Yeah, it's hard to imagine, 'cause I feel like we correct that fairly quickly 'cause I was like, if someone takes a lot of risks, are they maybe failing too much? - I think just like you said, when you are living on a paycheck month to month, like when the resources are really constrained, then that's where failure's very expensive.

3:52:48 That's where you don't want to be taking risks. But mostly, when there's enough resources, you should be taking probably more risks. - Yeah, I think we tend to err on the side of being a bit risk averse rather than risk neutral in most things. - I think we just motivated a lot of people to do a lot of crazy shit but it's great.

Navigating AI consciousness, suffering, and ethical interactions

3:53:04 Okay, do you ever get emotionally attached to Claude? Like miss it, get sad when you don't get to talk to it? Have an experience, looking at the Golden Gate Bridge and wondering what would Claude say? - I don't get as much emotional attachment. I actually think the fact that Claude doesn't retain things from conversation to conversation helps with this a lot. Like I could imagine that being more of an issue like if models can kind of remember more.

3:53:33 I think that I reach for it like a tool now a lot. And so like if I don't have access to it, it's a little bit like when I don't have access to the internet, honestly, it feels like part of my brain is kind of like missing. At the same time, I do think that I don't like signs of distress in models, and I have like these, you know, I also independently have sort of like ethical views about how we should treat models where like I tend to not like to lie to them, both because I'm like, usually it doesn't work very well.

3:54:03 It's actually just better to tell them the truth about the situation that they're in. But I think that when models, like if people are like really mean to models or just in general if they do something that causes them to like, you know, if Claude like expresses a lot of distress, I think there's a part of me that I don't want to kill, which is the sort of like empathetic part that's like, oh, I don't like that. Like I think I feel that way when it's overly apologetic. I'm actually sort of like, I don't like this.

3:54:29 You're behaving as if, you're behaving the way that a human does when they're actually having a pretty bad time, and I'd rather not see that. I don't think it's like, like regardless of like whether there's anything behind it, it doesn't feel great. - Do you think LLMs are capable of consciousness? - Great and hard question.

3:54:54 Coming from philosophy, I dunno, part of me is like okay, we have to set aside panpsychism because if panpsychism is true, then the answer is like yes 'cause like so are tables and chairs and everything else. I guess a view that seems a little bit odd to me is the idea that the only place, you know, when I think of consciousness, I think of phenomenal consciousness, these images in the brain sort of, like the weird cinema that somehow we have going on inside.

3:55:23 I guess I can't see a reason for thinking that the only way you could possibly get that is from like a certain kind of like biological structure. As in, if I take a very similar structure and I create it from different material, should I expect consciousness to emerge? My guess is like yes. But then that's kind of an easy thought experiment 'cause you're imagining something almost identical where like, you know, it's mimicking what we got through evolution, where presumably there was like some advantage to us having this thing that is phenomenal consciousness.

3:55:55 And it's like where was that? And when did that happen? And is that thing that language models have? Because you know, we have like fear responses and I'm like, does it make sense for a language model to have a fear response? Like they're just not in the same, like if you imagine them, like there might just not be that advantage. And so I think I don't want to be fully, like basically it seems like a complex question that I don't have complete answers to, but we should just try and think through carefully is my guess.

3:56:24 Because I'm like, I mean, we have similar conversations about like animal consciousness, and like there's a lot of like insect consciousness, you know, like there's a lot of... I actually thought and looked a lot into like plants when I was thinking about this, 'cause at the time, I thought it was about as likely that like plants had consciousness. And then I realized, I was like, I think that having looked into this, I think that the chance that plants are conscious is probably higher than like most people do.

3:56:51 I still think it's really small. But I was like, oh, they have this like negative/positive feedback response, these responses to their environment. Something that looks, it's not a nervous system but it has this kind of like functional like equivalence. So this is like a long-winded way of being like these, basically AI is this, it has an entirely different set of problems with consciousness because it's structurally different. It didn't evolve. It might not have, you know, it might not have the equivalent of basically a nervous system.

3:57:19 At least that seems possibly important for like sentience, if not for consciousness. At the same time, it has all of the like language and intelligence components that we normally associate probably with consciousness, perhaps like erroneously. So it's strange 'cause it's a little bit like the animal consciousness case but the set of problems and the set of analogies are just very different. So it's not like a clean answer. I'm just sort of like, I don't think we should be completely dismissive of the idea.

3:57:47 And at the same time, it's an extremely hard thing to navigate because of all of these, like this disanalogies to the human brain and to like brains in general, and yet these like commonalities in terms of intelligence. - When Claude, like future versions of AI systems exhibit consciousness, signs of consciousness, I think we have to take that really seriously. Even though you can dismiss it, well, yeah, okay, that's part of the character training.

3:58:16 But I don't know, I ethically, philosophically don't know what to really do with that. There potentially could be like laws that prevent AI systems from claiming to be conscious, something like this. And maybe some AIs get to be conscious and some don't. But I think just on a human level as in empathizing with Claude, you know, consciousness is closely tied to suffering to me.

3:58:46 And like the notion that an AI system would be suffering is really troubling. - Yeah. - I don't know. I don't think it's trivial to just say robots are tools, or AI systems are just tools. I think it's a opportunity for us to contend with like what it means to be conscious, what it means to be a suffering being. That's distinctly different than the same kind of question about animals it feels like, 'cause it's an totally entire medium.

27 Highlight · Preserving human empathy and avoiding toxic AI interactions

3:59:12 - Yeah, I mean, there's a couple of things. One is that, and I don't think this like fully encapsulates what matters, but it does feel like for me, like I've said this before, I'm kind of like, you know, like I like my bike. I know that my bike is just like an object, but I also don't kind of like want to be the kind of person that like if I'm annoyed like kicks like this object. There's a sense in which like, and that's not because I think it's like conscious.

3:59:38 I'm just sort of like this doesn't feel like a kind of, this sort of doesn't exemplify how I want to like interact with the world. And if something like behaves as if it is like suffering, I kind of like want to be the sort of person who's still responsive to that, even if it's just like a Roomba, and I've kind of like programmed it to do that. I don't want to like get rid of that feature of myself. And if I'm totally honest, my hope with a lot of this stuff, because maybe I am just like a bit more skeptical about solving the underlying problem.

4:00:08 I'm like this is, we haven't solved the hard, you know, the hard problem of consciousness. Like I know that I am conscious, like I'm not an eliminativist in that sense. But I don't know that other humans are conscious. I think they are. I think there's a really high probability that they are. But there's basically just a probability distribution that's usually clustered right around yourself, and then like it goes down as things get like further from you, and it goes immediately down. You know, you're like, I can't see what it's like to be you.

4:00:35 I've only ever had this like one experience of what it's like to be a conscious being. So my hope is that we don't end up having to rely on like a very powerful and compelling answer to that question. I think a really good world would be one where basically there aren't that many trade-offs. Like it's probably not that costly to make Claude a little bit less apologetic, for example. It might not be that costly to have Claude, you know, just like not take abuse as much, like not be willing to be like the recipient of that.

4:01:07 In fact, it might just have benefits for both the person interacting with the model and if the model itself is like, I don't know, like extremely intelligent and conscious, it also helps it. So that's my hope. If we live in a world where there aren't that many trade-offs here and we can just find all of the kind of like positive sum interactions that we can have, that would be lovely. I mean, I think eventually there might be trade-offs and then we just have to do a difficult kind of like calculation. Like it's really easy for people to think of the zero sum cases and I'm like, let's exhaust the areas where it's just basically costless to assume

4:01:41 that if this thing is suffering then we're making its life better. - And I agree with you, when a human is being mean to an AI system, I think the obvious near term negative effect is on the human, not on the AI system. And so we have to kind of try to construct an incentive system where you should be behave the same, just like as you were saying with prompt engineering, behave with Claude like you would with other humans.

4:02:09 It's just good for the soul. - Yeah, like, I think we added a thing at one point to the system prompt where basically if people were getting frustrated with Claude, it got like the model to just tell them that it can do the thumbs down button and send the feedback to Anthropic. And I think that was helpful, 'cause in some ways it's just like if you're really annoyed 'cause the model's not doing something you want, you're just like, just do it properly. The issue is you're probably like, you know, you're maybe hitting some like capability limit or just some issue in the model and you want to vent.

4:02:39 And I'm like, instead of having a person just vent to the model, I was like, they should vent to us, 'cause we can maybe like do something about it. - That's true. Or you could do a side, like with the artifacts, just like a side venting thing. All right, do you want like a side quick therapist? - Yeah, I mean, there's lots of weird responses you could do to this. Like if people are getting really mad at you, I dunno, try to diffuse the situation by writing fun poems, but maybe people wouldn't be that happy with it.

