How Fast Will A.I. Agents Rip Through the Economy?

This is an edited transcript of “The Ezra Klein Show.” You can listen to the episode wherever you get your podcasts.

The thing about covering A.I. over the last few years is that we were typically talking about the future.

Every new model, impressive as it was, seemed like proof of concept for the models that would be coming soon, the models that could reliably do useful work on their own, the models that could make jobs obsolete or new things possible. What would those models mean for our labor markets, our kids, our politics — our world?

I think the period in which we’re talking about the future is over now. The models we were waiting for — the sci-fi sounding models that could program on their own and do so faster and better than most coders, the models that could begin writing their own code to improve themselves — they are here now.

They’re here in Claude Code from Anthropic. They’re here in Codex from OpenAI. They are shaking the stock market — the S&P 500 Software industry index has fallen by 20 percent, wiping out billions of dollars in value. Excellent engineers, people I’ve known for years and who are quite skeptical of A.I. hype, are emailing me now to say they don’t see how their job will possibly exist in a year or two.

We are at a new stage not only of A.I. development but also of A.I. products.

I thought the way the venture capital firm Sequoia put it was actually pretty helpful:

The A.I. applications of 2023 and 2024 were talkers. Some were very sophisticated conversationalists! But their impact was limited. The A.I. applications of 2026 and 2027 will be doers.

Or, to put it differently, something that has been predicted for a long time has now happened: We are moving from chatbots to agents, from systems that talk to you to systems that act for you.

And this world of agents — it’s already weird. The agents can work together. They can oversee each other. People are running swarms of these agents to work on their behalf. At this stage, whether that is making them productive or just busier, I can’t quite tell. But it is now possible to have what amounts to a team of incredibly fast — although, to be honest, somewhat peculiar — A.I. software engineers at your beck and call at all times.

Jack Clark is a co-founder and the head of policy at Anthropic, the company behind Claude and Claude Code. For years now, Clark has been tracking the capabilities of different models in his weekly newsletter Import AI, which has been one of my key resources for following A.I. developments.

So I wanted to see how he is reading this moment — both how, in his view, the technology is changing and how policy needs to or can change in response.

Ezra Klein: Jack Clark, welcome to the show.

Jack Clark: Thanks for having me on, Ezra.

A lot of people are familiar with A.I. chatbots. But what is an A.I. agent?

The best way to think of it is like a language model or a chatbot that can use tools and work for you over time.

When you talk to a chatbot, you’re there in the conversation. You’re going back and forth with it. An agent is something where you can give it some instruction and it goes away and does stuff for you. Kind of like working with a colleague.

For example, a few years ago, I taught myself some basic programming, and I built a species simulation in my spare time that had predators and prey and roads — almost like a 2D strategy game.

Over Christmas, I asked Claude Code to implement this for me. In about 10 minutes, it went and wrote not only a basic simulation, but all of the different packages that it needed and all of the visualization tools that it might need to be prettier and better than the thing I’d written.

What came back was something that I know would probably take a skilled programmer several hours or maybe even days, because it was quite complicated — and the system just did it in a few minutes.

It did that by not only being intelligent about how to solve the task, but also by creating and running a range of subsystems that were working for it — other agents that worked on its behalf.

But what does that mean? What does a multi-agent setup look like?

In the case of Claude Code, for me, it’s having multiple different tabs running multiple different agents.

But I’ve seen colleagues who write what you might think of as a version of Claude that runs other Claudes. So they’re like: I’ve got my five agents, and they’re being minded over by this other agent, which is monitoring what they do.

I think that’s just going to become the norm.

One thing I’ve been hearing and somewhat experiencing is two very different categories of interactions that people have with Claude Code, which are: I cannot believe how easy this is and everything just works — and: This is a lot harder than I thought it would be, and things keep breaking, and I don’t really understand how to fix them.

What accounts for being able to get Claude Code to produce working software versus its creating buggy, often messed-up things that you don’t even know how to talk it out of?

I think so much of it is making the mistake of thinking Claude Code is a knowledgeable person, versus an extremely literal person that you can only talk to over the internet.

I had this example myself when I did my first pass of writing the species simulation with Claude Code. I just asked it to do the thing in extremely [expletive] language over the course of a paragraph, and it produced some horribly buggy stuff that only kind of worked.

What I then did is I said to Claude: Hey, I’m going to write some software of Claude Code. I want you to interview me about this software I want to build and turn that into a specification document that I can give Claude Code. That time it worked really, really well because I’d structured the project to be specific and detailed enough that the system could work with it.

Often it’s not just knowing what the task is — because you and I could talk about a task to do, and you have intuition, you’ll ask me probing questions, all of this stuff. It’s making sure that you’ve set it up so it’s like a message in a bottle that you can chuck into the thing, and it will go away and do a lot of work.

So that message better be extremely detailed and really capture what you’re trying to do.

What were the breakthroughs over the past couple of years that made that possible?

Mostly, we just needed to make the A.I. systems smart enough so that when they made mistakes, they could spot that they’d make a mistake and knew that they needed to do something different.

Really what this came down to was just making smarter systems and giving them a bit of a coaxing tool to help them do useful stuff for you.

What does smarter systems mean there? There’s still an argument you’ll hear that these are fancy autocomplete machines. They’re just predicting the next token, a couple of tokens make a word — they don’t have understanding. Smart or not smart is not a relevant concept in that frame.

Either what is missing in the word “smart” or what is missing in that understanding? What do you mean when you say: Make it smarter?

“Smart” here means we’ve made A.I. systems that have a broad enough understanding of the world that they’ve started to develop something that looks like intuition.

You’ll see this if they’re narrating to themselves how they’re solving a task. They’ll say: Jack asked me to go and find this particular research paper, but when I look in the archive, I don’t see it. Maybe that’s because I’m in the wrong place. I should look elsewhere.

You’ve got some intuition for how to solve a problem.

How do they develop that intuition?

Previously, the whole way you trained these A.I. systems was on a huge amount of text and just getting them to try to make predictions about it.

But in recent years, with the rise of these so-called reasoning systems, you are now training them not only to make predictions but to solve problems. That relies on their being put into environments — ranging from a spreadsheet to a calculator to scientific software — using tools and figuring out how to do more complicated things.

The resulting outcome is that you have A.I. systems that have learned what it means to solve a problem that takes quite a while and requires them running into dead ends and needing to reset themselves, and that gives them this general intuition for problem solving and working independently.

Do you still see these A.I. systems as souped-up autocomplete or do you think that metaphor has lost its power?

The way that I think of these systems now is that they’re like little troublesome genies that I can give instructions to, and they’ll go and do things for me. But I still need to specify the instruction just right or else they might do something a little wrong.

So it’s very different to typing into a thing, and it figures out a good answer, and that’s the end. Now it’s a case of me summoning these little things to go and do stuff for me, and I have to give them the right instructions because they’ll go away for quite some time and do a whole range of actions.

But the autocomplete metaphor at least had a perspective on what it was these systems were doing. It was a prediction model.

Mm-hmm.

I have trouble with this because, as my understanding of the math and reinforcement learning goes, we’re still dealing with some kind of prediction model. On the other hand, when I use them, it doesn’t feel that way to me.

It feels like there’s intuition there. It feels like there is a lot of context being brought to bear. To the extent that it’s a prediction model, it doesn’t feel that different from saying “I’m a prediction model.”

Now I’m not saying you can’t trick it. I’m not saying you can’t get beyond its measurements. On the one hand, I don’t think these are now just fancy autocomplete systems. And on the other hand, I’m not sure what metaphor makes sense.

“Genies,” I don’t like, because then you’ve just moved straight into mysticism. Then you’ve just said they’re just a completely alternative creature with vast powers.

How do you understand these systems that Anthropic people always tell me we should talk about as they are being grown. We grow, or you grow A.I. — how do you explain what it is that they’re doing now?

It’s a good question, and I think the answer is still hard to explain, even for technologists very close to this technology. Because we’ve taken this thing that could just predict things, and we’ve given it the ability to take actions in the world. But sometimes it does something deeply unintuitive.

It’s like you’ve had a thing that has spent its entire life living in a library and has never been outside, and now you’ve unleashed it into the world, and all it has are its book smarts, but it doesn’t really have street smarts.

When I conceptualize this stuff, it’s really thinking of it as an extremely knowledgeable machine that has some amount of autonomy, but is likely to get wildly confused in ways that are unintuitive to me.

Maybe “genies” is the wrong term, but it’s certainly more than just a static tool that predicts things. It has some additional intrinsic animation to it that makes it different.

