Stanford Professor Erik Brynjolfsson: Why Success Starts With Looking Like Failure

\[00:00:00\] **Johan:** I'm sitting on the campus of Stanford, California right now and I have to admit that there's something really special about this place. You can feel the weight of the volume of ideas and conversations that has really shaped how we think about technology and organization and our future. And in about 30 minutes I'm heading into one of those conversations that I've Jag har pratat med Erik Brynjolfsson, han har spenderat tre decennier på det här sättet studerande en fråga som jag känner att nästan alla ledare som jag pratar med just nu är verkligen krävande med.

Så när transformativa teknologier kommer på väg, varför skiljer några organisationer verkligen genom detta och andra inte? And his research on both AI and productivity has become really essential reading for anyone that's trying to navigate what's happening right now. Men det är något som jag har tänkt på i taxi över till campus som jag känner är den korta tensionen av vad jag vill utforska tillsammans med Erik idag.

Det är lätt att tänka på AI som teknik. Och det är en system 2 och en ekonomisk sätt att tänka. Det väldigt logiskt, vi uppmärksammar det som hur vi kan rädda kostnaderna rädda FTEs och göra saker mer effektiva. Men på andra hand har vi organisationer verkligen strävar med adoptionssidan av saker.

Personligen tror jag att det är för att vi lackar en vision av... Who do we want to become? What's the future that we actually want to build with this new technology? How can we make humans become truly exceptional and not just automated away? I looked for that in the conversation and I really hope that this will be one of those that I will remember for a long time.

Cheers.

\[00:02:02\] **Johan:** Erik Brynjolfson. So nice to meet you.

\[00:02:05\] **Erik:** Well, it's a pleasure meeting you and thank you for pronouncing my name correctly.

That doesn't happen as much in America.

\[00:02:10\] **Johan:** I'm a Swedish guy, so it probably counts a little bit easier. Exactly. It's exciting for me to be here today. We're, we're at Stanford campus right in California, so that's

\[00:02:18\] **Erik:** here We are heart of Silicon Valley. I came here about five years ago and I've been just loving it.

As you can see, the weather's almost always beautiful and yeah. The, uh, the atmosphere of all the entrepreneurship and tech is, is just very invigorating for me.

\[00:02:32\] **Johan:** And before that, you spent almost 20 years, uh, at MIT and

\[00:02:35\] **Erik:** yeah, over 20 years at MITI was a PhD student, I was a professor there. I, I love MIT, I love Cambridge and Harvard and all the amazingly smart people there.

I like to go back when I can, but there's just no place in the world like Silicon Valley. Yeah, it's really unique.

\[00:02:49\] **Johan:** So for those who don't know your work, what is it that you do and why does that feel important?

\[00:02:54\] **Erik:** Well,

I'm a professor here at Stanford and I've been focused on what are the economic implications of.

Information technology, and especially artificial intelligence. Mm. I actually, uh, uh, first started teaching a course on artificial intelligence. Uh, just when I graduated from college in, in the 19 ear, late 1980s, uh, built expert systems, if anybody remembers what they are, they have these rule-based systems.

And we started a company, uh, Todd Borough and I started a company called Foundation Technology. So I've been doing the built expert systems for companies. So we've been doing this for a long, long time. When I decided to get my PhD, I tried to do both AI and economics at the same time, but there was nobody who could really help me with both those things.

And so I had to choose. I chose economics, which was, um, fun for me. I wanted to think about how it's changing the world. So that's what I've been doing ever since. Thinking about the economic implications of ai, um, of course there's never been a time in history like the past few years where AI's just really.

Taking off. I also have a company called Work Helix. You know, I've started a number of different companies and, and Work Helix is the one I'm most excited about ever. Um, and what Work Helix does is it looks at the opportunities where AI can be applied in, in, uh, companies and helps, uh, identify a roadmap for doing that and then measure the benefits.

\[00:04:12\] **Johan:** Wow, that's interesting. Yeah. Measure the benefit is, is on top of everyone's mind. It feels like we're in a, in a kind of an interesting intersection right now where. Gen AI and agent AI is kind of forcing, almost like a moral question within leadership teams. Are we fundamentally, what are we building towards?

Are we building an automation play? We're trying to automate humans as much as we possibly can, and it kind of forces a question of, okay. Then what do we do? Or are we trying to augment humans towards being more and more exceptional? I, I would imagine that you spent quite some time thinking about this.

\[00:04:48\] **Erik:** I, I do.

That is such an important question. It's the heart of a lot of my research. Um, you know, a a lot of people, like I was inspired by the Turing test when I was a kid, when I first heard about this idea. You know, could you make AI that perfectly imitated a human so closely? You couldn't tell which was which.

And I thought that was an amazing idea. Um, now as I think more about it, I think it's a cool philosophical concept, but it's exactly the wrong strategy from the perspective of economics or from business that in fact there's a lot more value that can be created by. Augmenting humans doing things that humans can't do or allowing humans to do new things they never could have done before than there is by imitating them and simply matching humans.

I mean, on one hand, and, and with all due respect to my fellow humans, it's too low a ceiling to just match humans. I mean, imagine if, if Henry Ford, when he was building a car, had said, oh, we're going to make a vehicle that can walk or run as fast as a human. It would've been, would've been kind of a, a, a unambitious goal.

Um, so we wanna do new things that are better than have ever been done before and raise the ceiling. The other thing is, and it's a little subtler, is that if you simply imitate humans, um, then that is. That tends to lower the value of human labor. Machines become a substitute Yeah. For humans. And that drives down wages and at leads to more of a concentration of wealth and power among the people who control the technology and the capital.

And most of us would like to have a world not just of prosperity, of, but of widely shared prospect.

\[00:06:20\] **Johan:** Yeah, absolutely.

\[00:06:21\] **Erik:** So for both those reasons, um, I've been pushing for people to think more about how we can use machines to compliment or augment humans. I wrote a paper called The Turing Trap that lays out this argument in more detail that basically argues that the, the Turing test mentality of imitating humans can become a trap.

And we should really, whether you're a business person or a policymaker or just a a, a technologist, uh, working with the technology, uh, you should be looking to, uh, extend capabilities through augmenting humans.

\[00:06:52\] **Johan:** And what's the trap?

\[00:06:53\] **Erik:** Well, the trap is that if you, uh, imitate make machines that imitate humans, then that tends to drive down the economic wages, economic power.

So that's the trap. Yeah, exactly. And, and with economic loss of economic power becomes a loss of political power. And then you're in this very bad equilibrium where people aren't happy with their outcome, but they have no economic or political power to change things. You see, as long as people, as, as long as humans are valuable, then they have some bargaining power, they can say, Hey, we don't wanna go that way.

And throughout all of history, there's always been this tension between, you know, workers and capital and, uh, citizens and central power. And one of the leverage points that citizens had is they could say, Hey, we don't like this. We're gonna step back. And that gave them some leverage. But if, uh, if people have no economic power or leverage, uh, then they could be trapped in a bad equilibrium forever.

\[00:07:48\] **Johan:** And who gets to make that? Choice because it's interesting, it's happening so fast. Uh, also it's happening across every organization right now, and there's, it seems like a smoother, or at least a faster path that works better in like a quarter economics logic of, of going the automation path. Whereas like the augmentation, it's so much about human change.