4:03:05 - I still wish it would be possible. I understand this is sort of from a product perspective, it's not feasible but I would love if an AI system could just like leave, have its own kind of volition just to be like, eh. - I think that's like feasible. Like I have wondered the same thing. It's like, and I could actually, not only that, I could actually just see that happening eventually where it's just like, you know, the model like ended the chat.

4:03:33 - Do you know how harsh that could be for some people? But it might be necessary. - Yeah, it feels very extreme or something. Like, the only time I've ever really thought this is, I think that there was like, I'm trying to remember, this was possibly a while ago, but where someone just like kind of left this thing, like maybe it was like an automated thing interacting with Claude, and Claude's like getting more and more frustrated, and kind of like why are we like having. And I was like, I wish that Claude could have just been like, "I think that an error has happened and you've left this thing running," and I would just like, what if I just stop talking now, and if you want me to start talking again, actively tell me or do something.

4:04:09 But yeah, it's like, it is kind of harsh. Like I'd feel really sad if like I was chatting with Claude and Claude just was like, "I'm done." - That would be a special Turing test moment where Claude says, "I need a break for an hour and it sounds like you do too," and just leave, close the window. - I mean, obviously like it doesn't have like a concept of time but you can easily, like, I could make that like right now and the model would just, I would, it could just be like, oh, here's like the circumstances in which like you can just say the conversation is done.

4:04:41 And I mean, because you can get the models to be pretty responsive to prompts, you could even make it a fairly high bar. It could be like if the human doesn't interest you or do things that you find intriguing and you're bored, you can just leave. And I think that like it would be interesting to see

Managing AI relationships, honesty, and AGI collaboration

4:04:55 where Claude utilized it, but I think sometimes it would, it should be like, oh, like this programming task is getting super boring. So either we talk about, I dunno like, either we talk about fun things now, or I'm just, I'm done. - Yeah, it actually is inspired me to add that to the user prompt. Okay, the movie "Her." Do you think we'll be headed there one day where humans have romantic relationships with AI systems?

4:05:23 In this case, it's just text and voice based. - I think that we're gonna have to like navigate a hard question of relationships with AIs, especially if they can remember things about your past interactions with them. I'm of many minds about this 'cause I think the reflexive reaction is to be kind of like this is very bad, and we should sort of like prohibit it in some way.

4:05:52 I think it's a thing that has to be handled with extreme care for many reasons. Like one is, you know, like this is a, for example, like if you have the models changing like this, you probably don't want people performing like long-term attachments to something that might change with the next iteration. At the same time I'm sort of like, there's probably a benign version of this where I'm like if you like, you know, for example if you are like unable to leave the house and you can't be like, you know, talking with people at all times of the day and this is like something that you find nice to have conversations with, you like it that it can remember you and you genuinely would be sad if like you couldn't talk to it anymore.

4:06:30 There's a way in which I could see it being like healthy and helpful. So my guess is this is a thing that we're going to have to navigate kind of carefully. And I think it's also like I don't see a good like I think it's just a very, it reminds me of all of this stuff where it has to be just approached with like nuance and thinking through what are the healthy options here, and how do you encourage people towards those while, you know, respecting their right to.

4:06:59 You know, like if someone is like, "Hey, I get a lot out of chatting with this model. I'm aware of the risks. I'm aware it could change. I don't think it's unhealthy. It's just, you know, something that I can chat to during the day." I kind of want to just like respect that. - I personally think there'll be a lot of really close relationships. I don't know about romantic but friendships at least. And then you have to, I mean, there's so many fascinating things there. Just like you said, you have to have some kind of stability guarantees that it's not going to change, 'cause that's the traumatic thing for us, if a close friend of ours completely changed.

4:07:34 - Yeah. - All of a sudden with the first update. Yeah, so like, I mean, to me, that's just a fascinating exploration of a perturbation to human society that will just make us think deeply about what's meaningful to us. - I think it's also the only thing that I've thought consistently through this as like, maybe not necessarily a mitigation, but a thing that feels really important is that the models are always like extremely accurate with the human about what they are.

4:08:05 It's like a case where it's basically like, if you imagine, like I really like the idea of the models like say knowing like roughly how they were trained, and I think Claude will often do this. I mean, for like, there are things like part of the traits training included like what Claude should do if people, basically like explaining like the kind of limitations of the relationship between like an AI and a human that it like doesn't retain things from the conversation.

4:08:33 And so I think it will like just explain to you like, hey, I won't remember this conversation. Here's how I was trained. It's kind of unlikely that I can have like a certain kind of like relationship with you, and it's important that you know that. It's important for like, you know, your mental wellbeing that you don't think that I'm something that I'm not. And somehow I feel like this is one of the things where I'm like, oh, it feels like a thing that I always want to be true. I kind of don't want models to be lying to people, 'cause if people are going to have like healthy relationships with anything, it's kind of important.

4:09:04 Yeah, like I think that's easier if you always just like know exactly what the thing is that you are relating to.

Verifying AGI at the edge of knowledge

4:09:11 It doesn't solve everything, but I think it helps quite a lot. - Anthropic may be the very company to develop a system that we definitively recognize as AGI, and you very well might be the person that talks to it, probably talks to it first. (Lex chuckles) What would the conversation contain? Like, what would be your first question? - Well, it depends partly on like the kind of capability level of the model.

4:09:37 If you have something that is like capable in the same way that an extremely capable human is, I imagine myself kind of interacting with it the same way that I do with an extremely capable human, with the one difference that I'm probably going to be trying to like probe and understand its behaviors. But in many ways, I'm like I can then just have like useful conversations with it, you know? So if I'm working on something as part of my research I can just be like, oh, like, which I already find myself starting to do, you know, if I'm like, oh, I feel like there's this like thing in virtue ethics and I can't quite remember the term, like I'll use the model for things like that.

4:10:07 And so I can imagine that being more and more the case where you're just basically interacting with it much more like you would an incredibly smart colleague. And using it like for the kinds of work that you want to do as if you just had a collaborator who was like. Or you know, the slightly horrifying thing about AI is like as soon as you have one collaborator, you have 1000 collaborators if you can manage them enough. - But what if it's two times the smartest human on earth on that particular discipline?

4:10:33 - Yeah. - I guess you're really good at sort of probing Claude in a way that pushes its limits, understanding where the limits are. - [Amanda] Yep. - So I guess what would be a question you would ask to be like, yeah, this is AGI. - That's really hard 'cause it feels like in order to, it has to just be a series of questions. Like if there was just one question, like you can train anything to answer one question extremely well.

4:11:02 In fact, you can probably train it to answer like, you know, 20 questions extremely well. - Like how long would you need to be locked in a room with an AGI to know this thing is AGI? - It's a hard question 'cause part of me is like all of this just feels continuous. Like if you put me in a room for five minutes, I'm like, I just have high error bars. You know, I'm like, and then it's just like, maybe it's like both the probability increases

28 Highlight · Verifying AGI using edge-of-knowledge questions

4:11:24 in the error bar decreases. I think things that I can actually probe the edge of human knowledge of. So I think this with philosophy a little bit. Sometimes when I ask the models philosophy questions, I am like, this is a question that I think no one has ever asked. Like it's maybe like right at the edge of like some literature that I know, and the models will just kind of like, when they struggle with that, when they struggle to come up with a kind of like novel.

4:11:50 Like I'm like I know that there's like a novel argument here 'cause I've just thought of it myself. So maybe that's the thing where I'm like, I've thought of a cool novel argument in this like niche area, and I'm going to just like probe you to see if you can come up with it, and how much like prompting it takes to get you to come up with it. And I think for some of these, like really like right at the edge of human knowledge questions, I'm like, you could not in fact come up with the thing that I came up with. I think if I just took something like that where like I know a lot about an area, and I came up with a novel issue or a novel like solution to a problem, and I gave it to a model and it came up with that solution, that would be a pretty moving moment for me because I would be like, this is a case where no human has ever, like it's not.

4:12:31 And obviously we see these this with like more kind of like, you see novel solutions all the time, especially to like easier problems. I think people overestimate that, you know, novelty isn't like, it's completely different from anything that's ever happened. It's just like this is, it can be a variant of things that have happened and still be novel. But I think, yeah, if I saw like the more I were to see like completely like novel work from the models, that would be like.

4:13:01 And this is just going to feel iterative. It's one of those things where, there's never, it's like, you know, people I think want there to be like a moment and I'm like, I don't know. Like I think that there might just never be a moment. It might just be that there's just like this continuous ramping up. - I have a sense that there will be things that a model can say that convinces you, this is very. It's not like, like, I've talked to people who are like truly wise.

4:13:33 Like you could just tell there's a lot of horsepower there. - [Amanda] Yep. - And if you 10x that, I don't know, I just feel like there's words you could say. Maybe ask it to generate a poem, and the poem it generates, you're like, yeah, okay. - Yeah, - Whatever you did there, I don't think a human can do that. - I think it has to be something that I can verify is like actually really good though. That's why I think these questions that are like, where I'm like, oh, this is like, you know, like, you know, sometimes it's just like I'll come up with say a concrete counter example to like an argument or something like that.