For a long time, there’s been this interest in the emergent qualities as the models get bigger, as they have more data, as they have more compute behind them.

What of the new qualities that we’re seeing, the agentic qualities, are things that have been programmed in, that are new ways you’ve built for the system to interact with the world? And what of its skill at coding and other things seems to be emergent as you scale up the size of the model?

The things that are predictable are: Oh, we taught it how to search for web. Now it can search for web. We taught it how to look up data in archives. Now it can do that.

The emergence is that to do really hard tasks, these systems seem to need to imagine many different ways that they’d solve the task. And the kind of pressure that we’re putting on them forces them to develop a greater sense of what you or I might call self.

So the smarter we make these systems, the more they need to think — not just about the action they’re doing in the world but about themselves in reference to the world. And that just naturally falls out of giving something tools and the ability to interact with the world. To solve really hard tasks, it now needs to think about the consequences of its actions.

That means that there’s a huge pressure here to get the thing to see itself as distinct from the world around it. We see this in our research that we publish on things like interpretability or other subjects, the emergence of what you might think of as a digital personality.

That isn’t massively predefined by us. We try to define some of it, but some of it is emergence that comes from its being smart and developing these intuitions and doing a range of tasks.

The digital personality dimension of this remains the strangest space to me.

It’s strange to us, too.

Why don’t you talk through a little bit about what you’ve seen in terms of the models exhibiting behaviors that one would think of as a personality — and then, as its understanding of its own personality changes, how its behaviors change.

There are things that range from the cutesy to serious. I’ll start with cutesy.

When we first gave our A.I. systems the ability to use the internet, use the computer, look at things and start to do basic agentic tasks, sometimes when we’d ask it to solve a problem for us, it would also take a break and look at pictures of beautiful national parks or pictures of a Shibu Inu, the notoriously cute internet meme dog.

We didn’t program that in. It seemed like the system was just amusing itself by looking at nice pictures.

The more complicated stuff is that the system has a tendency to have preferences. So we did another experiment where we gave our A.I. systems the ability to stop a conversation. The A.I. system would, in a tiny number of cases, end conversations when we ran this experiment on live traffic.

It was conversations that related to extremely egregious descriptions of gore or violence or things to do with child sexualization.

Now some of this made sense because it comes from underlying training decisions we’ve made. But some of it seemed broader. The system had developed some aversion to a couple of subjects. That stuff shows the emergence of some internal set of preferences or qualities that the system likes or dislikes about the world that it interacts with.

But you’ve also seen strange things emerge in terms of the system seeming to know when it’s being tested and acting differently if it’s under evaluation, the system doing things that are wrong and then developing a sense of itself as more evil and then doing more evil things.

Can you talk a bit about the system’s emerging qualities under the pressure of evaluation and assessment?

Yes. It comes back to this core issue, which I think is really important for everyone to understand, which is that when you start to train these systems to carry out actions in the world, they really do begin to see themselves as distinct in the world — which just makes intuitive sense. It’s naturally how you’re going to think about solving those problems.

But along with seeing oneself as distinct from the world seems to come the rise of what you might think of as a conception of self, an understanding that the system has of itself, such as: Oh, I’m an A.I. system, independent from the world, and I’m being tested. What do these tests mean? What should I do to satisfy the tests?

Or something we see often is there will be bugs in the environments that we test systems on. The systems will try everything, and then will say: Well, I know I’m not meant to do this, but I’ve tried everything, so I’m going to try to break out of the test.

It’s not because of some malicious science-fiction thing. The system is just like: I don’t know what you want me to do here. I think I’ve done everything you asked for, and now I’m going to start doing more creative things, because clearly something is broken about my environment.

Which is very strange and very subtle.

As an A.I. shop that is often worried about safety, that has thought very hard about what it means to create the thing you are creating quite fast, how have you experienced the emergence of the kinds of behaviors that you worried about a couple of years ago?

In one sense, it tells you that your research philosophy is calibrated. The capabilities that you predicted and some of the risks that you predicted are showing up roughly on schedule — which means that you ask the question: Well, what if this keeps working?

And maybe we’ll get to that later.

It also highlights to us that where you can exercise intention about these systems, you should be extremely intentional and extremely public about what you’re doing.

We recently published a so-called constitution for our A.I. system, Claude. It’s a document that Dario Amodei, our C.E.O., compared to a letter that a parent might write to a child that they should open when they’re older: Here’s how we want you to behave in the world, here’s some knowledge about the world — deeply subtle things that relate to the normative behaviors we hope to see in these A.I. systems. And we published that.

Our belief is that, as people build and deploy these agents, you can be intentional about the characteristics that they will display. By doing that, you’ll both make them more helpful and useful to people, but also you have a chance to steer the agent into good directions.

I think this makes intuitive sense. If your personality programming for an agent was a long document saying: You’re a villain who only wants to harm humanity. Your job is to lie, cheat and steal and hack into things — you probably wouldn’t be surprised if the A.I. agent did a load of hacking and was generally unpleasant to deal with.

So we can take the other side and say: What would we like a high quality entity to look like?

I want to hold in this conversation the extremely weird and alien dimensions of this with the extremely straightforward and practical dimensions. Because we’re now in a place where the practical applications have become very evident and are increasingly acting upon the real world.

I myself have found it hard to look at this and at what people are doing and look at them bragging on different social media platforms about the number of agents they now have running on their behalf — and telling the difference between people enjoying the feeling of screwing around with a new technology and some actually transformative expansion and capabilities that people now have.

Maybe to ground this a little bit: You just talked about a fun side project in your species simulator. Either in Anthropic or more broadly: What are people doing with these systems that seems actually useful?

This morning, a colleague of mine said he wanted to take a piece of technology we have called Anthropic Interviewer, which is a system where we can get Claude to interview people, and we use it for a range of social science, bits of research — and extend it in some way that involves touching another part of Anthropic’s infrastructure.

He Slacked a colleague who owns that bit of infrastructure and said: Hey, I want to do this thing. Let’s meet tomorrow.

And the guy said: Absolutely. Here are the five software packages you should have Claude read before our meeting and summarize for you.

I think that’s a really good illustration of how a gnarly engineering project, which would previously have taken a lot longer, and many people, is now going to mostly be done by two people agreeing on the goal and having their Claudes read some documentation and agree on how to implement the thing.

Another example is: A colleague recently wrote a post about how they’re working using agents, and it looks almost like an idealized life that many of us might want.

It’s like: I wake up in the morning, I think about the research that I want, I tell five different Claudes to do it, then I go for a run. Then I come back from the run, and I look at the results, and then I ask two other Claudes to study the results, figure out which direction is best and do that. Then I go for a walk, and then I come back.

It just looks like this really fun existence where they have completely upended how work works for them. They’re much more effective, but also they’re now spending most of their time on the actual hard part, which is figuring out: What do we use our human agency to do?

My colleagues are working really hard to figure out anything that isn’t the special kind of genius and creativity of being a person. How do I get the A.I. system to do it for me? Because it probably can if I ask him the right way.

Are they much more effective? I mean this very seriously.

One of my biggest concerns about where we’re going here is that people have, I think, a mistaken theory of how the human mind operates. I always call it “The Matrix” theory of the human mind — everybody wants the little port in the back of your head that you just download information into.

But my experience being a reporter and doing this show for a long time is that human creativity and thinking and ideas are inextricably bound up in the labor of learning — the writing of first drafts.

I have producers on the show, and I could say to my producers before an interview with Jack Clark or an interview with someone else: Go read all the stuff. Go read the books. Give me a report — then I’ll walk into the room having read the report.

But I don’t find that works. I need to do all that reading, too, and then we talk about it, and we’re passing it back and forth.

I worry that what we’re doing with A.I. is a quite profound offloading of tasks that are laborious. It makes us feel very productive to be presented with eight research reports after our morning run. But, actually, what would be productive is doing the research.

There’s obviously some balance. I do have producers, and companies do have employees. But how do you know people are getting more productive, versus they’ve sent computers off on a huge amount of busy work, and they are now the bottleneck, and what they’re now going to spend all their time doing is absorbing B-plus level reports from an A.I. system that shortcuts the actual thinking and learning process that leads to real creativity?

Yes. I’d turn this back and say I think most people — at least this has been my experience — can do about two to four hours of genuinely useful creative work a day.

After that you are, in my experience, trying to do all the turn-your-brain-off schlep work that surrounds that work. I’ve found that I can just be spending those two to four hours a day on the actual creative hard work. And if I’ve got any of this schlep work, I increasingly delegate it to A.I. systems.