Company culture seems. Also more difficult to prove in the short term, I would imagine?

\[00:08:18\] **Erik:** Uh, it definitely is harder to measure the benefits. It requires more creativity. Ultimately, I think it's a lot more rewarding. So I talk to a lot of managers and I think that, you know, especially CFOs who are just looking at, at dollars and numbers, there's a temptation, a gravitational pull towards cost cutting.

How many, how much labor hours can we save? And I, I have to say that it, that can be very profitable at times. Um, but it, it, it doesn't give you much sustainable advantage. The real benefit comes from doing new things and raising the ceiling, having better customer service, new products and services, higher quality.

That requires a little bit more creativity as you're suggesting, but ultimately it's more sustainable. So, uh, I do encourage the CEOs and managers that I meet with to try to resist the easy temptation of just cutting costs. Yeah. And be more creative. Uh, I, I think that, um. With all new technologies. The real value almost always comes from doing these new things from reinventing the organization with electricity, there was a reinvention of factories.

With steam engine, there's reinvention of the whole economy. And with ai, you can get some wins by cost cutting. Hmm. But if a manager's creative and creates new products, services, ways of working, um, they're gonna get more benefit. Um, that said, you know, I, I, I do understand that, um, it, it's more natural to just look at replacing tasks.

It, I mean, it doesn't take a lot of creativity to look at the things you're already doing and think, okay, how could a machine do that or this or that? Uh, it takes more creativity to imagine entirely new things. But when I tell people this story, I, I gave a talk at, uh, icle, one of the big, uh, AI conferences, and it was, uh, and I was encouraging that the AI researchers there to, to think more broadly about.

Augmenting and complementarity, not just, uh, imitation. And so many of them came up to me after my talk and, and they said, you know, they literally said, oh, we didn't realize all the benefits of augmentation. Thank you for telling us. Mm-hmm. We've been focusing on the wrong things. Yeah. But now that you gave us a new target, we can do that.

This is something we can do. And, and, and I think they're right, that they're just been remarkably good, that when someone puts a target out there, that they focus on that and they succeed. You see all these benchmarks being saturated very quickly. And what we need is new benchmarks that are more focused on augmenting, not just automating work.

\[00:10:44\] **Johan:** Yeah. It's interesting. And we also have in, in kind of the private sphere. We struggle to implement the already existing technology in a sense. Mm-hmm. There's a lot of, I don't know, fear perhaps of, of are we automating ourselves away? And that's creating a, a kind of silent pushback in, in terms of engagement and so forth.

Mm-hmm. And I feel that I'm lacking a lot of the, the kind of positive story about what do my work look like as an augment the, the individual.

\[00:11:09\] **Erik:** That's right.

Yeah. I, I, I think that another benefit of the augmentation approach is it's a lot easier to get buy-in from the workforce.

\[00:11:17\] **Johan:** Yeah, exactly.

\[00:11:18\] **Erik:** You know, if they see this as a tool that's gonna replace them.

Of course they're gonna resist it. Hmm. But if they see it as a tool that allows 'em to do their jobs better and allows 'em to do things, um, that's awesome. I, I did this paper with, uh, de Yang, uh, Yia and several other researchers, and what we did was we, uh, uh, interviewed, uh, thousands of workers. We actually had a voice agent interview them, that helped us with the interviews, um, and asked them what kinds of.

Tasks in their job, would they lack help with? Where would they like to have automated? And it was very striking how the paper's called the Future of Work with AI Agents. By the way, if anybody wants to look at it, it was striking how, um, workers, they, they do want agents to help them with a lot of their tasks.

Um, and there's others that they would like to have agents just automate and replace, but the current technology doesn't match up very well with their desires. And if you actually take the time to talk to workers, I think you're much more likely to get acceptance and actually people welcoming it. And if you want, if you're a CEO, um, you know, you, you want your workforce to be buying into this.

Hmm. And so I encourage them to be more conscious about taking the worker's voice into account.

\[00:12:27\] **Johan:** When I did the research going into this interview, I, I stumbled across your J Curve. Mm-hmm. Can you explain what that is?

\[00:12:35\] **Erik:** Yeah. Well, um. Like most people, I'm blown away by the core AI capabilities. At the same time, I'm disappointed that we're not seeing real business value.

Yeah, the official productivity numbers aren't really picking up. Um, and this is a pattern actually. You see over and over with earlier technologies. I mentioned electricity earlier. It took about 20 to 30 years before you saw a big gain in productivity from electricity. And when I studied this as a graduate student, um, what we found was that the first factories electrified, they simply pasted electricity onto the old ways of working.

They didn't really rethink things and it didn't lead to much of a productivity gain or really any at all according to the official records. It was only after they kind of reinvented the factory. Instead of having all clustered around a central power source, they had it. Each piece of a machinery had its own separate electric motor, which allowed them to lay it out, um, over an acre or more.

Um, have, and have the equipment laid out based on the, uh, flow of machine flow of, uh, materials as opposed to who needed power. Um, when they did that new layout, they had a doubling of productivity, even a tripling of productivity in some of the factories. Hmm. So the lesson there is not that electricity was a dud or it was overhyped.

The lesson is that electricity, I itself didn't give you the big benefit. Now, coming to your question about the J curve, during that period when they were trying to figure out better ways of using it, they were inventing new business processes that are re-skilling the workforce. Um, there wasn't much of an increase in output, but there's a big.

Effort and time of management, even consulting work that showed up as more inputs, no increase in output. In other words, lower productivity. Yeah. Later, once they had it figured out, then that productivity took off and if you map it, it looks like a J at first it's kind of flat or down even, and then it takes off.

We saw this pattern over and over with different technologies and this paper, the productivity J curve, we mapped out, you know, formally and mathematically why you would expect this pattern and to bring it back to ai. Right now, I think we're in the early stages or actually the, the turning point of that J curve.

I think we're just beginning to have some takeoff now after a little bit of a lull.

\[00:14:51\] **Johan:** How do you know if you are on that path versus that you are kind of just floundering around doing bad experimentation and bad implementation?

\[00:15:01\] **Erik:** Yeah, well. In, in the practical world, you know, both things are happening. Yeah, I've certainly seen some companies that are founding around not using it effectively.

I've seen others that are doing genuine reinvention. Uh, I do feel like we're doing things a lot better and faster. Today in 2025 than happened with electricity or with the steam engine? Some of the earlier technologies, uh, part of it is the technology itself is just easier to implement. It doesn't require as much special equipment.

You know, you can run, uh, you know, chat GPT or Cloud on or Gemini on a, on an iPhone. Um, and, uh, you can roll it out very quickly. You don't need a lot of special training, so that helps a lot. The other thing is just, um, you know, managers are, are much more scientific, I think, and careful about implementing things.