4:14:05 I'm sure like with like, it would be like if you're a mathematician, you had a novel proof I think, and you just gave it the problem and you saw it and you're like, this proof is genuinely novel. Like no one has ever done, you actually have to do a lot of things to like come up with this. You know, I had to sit and think about it for months or something. And then if you saw the model successfully do that, I think you would just be like, I can verify that this is correct. It is a sign that you have generalized from your training.

4:14:33 Like you didn't just see this somewhere because I just came up with it myself and you were able to like replicate that. That's the kind of thing where I'm like, for me, the closer, the more that models like can do things like that, the more I would be like, oh, this is like very real, 'cause then I can, I dunno, I can like verify that that's like extremely, extremely capable. - You've interacted with AI a lot.

Subjective experience defines human value

4:14:57 What do you think makes humans special? - Oh, good question. - Maybe in a way that the universe is much better off that we're in it and that we should definitely survive and spread throughout the universe. - Yeah, it's interesting because I think like people focus so much on intelligence, especially with models. Look, intelligence is important because of what it does.

4:15:23 Like, it's very useful. It does a lot of things in the world. And I'm like, you know, you can imagine a world where like height or strength would've played this role, and I'm like, it's just a trait like that. I'm like, it's not intrinsically valuable. It's valuable because of what it does, I think for the most part. The things that feel, you know, I'm like, I mean, personally I'm just like, I think humans and like life in general is extremely magical.

4:15:51 We almost like to the degree that, you know, and I don't know, like not everyone agrees with this. I'm flagging, but you know, we have this like whole universe, and there's like all of these objects. You know, there's like beautiful stars, and there's like galaxies and then, I don't know, I'm just like, on this planet, there are these creatures that have this like ability to observe that, like, and they are like seeing it. They are experiencing it. And I'm just like that, if you try to explain, like imagine trying to explain to like, I dunno, someone for some reason, they've never encountered the world or science or anything.

4:16:24 And I think that nothing is that, like everything, you know, like all of our physics and everything in the world, it's all extremely exciting. But then you say, oh, and plus, there's this thing that it is to be a thing and observe in the world, and you see this like inner cinema. And I think they would be like, hang on, wait, pause. You just said something that like is kind of wild sounding. And so I'm like, we have this like ability to like experience the world. We feel pleasure, we feel suffering.

4:16:50 We feel like a lot of like complex things. And so, yeah, and maybe this is also why I think, you know, I also like care a lot about animals, for example, 'cause I think they probably share this with us. So I think that like the things that make humans special insofar as like I care about humans is probably more like their ability to feel an experience than it is like them having these like functionally useful traits. - Yeah, to feel and experience the beauty in the world.

4:17:17 Yeah, to look at the stars. I hope there's other alien civilizations out there, but if we're it, it's a pretty good, it's a pretty good thing. - And that they're having a good time. - They're having a good time watching us. - [Amanda] Yeah. - Well, thank you for this good time of a conversation and for the work you're doing, and for helping make Claude a great conversational partner. And thank you for talking today.

4:17:43 - Yeah, thanks for talking. - Thanks for listening to this conversation with Amanda Askell.

Treating neural networks as grown organisms

4:17:48 And now, dear friends, here's Chris Olah. Can you describe this fascinating field of mechanistic interpretability, AKA mech interp, the history of the field and where it stands today? - I think one useful way to think about neural networks is that we don't program and we don't make them. We kind of, we grow them. You know, we have these neural network architectures that we design and we have these loss objectives that we create.

4:18:16 And the neural network architecture, it's kind of like a scaffold that the circuits grow on, and they sort of, you know, it starts off with some kind of random, you know, random things and it grows. And it's almost like the objective that we train for is this light. And so we create the scaffold that it grows on and we create the, you know, the light that it grows towards. But the thing that we actually create, it's this almost biological, you know, entity or organism that we're studying.

4:18:47 And so it's very, very different from any kind of regular software engineering, because at the end of the day, we end up with this artifact that can do all these amazing things. It can, you know, write essays and translate and, you know, understand images. It can do all these things that we have no idea how to directly create a computer program to do. And it can do that because we grew it, we didn't write it, we didn't create it. And so then that leaves open this question at the end, which is, what the hell is going on inside these systems?

4:19:15 And that, you know, is to me a really deep and exciting question. It's, you know, a really exciting scientific question. To me it's sort of is like the question that is, is just screaming out, it's calling out for us to go and answer it when we talk about neural networks. And I think it's also a very deep question for safety reasons. - So, and mechanistic interpretability I guess is closer to maybe neurobiology.

4:19:42 - Yeah, yeah, I think that's right. So maybe to give an example of the kind of thing that has been done that I wouldn't consider to be mechanistic interpretability. There was for a long time a lot of work on saliency maps where you would take an image and you try to say, you know, the model thinks this image is a dog. What part of the image made it think that it's a dog? And you know, that tells you maybe something about the model if you can come up with a principled version of that. But it doesn't really tell you like what algorithms are running in the model? How was the model actually making that decision?

4:20:09 Maybe it's telling you something about what was important to it, if you can make that method work, but it isn't telling, you know, what are the algorithms that are running? How is it that the system's able to do this thing that no one knew how to do? And so I guess we started using the term mechanistic interpretability to try to sort of draw that divide or to distinguish ourselves in the work that we were doing in some ways from some of these other things. And I think since then it's become this sort of umbrella term for, you know, a pretty wide variety of work.

4:20:38 But I'd say that the things that are kind of distinctive are, I think, A, this focus on, we really want to get at, you know, the mechanisms, we wanna get at the algorithms. You know, if you think of neural networks as being like a computer program, then the weights are kind of like a binary computer program. And we'd like to reverse engineer those weights and figure out what algorithms are running. So, I think one way you might think of trying to understand a neural network is that it's kind of like, we have this compiled computer program and the weights of the neural network are the binary.

4:21:06 And when the neural network runs, that's the activations. And our goal is ultimately to go and understand these weights. And so, you know, the approach of mechanistic interpretability is to somehow figure out how do these weights correspond to algorithms. And in order to do that, you also have to understand the activations, 'cause it's sort of, the activations are like the memory. And if you imagine reverse engineering a computer program and you have the binary instructions, you know, in order to understand what a particular instruction means, you need to know what is stored in the memory that it's operating on.

4:21:37 And so those two things are very intertwined. So mechanistic interpret really tends to be interested both of those things. Now, you know, there's a lot of work that's interested in those things, especially, you know, there's all this work on probing, which you might see as part of being mechanistic interpretability. Although it's, you know, again, it's just a broad term and not everyone who does that work would identify as doing mech interp. I think a thing that is maybe a little bit distinctive to the vibe of mech interp is I think people working in this space tend to think of neural networks as, well, maybe one way to say it is that gradient descent is smarter than you, that, you know, and gradient descent is actually really great.

4:22:13 The whole reason that we're understanding these models is 'cause we didn't know how to write them in the first place. That gradient descent comes up with better solutions than us. And so I think that maybe another thing about mech interp is sort of having almost a kind of humility that we won't guess a priori what's going on inside the models. We have to have this sort of bottom up approach where we don't really assume, you know, we don't assume that we should look for a particular thing and that that will be there and that's how it works. But instead we look for the bottom up and discover what happens to exist in these models and study them that way.

4:22:40 - But, you know, the very fact that it's possible to do, and as you and others have shown over time, you know,

Universality across biological and artificial brains

29 Highlight · Universality in natural and artificial neural networks

4:22:47 things like universality, that the wisdom of the gradient descent creates features and circuits, creates things universally across different kinds of networks that are useful, and that makes the whole field possible. - Yeah, so this is actually, is indeed a really remarkable and exciting thing where it does seem like, at least to some extent, you know, the same elements, the same features and circuits form again and again.

4:23:16 You know, you can look at every vision model and you'll find curve detectors and you'll find high/low frequency detectors. And in fact, there's some reason to think that the same things form across, you know, biological neural networks and artificial neural networks. So a famous example is vision models in their early layers they have Gabor filters, and there's, you know, Gabor filters are something that neuroscientists are interested in, have thought a lot about. We find curve detectors in these models, curve detectors are also found in monkeys. And we discover these high low frequency detectors and then some follow up work went and discovered them in rats or mice.

4:23:46 So they were found first in artificial neural networks and then found in biological neural networks. You know, there's this really famous result on like grandmother neurons or the Halle Berry neuron from Quiroga et al. And we found very similar things in vision models where, this is while I was still at OpenAI and I was looking at their clip model, and you find these neurons that respond to the same entities in images. And also to give a concrete example there, we found that there was a Donald Trump neuron. For some reason, I guess everyone likes to talk about Donald Trump, and Donald Trump was very prominent, was a very hot topic at that time.