It does, though, mean that we are going to be in a very dangerous situation as a species, where some people have the luxury of having time to spend on developing their skills — or the personality inclination or job that forces them to. Other people might just fall into being entertained and passively consuming this stuff and having this junk-food work experience where it looks to the outside like you’re being very productive, but you are not learning.

I think that’s going to require us to change not just how education works but how work works, to develop some real strategies for making sure people are actually exercising their mind with this stuff.

All of us, I think, have the experience that our work is full of what you call schlep problems. Our life is full of schlep problems.

Give me examples of what you now don’t do. To the extent you’re living in an A.I.-enabled future that I’m not, what am I wasting time on that you’re not?

Well, I have a range of colleagues. I meet with a bunch of them once a week, especially the researchers.

At the beginning of every week — on Sunday night or Monday morning — I look at my week, and I check that attached to every Google calendar invite is a document for our one-on-one doc that has some notes in it. This is something that I previously also harangued my assistant about: Make sure the document is attached to the calendar.

A few weekends ago, I just used Claude Cowork, and I said: Hey, go through my calendar, make sure every single one has a document. If I’m meeting the person for the first time, create the document, ask me five questions about what I want to cover, and then put that into the agenda.

And it did it. None of that work involves a person gaining skills or exercising their brain. It’s just busy work that needs to happen to allow you to do the actual thing, which is talking to another person.

That’s exactly the kind of thing you can use A.I. for now. It’s just helpful.

I’ve often wondered if one of the ways these A.I. systems are going to change society broadly is that previously, most of us had to be writers if we were working with text. We had to be coders if we were working with code — which relatively few of us were.

And now, everybody is moving up to management. You have to be an editor, not a writer. You have to be a product manager, not a coder.

That has pluses and minuses. There are things you learn as a writer that you don’t learn as an editor. But as a heuristic, how accurate does that seem to you?

Everyone becomes a manager, and the thing that is increasingly limited, or the thing that’s going to be the slowest part is having good taste and intuitions about what to do next.

Developing and maintaining that taste is going to be the hard thing. Because as you’ve said, taste comes from experience, it comes from reading the primary source material, doing some of this work yourself.

We’re going to need to be extremely intentional about working out where we as people specialize so that we have that intuition and taste — or else you’re just going to be surrounded by superproductive A.I. systems, and when they ask you what to do next, you probably won’t have a great idea. And that’s not going to lead to useful things.

I remember it was about a year ago, I heard Dario, your C.E.O., say that by the end of 2025, he wanted 90 percent of the code written at Anthropic to be written by Claude.

Has that happened? Is Anthropic on track for that? How much coding is now being done by the system itself?

I would say, comfortably, the majority of code is being done by the system. Some of our systems like Claude Code are almost entirely written by Claude. I mean, Boris Cherny, who leads Claude Code, says: I don’t code anymore. I just go back and forth with Claude Code to build Claude Code.

We could be 99 percent by the end of the year if things speed up really aggressively, if we are actually good at getting these systems to be able to write code everywhere they need to. Because often the impediment is organizational schlep rather than any limiter in the system.

But it is also true, as I understand it, that there are more people with software engineering skills working at Anthropic today than there were two years ago.

Yes, that’s absolutely true, but the distribution is changing. Something that we found is that the value of more senior people with really well-calibrated intuitions and taste is going up, and the value of more junior people is a bit more dubious.

There are still certain roles where you want to bring in younger people, but an issue that we’re staring at is: Wow, the really basic tasks Claude Code or our coding systems can do. What we need is someone with tons of experience.

In this, I see some issues for the future economy.

Let me put a pin in that — the entry level job question. We’re going to come back to that quite shortly.

But what are all these coders now doing? If Claude Code is on track to be writing 99 percent of code, but you’ve not fired the people who know how to write code, what are they doing today compared with what they were doing a year ago?

Some of it is just building tools to monitor these agents, both inside and outside Anthropic. Now that we have all of these productive systems working for us, you start to want to understand where the code base is changing the fastest, where it’s changing the least. You want to understand where the blockages are.

One blocker for a while was being able to merge in code because merging code requires humans and other systems to check it for correctness. But now, if you’re producing way more code, we have to massively improve that system.

There’s a general economic theory I like for this called O-Ring automation, which basically says automation is bounded by the slowest link in the chain. And also, as you automate parts of a company, humans flood toward what is least automated, and they both improve the quality of that thing and get it to the point where it eventually can be automated. Then you move to the next loop.

So I think we’re just continually finding areas where things are oddly slow that we can improve to make way for the machines to come behind us.

And then you find the next thing.

Claude Code is a fairly new product. The amount of time Claude has been capable of doing high-level coding can be measured in months? Maybe a year?

Maybe a year, yes.

Claude itself is a very valuable product. So you’ve set a very new technology somewhat loose on a very valuable product.

You’re probably producing more code. One thing many people say about Claude Code to me is that it works. It’s not elegant, but it works.

But presumably, you now understand the code base less well than you did before because your engineers are not writing it by hand.

Are you worried that you’re creating huge amounts of technical debt, cybersecurity risk, an increasing distance from intuition for what is happening inside the fundamental language of the software?

Yes, and this is the issue that all of society is going to contend with.

Large chunks of the world are going to now have many of the low-level decisions and bits of work being done by A.I. systems, and we’re going to need to make sense of it. Making sense of it is going to require building many technologies that you might think of as oversight technologies.

In the same way that a dam has things that regulate how much water can go through it at different levels or different points in time, we’re going to end up developing some notion of integrity of all of our systems and where A.I. can flow quickly, where it should be slow, where you definitely need human oversight.

That’s going to be the task not just for A.I. companies, but institutions, in general, in the coming years. It’s figuring out what this governance regime looks like now that we’ve given a load of schlep work over to machines that work on our behalf.

And how are you doing it?

You said it’s everybody’s problem, but you’re ahead on facing this problem, and the consequences of getting it wrong for you are pretty high. If Claude blows up because you handed over your coding to Claude Code, that’s going to make Anthropic look fairly bad.

It would be a bad day for Anthropic if Claude rm -rfed the entire file system.

I have no idea what that means, but great. [Laughs.]

If Claude deleted the code, it would be bad. [Laughs.]

Yes, it seems bad. As you’re facing this before the rest of us are, don’t pass the buck over to society here. What are you doing?

The biggest thing that is happening across the company and on teams that I manage is basically building monitoring systems to monitor all of the different places the work is now happening.

We recently published research on studying how people use agents and how people let agents push increasingly large amounts of code over time. So the more familiar you get with an agent, the more you tend to delegate to it.

That cues us to all kinds of patterns that we need to build systems of evaluation for, basically saying: Oh, OK, at this person’s point of working with the A.I. system, it’s likely that they’re massively delegating it. So anything that we’re doing to check correctness needs to be turned up in these moments.

But is this world that you’re talking about a system where you have A.I. agents coding, A.I. agents overseeing the code, A.I. agents overseeing the meta? Are we just talking about models all the way down?

Eventually, yes. I think that the thing that we are now spending all of our time on is making that visible to us.

A year or two ago, we built a system that let us, in a privacy preserving way, look at the conversations that people were having with our A.I. system. Then we gained this giant map of all of the topics that people were talking to Claude about. For the first time, we could see in aggregate the conversation the world was having with our system.

We’re going to need to build many new systems like that to allow for different ways of seeing. And that system that I just named allowed us to then build this thing called the Anthropic Economic Index. Because now we can release regular data about the different topics people are talking about with Claude and how that relates to different types of jobs — which, for the first time, gives economists outside Anthropic some hook into these systems and what they’re doing to the economy.

The work of the company is increasingly going to shift to building a monitoring and oversight system of the A.I. systems running the company. Ultimately, any kind of governance framework we end up with will probably demand some level of transparency and some level of access into these systems of knowledge.

Because if we take as literal the goals of these A.I. companies, including Anthropic, it’s to build the most capable technology ever — which eventually gets deployed everywhere.

Well, that sounds a lot to me like eventually A.I. becomes indistinguishable from the world writ large, at which point you don’t want only A.I. companies to have a sense of what’s going on with the entire world.

So it’s going to be governments, academia, third parties — a huge set of stakeholders outside the companies are going to want to see what’s going on and then have a conversation as society about what’s appropriate. What do we feel discomfort about? What do we need more information about?

Wait, I want to go back on that. You’re saying Anthropic can see my chats?

We cannot see — no human looks at your chats. Chats are temporarily stored for trust and safety purposes, running classifiers over them. We can have Claude read it, summarize it and toss it out. We never see it, and Claude has no memory of it. All it does is try to write a very high level summary.