They've got consulting companies that help help them out. Um, they've got business schools, you know, they're just being much more disciplined about it. Back in, you know, the late 18 hundreds or the late 17 hundreds, um, there wasn't as much of a science of management. So for both those reasons we're speeding things up.

\[00:16:06\] **Johan:** Yeah.

How would you measure the actual quality that you get out of ai if, if we assume that just having the C-O-C-F-O. Numbers game is so easy to kind of gravitate down towards. We, we eliminate hours. Mm-hmm. But that's mm-hmm. As, as you were talking about earlier.

\[00:16:23\] **Erik:** Yeah. I, I think in general you want have a, a, a dashboard, a suite of different KPIs.

Yeah. I mean, to be concrete, uh, you know, one, one of the studies we did that just came out earlier this year in the quarterly Journal of Economics was called, uh, generative AI at Work. Mm. And we did a really in-depth study of the rollout of Gen AI in a call center, a contact center. Um, and we measured about a dozen different KPIs.

Uh, we measured, uh, average handle time of the call. We measured customer satisfaction. With that promoter scores, we measured customer sentiment. That is, if you looked at all of the transcripts of the discussions, how many happy words, yeah. How many angry words we looked at employee turnover, uh, et cetera.

And you could see these different metrics. Some of them improved a lot, some of them improved a little. Actually, I. I was surprised how almost all of them improved significantly. I had, I've done a lot of these studies, you know, since I was a, a student and I hadn't seen any case where you had so many metrics improve so rapidly.

Um, across the board, just within four or five months, you had double digit productivity gains on almost all these metrics and also these different groups, you know, the, the productivity, the stockholders were doing better. Um, in terms of more efficient company. The customers were happier. It wasn't coming at the expense of customer satisfaction.

Actually, they had higher customer sentiment. Um, and even the employees were happier. It wasn't like a, this was an electronic sweat shot that was just squeezing the workers. Uh, employee turnover actually went down. So this was a case where there's a real win-win, but you need to do these, uh, these multiple metrics.

The other thing that's important is you need to. Do causal estimates, not just correlations.

Okay.

So in, in, in economics, there's been, over the past 10 years or so, there's been something that's called the credibility revolution.

Mm-hmm.

And the idea is to take the, we've all had the saying, you know, correlation is not causality.

Yeah. Um, but most of economics for the previous a hundred years was just looking at correlations. Most of business was looking at correlations. Uh, the credibility revolution helps show when you can actually get causal estimates out of it. Uh, a controlled experiment, uh, uh, randomized controlled treatment, uh, study, um, that can give you causal estimates.

That's kind of the gold standard.

Um, unfortunately, business, it hasn't really caught on in business yet. Mm-hmm. And one of my missions is to bring the credibility revolution to business. My company work helps companies do that kind of careful causal estimation.

Um, I can give you an example Yeah. Of a company that didn't, I think it, I mean, just to make it a little more concrete, we visited one company. And, uh, they were very happy that, uh, using an AI tool and they found that the workers using the AI tool were about 41% more productive on several different measures than the people not using the AI tool.

And they thought this was a big win. But then we looked at the data a little bit more closely and we saw that the workers who were using the AI tool were actually more productive even before they got the tool.

\[00:19:21\] **Johan:** Okay.

\[00:19:22\] **Erik:** So they had already been more productive. Yeah. This was not like a, a random sample of people, what we call selection bias.

Yeah. Um, your most forward looking workers were the ones who first adopted the technology. So when we went back and did things more carefully and we controlled for it. Luckily it was still productive, but it was only 11%, not 40, uh, 1%. Um, so most of the effect was due to the selection, but some of it was due to the tool in fairness.

Um, but that's the kind of thing you have to do carefully to really understand where the benefits are. And, and by the way, theoretically, and in some cases it's even possible for that number to be reversed. You could actually have it, the tool have a negative effect in reality, but look like it's positive if the selection bias is high enough.

There's all sorts of errors like that that you have to be, be careful about.

\[00:20:05\] **Johan:** I've always been kind of interested in how come we we're seeing. In relative terms. So low productivity gains, we're talking, if you're talking about 10, 12, 15%. Mm-hmm. It seems like the potential in the already existing technology is quite massively larger than 11%.

\[00:20:23\] **Erik:** Well, you know, i I, it is sort of a, a, a mindset difference, uh, for economists. If we do one or 2%. Yeah, that's Sure. We get very excited. I have a bet with Bob Gordon. Um, the Congressional budget office in the United States has been predicting about 1.4% productivity growth for the rest per year, for the rest of the, uh, the decade.

And I made a bet with Bob Gordon, there's gonna be 1.8%, and he thought that was like, oh my God, that's so high. I'll take the other side of the bet. So, um, you know, that's four tenths of a percent. I think actually we're already on track. It's gonna be more like double more like 3%. Mm. So it is sort of a, a matter of calibrating what you think is a big deal.

Um, there are certainly some particular implementations that have, you know, 30, 40, 50% or a hundred percent productivity gains in, in narrow things. But when. Go across the entire organization, um, you know, getting, you know, 15% productivity is actually, is actually pretty good. Um, and so of course, hopefully you can do that year after year and continue to compound.

Um, I'm, compared to most economists, I'm fairly bullish. Hmm. Uh, saying productivity could double to 3% per year. Um, but a, a lot of economists think it's much less. I, I do notice a difference when I talk to technologists versus economists or when I talk to people in the Bay Area in Silicon Valley versus people on the East Coast or in Europe.

\[00:21:39\] **Johan:** Are you optimistic for the kind of future of jobs? Is it a, a. So what's your kind of bet on where we actually will end up?

\[00:21:47\] **Erik:** Yeah. Well, I'm very optimistic about productivity. Yeah. Um, at least compared to most economists.

And I'm optimistic about a lot of wealth creation. I'm optimistic about the potential of AI to really transform the economy. I'm worried about jobs. Yeah. Um, I, you know, I, I think that there's a scenarios where it becomes very troubling. Uh, I did a paper recently where, where we found some significant negative effects on jobs.

Um, I would say that, um, a lot of this is gonna depend on our choices that we can. Uh, design systems that are more likely to augment and lead to, to job growth. Uh, we can also design systems that lead to a lot of, uh, you know, job loss and falling wages. Um, so this is something I, I'm pretty worried about and there's no guarantee that the next five or 10 years are gonna be good for, for most workers.

It's something we're gonna have to work on.

\[00:22:36\] **Johan:** What are the most important of those choices?

\[00:22:40\] **Erik:** Well, one of them is, is augmenting versus automate. We already talked about, and let me just show you how the data turns out. So we did this paper, um, we call it canaries in the coal mine, you know, the early warning.

And we looked at what was happening to different job categories. And what we found was that overall the labor force was kind of noisy and, you know, having ups and downs, but if you zoomed in on certain categories, we found falling employment for early career workers, especially in highly exposed occupations like coding and call centers.

We found rising employment in the occupations that were not exposed, like, uh, home health aids, but most interestingly, um. If you divided the workers by how they were using ai, where they were using it to automate tasks versus to augment or complement tasks, you got very different trajectories. The people who were using it to automate tasks saw falling employment.