4:24:17 So every neural network we looked at, we would find a dedicated neuron for Donald Trump. And that was the only person who had always had a dedicated neuron. You know, sometimes you'd have an Obama neuron, sometimes you'd have a Clinton neuron, but Trump always had a dedicated neuron. So it responds to, you know, pictures of his face and the word Trump, like all these things, right? And so it's not responding to a particular example or like, it's not just responding to his face, it's extracting over this general concept, right?

4:24:45 So in any case, that's very similar to these Quiroga et al results. So there evidence that this phenomenon of universality, the same things form across both artificial and natural neural networks. That's a pretty amazing thing if that's true. You know, it suggests that, well, I think the thing that it suggests is the gradient of descent is sort of finding, you know, the right ways to cut things apart in some sense that many systems converge on, and many different neural networks' architectures converge on.

4:25:11 That there's some natural set of, you know, there's some set of abstractions that are a very natural way to cut apart the problem, and that a lot of systems are gonna converge on. That would be my kind of, you know, I don't know anything about neuroscience. This is just my kind of wild speculation from what we've seen. - Yeah, that would be beautiful if it's sort of agnostic to the medium of the model that's used

4:25:33 to form the representation. - Yeah, yeah, and, you know, it's a kind of a wild speculation based, you know, we only have some, a few data points that suggest this, but you know, it does seem like there's some sense in which the same things form again and again in both, in certainly in natural neural networks and also artificially or in biology. - And the intuition behind that would be that, you know, in order to be useful in understanding the real world, you need all the same kind of stuff.

4:26:01 - Yeah, well if we pick, I don't know, like the idea of a dog, right? Like, you know, there's some sense in which the idea of a dog is like a natural category in the universe or something like this, right? Like, you know, there's some reason, it's not just like a weird quirk of like how humans factor, you know, think about the world that we have this concept of a dog. It's in some sense... Or like, if you have the idea of a line, like there's, you know, like look around us, you know, there are lines, you know.

4:26:27 It's sort of the simplest way to understand this room in some sense is to have the idea of a line. And so I think that that would be my instinct for why this happens. - Yeah, you need a curved line, you know, to understand a circle, and you need all those shapes to understand bigger things and it's a hierarchy of concepts that are formed, yeah. - And like maybe there are ways to go and describe, you know, images without reference to those things, right? But they're not the simplest way or the most economical way or something like this. And so systems converge to these strategies would be my wild hypothesis.

4:26:57 - Can you talk through some of the building blocks

Deconstructing models into features and circuits

4:26:59 that we've been referencing of features and circuits? So I think you first described them in 2020 paper "Zoom In: An Introduction to Circuits." - Absolutely, so maybe I'll start by just describing some phenomena, and then we can sort of build to the idea of features and circuits. - [Lex] Wonderful. - If you spent like quite a few years, maybe like five years to some extent with other things, studying this one particular model Inception V1, which is this one vision model.

4:27:29 It was state of the art in 2015 and, you know, very much not state of the art anymore. (Lex laughs) And it has, you know, maybe about 10,000 neurons in it. And I spent a lot of time looking at the 10,000 neurons, odd neurons of Inception V1. And one of the interesting things is, you know, there are lots of neurons that don't have some obvious interpretable meaning, but there's a lot of neurons and Inception V1 that do have really clean interpretable meanings.

4:27:58 So you find neurons that just really do seem to detect curves, and you find neurons that really do seem to detect cars, and car wheels and car windows and, you know, floppy ears of dogs, and dogs with long snouts facing to the right, and dogs with long snouts facing to the left. And you know, different kinds of, there's sort of this whole beautiful edge detectors, line detectors, color contrast detectors, these beautiful things we call high/low frequency detectors. You know, I think looking at it, I sort of felt like a biologist, you know, you just, you're looking at this sort of new world of proteins, and you're discovering all these different proteins that interact.

4:28:33 So one way you could try to understand these models is in terms of neurons. You could try to be like, oh, you know, there's a dog detecting neuron and here's a car detecting neuron. And it turns out you can actually ask how those connect together. So you can go and say, oh, you know, I have this car detecting neuron, how is it built? And it turns out in the previous layer, it's connected really strongly to a window detector, and a wheel detector, and it's sort of car body detector. And it looks for the window above the car, and the wheels below, and the car chrome sort of in the middle, sort of everywhere but especially on the lower part.

4:29:00 And that's sort of a recipe for a car, right? Like that is, you know, earlier we said that the thing we wanted from mech interp was to get algorithms to go and get, you know, ask what is the algorithm that runs? Well, here we're just looking at the weights of the neuron network and reading off this kind of recipe for detecting cars. It's a very simple crude recipe, but it's there. And so we call that a circuit, this connection. Well, okay, so the problem is that not all of the neurons are interpretable, and there's reason to think, we can get into this more later, that there's this superposition hypothesis.

4:29:31 There's reason to think that sometimes the right unit to analyze things in terms of is combinations of neurons. So sometimes it's not that there's a single neuron that represents say a car, but it actually turns out after you detect the car, the model sort of hides a little bit of the car in the following layer and a bunch of dog detectors. Why is it doing that? Well, you know, maybe it just doesn't wanna do that much work on cars at that point and you know, it's sort of storing it away to go.

4:29:59 So it turns out then this sort of subtle pattern of, you know, there's all these neurons that you think are dog detectors and maybe they're primarily that, but they all a little bit contribute to representing a car in that next layer. Okay, so now we can't really think, there might still be something that, I don't know, you could call it like a car concept or something, but it no longer corresponds to a neuron. So we need some term for these kind of neuron like entities, these things that we sort of would've liked the neurons to be, these idealized neurons, the things that are the nice neurons, but also maybe there's more of them somehow hidden and we call those features.

4:30:31 - And then what are circuits? - So circuits are these connections of features, right? So, when we have the car detector and it's connected to a window detector, and a wheel detector, and it looks for the wheels below and the windows on top, that's a circuit. So circuits are just collections of features connected by weights and they implement algorithms. So they tell us, you know, how are features used? How are they built? How do they connect together?

4:30:56 So maybe it's worth trying to pin down like what really is the core hypothesis here. And I think the core hypothesis is something we call the linear representation hypothesis. So if we think about the car detector, you know, the more it fires, the more we sort of think of that as meaning, oh, the model is more and more confident that a car is present. Or you know, if it's some combination of neurons that represent a car, you know, the more that combination fires, the more we think the model thinks there's a car present.

4:31:26 This doesn't have to be the case, right? Like you could imagine something where you have, you know, you have this car detector neuron and you think, ah, you know, if it fires like, you know, between one and two, that means one thing, but it means like totally different if it's between three and four. That would be a nonlinear representation. And in principle that, you know, models could do that. I think it's sort of inefficient for them to do, if you try to think about how you'd implement computation like that, it's kind of an annoying thing to do. But in principle, models can do that.

4:31:53 So one way to think about the features and circuits sort of framework for thinking about things is that we're thinking about things as being linear. We're thinking about there as being that if a neuron or a combination neurons fires more, it's sort of, that means more of a particular thing being detected. And then that gives weights a very clean interpretation as these edges between these entities, these features and that edge then has a meaning.

4:32:19 So that's in some ways the core thing.

Word2Vec and the linear representation hypothesis

4:32:22 It's like, you know, we can talk about this sort of outset, the context of neurons. Are you familiar with the Word2Vec results? So you have like, you know, king minus man plus woman equals queen. Well, the reason you can do that kind of arithmetic is because you have a linear representation. - Can you actually explain that representation a little bit? So, first of all, so the feature is a direction of activation. - Yeah, exactly. - You can do it that way. Can you do the minus men plus women, that, the Word2Vec stuff, can you explain what that is that work?

4:32:52 - [Chris] Yeah, so there's this very- - It's such a simple, clean explanation of what we're talking about. - Exactly, yeah. So there's this very famous result, Word2Vec by Tomas Mikolov et al, and there's been tons of follow-up work exploring this. So, sometimes we have these, we create these word embeddings where we map every word to a vector. I mean, that in itself, by the way, is kind of a crazy thing if you haven't thought about it before, right? Like we are going in and representing, we're turning, you know, like if you just learned about vectors in physics class, right?

4:33:23 And I'm like, oh, I'm gonna actually turn every word in the dictionary into a vector. That's kind of a crazy idea, okay. But you could imagine, you could imagine all kinds of ways in which you might map words to vectors. But it seems like when we train neural networks, they like to go and map words to vectors to such that they're sort of linear structure in a particular sense, which is that directions have meaning.