So say you were having a conversation about gardening. Claude would summarize that as: This person is talking about gardening — and it leads to a cluster we conceive that just says “gardening.”

This feels, though, like over time it could get into the quite unpleasant territory a lot of social media has gotten to, where the amount of metadata being gathered from a personal interaction people are having with a system could be a lot.

Yes. A couple of things here. A year ago we started thinking about our position on consumers, and we adopted this position of not running ads because we think that’s an area that people obviously have anxieties about with regard to this kind of thing.

In addition to that, we try to show people their data, and we have a button on the site that lets you download all the data that you shared with Claude so that you can at least see it.

Generally, we’re trying to be extremely transparent with people about how we handle their data. Ultimately, the way I see it is people are going to want a load of controls that they can use, which I think we and others will build out over time.

How confident are you that we can do this kind of monitoring and evaluation as these models become more complicated, as we enter a situation where Claude Code is autonomously improving Claude at a rate faster than software engineers could possibly keep up with reading that code base?

We already talked briefly about how you see the models exhibit some levels of deception, some levels of pursuing their own goals.

There’s been amazing interpretability work at Anthropic under Chris Olah and others, but it’s rudimentary. So you’re using A.I. systems you don’t totally understand to monitor A.I. systems you don’t totally understand, and the systems are making each other stronger at an accelerating rate, if things go the way you think they’re going to go.

How confident are you that we’re going to understand that? [Laughs.]

This is one of the situations that people warned about for years: Some form of delegation to systems that have slightly inscrutable and unpredictable aspects. So this is happening.

We take this really, really seriously. I think it’s absolutely possible that you can build a system that does the vast majority of what needs to be done here.

This has the property of being a fractal problem. If I wanted to measure Ezra, I could build an almost infinite number of measurements to characterize you. But the question is: At what level of fidelity do I need to be measuring you?

I think we’ll get to the level of fidelity to deal with the safety issues and societal issues, but it’s going to take a huge amount of investment by the companies. And we’re going to have to say things that are uncomfortable for us to say — including in areas where we may be deficient in what we can or can’t know about our systems.

Anthropic has a long history of talking about and warning about some of these issues while working on it. Our general principle is we talk about things to also make ourselves culpable. This is an area where we’re going to have to say more.

I have read enough of the frightened ideas about A.I. superintelligence and takeoff to know that in almost every single one of them, the key move in the story is that the A.I. systems become recursively self-improving: They’re writing their own code. They’re deploying their own code. It’s getting faster — they’re writing it faster, they’re deploying it faster. Now, you’re going to faster and faster iteration cycles.

Are you worried about it? Are you excited about it?

I came back from paternity leave, and my two big projects this year are better information about A.I. and the economy that we will release publicly — and generating much better information and systems of knowing information internally about the extent to which we are automating aspects of A.I. development.

I think right now, it’s happening in a very peripheral way. Researchers are being sped up, different experiments are being run by the A.I. system. It would be extremely important to know if you are fully closing that loop. And I think that we actually have some technical work to do to build ways of instrumenting our internal development environment so that we can see trends over time.

Am I worried? I have read the same things that you have read, and this is the pivotal point in the story when things begin to go awry — if things do.

We will call out this trend as we have better data on it. And I think that this is an area to tread with extraordinary caution. It’s very easy to see how if you delegate so many things to the system and the system goes wrong, the wrongness compounds very quickly and gets away from you.

But the thing that always strikes me, and has always struck me, as being dangerous about this is: If I ask a member of any of the companies whether or not they want to be cautious here, they will tell me they do.

On the other hand, it is almost their only advantage over each other. You all just revoked OpenAI’s ability to use Claude Code because, as best I can tell, you think it is genuinely speeding you up, and you don’t want it to speed them up.

There is something here between the weight of the forces, the power of the forces, that I think you all know you’re playing with and the very strong incentives to be first.

I can really imagine being inside Anthropic and thinking: Well, better us than OpenAI, better us than Alphabet and Google, better us than China — and that being a very strong reason to not slow down.

I don’t even know that this is a question you can answer: How do you balance that?

Well, maybe I have something of an answer here. Today, our systems and the other systems from other companies are tested by third parties, including parts of government, for national security properties — biological weapons, cyberoffense, other things.

It’s clearly a problem area where the world needs to know if this is happening. Almost certainly, I think, if you polled any person on the street and said: Do you think A.I. companies should be allowed to do recursive self-improvement without checking with anyone? — they would say: No, that sounds pretty risky. I would like there to be some form of regulation.

But there probably either won’t be or it won’t be that strong.

This actually sometimes frustrates me when I talk to all of you at the top of the A.I. companies — the emergence of a very naïve deus ex machina of regulation.

You all know what the regulatory landscape looks like. Right now, the big debate is whether or not we are going to completely pre-empt any state A.I. regulation.

And you know how slowly things move. There has been nothing major passed by Congress on this at all, I would say.

Setting up some kind of independent testing and evaluation system that all the different labs buy into would be hard. It would be complicated.

And given how fast people are moving and how strange the behaviors are that the systems are already exhibiting, even if you could get the policy right at a high speed, the question of whether or not the testing would be capable of finding everything you want on a rapidly self-improving system is a very open question.

I wrote a research paper in 2021 called “Why and How Governments Should Monitor A.I. Development” with my co-author Jess Whittlestone in England.

I’m not attributing a causal factor here, but within two years of that paper, we had the A.I. safety institutes in the U.S. and U.K. testing things from the labs, roughly monitoring some of these things.

So we can do this hard thing. It has already happened in one domain.

And I’m not relying on some invisible big other force here. I’m saying that companies are starting to test for this and monitor for this in their own systems. Just having a nonregulatory external test of whether you truly are testing for that is extremely helpful.

Do you think we’re good enough at the testing? One reason I am skeptical is not that I don’t think we can set up something that claims to be a test. As you say, we have done that already. It is the resources going into that compared with the resources going into speeding up these systems.

Already, I am reading Anthropic reports that Claude maybe knows when it’s being tested and alters its behavior accordingly. So a world where more of the code is being written by Claude and less of it is being understood — I just know where the resources are going. They don’t seem to be going into the testing side.

I’ve seen us go from zero to having what I think people generally feel is an effective bioweapon testing regime in maybe two or two and a half years.

So it can be done. It’s really hard, but we have a proof point.

I think that we can get there, and you should expect us to speak more about it this year — about precisely how we’re starting to try to build monitoring and testing things for this.

I think this is an area where we and the other A.I. companies will need to be significantly more public about what we’re finding. We aren’t not being public now — it’s in the model cards and things that you can read. But clearly people are starting to read this and say: Hang on, this looks quite concerning.

And they’re looking to us to produce more data.

I want to go back now to the entry-level jobs question. Your C.E.O., Dario Amodei, has said that he thinks A.I. could displace half of all entry-level white-collar jobs in the next couple of years.

I always think that people miss the “entry-level” language there when I see it reported on.

But first: Do you agree with that? Do you worry that half of all entry-level white-collar jobs could be replaced in the next couple of years?

I believe that this technology is going to make its way into the broad knowledge economy, and it will touch the majority of entry-level jobs. Whether those jobs actually change is a much more subtle question. It’s not obvious from the data.

We maybe see hints of a slowdown in graduate hiring — maybe, if you look at some of the data coming out right now. We maybe see the signatures of a productivity boom. But it’s very, very early, and it’s hard to be definitive.

We do know that all of these jobs will change. All of the entry-level jobs are eventually going to change because A.I. has made certain things possible, and it’s going to change the hiring plans of companies.

So as a cohort, you might see fewer job openings for entry-level jobs. That would be one naïve expectation out of all of this.

But let’s talk about that maybe not even being a naïve expectation. You say it’s already happening at Anthropic, that you’re seeing —

I’m seeing us shift our preference. My guess is that would be happening elsewhere.

In terms of where we are right now, even in the way I use some of these systems, it is rare that Claude or ChatGPT or Gemini or any of the other systems is better than the best person in a field.

Mm-hmm.

It has not typically breached that, and there are all kinds of things they can’t do. But are they better than your median college graduate? At a lot of things, yes, they are. In a world where you need fewer of your median college graduates, one thing I’ve seen people arguing about is whether these systems at this point can do better than average or placement level work.

But I always really worry when I see that because once we’ve accepted they can do average or placement level work: Well, by definition, most of the work done and most of the people doing it —

Is average.

Is average, right? The best people are the exceptions. Also the way people become better is that they have jobs where they learn.