The people who were using it to augment work, saw growing employment and growing productivity. Hmm. Um, so that's an example of a choice. Augment versus automate that leads to different outcomes.

\[00:23:47\] **Johan:** It's interesting. We got, we got a couple of, uh, brain studies coming out of over this summer as well. I think it's kind of a similar thing happening there.

Are you choosing to completely, um. Delegate your thinking to ai or are you using it as a central part of the kind of creative process? So its the same, it's the same choice really.

\[00:24:06\] **Erik:** It is the same choice. And, uh, I, I think the outcomes are a lot better when you, when the human stays involved. The other thing I have to say is that to some extent, although I, I love, you know, age agentic, AI and, and the power of AI and lots of things, in some ways it's been oversold as being able to do everything.

Mm. And what we found was that in many cases, um, AI still able to do a lot, but there are just some things that really are best left to humans.

\[00:24:32\] **Johan:** Is that a, a question of AI in the fall of 2025? Meaning that, well, it, it's gonna, the technology is gonna so rapidly evolve, so, so that's just a, a, yeah. A one year statement, or is it always gonna be true, do you think?

\[00:24:46\] **Erik:** Well, I wouldn't say either that it's a one year statement or that it's always gonna be true. It's somewhere in between. Uh, I, I, I do think that, um, over time, AI will, is becoming more powerful and be able to do more and more tasks, uh, is not, it's gonna take a lot more than one year for it to be able to do the full set of tasks that humans are gonna be able to do.

I don't know whether it's it's a decade or, or, or more, um, at, at the same time. Um, it, it, it is the reality that right now when companies roll these systems out, um, most of the times they're gonna have more robustness by keeping humans in the loop or humans on the loop. Um, AI is great, especially machine learning when you have lots of training data.

Yeah. Um, and so lots of examples, but if it's something that hasn't appeared in the training data before, it's an exception, um, then. Machines have a lot of trouble. I have to say humans aren't so great either, but, but, but we're better than machines. Yeah. You know, um, improvising is one of our superpowers.

We're able to do things that, that you just figure things out on the fly and in the real world you end up seeing a lot of these long tail one off exceptional tasks, and that's where humans can be stronger. So there's kind of a, a division of labor there.

\[00:25:56\] **Johan:** Do you see in, in your mental model of, of kind of the augmentation human plus machine E equals what we're, what we really want to get to, do you see that there are, are, besides, um, handling the, the exceptions, are there other things that humans should focus on?

If we almost give career advice to, to people listening, like, what are the skills that I currently possess that I should kind of double down on because they're future proof as a consequence of the technology.

\[00:26:24\] **Erik:** Yeah, I can't promise anything that's completely futureproof. Mm. I, I wrote a book with Andy McAfee called The Second Machine Age, and we described some of the skills, and, and I think, and mostly I feel like it held up reasonably well, but at the same time, things have changed since we wrote that book and, and there's this constant evolution.

So I don't think anyone could write down a list that will be true forever. Um, but I can tell you circa 2025 some of the things that, that I'm seeing. Hmm. Uh, one way to think about it is if you divide. Um, most tasks can be divided into three C3 parts. Uh, there's defining the problem, asking the right questions.

There's executing on that to answer that question or address that need. And then there's verifying checking. Is this what you really want? How do we need to modify it? Humans can do all three of those parts, and we've been doing them for, since we've been humans. Um, but machines are getting very good at the middle one.

Yeah. Once it's well-defined, they can execute it. And so they're blowing away some of these tests and that suggests that humans are gonna disproportionately have a comparative advantage in the first and the third. Mm. So we need to be better at asking the right questions to defining the problem. You know, speaking of second machine age, there was a quote in that book that I like a lot.

It was from, uh, Pablo Picasso, and, um. He was shown an early computer, and he looked at it for a while and then he said, huh, you know, that's not very interesting. All it does is it gives you answers. Mm. And, uh, and you know, there's some truth to that, that that really the most important thing is asking the right questions.

Mm-hmm. And that's becoming even more important. Now. It's not just prompt engineering, but it's just more, you know, and the grander scale. What is it that we wanna aim this incredibly powerful technology at? And then on the other side, we all know that machines, you know, they, they still hallucinate and they, sometimes you misalign, they, they aren't doing exactly what you meant.

Maybe they're doing what you literally asked for, but then you realize that's not really what I wanted. Um, and so there's, it becomes kind of an iterative cycle where after you, you get the output, you verify it, you, you iterate. I, I do this all the time. When I use work with, you know, cha EPT or Clot or, or Gemini, I, um, I will, you know.

Have it answer a question, I'll realize that's not quite it and I'll go back. Yeah. Sometimes I'll, I'll notice that it's made a mistake, you know, that I, I know some of the economic literature. I was recently talking to it and he gave me a, a citation that was exactly what I was looking for, and then I, it, it, it was from Eric Olsen.

I said, are you sure that Eric Olson wrote that? He goes, well, okay. No, it, it didn't, Eric didn't write because I knew I hadn't written it. Yeah. He said that's the kind of thing that he would've written. I was like, yeah, it is the kind of thing I wouldn't written, but I haven't written it yet.

\[00:28:52\] **Johan:** So it was good though.

So I'll have written in the next book.

\[00:28:54\] **Erik:** I, I know, I know, but you have to like, you know, be familiar with the, the, the topic area to be able to push back sometimes. But that kind of give and take, I find that you can get a lot further than you could by just throwing it over the wall and saying, okay, ai, you do it.

And, and many, many, uh, problems are like that. And, and I think these are. Skills that can be learned. We can have, uh, students, kids, adults learn how to get better at asking the right questions, how to be creative, how to, how to be careful about verifying things and uh, and working in teams. Those are some of the things.

I mean, there are other areas, uh, just to throw in a few others skills. You know, I think a lot of interpersonal interactions, a lot of people prefer to interact with other humans. Um, I think. The whole entertainment economy where we're like, um, watching sports or chess with other real humans is something that people will have a preference for.

Um, so there are, there are aspects of the economy that I think we'll see growing over time that really require the human touch. Um, but I guess my last piece of advice on that is we just have to be very nimble and keep updating because it, it's constantly changing. Hmm. The, the era where you could learn some skills, you know, when you're 18 years old and then just do the same thing for the next 40 years.

Yeah. You know, that era is gone and gonna have to be much more nimble.

\[00:30:10\] **Johan:** I think it's so fascinating how fundamentally a, a question of technology is, at least for the people that I speak to, that I, that I respect. Know, they know their things around ai. Mm-hmm. It almost always comes down to a question of, of the essence of humanity.

Yeah. And I think it's fascinating that kind of a technology, uh, technological question drives that discussion. So, so

\[00:30:34\] **Erik:** it is, it is, it is really interesting. You know, philosophers for a long time have been debating, you know, what is the essence of humanity? What is a good life? Yeah. What is our ultimate goal?