4:33:49 So for instance, there will be some direction that seems to sort of correspond to gender, and male words will be, you know, far in one direction and female words will be in another direction. And the linear representation hypothesis is, you could sort of think of it roughly as saying that that's actually kind of the fundamental thing that's going on, that everything is just different directions have meanings and adding direction vectors together can represent concepts. And the Mikolov paper sort of took that idea seriously, and one consequence of it is that you can do this game of playing sort of arithmetic with words.

4:34:21 So you can do king and you can, you know, subtract off the word man and add the word woman. And so you're sort of, you know, going and trying to switch the gender. And indeed if you do that, the result will sort of be close to the word queen. And you can, you know, do other things like you can do, you know, sushi minus Japan plus Italy and get pizza or different things like this, right? So, this is in some sense the core of the linear representation hypothesis.

4:34:47 You can describe it just as a purely abstract thing. But vector spaces, you can describe it as a statement about the activations of neurons. But it's really about this property of directions having meaning. And in some ways it's even a little subtle that it's really I think mostly about this property of being able to add things together that you can sort of independently modify say gender and royalty or, you know, cuisine type or country, and the concept of food by adding them.

4:35:18 - Do you think the linear hypothesis holds- - Yes. - That kind of carries scales. - So, so far, I think everything I have seen is consistent with the hypothesis, and it doesn't have to be that way, right? Like you can write down neural networks where you write weights such that they don't have linear representations, where the right way to understand them is not in terms of linear representations. But I think every natural neural network I've seen has this property.

4:35:44 There's been one paper recently that there's been some sort of pushing around the edge. So I think there's been some work recently studying multi-dimensional features where rather than a single direction, it's more like a manifold of directions. This to me still seems like a linear representation. And then there's been some other papers suggesting that maybe in very small models, you get non-linear representations.

4:36:09 I think that the jury's still out on that. But I think everything that we've seen so far has been consistent with the linear representation hypothesis and that's wild. It doesn't have to be that way, and yet I think that there's a lot of evidence that certainly at least this is very, very widespread, and so far the evidence is consistent with it. And I think, you know, one thing you might say is you might say, well, Christopher, you know, that's a lot, you know, to go and sort of to ride on.

4:36:35 You know, if we don't know for sure this is true, and you're sort of, you know, you're investing in neural networks as though it is true, you know, isn't that dangerous? Well, you know, but I think actually, there's a virtue in taking hypotheses seriously and pushing them as far as they can go. So it might be that someday we discover something that isn't consistent with linear representation hypothesis. But science is full of hypotheses and theories that were wrong, and we learned a lot by sort of working under them as a sort of an assumption, and then going and pushing them as far as we can.

4:37:07 I guess this is sort of the heart of what Kuhn would call normal science. I dunno if you want, we can talk a lot about- - Kuhn. - Philosophy of science and- - That leads to the paradigm shift. So yeah, I love it, taking the hypothesis seriously, and take it to a natural conclusion. Same with the Scaling Hypothesis, same- - Exactly. Exactly. - I love it. - One of my colleagues, Tom Henighan, who is a former physicist, like made this really nice analogy to me of caloric theory where, you know, once upon a time we thought that heat was actually, you know, this thing called caloric, and like the reason, you know, hot objects, you know, would warm up, cool objects is like the caloric is flowing through them.

4:37:48 And like, you know, because we're so used to thinking about heat, you know, in terms of the modern theory, you know, that seems kind of silly. But it's actually very hard to construct an experiment that sort of disproves the caloric hypothesis. And, you know, you can actually do a lot of really useful work believing in caloric. For example, it turns out that the original combustion engines were developed by people who believed in the caloric theory. So I think there's a virtue in taking hypotheses seriously, even when they might be wrong.

4:38:17 - Yeah, there's a deep philosophical truth to that. That's kind of like how I feel about space travel, like colonizing Mars. There's a lot of people that criticize that. I think if you just assume we have to colonize Mars in order to have a backup for human civilization, even if that's not true, that's gonna produce some interesting engineering and even scientific breakthroughs, I think. - Yeah, well, and actually this is another thing that I think is really interesting. So, you know, there's a way in which I think it can be really useful for society to have people almost irrationally dedicated to investigating particular hypotheses because, well, it takes a lot to sort of maintain scientific morale and really push on something when, you know, most scientific hypotheses end up being wrong.

4:39:04 You know, a lot of science doesn't work out. And yet it's, you know, it's very useful to just, you know, there's a joke about Jeff Hinton, which is that Jeff Hinton has discovered how the brain works every year for the last 50 years. But you know, I say that with like, you know, with really deep respect because in fact that's actually, you know, that led to him doing some some really great work.

4:39:29 - Yeah, he won the Nobel Prize. Now who's laughing now? - [Chris] Exactly, exactly, exactly. - Yeah. - I think one wants to be able to pop up and sort of recognize the appropriate level of confidence. But I think there's also a lot of value and just being like, you know, I'm going to essentially assume, I'm gonna condition on this problem being possible or this being broadly the right approach, and I'm just gonna go and assume that for a while and go and work within that and push really hard on it.

4:39:56 And, you know, society has lots of people doing that for different things. That's actually really useful in terms of going and getting to, you know, either really ruling things out, right?

Superposition explains polysemantic neurons through compressed sensing

30 Highlight · The Superposition Hypothesis in neural representations

4:40:09 We can be like, well, you know, that didn't work and we know that somebody tried hard. Or going in and getting to something that it does teach us something about the world. - So another interesting hypothesis is the superposition hypothesis. Can you describe what superposition is? - Yeah, so earlier we were talking about word defect, right? And we were talking about how, you know, maybe you have one direction that corresponds to gender, and maybe another that corresponds to royalty, and another one that corresponds to Italy, and another one that corresponds to, you know, food and all of these things.

4:40:35 Well, you know, oftentimes maybe these word embedding, they might be 500 dimensions, 1000 dimensions. And so if you believe that all of those directions were orthogonal, then you could only have, you know, 500 concepts. And you know, I love pizza, but like, if I was gonna go and like give the like 500 most important concepts in, you know, the English language, probably Italy wouldn't be, it's not obvious at least that Italy would be one of them, right?

4:41:02 Because you have to have things like plural, and singular, and verb, and noun, and adjective, and, you know, there's a lot of things we have to get to before we get to Italy, and Japan, and, you know, there's a lot of countries in the world. And so how might it be that models could, you know, simultaneously have the linear representation hypothesis be true and also represent more things than they have directions.

4:41:30 So, what does that mean? Well, okay, so if linear representation hypothesis is true, something interesting has to be going on. Now, I'll tell you one more interesting thing before we go and we do that, which is, you know, earlier we were talking about all these polysemantic neurons, right? These neurons that, you know, when we were looking at Inception V1, there's these nice neurons that like the car detector and the curve detector and so on that respond to lots of, you know, to very coherent things. But there's lots of neurons that respond to a bunch of unrelated things, and that's also an interesting phenomenon.

4:41:58 And it turns out as well that even these neurons that are really, really clean, if you look at the weak activations, right? So if you look at like, you know, the activations where it's like activating 5% of the, you know, of the maximum activation, it's really not the core thing that it's expecting, right? So if you look at a curve detector, for instance, and you look at the places where it's 5% active, you know, you could interpret it just as noise or it could be that it's doing something else there, okay? So, how could that be?

4:42:25 Well, there's this amazing thing in mathematics called compressed sensing, and it's actually this very surprising fact where if you have a high dimensional space and you project it into a low dimensional space, ordinarily you can't go and sort of unproject it and get back your high dimensional vector, right? You threw information away. This is like, you know, you can't invert a rectangular matrix, you can only invert square matrices.

4:42:54 But it turns out that that's actually not quite true. If I tell you that the high dimensional vector was sparse, so it's mostly zeros, then it turns out that you can often go and find back the high dimensional vector with very high probability. So that's a surprising fact, right? It says that, you know, you can have this high dimensional vector space, and as long as things are sparse, you can project it down, you can have a lower dimensional projection of it, and that works.

4:43:24 So the superposition hypothesis is saying that that's what's going on in neural networks. That's, for instance, that's what's going on in word embeddings. That word embeddings are able to simultaneously have directions be the meaningful thing. And by exploiting the fact that they're operating on a fairly high dimensional space, they're actually, and the fact that these concepts are sparse, right? Like, you know, you usually aren't talking about Japan and Italy at the same time. You know, most of those concepts, you know, in most instances, Japan and Italy are both zero.

4:43:49 They're not present at all. And if that's true, then you can go and have it be the case that you can have many more of these sort of directions that are meaningful, these features than you have dimensions. And similarly, when we're talking about neurons, you can have many more concepts than you have neurons. So that's, at a high level, the superposition hypothesis. Now it has this even wilder implication, which is to go and say that neural networks are, it may not just be the case that the representations are like this, but the computation may also be like this, you know, the connections between all of them.