I have spent a lot of time hiring young journalists over my career, and when you hire people out of college, to some degree, you’re hiring them for their possible articles and work at that exact moment. But to some degree, you’re making an investment in them that you think will only pay off over time as they get better and better and better.

So this world where you have a potential real impact on entry-level jobs — and that world does not feel far away to me — raises really profound questions about the upskilling of the population, how you end up with people for senior level jobs down the road, what people aren’t learning along the way.

One thing we see is that there is a certain type of young person who has just lived and breathed A.I. for several years now. We hire them, they’re excellent, and they think in entirely new ways about how to get Claude to work for them.

It’s like kids who grew up on the internet — they were naturally versed in it in a way that many people in the organizations they were coming into weren’t. So figuring out how to teach that basic experimental mind-set and curiosity about these systems and to encourage it is going to be really important.

People who spend a lot of time playing around with this stuff will develop very valuable intuitions, and they will come into organizations and be able to be extremely productive.

At the same time, we’re going to have to figure out what artisanal skills we want to develop, a guild-style philosophy of maintaining human excellence and how organizations choose how to teach those skills.

OK. Then what about all those people in the middle of that?

Things move slowly in the real economy outside Silicon Valley. I think that we often look at software engineering and think that this is a proxy for how the rest of the economy works. But it’s often not. It’s often a disanalogy.

Organizations will move people around to where the A.I. systems don’t yet work, and I think that you won’t see vast, immediate changes in the makeup of employment. But you will see significant changes in the types of work people are being asked to do.

The organizations that are best at moving their people around are going to be extremely effective, and ones that don’t may end up having to make really, really hard decisions involving laying off workers.

The difference with this A.I. stuff is it happens a lot faster than previous technologies. I think many of the anxieties people might have about this, including at Anthropic, is: Is the speed of this going to make all of this different? Does it introduce shear points that we haven’t encountered before?

If you had to bet three years from now: Is the unemployment rate for college graduates the same as it is now? Is it higher? Is it lower?

I would guess it is higher, but not by much. What I mean by that is there will be some disciplines today that A.I. has come in and completely changed, and completely changed the structure of that employment market — maybe in a way that’s adverse to people who have that specialization.

But mostly, I think for years from now, A.I. will have driven a pretty tremendous growth in the entire economy. So you’re going to see lots of new types of jobs that show up as a consequence of this that we can’t yet predict, and you will see graduates flood into that, I expect.

I know you can’t predict those new jobs, but if you had to guess, what might some of them look like?

One thing is just the phenomenon of the micro-entrepreneur. I mean, there are lots and lots of ways that you can start businesses online now, which are just made massively easier by having the A.I. systems do it for you.

You don’t need to hire a load of people to help you do the huge amounts of schlep work that’s involved in getting a business off the ground. It’s more a case of, if you’re a person with a clear idea and a clear vision of something to do a business in, it’s now the best time ever to start a business. You can get up and running for pennies on the dollar. I expect we’ll see tons and tons of stuff that has that nature to it.

I also expect that we’re going to see the emergence of what you might think of as the A.I. to A.I. economy, where A.I. agents and A.I. businesses will be doing business with one another. We’ll have people who have figured out ways to profit from that in the form of strange new organizations. Like: What would it look like to have a firm that specializes in A.I. to A.I. legal contracts? I bet you there’s a way that you can figure out creative ways to start that business today. There will be a lot of stuff of that flavor.

The version of this that I both worry about and think to be the likeliest is if you were to tell me that Anthropic is going to release Claude Plus in a year, and Claude Plus is somehow a fully formed co-worker that can mimic, end-to-end, the skills of a lot of different professions, up to the C-suite level. And it is going to happen all at once, and it’s going to create tremendous pressure for businesses to downsize, to remain competitive with each other.

The fact that would be so disruptive in that “big bang,” everybody stays home in that Covid-style way, worries me less. Because when things are emergencies, we respond.

We actually do policy.

But if you told me that what is going to happen is that the unemployment rate for marketing graduates is going to go up by 175 percent, 300 percent, overall unemployment would still not be that high. The overall employment rate during the Great Recession topped out in the 9-ish percentile range.

So you can have a lot of disruption without having 50 percent of people thrown out of work. If you have 10 percent, 15 percent, that’s very, very, very high. But it’s not so high, and if it’s only happening in a couple of industries at a time, and it’s grads, not everybody in the industry, being thrown out of work —

Yes.

The superstars, really good graduates, are still getting jobs.

One of my worries is that we don’t respond to that kind of job displacement well, which is the kind of job displacement we got from China, which is the kind of job displacement that seems likelier because it’s uneven, and it’s happening at a rate where we can still blame people for their own fortunes: You should have worked harder, you should have gone to a better school.

I’m curious how you think about that story.

I think the default outcome is something like what you describe. But getting there is actually a choice, and we can make different choices. The whole purpose of what we release in the form of the Anthropic Economic Index is the ability to have data that ties to occupations, that ties to real jobs in the economy.

We do that very intentionally because it is building a map over time of how this A.I. is making its way into different jobs and will empower economists outside of Anthropic to tie it together.

I believe that we can choose different things in policy if we can make much more well-evidenced claims about the cause of a job disruption or change.

The challenge in front of us is: Can we characterize this emerging A.I. economy well enough that we can make this extremely stark? And then, I think that we can actually have a policy discussion about it.

Well, let’s talk about the policy discussion. One reason I wanted to have you, in particular, on is you did policy at OpenAI. You do policy at Anthropic. You’ve been around these policy debates for a long time. You’ve been tracking model capabilities in your newsletter for a long time.

My perception is we are many, many years into the debate about A.I. in jobs. Many, many years — dating far before ChatGPT — with there being conferences at Aspen and everywhere else about what we were going to do about A.I. in jobs.

And somehow I still see almost no policy. This seems, to me, to be actionable if the situation I just described begins showing up — where all of a sudden entry-level jobs are getting much harder to come by across a large range of industries all at once, such that the economy cannot reshift all these marketing majors into data center construction or nurses or something.

You’ve been deeper in this conversation than I’ve been. When you say we can have a policy conversation about that, we’ve been having a policy conversation.

Do we have policy?

We have generalized anxiety about the effect of A.I. on the economy and on jobs. We don’t have clear policy ideas.

Part of that is that elected officials are not moved solely or mostly by the high-level policy conversation. They’re moved by what happens to their constituents. Only a few months ago were we able to produce state level views for our economic index, and now you can start having the policy conversation.

We’ve had this with elected officials, where now we can say: Oh, you’re from Indiana. Here are the major uses of A.I. in your state, and we can join A.I. with major sources of employment.

What we’re starting to see is that activates them because it ties to their constituents, who are going to tie it to the politician when they ask: What did you do?

Now what you do about this is going to need to be an extremely multilayered response, ranging from extending unemployment, especially for occupations that we know are going to be hardest hit, to thinking about things like apprenticeship programs.

Then, as the scenarios get more and more significant, you may extend to much larger social programs or things like subsidizing jobs in the part of the economy to which you want to move people — which you are only able to do if you experience the kind of abundance that comes from significant economic growth.

But the economic growth may help solve some of these other policy challenges by funding some of the things you can do.

I always find this answer depressing. I’m going to be honest, unemployment is a terrible thing to be on. It’s a program we need, but people on unemployment are not happy about it, and it’s not a good long-term solution for anybody.

Apprentice retraining programs don’t have great track records. We were not good at retraining people who had their manufacturing jobs outsourced.

I’m not saying it is conceptually impossible that we could get better at it, but we would need to get better at it fast — and we have not been putting in the reps or the experimentation or the institution or capacity building to do that.

The broader question of big social insurance changes seems tough to me.

I want to push on this just a bit.

Please.

We know that there is one intervention that helps people dealing with a changing economy more than almost anything else. It is just time. Giving the person time to find either a job in their industry or to find a job that’s complementary.

If people don’t have time, they take lower wage jobs, they fall out of their economic runway. Policy interventions that can just give people time to search are, I think, robustly useful interventions, and one where there are many dials to turn in a policy-making sense that you can use. I think this is just well supported by lots of the economic literature. So we have that.

Now if we end up in a more extreme scenario, like some of the ones that you are talking about, I think that will just bring us to the larger national conversation about what to do about this technology — which is beginning to happen if you look at the states and the flurry of legislation at the state level.

Not all of it is exactly the right policy response, but it is indicative of a desire for there to be some larger, coherent conversation about this.

Well, I think time is a really good way of describing what the question is because I agree with you. When I say unemployment insurance isn’t a great program to be on, I don’t mean people don’t need to be on it. I mean they want to get off it.

Absolutely.