And it used to be just sort of a, a philosophical discussion. Not, no, no insult intended, but something that philosophers would talk to each other. Now it's really becoming a very practical Yeah. Question that you go to the AI labs and they have people working on alignment research and it becomes really relevant.

What is the goal that we're trying to get these machines to do? Um, if we had something very capable of carrying out, executing our, our, our desires, we better be clear about what those desires are, what our values are.

\[00:31:12\] **Johan:** I think it's, uh, it went around on LinkedIn a, a year ago. Uh, but it was something along the lines of I expected like when, when the robot entered to, to not having to do the dishes and clean up, but to be fully focused on my creative side.

But what ended up happening is that my creative side get automated, but I still do the dishes.

\[00:31:30\] **Erik:** Exactly. That is, I did see that. It it is, uh, it is ironic. Yeah. That sometimes it's not matched up. And that paper we wrote, the future of work with AI agents highlights that we had kind of a, a, a two by two there.

What are the things that AI is good at and what are the things that people want them to do? Yeah, exactly. And you could have high, low and, and on each dimension. And so there's sort of four quadrants. And the sad thing was that the, the data was almost evenly distributed through all four quadrants. There were certainly examples of things that AI was good is what we wanted it to do, but there are also examples of AI doing things that we didn't want it to do, and people wanting AI to do things that, that it wasn't yet good at.

So, uh, uh, we're hoping that that research will help align people as they de develop AI to focus it more on the things we really care about and not the things we don't.

\[00:32:19\] **Johan:** I would imagine that you have plenty of discussions with the really him high impact people, people with power. Mm-hmm. How do they feel?

Or they do, they reflect a lot around the responsibility for this question that we're currently in? Like, do, do we go the automate or augment path and, and like, what, what's my role in, in building a future that we all wanna,

\[00:32:40\] **Erik:** I think they do. I, I, I, I think that that. They, they, you know, when you're, you're talking about them informally, that they, a lot of them do care about these things.

At the same time, I think there's some economic pressures that drive them in certain directions and, and sometimes I feel like their actions don't match up with their words or maybe even with their own desires. Um, I think, uh, I've got the sense none of them have said this to me explicitly, but that they feel a little bit trapped in a, um.

Uh, a, a race to the bottom. Yeah. Or a prisoner's dilemma where they would like it if everybody kind of agreed to work on some of the higher purposes, but they're being pulled in another direction. I mean, to give you, actually, one person did say something a little, you didn't say it this way, but, but what, what I, I'll try and quote him accurately.

Uh, de um, at the Paris Action Summit, Hmm. Uh, everyone had like a little five minute chan a chance to give a five minute mini speech about what they thought was important. And it was striking to me that Demis, you know, the head of, uh. Google DeepMind, you know, Nobel Prize winner, one of the best researchers in the field.

Um, took his five minutes. He used the whole time to advocate for kind of a, a, a CN entity. Uh, you know, the, uh, the particle physics laboratory where, where everyone cooperated. And he was hoping that just as the particle physicists worked together from all the different countries, all the different labs, to try to understand the nature of the universe and the nature of particle physics, that something similar could be created for ai.

That the Frontier labs, the governments, the researchers could come together and coordinate on research. Um, it seems like we're pretty far from that actually really happening. Mm. Um, but it wasn't just. Demis who said that when I talked to a lot of other researchers, informally, I won't say their names, um, they expressed a similar desire and sentiment.

\[00:34:28\] **Johan:** So what do you think that it's unlikely to happen? Is it because a lot of the actual research is, is not done by state funded, uh, schools. So it's more done by, by like. Open AI or stuff like that? Or is it more

\[00:34:42\] **Erik:** Yeah, I question think it's impossible. I mean, so I don't wanna imply that there's never going to happen because these, a lot of these leading folks would like to see it happen at the same time.

There's a really strong economic incentive Yeah. To, you know, if there's literally hundreds of billions, maybe trillions of dollars at stake. If, if you can have a more powerful AI system and, and you don't share it with other people, um, there's a lot of economic value in that. There's a lot of literal power in that, you know, uh, power to an individual, power to a government.

Um, the reality is, is that these technologies are generally dual use, meaning they have amazing civilian uses that can make us, you know, wealthier, healthier, um, you know, happier. Cleaner environment. They also have military uses. Um, they, you know, right now I'm told that, uh, um, the majority of the deaths on the battlefield in Ukraine are from drones.

Mm. And of course, there's cyber technology. Uh, I'm very worried about, um, uh, biological Yeah, sure. You know, viruses and other pathogens. Um, so as you get more powerful technologies for these good uses, it's often not very hard to also apply them to, uh, some of the military dangerous uses. And that leads to, uh, uh, you know, a geopolitical tension as well.

\[00:35:58\] **Johan:** Yeah.

And the, and the modern bad guy is also a knowledge worker in a sense.

\[00:36:02\] **Erik:** Yeah.

Uh, yeah, exactly. And, and so, I mean, with nuclear weapons, it, it, it's pretty hard to make a nuclear bomb. Um, and so you have to have a lot of specialized materials and very expensive equipment, uh, with ai. We may be giving the capability for really vast levels of destruction to, you know, somebody with a, a laptop computer, um, to make a new virus or, or something like that.

Um, that is a kind of a scary possibility.

\[00:36:31\] **Johan:** Yeah.

How do you see kind of your role in this? What, what do you wanna be the force for?

\[00:36:37\] **Erik:** Well, my comparative advantage, I think is in studying the economic side of it. I care a lot about some of the more catastrophic risks and I support the people who are, are working on that.

Um, you know, Joshua Bengio, um, but I think he's now the most cited, uh, scientist in all of history. Um, has devoted a lot of his life to that. And I was just had dinner with him the other night and, and support, uh, what he's doing as are a lot of other top researchers. But for me, I'm focusing on the economic implications.

Right now, what I see is our capabilities just taking off. Yeah. We've never seen such rapid progress to have really just immense power. And, and that's in part because we have so many resources going into it. Literally hundreds of billions of dollars and you know, I don't know how many IQ points. Mm-hmm. A lot of small, smart people working on it gravitating at the same time.

Look at the economic side, it's barely, you know, moving. There's very few people. Yeah. Certainly not billions of dollars. Um, so my mission is for me, myself to, to study that and bring the people here at the Stanford Digital Economy Lab to study it. We've got some amazing researchers who are working on it.

I'm also trying to change the field, uh, two weeks ago. Along with Ajay Agarwal and, uh, Anton Knick, I organized, uh, the first ever NBER workshop on the economics of transformative ai. And NBER is the National Bureau of Economic Research. It's kind of the premier gathering of economists on different topics, and it was a big milestone when they were willing to support us having a, a, uh, a workshop on the economics of transformative ai, not just regular ai.

We call it transformative ai, and we define that as AI that's powerful enough to transform the economy the way that the agricultural revolution became the industrial Revolution. Or to be more concrete, we borrowed, uh, Dario Am Moe's, uh, definition of powerful ai. Dario is the, uh, founder of, uh, anthropic.