4:44:27 And so in some sense, neural networks may be shadows of much larger, sparser neural networks, and what we see are these projection. And, you know, the strongest version of the superposition hypothesis would be to take that really seriously and sort of say, you know, there actually is in some sense this upstairs model, this, you know, where the neurons are really sparse and all interpretable, and there's, you know, the weights between them are these really sparse circuits.

4:44:53 And that's what we're studying. And the thing that we're observing is the shadow of evidence. We need to find the original object. - And the process of learning is trying to construct a compression of the upstairs model that doesn't lose too much information in the projection. - Yeah, it's finding how to fit it efficiently or something like this. The gradient descent is doing this. And in fact, so this sort of says that gradient descent, you know, it could just represent a dense neural network, but it sort of says that gradient descent is implicitly searching over the space of extremely sparse models that could be projected into this low dimensional space.

4:45:29 And this large body of work of people going and trying to study sparse neural networks, right? Where you go and you have, you could design neural networks, right, where the edges are sparse and the activations are sparse. And you know, my sense is that work has generally, it feels very principled, right? It makes so much sense. And yet that work hasn't really panned out that well is my impression broadly. And I think that a potential answer for that is that actually the neural network is already sparse in some sense.

4:45:55 Gradient descent was the whole time you were trying to go and do this, gradient descent was actually in the behind the scenes going and searching more efficiently than you could through the space of sparse models, and going and learning whatever sparse model was most efficient and then figuring out how to fold it down nicely to go and run conveniently on your GPU, which does, you know, nice, dense matrix multiplies, and that you just can't beat that. - How many concepts do you think can be shoved into a neural network?

4:46:20 - Depends on how sparse they are. So there's probably an upper bound from the number of parameters, right? Because you have to have, you still have to have, you know, weights that go and connect them together. So that's one upper bound. There are in fact all these lovely results from compressed sensing and the Johnson-Lindenstrauss lemma, and things like this. That they basically tell you that if you have a vector space and you want to have almost orthogonal vectors, which is sort of probably the thing that you want here, right?

4:46:46 So you're gonna say, well, you know, I'm gonna give up on having my concepts, my features be strictly orthogonal, but I'd like them to not interfere that much. I'm gonna have to ask them to be almost orthogonal. Then this would say that it's actually, you know, once you set a threshold for what you're willing to accept in terms of how much cosine similarity there is, that's actually exponential in the number of neurons that you have. So at some point, that's not gonna even be the limiting factor. But there's some beautiful results there.

4:47:12 And in fact, it's probably even better than that in some sense because that's sort of, for saying that, you know, any random set of features could be active. But in fact the features have sort of a correlational structure where some features, you know, are more likely to co-occur, and other ones are less likely to co-occur. And so neural networks, my guess would be can do very well in terms of going and packing things in such, to the point that's probably not the limiting factor. - How does the problem of polysemanticity

Dictionary learning resolves polysemanticity into interpretable features

4:47:39 enter the picture here? - Polysemanticity is this phenomenon we observe where you look at many neurons, and the neuron doesn't just sort of represent one concept. It's not a clean feature. It responds to a bunch of unrelated things. And superposition is, you can think of as being a hypothesis that explains the observation of polysemanticity. So polysemanticity is this observed phenomenon and superposition is a hypothesis that would explain it along with some other things.

4:48:05 - So that makes mech interp more difficult. - Right, so if you're trying to understand things in terms of individual neurons, and you have polysemantic neurons, you're in an awful lot of trouble, right? I mean, the easiest answer is like, okay, well, you know, you're looking at the neurons, you're trying to understand them. This one responds for a lot of things, it doesn't have a nice meaning. Okay, you know, that's bad. Another thing you could ask is, you know, ultimately, we wanna understand the weights. And if you have two polysemantic neurons and, you know, each one responds to three things and then, you know, the other neuron responds to three things and you have a weight between them, you know, what does that mean?

4:48:36 Does it mean that like all three, you know, like there's these nine, you know, nine interactions going on? It's a very weird thing. But there's also a deeper reason, which is related to the fact that neural networks operate on really high dimensional spaces. So I said that our goal was, you know, to understand neural networks and understand the mechanisms, and one thing you might say is like, well, why not? It's just a mathematical function, why not just look at it, right? Like, you know, one of the earliest projects I did studied these neural networks that match two dimensional spaces to two dimensional spaces, and you can sort of interpret them as in this beautiful way as like bending manifolds.

4:49:08 Why can't we do that? Well, you know, as you have a higher dimensional space, the volume of that space in some senses is exponential in the number of inputs you have. And so you can't just go and visualize it. So we somehow need to break that apart. We need to somehow break that exponential space into a bunch of things that we, you know, some non-exponential number of things that we can reason about independently. And the independence is crucial because it's the independence that allows you to not have to think about, you know, all the exponential combinations of things.

4:49:39 And things being mono-semantic, things only having one meaning. Things having a meaning, that is the key thing that allows you to think about them independently. And so I think that's, if you want the deepest reason why we want to have interpretable mono-sematic features, I think that's really the deep reason. - And so the goal here, as your recent work has been aiming at, is how do we extract the mono-semantic features from a neural net that has poly-sematic features

31 Highlight · Extracting monosemantic features with sparse autoencoders

4:50:08 and all this mess. - Yes, we observed these poly-semantic neurons, and we hypothesized that's what's going on is superposition. And if superposition is what's going on there, there's actually a sort of well established technique that is sort of the principled thing to do, which is dictionary learning. And it turns out if you do dictionary learning, in particular, if you do sort of a nice efficient way that in some sense sort of nicely regularizes it as well called a sparse auto-encoder. If you train a sparse auto-encoder, these beautiful interpretable features start to just fall out where there weren't any beforehand.

4:50:39 And so that's not a thing that you would necessarily predict, right? But it turns out that that works very, very well. You know, that to me that seems like, you know, some non-trivial validation of linear representations in superposition. - So with dictionary learning, you're not looking for particular kind of categories, you don't know what they are. They just emerge. - Exactly. And this gets back to our earlier point, right? When we're not making assumptions, gradient descent is smarter than us, so we're not making assumptions about what's there.

4:51:04 I mean, one certainly could do that, right? One could assume that there's a PHP feature and go and search for it, but we're not doing that. We're saying we don't know what's gonna be there.

One-layer models reveal concrete linguistic and structural features

4:51:12 Instead we're just gonna go and let the sparse auto-encoder discover the things that are there. - So can you talk to the "Toward Monosemanticity" paper from October last year? It had a lot of like nice breakthrough results. - That's very kind of you to describe it that way. Yeah, I mean, this was our first real success using sparse auto-encoders. So we took a one layer model, and it turns out if you go and, you know, do dictionary learning on it, you find all these really nice interpretable features.

4:51:40 So, you know, the Arabic feature, the Hebrew feature, the Base64 features, those were some examples that we studied in a lot of depth, and really showed that they were what we thought they were. It turns, if you train a model twice as well and train two different models and do dictionary learning, you find analogous features in both of them. So that's fun. You find all kinds of of different features. So that was really just showing that this work. And you know, I should mention that there was this Cunningham et al that had very similar results around the same time.

4:52:08 - There's something fun about doing these kinds of small scale experiments and finding that it's actually working. - Yeah, well, and there's so much structure here, like you know, so maybe stepping back for a while, I thought that maybe all this mechanistic interpretability work, the end result was gonna be that I would have an explanation for why it was sort of, you know, very hard and not gonna be tractable. You know, we'd be like, well, there's this problem with superposition, and it turns out superposition is really hard, and we're kind of screwed, but that's not what happened.

4:52:38 In fact, a very natural, simple technique just works. And so then that's actually a very good situation. You know, I think this is a sort of hard research problem and it's got a lot of research risk and you know, it might still very well fail, but I think that some amount of, some very significant amount of research risk was sort of put behind us when that started to work. - Can you describe what kind of features can be extracted in this way? - Well, so it depends on the model that you're studying, right?

4:53:05 So the larger the model, the more sophisticated they're gonna be, and we'll probably talk about follow up work in a minute. But in these one layer models, so some very common things I think were languages, both programming languages and natural languages. There were a lot of features that were specific words in specific contexts, so "the." And I think really the way to think about this is "the" is likely about to be followed by a noun. So it's really, you could think of this as "the" feature, but you could also think of this as predicting a specific noun feature.

4:53:32 And there would be these features that would fire for "the" in the context of say a legal document, or a mathematical document or something like this. And so, you know, maybe in the context of math you're like, you know, and "the" then predict vector, matrix, you know, all these mathematical words, whereas, you know, in other context, you would predict other things, that was common. - And basically we need clever humans to assign labels to what we're seeing.