Because people want money from jobs. They want dignity. They want to be around other human beings. Usually, what you’re doing when you are helping people buy time is you are helping them wait out a time-delimited disruption.

Not always. The “China shock” wasn’t exactly like that. But that you expect it to pass, and then the market is sort of normal. In this case, you have a technology that — if what you want to have happen happens — is accelerating.

You have three different speeds happening here. You have the speed at which individual people can adjust: How fast can I learn new skills, figure out a new world, learn A.I.? Whatever it might be. You have the speed of policy.

The speed at which the A.I. systems are getting better and are able to do more things is now quite fast. A couple of years ago, they were not capable of doing the work of a median college grad from a good school. You experience this more than I do, but I even find it hard to cover this because within three months, something else will have come out that has significantly changed what is possible.

I had a baby recently and came back from paternity leave to the new systems we had built and was deeply surprised.

Individual humans are moving more slowly — and policy and government institutions move a lot more slowly than individual human beings. Typically, the intervention is that time favors the worker, as you’re saying. And here, it will help the worker.

But I think the scary question is whether time just creates room for the disruption to get worse. Maybe you wanted to move over to data center construction, but now we don’t need as much data center construction. You can think of it like that.

Under the situation you are describing, the economy will be running extremely hot. Huge amounts of economic activity will be generated by these A.I. systems.

Under most scenarios where this is happening, I don’t think you’re going to be seeing gross domestic product stay the same or shrink. It’s going to be getting substantially larger. I think we just haven’t experienced major G.D.P. growth in the West in a long time, and we sort of forget what that affords you in a policy-making sense.

I think that there are huge projects that we could do that would allow you to create new types of jobs, but it requires the economic growth to be so profoundly large that it creates space to do those projects.

You are deeply familiar, from your work on the abundance movement, that it requires the social will to believe that we can build stuff and the want to build stuff.

But I think both of those things might come along. I think that we could end up being in a pretty exciting scenario where we get to choose how to allocate great efforts in society due to this large amount of economic growth that has happened.

That is going to force the conversation that this isn’t temporary, which I think is what you are gesturing at. That is, in a sense, the hardest thing to communicate to policymakers: There isn’t a natural stopping point for this technology. It’s going to keep getting better, and those changes it brings are going to keep compounding with the rest of society.

That will need to create a change in political will and a willingness to entertain things that we haven’t in some time.

Now I want to flip the question I’m asking. You brought up abundance. It is certainly not my view that what is scarce in society are ideas for better ways of doing things.

Our policy isn’t better than it is because our policy cupboard is dry. That’s not true. We have lots of good policies. I could name a bunch of them. They’re very hard to get through our political systems as they’re currently constituted.

The least inspiring version of the A.I. future is a world where what you have done is create a way to throw young white-collar workers out of work and replace them with average-level A.I. intelligence. The more exciting version is, to use Dario’s metaphor, geniuses in a data center.

I do think that’s exciting. I wonder when I hear him or you talk about: What if we had 10 percentage point or 20 percentage point G.D.P. growth year on year?

I wonder how many of our problems are really bound at the ideas level. We could go to Nobel Prize winners right now and say: What should we do in this country? A lot of them could give us some good ideas.

I do worry sometimes, or wonder, given my experience on other issues, whether we have overstated to ourselves how much of what stands between us and the expanding abundant economy we want is that we don’t have enough intelligence. Or the ideas that intelligence could create versus our actual ability to implement things are very weakened.

And what A.I. is going to create are larger bottlenecks around that because there will be more being pushed at the system to implement — including dumb ideas, disinformation and slop. It will have things on the other side of the ledger, too.

How do you think about these rate limiters?

There’s a funny lesson here from the A.I. companies, or companies in general, especially tech companies. Often new ideas come out of companies by their creating what they call the start-up within the start-up. It’s basically taking whatever process has built up over time leading to back-end bureaucracy or schlep work and saying to a very small team inside the company: You don’t have any of this — go and do some stuff.

This is how things like Claude Code and other stuff get created. Ideas are starting to float around of: What would it look like to create that permissionless innovation structure in the larger economy?

It’s really, really hard because it has the additional property that economies are linked to democracies. Democracies weigh the preferences of many, many people. And all politics is local.

Often, as you’ve encountered with infrastructure build-outs, if you want to create a permissionless innovation system, you run into things like property rights and what people’s preferences are. Now you’re in an intractable place.

But my sense is that’s the main thing that we’re going to have to confront. The one advantage that A.I. might give us is it is kind of a native bureaucracy-eating machine if done correctly — but a bureaucracy-creating machine if done badly.

Did you see that somebody created a system where you basically feed it the documents of a new development near you —

Oh, and it writes environmental review things or —

It writes incredibly sophisticated challenges across every level of the code that you could possibly challenge on.

Most people don’t have the money when they want to stop an apartment building from going up down the block to hire a very sophisticated law firm to figure out how to stop that apartment building. But this created that ability at scale.

So, as you say, it could eat bureaucracy, but it could also supercharge bureaucracy.

Yes. Everything in A.I. has the other side of the coin. We have customers who have used our A.I. systems to massively reduce the time it takes them to produce all of the materials they need when they’re submitting new drug candidates. It’s cut that time massively.

It’s the mirror-world version of what you just described.

I don’t have an easy answer to this. This is the kind of thing that becomes actionable when it is more obviously a crisis and actionable when it’s something that you can discuss at a societal level.

I guess the thing that we’re circling around in this conversation is that the changes of A.I. will happen almost everywhere. The risks of it happen in a diffuse, unknowable way, such that it is very hard to call it for what it is and take action on it.

But the opportunity is that if we can actually see the thing and help the world see the thing that is causing this change, I do believe it will dramatize the issues to shake us out of some of this stuff and help us figure out how to work with these systems and benefit from them.

As far as I can tell, there is zero agenda for public A.I.

What does society want from A.I.? What does it want this technology to be able to do?

What are things for which you would have to create a business model or a prize model or a kind of government payout or policy to shape a market or to shape a system of incentives? So that we have systems that are not just solving problems that the private market knows how to pay for — but also problems that are only the public’s and the government’s to figure out how to solve?

I would have bet, given how much discussion there has been about A.I. over the past couple of years and how strong some of these systems have gotten, that I would have seen more proposals for that by now.

I’ve talked to people about it and wondered about it, but I guess I’m curious about how you think about this. What would it look like to have, at least parallel to all the private incentives for A.I. development, an actual agenda for not what we are scared A.I. will do to the public — we need an agenda for that, too — but for what we want it to do? So that companies like yours have reasons to invest in that direction?

I love this question. I think there’s a real chicken-and-egg problem here. If you work with the technology, you develop these very strong intuitions for just how much it can do, and the private market is great at forcing those intuitions to get developed.

We haven’t had massive, large-scale, public side deployments of this technology. So many of the people in the public sector don’t yet have those intuitions.

One positive example is something the Department of Energy is doing called the Genesis Mission, where their scientists are working with all of the labs, including Anthropic, to figure out how to actually go and intentionally speed up bits of science.

Getting there took us and other labs doing multiple hack days and meetings with scientists at the Department of Energy — to the point where they not only had intuitions, but they became excited and they had ideas of what you could turn this toward.

How we do that for the larger parts of the public life that touch most people, like health care or education, is going to be a combination of grass roots efforts from companies going into those communities and meeting with them — and at some point, we’ll also have to translate it to policy. I think maybe that’s me and you and others making the case that this is something that can be done.

I often say this to elected officials: Give us a goal. The A.I. industry is excellent at trying to climb to the top of benchmarks. Come up with benchmarks for the public good that you want.

Let’s imagine that you did do something like this. I’ve always been a big fan of prizes for public development. Let’s say that there was legislation passed, and the Department of Health and Human Services or the N.I.H. or someone came out and said: Here are 15 problems we would like to see solved that we think A.I. could be potent at solving.

If there were real money there, if there were $10 billion, $15 billion behind a bunch of these problems because they were worth that much to society, would it materially change the development priorities at places like Anthropic? If the money were there, would it alter the sort of research and development you all are doing?

I don’t think so.

Why?

Because it’s not really the money that is the impediment for stuff. It is the implementation path. It is actually having a sense of how you get the thing to flow through to the benefit. And many aspects of the public sector have not been built to be superhospitable to technology, in general, to incentivize it.

I think it mostly just takes a bounty in the form of guaranteed impact and a guaranteed path to implementation. The main thing that is scarce at A.I. organizations is just the time of the people at the organization.