Hmm. Um. CEO. And he wrote a, a terrific paper, uh, last year called Machines of Loving Grace. And I gave him a little bit of advice as he, as he wrote that. But he had a definition of powerful ai, which was, it was like taking, creating a country of geniuses on a data center. Hmm. So you have all these geniuses, Einstein level, you know, Nobel, PhD level geniuses in physics, in chemistry and management, marketing, logistics, economics, all these different topics.

And you make millions of instances of each of them, and you have them each running, you know, 10 or a hundred times faster than a human can think. Yeah. Just imagine what that does for the economy. Uh, Dario thinks it's only a couple of years away. I, I'm not sure how long it will take, but I'm already seeing glimmers of it.

And so our mission, our, our, sorry, our, um, uh, what would you call it? Our charge to all of the economists who are gathered, uh. Here at Stanford, we had, uh, about 16 presentations was to imagine that we had this country of geniuses in the data center. What would that mean for whatever topic area you're working on?

Hmm. And we, and then the following week we came out with a paper called the Economics of Transformative AI or Research Agenda where we, where we laid out a set of questions for nine big areas like economic growth and productivity, uh, inequality, concentration of power, how you measure, and what the meaning of, well, you know, wellbeing and, and welfare is catastrophic risk.

And now people are working on all of those topics.

\[00:40:10\] **Johan:** I think it's fascinating how we as humans, we're, we're kind of really bad at understanding those kind of exponential curves, and it must be kind of thrilling to be part of the thought experiments of what could we, like, what are the problems that we could tackle?

\[00:40:22\] **Erik:** That's right. Yeah. We need to get, start doing it in advance. We can't wait until we're there. Yeah. Uh, you know, uh, Ernest Hemingway, uh, once described how people go bankrupt. He said, slowly at first, and then suddenly. Suddenly. Yeah. Um, and that's the nature of the exponential curve. And, and we have to be careful that we don't fall into that trap where we, things happen slowly at first and then suddenly before we're ready.

There's a complete transformation. Big changes in, I don't know, unemployment or wealth, um, economic power, um. And that's why we want the researchers to work on it today. And, uh, we're sounding the alarm that they have to start paying attention to it. Um, that gap between the capabilities and our understanding is just getting bigger and bigger and we need to, you know, speed up the, uh, economics of it.

So, so my mission is, is to help transform the economics profession, lead that charge towards, uh, understanding the economics of transformative ai.

\[00:41:19\] **Johan:** That's cool.

That's a cool mission. What do you see in terms of, one of the things that I worry about is like this, the rate of change would probably mean that, that.

Certain level of, of professionals will be quite all of a sudden out of jobs. And it, and it's different, uh, types of people than are normally out of jobs. If you'd say it's knowledge workers who went to college and and stuff like that. And, and what is the actual impact on, on our societies and in what timescale and what's gonna happen after that.

How do you see this playing out from, from like the macro perspective?

\[00:41:56\] **Erik:** Well. Current AI certainly has a very different set of capabilities than the technology of 20 years ago, or 50 years ago, and it's affecting different tasks and different skills. And you could see that, for instance, one of my, um, early papers, uh, I did with Tom Mitchell was what can Machine Learning do?

And we kind of ranked all the tasks, 18,000 tasks in O net as to what, uh, think could be Daniel Rock, um, also, uh, on that. And then later, Daniel Rock and a team at OpenAI did a similar analysis for, um, LLMs in particular. Mm. Um, and what you could see is exactly what you're saying, that a lot of the effects were in more high paid jobs, uh, professionals.

Um, so there's kind of an inversion where what we may see is that the, the doctors are increasingly replaced, but the nurses are still needed. And the, you know, uh, lawyers and the accountants, um, find that they have trouble, um, adding value, but the carpenter and the plumber becomes more and more valuable.

Um. And that's partly because the cognitive tasks are having much more AI's having much more success at doing cognitive tasks than physical tasks. Now, robotics may be coming along as well a few years later. And then, you know, it, it's a broader effect. And so the initial effect will be some disruption of the labor market.

Some jobs becoming more valuable, others jobs becoming less valuable over time. It could spread across the entire labor market. And we need to think about a new economic system that isn't so based on labor income, but based on other things, maybe universal basic income or, or, uh, Nicholas Berg has an idea of universal basic capital.

Yeah. Uh, these are some alternative approaches.

\[00:43:37\] **Johan:** That's what do you see yourself doing in, in like 10, 15 years?

\[00:43:45\] **Erik:** Well, I hope I can help us manage this transition because, I mean, the good news is we have a. Amazing technology that allows us to change the world in ways we've never seen before. And because of these exponentials, a lot of it's gonna be hard to predict.

Hmm.

So we're trying to have a, a, a dashboard that measures the changes in close to real time. And then I hope we can help guide it towards some of the good outcomes if we do it right. I think the next 10 years could be the, the best decade in all of human history. Uh, but if we do it wrong, it could be literally the worst.

Mm. It could be just, uh, terrible economically and on other dimensions. So I don't think we've ever seen such a sharp, um, divide between the, the good outcomes and the bad outcomes as we have right now. In a way, it really relates back to this fact that, um. We have very powerful technologies and, and the more powerful these tools are, the more we can change the world.

I mean, almost by definition. And that may, you know, there's only so much you can do with a spear or a shovel. You can do more with a, uh, uh, a bulldozer or a nuclear weapon, and you can do even more with, with artificial intelligence. And, uh, if we guide them correctly, we're going to have a lot of good outcomes.

Uh, I, my mission, as I mentioned, is to understand the economic implications of this. Hmm. It starts with basically science. So I'm not as focused on the policy right now. Hmm. I'm focused on the understanding the economic implications and, um, what the leverage points are. And then from that should follow some policy implications.

\[00:45:18\] **Johan:** Do

\[00:45:18\] **Erik:** you have kids?

Yes, I've got, uh, I've got four kids.

\[00:45:21\] **Johan:** How old are they?

\[00:45:23\] **Erik:** They're all in their twenties and early thirties now.

\[00:45:25\] **Johan:** Ah, okay. Because I've been thinking, so my kids, how you much younger, so. Mm. Happily, I don't have to answer this question just yet. Mm-hmm. Like, what is the recommendation I would give a 15-year-old in terms of Yeah.

\[00:45:36\] **Erik:** It's tough. It's really tough. I mean, look, the thing I think you should always tell your kids is, you know, do things you really enjoy. Yeah. Be passionate about it. And I think it's even more true now that if you're just, you know, clocking in, doing something not excited about, you're not gonna be that good at it, and you're not gonna really make a difference.

Um, in order to make a difference, you have to be like, passionate about something. And in a world of ai, I think, uh, a lot of the returns are gonna look more like a power law that is, there's gonna be a long tail of like sort of average or low performers. And there's gonna be a few that really make a difference.

And to be in that top of that tail, you have to pick some area. And there's many different areas you can do that in. Um, you have to be something that you're, you're, you're passionate about and you can make a difference. Uh, I think, and this is not, uh, the deepest thought, but using AI is a great way to, to help explore that.