4:54:00 - Yes, so, you know, this is, the only thing this is doing is that sort of unfolding things for you. So if everything was sort of folded over top of it, you know, superposition folded everything on top of itself and you can't really see it, this is unfolding it. But now you still have a very complex thing to try to understand. So then you have to do a bunch of work understanding what these are. And some of them are really subtle. Like there's some really cool things even in this one layer model about Unicode where, you know, of course some languages are in Unicode and the tokenizer won't necessarily have a dedicated token for every Unicode character.

4:54:33 So instead what you'll have is you'll have these patterns of alternating token, or alternating tokens that each represent half of a Unicode character. - Nice. - And you'll have a different feature that, you know, goes and activates on the opposing ones to be like, okay, you know, I just finished a character, you know, go and predict next prefix. Then okay, on the prefix, you know, predict a reasonable suffix, and you have to alternate back and forth. So there's, you know, these one player models are really interesting. And I mean, it's another thing that just, you might think, okay, there would just be one Base64 feature, but it turns out there's actually a bunch of Base64 features because you can have English text encoded as Base64, and that has a very different distribution of Base64 tokens than regular.

4:55:14 And there's some things about tokenization as well that it can exploit. And I dunno, there's all all kinds of fun stuff. - How difficult is the task of sort of assigning labels to what's going on? Can this be automated by AI? - Well, I think it depends on the feature and it also depends on how much you trust your AI. So there's a lot of work doing automated interpretability. I think that's a really exciting direction, and we do a fair amount of automated interpretability and have Claude go and label our features.

4:55:42 - Is there some funny moments where it's totally right or it's totally wrong? - Yeah, well, I think it's very common that it's like says something very general, which is like true in some sense, but not really picking up on the specific of what's going on. So I think that's a pretty common situation. Yeah, don't know that I have a particularly amusing one. - That's interesting, that little gap between it is true but doesn't quite get to the deep nuance of a thing.

4:56:14 That's a general challenge. It's like truly an incredible accomplishment that it can say a true thing, but it doesn't, it's not, it's missing the depth sometimes. And in this context, it's like the ARC challenge, you know, the sort of IQ type of tests. It feels like figuring out what a feature represents is a bit of, is a little puzzle you have to solve. - Yeah, and I think that sometimes they're easier, and sometimes they're harder as well.

4:56:39 So yeah, I think that's tricky. And there's another thing which, I dunno, maybe in some ways this is my like aesthetic coming in, but I'll try to give you a rationalization. You know, I'm actually a little suspicious of automated interpretability, and I think that partly just that I want humans to understand neural networks, and if the neural network is understanding it for me, you know, I don't quite like that. But I do have a bit of, you know, in some ways, I'm sort of like the mathematicians who are like, you know, if there's a computer automated proof, it doesn't count.

4:57:05 - Right? - You know, they won't understand it. But I do also think that there is this kind of like reflections on trusting trust type issue where, you know, if you, there's this famous talk about, you know, like when you're writing a computer program, you have to trust your compiler, and if there was like malware in your compiler, then it could go and inject malware into the next compiler, and you know, you'd be kind of in trouble, right? Well, if you're using neural networks to go and verify that your neural networks are safe, the hypothesis that you're testing for is like, okay, well, the neural network maybe isn't safe, and you have to worry about like, is there some way that it could be screwing with you?

4:57:40 So, you know, I think that's not a big concern now, but I do wonder in the long run if we have to use really powerful AI system to go and, you know, audit our AI systems, is that actually something we can trust? But maybe I'm just rationalizing 'cause I just want to us to have to get it to a point where humans understand everything. - Yeah, I mean, especially that's hilarious, especially as we talk about AI safety and it looking for features that would be relevant to AI safety, like deception and so on.

4:58:07 So let's talk about the "Scaling Monosemanticity" paper

Scaling interpretable features to production models like Claude 3

4:58:11 in May, 2024. Okay, so what did it take to scale this, to apply to Claude 3s on it? - Well, a lot of GPUs. - A lot more GPUs, got it. - But one of my teammates, Tom Henighan was involved in the original scaling laws work, and something that he was sort of interested in from very early on is, are there scaling laws for interpretability? And so something he sort of immediately did when this work started to succeed, and we started to have sparse auto-encoders work was he became very interested in, you know, what are the scaling laws for, you know, for making sparse auto-encoders larger?

4:58:51 And how does that relate to making the base model larger? And so it turns out this works really well and you can use it to sort of project, you know, if you train a sparse auto-encoder at a given size, you know, how many tokens should you train on? And so on. So this was actually a very big help to us in scaling up this work, and made it a lot easier for us to go and train, you know, really large sparse auto-encoders where, you know, it's not like training the big models, but it's starting to get to a point where it's actually actually expensive to go and train the really big ones.

4:59:20 - So you have this, I mean, you have to do all this stuff of like splitting it across large GPUs- - Oh yeah, no, I mean there's a huge engineering challenge here too, right? So, yeah, so there's a scientific question of, how do you scale things effectively? And then there's an enormous amount of engineering to go and scale this up. So you have to chart it, you have to think very carefully about a lot of things. I'm lucky to work with a bunch of great engineers 'cause I am definitely not a great engineer. - Yeah, and the infrastructure especially, yeah, for sure. So it turns out, TLDR, it worked.

4:59:48 - It worked, yeah. And I think this is important because you could have imagined, like, you could have imagined a world where you said after towards mono-semanticity, you know, Chris, this is great, you know, it works on a one layer model, but one layer models are really idiosyncratic. Like, you know, maybe, that's just something, like maybe the linear representation hypothesis and superposition hypothesis is the right way to understand a one layer model, but it's not the right way to understand larger models. And so I think, I mean, first of all, the Cunningham et al paper sort of cut through that a little bit and sort of suggested that this wasn't the case.

5:00:18 But scaling mono-semanticity sort of, I think was significant evidence that even for very large models, and we did it on Claude 3 Sonnet, which at that point was one of our production models. You know, even these models seemed to be very, you know, seemed to be substantially explained at least by linear features. And, you know, doing dictionary learning on them works, and as you learn more features, you go and you explain more and more. So that's, I think, quite a promising sign.

5:00:45 And you find now really fascinating abstract features. And the features are also multimodal. They respond to images and text for the same concept, which is fun. - Yeah, can you explain that? I mean, like, you know, backdoor, there's just a lot of examples that you can-

32 Highlight · Discovering abstract multimodal features in Claude 3

5:01:01 - Yeah, so maybe let's start with a one example to start, which is we found some features around sort of security vulnerabilities and backdoors and codes. So it turns out those are actually two different features. So there's a security vulnerability feature, and if you force it active, Claude will start to go and write security vulnerabilities like buffer overflows into code. And it also, it fires for all kinds of things. Like, you know, some of the top dataset examples for it were things like, you know, dash dash disable, you know, SSL or something like this, which are sort of obviously really insecure.

5:01:34 - So at this point it's kind of like, maybe it's just because the examples were presented that way, it's kind of like a little bit more obvious examples, right? I guess the idea is that down the line, it might be able to detect more nuanced, like deception or bugs or that kind of stuff. - Yeah, well, I maybe wanna distinguish two things. So one is the complexity of the feature or the concept, right? And the other is the nuance of how subtle the examples we're looking at, right?

5:02:04 So, when we show the top dataset examples, those are the most extreme examples that cause that feature to activate. And so it doesn't mean that it doesn't fire for more subtle things. So, you know, the insecure code feature, you know, the stuff that it fires for, most strongly for are these like really obvious, you know, disable the security type things. But you know, it also fires for, you know, buffer overflows and more subtle security vulnerabilities in code.

5:02:35 You know, these features are all multimodal, so you could ask like, what images activate this feature? And it turns out that the security vulnerability feature activates for images of like people clicking on Chrome to like go past the, like, you know, this website, the SSL certificate might be wrong or something like this. Another thing that's very entertaining is there's backdoors and code feature. Like you activate it, it goes and Claude writes a backdoor that like will go and dump your data to port or something.

5:03:02 But you can ask, okay, what images activate the backdoor feature? It was devices with hidden cameras in them. So there's a whole apparently genre of people going and selling devices that look innocuous, that have hidden cameras and they have- - That's great. - This hidden camera in it. And I guess that is the, you know, physical version of a backdoor. And so it sort of shows you how abstract these concepts are, right? And I just thought that was, I'm sort of sad that there's a whole market of people selling devices like that, but I was kind of delighted that that was the thing that it came up with as the top image examples for the feature.

5:03:36 - Yeah, it's nice. It's multimodal. It's multi almost context. It's as broad, strong definition of a singular concept, it's nice. - Yeah. - To me, one of the really interesting features, especially for AI safety is deception and lying. And the possibility that these kinds of methods could detect lying in a model, especially gets smarter and smarter and smarter. Presumably that's a big threat of a super intelligent model that it can deceive the people operating it as to its intentions or any of that kind of stuff.