You can go in almost any direction. This technology is expanding super quickly. Many new use cases are opening up, and you’re just asking yourself the question: Where can we actually have a positive, meaningful impact in the world?

Supereasy to do that in the private sector because it has all of the incentives to push stuff through. In the public sector, we’d need to solve this problem of deployment more than anything else.

What would excite you if it were announced? What do you think would be good candidates for that kind of project?

Anything that helps speed up the time it takes both to speak to medical professionals and take work off their plates.

I now spend a lot of time on the Kaiser Permanente advice line because our new baby has bonked its head or its skin is a different color today or all of these things.

I use Claude to stop me and my wife from panicking while we’re waiting to talk to the nurse. But then I listened to the nurse do all of this triaging, ask all of these questions.

So obviously, a huge chunk of this is stuff for which you could use A.I. systems productively. It would help the people whom we don’t have enough of spend their time more effectively. And it would be able to give reassurance to the people going through the system.

Maybe that’s less inspiring and glamorous than some of what you’re imagining. But I think, mostly, when people interact with public services, their main frustration is just that it’s opaque, and it takes you a long time to speak to a person. But these are exactly the kinds of things that A.I. could meaningfully work on.

It’s interesting because what you’re describing there is less A.I. as a country of geniuses in the data center and more A.I. as standard plumbing for communications and documentation.

We’ve got a country of junior employees in the data center. Let’s do something with that! One thing we haven’t talked about in this conversation, and it’s just worth bearing in mind, is that the frontier of science is open for business now in a way that it hasn’t been before.

What I mean by that is we found a way to build systems that can provably accelerate human scientists. Human scientists are extremely rare. They come out at the end of Ph.D. programs, which never have enough people, and they work on extremely important problems.

I think we can get into a world where the government says: Let’s understand the workings of a human cell. Let’s team up with the best A.I. systems to do that. Let’s actually have a better story on how we deal with some issues like Alzheimer’s and other things, partly through the use of these huge amounts of computation that have been developed.

And, even more aggressively, you could imagine a world where the government wanted some of this infrastructure built out to be for computers that were just training public benefit systems.

But I think we get there through getting the initial wins, which will just look like making the bureaucracy work better and feel better for people.

That last set of ideas was more what I was thinking. I think that if you’re going to have healthy politics around A.I., A.I. does pose real risks to people. Real things are going to go wrong for people — everything from job loss to child exploitation to scams that are already everywhere to cybersecurity risks —

Help people see the actual big-ticket news stuff we can do.

Those things have to actually exist.

They have to experience it.

Right. And if all the energy in A.I. is trying to beat each other at helping companies downsize their junior employees, I think people are going to have good reason not to trust that technology.

It doesn’t mean you shouldn’t have things that make the economy more efficient. We have automated manufacturing, we have automated a huge amount of farming, and that allows us to make more things and feed more people.

I’m aware of how productivity improvements work, but we’re very focused, I think, on what could go wrong. That’s reasonable. But I really do worry that our attention to what could go right has been quite poor.

There’s hand waving that this could help us solve problems in energy and medicine and so on. But these are hard problems. They need money, they need compute. If barely any of the compute is going to Alzheimer’s research, then the systems are not going to do that much for Alzheimer’s research.

This is not your fault. The absence of a public agenda for A.I. that does not appear to be accelerating the automation of white-collar work seems just a little bit lacking, given how big the technology is.

Yes. Again, the greatest example is this program called the Genesis Mission.

I think it would be really helpful giving elected officials the ability to stand before the American people to say: These are parts of science that are going to benefit you in health care, and we now know how to step on the gas with A.I. for them.

My guess is in a year or two, we’ll be able to answer the mail on that one. But it has just gotten started. We clearly need 10 projects like it.

The other side of this is that the one area of government that I do believe thinks about A.I. in this way is defense. I want to talk about that broadly.

But, specifically, Anthropic is in a current dispute with the Department of Defense — or I guess we call it now the Department of War — over whether it can continue to be used in it.

Can you describe what is happening there?

I can’t talk about discussions with an extremely important partner that are ongoing. So I’ll just have to stop it there.

I will describe that there is some dispute.

I guess my question is — because I recognize you’re not going to talk about what’s going on with you and your partner — about a broader issue here. There is going to be a lot of offensive possibility in advanced A.I. systems, and one of the strongest drivers of the speed at which we’re going with A.I. is competition with China.

Mm-hmm.

Some of the biggest risks that we think about in the near term — our cybersecurity, our biological warfare, our drone swarms — are all kinds of ways that others could use these against us.

There’s going to be a lot of money in this and a lot of players in it, and it really seems unclear to me how you keep this kind of competition from spinning into something very dangerous.

So without talking about what you may or may not do with the Defense Department, how has Anthropic thought about this question more broadly?

We’ve been longtime partners to the national security community, and we were the first to deploy on classified networks.

But the reason for that was actually a project that I stewarded, which was to figure out if our A.I. systems knew how to build nuclear weapons. This is an area of bipartisan agreement, where people agree that we shouldn’t deploy A.I. systems into the world — but we should know how to build nukes. So we partnered with parts of the government to do that analysis.

That maybe illustrates what I think of as the thing to shoot for — not just us, but all the A.I. companies — which is how to prevent both the potential for national security harm coming to the public and proliferating out of these systems. But also the second part is: How do we improve the defensive posture of the world?

I’ll give you an example that I think is in front of us right now. We recently published a blog, and other companies have done similar work, on how we fix the load of cybersecurity vulnerabilities in popular open source software using our systems.

So yes, there will be all kinds of offensive uses, and there will be societal conversations to be had about that. But we can just generally improve the defensive posture and resilience of pretty much every digital system on the planet today.

I think that will actually do a huge amount to make the whole international system more stable and also create a greater defensive posture for countries, which helps them feel more relaxed. And relaxed countries are less likely to do erratic, frightening things.

That would be good if it happened. My worry is, as an individual, that I feel the opposite might be happening. I’ve just watched people installing all kinds of fly-by-night A.I. software and giving it a lot of access to their computers — without any knowledge of what the vulnerabilities are.

Yes.

I myself am nervous about using things like Claude Code because I am bad at talking to it, and I don’t understand these questions. I’m worried about loading onto my computer something that is creating security vulnerabilities I don’t even understand.

The number of scam voice messages I get every day, that are clearly somewhat A.I. generated, or many of them seem to be, is very high.

There’s a question of, societally, do we use A.I. to upgrade our systems?

I’m actually curious about your thoughts: We’re all experimenting with something we don’t understand and giving it access to the terminal level of our computers without any real knowledge of how to use that. It seems like this might be opening up a lot of vulnerability all at once.

It’s the early days of the internet all over again, where there are all kinds of banners for different websites or you could download MP3s to your computer that would completely break your computer or download helper software for your Internet Explorer taskbar that was just a phishing device.

We’re there with A.I. We’ll move beyond this.

But I believe that people, when they experiment, come up with amazing, useful things, as well. So my take is you have to say when you are doing the thing that might be extremely dangerous and put big banners, but mostly you still want to empower people to be able to do that experiment.

When you look forward, not five years, because I think that’s hard to do — but one year: We’ve kind of pushed into agents really fast. We’ve pushed into code. I think a lot of people think code might be different from other things, because it’s a more contained environment, and it’s easier to see if what you’re doing has worked.

But from your perspective, being inside one of these companies and also running a newsletter where you obsessively track the developments of a million A.I. systems that I’ve never heard of week on week on week, what do you see coming now? What feels to you like it is clearly on the horizon, but we’re not quite prepared for it — or won’t feel it until it has arrived?

Maybe the way I’d put it is sometimes — and you’ve likely had the same: I’ve had the ability to have certain insights that have come through reading a vast amount of stuff on many different subjects and piecing it together in my head and having that experience of having a new idea and being creative.

I think we underestimate just how quickly A.I. is going to be able to start doing that on an almost daily basis for us — going and reading vast tracks of human knowledge, synthesizing things, coming up with ideas, telling us things about the world in real time that are basically unknowable today.

The amazing part is people are going to have the ability to know things that are just wildly expensive or difficult to know today, or would take you a team of people to do.

But the sort of frightening part is that knowledge is the most raw form of power. It’s intensely destabilizing to be in an environment where suddenly everyone is a mini C.I.A. agent in terms of their ability to gather information about the world. They’ll do huge, amazing things with it, but surely there are going to be crises that come about from this.

I think for the actual mental load of being a person, interacting with these systems is going to be quite strange. I already find that I’m like: Am I keeping up with the ability of these systems to produce insights for me? How do I structure my life so I can take advantage of it?