Yeah. Overcharge yourself. Yeah. Um. People who work with AI are able to just accomplish so much more. I got a, an email from a friend of mine, uh, recently, and he said he had, uh, just the past two or three days, he had written 50,000 lines of code Yeah. With, uh, Claude Code. And, you know, he could never have done that before.

And he, he said, you know, I'm a true believer now. I, I see, I see the potential of it. And it's not just writing code. There are so many things where you can be tremendously leveraged. So I think it's, it's something we really need to, that I, you know, I encourage my kids to, to, to dive in and use the tools because it helps 'em fulfill their other interests.

\[00:47:04\] **Johan:** Yeah. It's interesting.

I had a period during, uh, during the spring where I found that working with ai. Was kinda hyper charging my ability to go deep into different subjects quite quickly. Yeah. But I was also context switching so much faster because I, I started a deep research agent and while that was doing something, I had to jump down into something else and I realized it was, it was quite exhausting working that way.

'cause my brain works best in, in like longer periods of deep focus. So I had to mm-hmm. Really think about mm-hmm. While I charge fully ahead mm-hmm. Diving deep into these tools, I also need to, to really guard like the biological side of me and, and creating like biological

\[00:47:45\] **Erik:** I agree. You know,

I think that these we're not used to working with these tools and they have some really awesome.

Benefits, but they also can have some pathologies. And you have to be careful. It's like being in a candy store and there's, you know, there's sugar and fat and you know, protein, whatever. And if you're not careful, you end up getting sucked into the, the, the things that are unhealthy or our brains aren't necessarily wired for a world where AI and social media tug our attention in lots of different directions.

So you have to develop discipline. I mean, you know. Kahneman talks about system one versus system two. I, I find that a very useful framework. You know, system one is your quick instinct. You grab for the cookie, you go click on social media. Um, you see something fun in chat GPT, um, it's satisfying in the moment, but it doesn't really lead to deeper thinking.

System two is more deliberative. Sometimes it's a little bit painful. Mm. But ultimately it's more satisfying. And I think that if you use AI correctly, it can help you explore those system two opportunities. I had a fun conversation recently with, um, bank Holmstrom. He's a Nobel Prize winner in economics, and, uh, he's at MIT, friend of mine there, but he's coming to visit Stanford and we had took a walk around the campus and he told me that every morning, uh, before he even gets out of bed, he sits down and he has a long conversation with Chachi pt.

Yeah, I thought that was pretty funny. I thought maybe he was joking, but then he showed me all of the. Conversations and he was asking chat GPT about, you know, how do quantum computers work and what are the implications for financial markets and all these, exploring all of these different thoughts going back and forth.

And I think that, you know, if you're someone like Ben Holmstrom, this is, uh, just an amazing technology. He's already a Nobel Prize winner. Mm-hmm. But I could see from his conversations, he was diving deeper into all sorts of other topics. So it, it, it, it, it allows people to live their best lives and become, uh, have a better understanding of the world.

But obviously it can and is being used also in lots of negative ways. Uh, people are using it to hijack attention. Yeah. Uh, redirect, you know, to do social manipulation of other people in ways that aren't, you know, aren't very healthy.

\[00:49:50\] **Johan:** Yeah, it's an interesting time to be a curious individual, and that's such a competitive advantage probably in the coming years.

\[00:49:57\] **Erik:** Absolutely. You know, one of the things that we're seeing both for individuals and for organizations is that rather than being a great leveler, it's often something that. Exaggerates differences. So that

\[00:50:09\] **Johan:** Yeah, that's a good point.

\[00:50:10\] **Erik:** Yeah. Yeah. When it comes to companies, we're seeing the top, you know, 10% of companies, uh, kind of pulling away from everybody else, these superstar companies, and, uh, using the technology more and more effectively, taking over market share and industries, and I think for individuals as well.

You know, my son, um, really loves playing with these tools. He was used to play with Khan Academy, and, uh, I remember one, one time when he was a little younger, I said, you know, go and, you know, go, why don't you go spend half an hour in Khan Academy and, and learn, you know, your algebra in high school. And so he did that.

And then I came back to him like two or three hours later, and he was there studying about the Civil war and biology and how cells work. He was just sucked in and absorbing all this information. At the same time, there's other people who, you know, they just don't like learning and they, and, and then. If you, if you don't have the structure of a classroom, uh, they can fall even further behind.

\[00:51:01\] **Johan:** Yeah. I had an interesting conversation with a, with a Swedish entrepreneur who's building a, based on Nick Bloom's, uh, 1984 study, the Two Sigma problem, which was about private tutorings effect on, on student outcomes. Right. So you had two standard deviations improvements based on, on, on private tutoring.

So it's trying to build an AI company being

\[00:51:20\] **Erik:** Absolutely.

That's a great example. Yeah. I mean, um, when people are in the classroom, everyone has to kind of learn at the same pace. Yeah. But if you have an individual tutor, you can, like you said, be two Sigma better. It, it's not cost effective. And, and for, you know, most people to hire an individual tutor for each person, but with AI you can do that.

And Tom Mitchell and I have been playing around with that. Tom Mitchell's a, a head of machine learning at Carnegie Mellon or was, and uh, he's just doing some great work on, uh, these AI tutors. Um, he showed me one. Of course it, it knows the material very well and it can, instead of, if somebody gets the answer wrong, it figures out how they got it wrong.

And instead of, you know, saying, oh, that's wrong, here's the right answer, it'll ask them a new question. Yeah.

\[00:52:02\] **Johan:** It's a Socratic method.

\[00:52:03\] **Erik:** The Socratic method to help them kind of see for themselves. Yeah. Which of course is much more likely to last. But the other thing, you know, and this was a, a a, an eyeopener for me was he could tune it to be incredibly entertaining.

Okay. And so he had it, you know, talking in the style of a pirate. Mm. Or you know, whatever the person, you know, resonates best with the person. And I was just laughing on the floor 'cause it was so funny what he had, what the, what the AI was, was doing. And I could see this being incredibly engaging. We've all had these teachers who are both good at conveying the materials, but also just very fun and entertaining.

Yeah. And so the hope is that everybody will have one of those customized to exactly their interests. Yeah.

\[00:52:42\] **Johan:** That's so true. And I think most people who, who end up being kind of. Enthralled by knowledge, and I want to live in that world and, and, uh, have that experience at some part of their life. There there was somebody who was probably a teacher, or a father or a mother Yeah.

Or something like that, who, who really shaped their way of thinking.

\[00:53:00\] **Erik:** And I think actually if done right, almost everyone can be reached. Yeah. In that way. You know, you've got kids and you know that when they're. When they're two or three years old, if you take a pilot blocks and put 'em in front of 'em on the floor, what are they gonna do?

Yeah. They're immediately gonna start building something. You don't have to say, oh, you have to build something. I, I'll give you, you know, a dollar if you build something. Yeah. You know, it's just, we're wired to enjoy building. Or you give 'em, you know, a paper and pen, they'll start drawing. Yeah. People love creating things and then they go to school and they're told, oh no, you can't draw.