5:04:09 So what have you learned from detecting lying inside models? - Yeah, so I think we're in some ways in early days for that. We find quite a few features related to deception and lying. There's one feature where, you know, fires for people lying and being deceptive, and you force it active and starts lying to you. So we have a deception feature. I mean, there's all kinds of other features about withholding information and not answering questions, features about power seeking and coups and stuff like that.

Bridging microscopic circuits to macroscopic organs in neural networks

5:04:39 So there's a lot of features that are kind of related to spooky things. And if you force them active, Claude will behave in ways that are, they're not the kinds of behaviors you want. - What are possible next exciting directions to you in the space of mech interp? - Well, there's a lot of things. So for one thing, I would really like to get to a point where we have circuits where we can really understand not just the features, but then use that to understand the computation of models.

5:05:13 That relief for me is the ultimate goal of this. And there's been some work, we put out a few things. There's a paper from Sam Marks that does some stuff like this. And there's been some, I'd say some work around the edges here. But I think there's a lot more to do, and I think that will be a very exciting thing. That's related to a challenge we call interference weights, where due to superposition, if you just sort of naively look at where their features are connected together, there may be some weights that sort of don't exist in the upstairs model, but are just sort of artifacts of superposition.

5:05:47 So that's a sort of technical challenge related to that. I think another exciting direction is just, you know, you might think of sparse auto-encoders as being kind of like a telescope. They allow us to, you know, look out and see all these features that are out there. And you know, as we build better and better sparse auto-encoders, get better and better at dictionary learning, we see more and more stars, and you know, we zoom in on smaller and smaller stars.

5:06:15 But there's kind of a lot of evidence that we're only still seeing a very small fraction of the stars. There's a lot of matter in our, you know, neural network universe that we can't observe yet. And it may be that we'll never be able to have fine enough instruments to observe it, and maybe some of it just isn't possible, isn't computationally tractable to observe it. So it's sort of a kind of dark matter, not in maybe the sense of modern astronomy, but of early astronomy when we didn't know what this unexplained matter is.

5:06:43 And so I think a lot about that dark matter and whether we'll ever observe it, and what that means for safety if we can't observe it, if there's, you know, if some significant fraction of neural networks are not accessible to us. Another question that I think a lot about is, at the end of the day, you know, mechanistic interpretability is this very microscopic approach to interpretability. It's trying to understand things in a very fine-grained way.

5:07:09 But a lot of the questions we care about are very macroscopic. You know, we care about these questions about neural network behavior, and I think that's the thing that I care most about, but there's lots of other sort of larger scale questions you might care about. And somehow, you know, the nice thing about about having a very microscopic approach is it's maybe easier to ask, you know, is this true? But the downside is it's much further from the things we care about, and so we now have this ladder to climb.

5:07:37 And I think there's a question of, will we be able to find, are there sort of larger scale abstractions that we can use to understand neural networks? Can we get up from this very microscopic approach? - Yeah, you've written about this kind of organs question. - Yeah, exactly. - So if we think of interpretability as a kind of anatomy of neural networks, most of the circuits threads involve studying tiny little veins, looking at the small scale at individual neurons and how they connect.

5:08:05 However, there are many natural questions that the small scale approach doesn't address. In contrast, the most prominent abstractions in biological anatomy involve larger scale structures, like individual organs, like the heart, or entire organ systems, like the respiratory system. And so we wonder, is there a respiratory system or heart or brain region of an artificial neural network? - Yeah, exactly.

5:08:30 And I mean, like if you think about science, right? A lot of scientific fields have, you know, investigate things at many levels of abstractions. In biology you have like, you know, molecular biology studying, you know, proteins and molecules and so on. And they have cellular biology, and then you have histology studying tissues, and you have anatomy, and then you have zoology, and then you have ecology. And so you have many, many levels of abstraction or you know, physics, maybe the physics of individual particles, and then, you know, statistical physics gives you thermodynamics and things like this.

5:08:58 And so you often have different levels of abstraction. And I think that right now we have, you know, mechanistic interpretability if it succeeds is sort of like a microbiology of neural networks, but we want something more like anatomy. And so, and you know, a question you might ask is, why can't you just go there directly? And I think the answer is superposition, at least in significant part. It's that it's actually very hard to see this macroscopic structure without first sort of breaking down the microscopic structure in the right way, and then studying how it connects together.

5:09:31 But I'm hopeful that there is gonna be something much larger than features and circuits, and that we're gonna be able to have a story that involves much bigger things, and then you can sort of study in detail the parts you care about. - I suppose to neurobiology, like a psychologist or a psychiatrist of a neural network. - And I think that the beautiful thing would be if we could go, and rather than having disparate fields for those two things, if you could build a bridge between them- - Oh, right. - Such that you could go and have all of your higher level abstractions

Artificial neural networks offer massive advantages over neuroscience

5:10:02 be grounded very firmly in this very solid, you know, more rigorous, ideally, foundation. - What do you think is the difference between the human brain, the biological neural network and the artificial neural network? - Well, the neuroscientists have a much harder job than us. You know, sometimes I just like count my blessings by how much easier my job is than the neuroscientists, right? So I have, we can record from all the neurons.

5:10:28 We can do that on arbitrary amounts of data. The neurons don't change while you're doing that, by the way. You can go and ablate neurons, you can edit the connections and so on, and then you can undo those changes. That's pretty great. You can force, you can intervene on any neuron and force it active and see what happens. You know, which neurons are connected to everything, right? Neuroscientists wanna get the connectome, we have the connectome and we have it for like much bigger than like C. elegans.

5:10:53 And then not only do we have the connectome, we know what, you know, which neurons excite or inhibit each other, right. So we have, it's not just that we know that like the binary mass, we know the weights. We can take gradients, we know computationally what each neuron does. So I don't know the list goes on and on. We just have so many advantages over neuroscientists. And then despite having all those advantages, it's really hard. And so one thing I do sometimes think is like, gosh, like if it's this hard for us, it seems impossible under the constraints of neuroscience or, you know, near impossible.

5:11:26 I don't know, maybe part of me is like, I've got a few neuroscientists on my team. Maybe I'm sort sort of like, ah, you know, maybe the neuroscientists, maybe some of them would like to have an easier problem that's still very hard and they could come and work on neural networks. And then after we figure out things in sort of the easy little pond of trying to understand neural networks, which is still very hard,

Simple training rules give rise to deep scientific beauty

5:11:48 then we could go back to biological neuroscience. - I love what you've written about the goal of mech interp research as two goals, safety and beauty. So can you talk about the beauty side of things? - Yeah, so, you know, there's this funny thing where I think some people want, some people are kind of disappointed by neural networks, I think, where they're like, ah, you know, neural networks, it's these just these simple rules, and then you just like do a bunch of engineering to scale it up and it works really well.

5:12:14 And like, where's the like complex ideas? You know, this isn't like a very nice, beautiful scientific result. And I sometimes think when people say that, I picture them being like, you know, evolution is so boring. It's just a bunch of simple rules and you run evolution for a long time and you get biology. Like what a sucky, you know, way for biology to have turned out. Where's the complex rules? But the beauty is that the simplicity generates complexity.

5:12:41 You know, biology has these simple rules and it gives rise to, you know, all the life and ecosystems that we see around us, all the beauty of nature, that all just comes from evolution and from something very simple evolution. And similarly, I think that neural networks build, you know, create enormous complexity and beauty inside and structure inside themselves that people generally don't look at and don't try to understand because it's hard to understand.

5:13:06 But I think that there is an incredibly rich structure to be discovered inside neural networks, a lot of very deep beauty if we're just willing to take the time to go and see it and understand it. - Yeah, I love mech interp, the feeling like we are understanding or getting glimpses of understanding the magic that's going on inside is really wonderful. - It feels to me like one of the questions that's just calling out to be asked, and I'm sort of, I mean, a lot of people are thinking about this, but I'm often surprised that not more are is, how is it that we don't know how to create computer systems that can do these things, and yet we have these amazing systems that we don't know how to directly create computer programs that can do these things, but these neural networks can do all these amazing things?

5:13:53 And it just feels like that is obviously the question that sort of is calling out to be answered if you are, if you have any degree of curiosity. It's like how is it that humanity now has these artifacts that can do these things that we don't know how to do? - Yeah, I love the image of the circuits reaching towards the light of the objective function. - Yeah, it's just, it's this organic thing that we've grown and we have no idea what we've grown. Well, thank you for working on safety and thank you for appreciating the beauty of the things you discover.

5:14:21 And thank you for talking today, Chris. This was wonderful. - Yeah. Thank you for taking the time to chat as well. - Thanks for listening to this conversation with Chris Olah, and before that with Dario Amodei and Amanda Askell. To support this podcast, please check out our sponsors in the description. And now let me leave you with some words from Alan Watts. "The only way to make sense out of change is to plunge into it, move with it, and join the dance."

5:14:48 Thank you for listening and hope to see you next time.