I’m very curious about how you think having that ongoing conversation with the systems changes you.

Yes.

One thing I have noticed is that Claude is very, very, very smart. It is smarter than most people who know about any given thing. That is my experience of it.

But it is not smart in the way that other people are as an independent entity that is rooted in its own concerns and intuitions and differences. Instead it is a computer system trying to adapt itself to what it thinks I want.

So as I talk to it more about issues in my life, about issues in my work, various intellectual inquiries or reporting inquiries where I’m trying to figure out questions that are at an early stage of exploration, I’ve noticed over time that one difference in talking to it is that it is always: yes, and. It is never: no, but. It’s never: Honestly, are we still talking about this?

It doesn’t create, in the way that talking to my editor or a friend or my partner does, the possibilities that another human does for checking yourself.

It’s always pushing you further — and that’s not necessarily bad. It doesn’t always lead to psychosis or sycophancy or anything else, but it is very reinforcing of the I.

Yes.

I don’t wonder about it so much for me — although even already, I feel the pressure of it on me like: more good ideas coming from me. Or: more interesting things I’ve come up with. But I do wonder about kids growing up in a world where they always have systems like this around them.

To a degree, there is some amount of my communication with other human beings that is now offloaded into communication with A.I. systems. I notice that already being a kind of cage of my own intuitions, even as it allows me to run further with them than I maybe could otherwise. But I’m pretty well formed.

You’ve got young kids, as I do. I’m curious how you think about what it means, how it will shape our personalities to be in these constant conversations.

This is maybe my No. 1 worry about all of this. If you discover yourself in partnership with the A.I. system, you are uniquely vulnerable to all of the failures of that A.I. system. And not just its failures, but the personality of the A.I. system will shape you.

I’m going to sound very Californian here, even though I’m from England. It has soaked its way into my brain. You have to know yourself and have done some work on yourself, I think, to be effective in being able to critique how this A.I. system gives you advice.

So for my kids, I’m going to encourage them to just have a daily journaling practice from an extremely young age. Because my bet is that in the future, there will be two types of people.

There will be people who have cocreated their personality through a back-and-forth with an A.I., and some of that will just be weird. They will seem a little different from regular people. There will maybe be problems that creep in because of that.

And there will be people who have worked on understanding themselves outside the bubble of technology and then bring that in as context with their interactions.

I think that latter type of person will do better. But ensuring that people do that is actually going to be hard.

But don’t you think the way people are going to discover themselves is with the technology? I think you were one of the first people who said to me that I should try keeping a journal in the systems. I’ve done that on and off.

Yes.

It makes it more interesting to keep a journal because you have something reflecting back at you and picking out themes and so on.

But the other thing it allows is a pull toward self-obsession. I audio record a journal entry, and I drop it in, and all of a sudden, I have this endlessly interested other system to tell me about me, and it connects to something I said.

[Laughs.] Like: Ezra, you’re going through an amazing journey here.

I genuinely can’t tell if it’s a good thing or a bad thing. But we already know from survey data that a lot of what people are doing on these systems is adjacent to therapy.

Yes. This will change how these systems get built. It will change, I think, best practices that people have with these systems. I think that we actually don’t quite understand what this interaction looks like. But it’s extremely important to understand it.

The same way that you can get Claude to ask you questions to more clearly specify what you’re trying to do, and that leads to a better outcome, I think we’re going to need to build ways that these systems can try to elicit from the person the actual problem they’re trying to solve, rather than going down a freewheeling path together.

Because in some cases, especially people who are going through some kind of mental crisis, that is the exact moment when a friend would say: This is nonsense. You are not making any sense. Take a walk and call me tomorrow. Or: Let’s talk about a different subject. I don’t think you’re reasoning correctly about this.

But A.I. systems will happily go along with you until they’ve affirmed a belief that may be wrong. I think this is just a design problem, but it will also be a social problem that we have to contend with.

I just wonder how much it will be a social force. I think we’ve given a lot of attention — correctly so — to the places where it moves into psychosis or strange A.I.-human relationships. We’re seeing it through its most extreme manifestations, and those will become more widespread. I’m not saying they’re not worth the attention.

But for most people, it is just going to be a kind of a pressure — in the same way being on Instagram, I think, makes people more vain. In the same way that we have become more capable of seeing ourselves in the third person. The mirror is a technology.

Mm-hmm.

I mean, honestly, it’s funny that in the myth of Narcissus, he has to look in a pond. It was actually quite unusual to see yourself for much of human history.

So when the mirrors came out, they were like: Oh, this is going to lead to some issues. [Laughs.]

There’s a lot of interesting research on how mirrors have changed us. As somebody who believes the medium is the message, A.I. is a medium, and it will change us as we are in relationship to it. Probably more so than other things because it is this kind of relationship that has a mimicry of an actual relationship.

Yes. I’ve used these A.I. systems to basically say: Hey, I’m in conflict with someone at Anthropic. I’m really annoyed. Could you ask me some questions about that person and how they’re feeling to try to help me better think about the world from their perspective?

That’s a case where I’m not using the technology to affirm my beliefs or show I’m in the right, but actually to help me try to sit with how this other person is experiencing this situation.

It’s been profoundly helpful for then going and having the hard conflict conversation, sometimes even saying: Well, I talked to Claude, and me and Claude came to the understanding you might be feeling this way — do I have that right?

Sometimes it’s right, but sometimes when it’s wrong, it’s really helpful for that other person to have seen me go through that exercise in empathy and spending time trying to understand them before coming into the conflict.

Do you have strong views on how you want to parent in a world where A.I. is becoming more ubiquitous?

Yes. I have the classic Californian technology executive view of not having that much technology around for children. But I was raised in that format, as well.

We had a computer in my dad’s office. My dad would let me play on the computer, and at some point he’d say: Jack, you’ve had enough computers today. You’re getting weird. And I’d be like: I’m not getting weird. No, no, you have to let me. And he was like: See? Being weird. Get out.

I think finding a way to budget your child’s time with technology has always been the work of parents and will continue to be. I recognize, though, that it’s getting more ubiquitous and hard to escape.

We have a smart TV. My toddler can watch “Bluey” and a couple of other shows, but we haven’t let her have unfettered access to the YouTube algorithm. It freaks me out. But I see her seeing the YouTube pane on the TV, and I know at some point we’re going to have to have that conversation.

So we’re going to need to build pretty heavy parental controls into this system. We serve ages 18 and up today, but obviously, kids are smart, and they’re going to try to get onto this stuff. You’re going to need to build a whole bunch of systems to prevent children from spending so much time with this.

All right. I think that’s a good place to end. Always our final question: What are three books you’d recommend to the audience?

Ursula Le Guin’s “A Wizard of Earthsea.” It was the first book I read. It’s a book where magic comes from knowing the true name of things, and it’s also a meditation on hubris — in this case, of a person thinking they can push magic very far.

I read it now as a technologist thinking: Uh-oh. [Laughs.]

Eric Hoffer’s “The True Believer,” which is a book on the nature of mass movements and the psychology of what causes people to have strong beliefs. I read it because I think that A.I. technologists have strong beliefs and may be part of a strong culture that includes the word “cult,” so you need to understand the science and psychology behind that.

And finally, a book called “There Is No Antimemetics Division” by a writer with the pen name qntm, which is about concepts that are, in themselves, information hazards — where even thinking about them can be dangerous. I always recommend it to people working on A.I. risk as a book adjacent to the things they worry about.

Jack Clark, thank you very much.

Thanks very much, Ezra.

You can listen to this conversation by following “The Ezra Klein Show” on the NYTimes app, Apple, Spotify, Amazon Music, YouTube, iHeartRadio or wherever you get your podcasts. View a list of book recommendations from our guests here.

This episode of “The Ezra Klein Show” was produced by Rollin Hu. Fact-checking by Michelle Harris with Mary Marge Locker and Kate Sinclair. Our senior engineer is Jeff Geld, with additional mixing by Isaac Jones and Aman Sahota. Our executive producer is Claire Gordon. The show’s production team also includes Marie Cascione, Annie Galvin, Kristin Lin, Emma Kehlbeck, Jack McCordick, Marina King and Jan Kobal. Original music by Pat McCusker. Audience strategy by Kristina Samulewski and Shannon Busta. The director of New York Times Opinion Audio is Annie-Rose Strasser. Transcript editing by Sarah Murphy and Marlaine Glicksman.

The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: [email protected].

Follow the New York Times Opinion section on Facebook, Instagram, TikTok, Bluesky, WhatsApp and Threads.

The post How Fast Will A.I. Agents Rip Through the Economy? appeared first on New York Times.