You sit quietly, do this. And it's almost like stamping out creativity. But back to what we were saying earlier about these three different stages, defining the question, executing, uh, uh, evaluating it. Um, you know, we can nurture that creativity. It's, it's already there. And if we can nurture it more, everybody can think more about.

Being creative and then Theis can, can execute and they can evaluate, oh, that's what I wanted. That's not what I wanted. Yeah. You know, the other day I wrote a poem for somebody, um, they were retiring and, uh, I had a lot of like stories and I, I was kind of struggling to structure them all. And so I gave my different stories and things into, uh, you know, catchy t Yeah.

And it, it came up, yeah. I had to iterate a few times to get it right. Um, sometimes it didn't totally understand what I had asked for. Um, but it was, it was actually kind of fun and, and the end product was, was really funny. Yeah. And it was really useful. Um, but it was a, it was a good kind of partnership of.

Taking a, a goal I had and some creativity, and then executing it and iterating.

\[00:54:31\] **Johan:** Yeah, I agree. And it's so fun too, if you split up like the content of what you're trying to say from the tone of what you're trying to say. So first I focus on, on getting the right content. I, I, I write all my stories and whatever, and then you quickly iterate between different tonality styles and, uh, what if you write a, a rhyme on it or what if this is a exactly century poem or something like that.

\[00:54:49\] **Erik:** I tried a few different poem styles. Oh, that's not quite right. And, you know, that's try to be a little funny or whatever. And, and, you know, the process itself was kind of enjoyable for me too. So that, that was, that was part of it as well.

\[00:55:02\] **Johan:** And now you have to run soon, but what is the, the kind of best possible outcome as the, as the last question?

What, what, uh, what does in, in 20 years time mm-hmm. What do we, uh, what does humanity spend our time on?

\[00:55:17\] **Erik:** Well, I do think we are in range of essentially eliminating. Extreme poverty. Yeah. I mean, with AI technology, there should be no reason that anyone is starving or goes without food, clothing, or other basic needs.

Now, of course, there's always gonna be extreme things. You know, I wanna fly to Pluto or something. You know, you can't, you can't satisfy everything. But at least our basic needs, the, the core of Maslow's hierarchy, I think can get taken care of. And it'll be interesting to see how people evolve in terms of what they care about.

I suspect that, uh, status and status hierarchies will continue to be important. Mm. I was talking to Reed Hoffman recently and he said that's, you know, he thinks that's an important part of how we're we're wired. Yeah. You know, from, from caveman days. Um, so be it. Hopefully we can channel that status into something productive, you know, how we help other people.

Um, or even just have fun, you know, video games and not into something that's a negative sum Yeah. Or destructive status game. And so we'll have to think about, uh. If you call it an economy where we, where we channel that in, in a productive way, um, but we could have a much lighter footprint on the environment, on the, on the world.

Um, we could have much better health, probably much longer health spans, maybe a lot of, or most diseases could be cured, have significant extension of life expectancy.

\[00:56:38\] **Johan:** Do you think that's realistic in the 20 year timeframe? Like have that type of massive breakthroughs in, in uh, uh,

\[00:56:45\] **Erik:** you know, I'm not sure. I, I, I know that, you know, I talked to de Abba about it.

He's pretty optimistic about it. Um, you know, maybe there's some selection bias there that he's chosen. But he, he, he did, you know, he had a postdoc at MIT in, in cognitive science. He knows a lot about biology. He got a Nobel Prize. Um, so, um, I wouldn't completely discount it. He may be on the optimistic end.

Hmm. And even if we don't have, you know. Massive increases. I think that there could be a, a lot of things that, that we prove, you know, alpha fold and, and understanding of protein folding is giving us some of the tools. I don't think it's translated into a lot of practical benefits yet, but, but that will come, I think.

Um, so those are some really promising outcomes. What's gonna be tougher is sort of the geopolitics and the psychology. These tools can also be used, you know, as I mentioned earlier, in very destructive ways, militarily, uh, they can be used in ways for social manipulation. Uh, I think it was Sam Altman who said that, uh, super persuasive, super intelligence and persuasion could come before super intelligence and, and sort of ity.

Yeah, that's interesting. And uh, that obviously there's people who would abuse that, you know, for marketing or for. Political or dego, uh, reasons. So this is gonna be tough, tough to navigate. Um, uh, uh, the reason I'm doing what I'm doing is I wanna at least help on the economic side and set up an economic system.

We will, I think, ultimately have to have a, a new economic, uh, system that is much less dependent on labor income. Hmm. Um, and find other ways for people to, to benefit. I think it has to start with the premise that humans are ends, not means, and through most of history, this is what Kant said, that, you know, you don't wanna think of humans as just a means to an end, but an end in of itself.

Itself. Yeah. And so we need to think about, okay, we want, you know, to just, uh, have humans benefit just for the sake of being humans and not just because they're a tool to some other goal. Um, and that's a different philosophy than, uh, we have had through most of history. It can only work that changes everything.

\[00:58:45\] **Johan:** That's super interesting. What, what are you really optimizing for? And within. Kind of optimizing for GDP growth, I would assume. Yeah.

\[00:58:52\] **Erik:** Yeah. And I don't think we need to optimize for GDP growth if we have, uh, machines that can create at least basic abundance. Yeah. And some people, you know, one of my favorite books is Siddhartha.

I dunno if you read it. You know, some people may decide that, you know, enough is enough. Hmm. Um, I feel like I have most of the wealth that I need, you know, you know, I'm lucky. Um, and I spend most of my time pursuing, you know, knowledge and, and I find that more, more satisfying. I'm sure there'll be other people who, who wanna pursue other goals.

Um, so we, we'll have to, we'll have to come up with a, a, a society and economic system where people are, are comfortable, uh, with. Uh, having hopefully a, a broad base of economic shared prosperity and then pursue other goals in ways that are constructive rather than, than destructive. But it's gonna require like a real fundamental reinvention.

And it's, to be fair, it's not the first time you, we, when we, the agricultural revolution change just from a hunter gatherer society in the industrial revolution, at one point 90% of people were basically subsistence farmers. And, and, and now we do a lot better than that. Um, so this will be a transition comparable to that.

\[01:00:00\] **Johan:** Fantastic. Thank you so much for taking the time. It's been an honor to, to get to visit your Stanford.

\[01:00:04\] **Erik:** Oh, such a pleasure. I really enjoyed

the conversation with you. Those are, those are great questions. And these are just very interesting topics, aren't they?

\[01:00:10\] **Johan:** Yeah, they are. And I, I wish you the best of luck and I be, and I wish us collectively the, the best of luck there.

There are truly important questions right now that we haven't figured out, and that's partly super interesting, partly super frightening.

\[01:00:24\] **Erik:** Yeah. It's an amazing time to be alive. Yeah. And, uh, and uh, it's, uh, it's fun to be, uh, one of the people in the middle of all this.

\[01:00:31\] **Johan:** Yeah, I can imagine. It's a good note to end on.