How confident are we that productivity gains from AI actually accrue to workers who can then spend some money on whatever product or service is scarce at the moment, or important to them? I would say not that confident. There's several scenarios out there, and I think that I feel like a lot of economists and just people in general, I think aren't talking enough about is speed. You talk about that if things are fast we need public policy. We need the, the, the new jobs aren't going to come fast enough. Training isn't going to happen fast enough where you're going to get, you know, things are going to get fully automated very quickly and people are going to become unemployed. There's not going to be enough time in the economy to see that pretty little graph of agriculture shrinking and services increasing. That took a long time. Right. This is decades. If we're on the order of like years or like 5 or 6 years, we're not going to have time to see that pretty little graph. We are going to need to think about how do we support the people who are becoming unemployed. And you know, many very smart people have made suggestions on how to do that. I think my personal, I wouldn't say favor, but I think the thing that makes more sense to me is somehow expanding the ownership of capital. If labor is replaced by capital, then what's going to help people is formerly you were a labor in labor now universal basic universal basic ETF. It is UTC.
Hello and welcome to another episode of the Odd Lots podcast, I'm Joe Weisenthal.
Tracy, it may have changed a little bit in recent weeks or months, but I think by and large, if you talk to economists about the long term impact of AI, particularly on jobs, by and large it seems like they point to history. And they say there have been many technologies in the past that people that were going to be very disruptive and destroy all kinds of jobs. And in many cases they did. But technologies create new jobs. We can't necessarily anticipate them before beforehand what they're going to be. And AI is like kind of no different. Ultimately, yes. But then to your point, you ask, like, well, what specific jobs do you have in mind? And I get that, you know, it's hard to tell. It's only the visible those. Right. But it's so frustrating. Right. Because here's this big new technology. It's supposed to be a productivity boost. And yet no one is actually sure what new jobs it's going to create from that productivity. This. I love him to death, but Adam Osmek wrote a piece, several weeks ago, and he was like, well, you know, the player piano disrupted the existence of piano players, but hotels still pay money for a human who will have a piano player human, an actual piano player in the lobby rather than a player. Which is true, but like not many people have jobs that are equivalent. And I think that like it's like, oh, you know, it's like I want to get like this insurance form reimbursed or whatever. This insurance reimbursed. Like I don't care about the human touch. That per se. I think there's something very happy to have the equivalent of the player piano there. There's something very dissatisfying about the idea that we're all just going to become, like, performative in a way. But I actually think that's kind of where we might be heading where like the sort of social skills I've said before, the looks maxing the personal branding, the multitasking, I guess, like, becomes more important. So the future is performative humanity. OpenAI I just spent a ton of money on TBN. I really love those guys. They're both very good looking guys. And so I sort of feel like, okay, this is this is, this is the biggest AI company in the world, sort of making a bet on these, like, great buying two very nice and charismatic humans. Yeah, yeah, yeah. So maybe that is the future. Just being nice and charismatic. Anyway, we need to talk more seriously about this, because I don't I don't know, I kind of feel maybe this is not just going to be like the steam engine. It might be very it might be very different. Maybe we won't have jobs. Maybe there will be new jobs. Anyway, someone who's been talking and thinking a lot about this and why I might be different, we're going to be speaking really of the perfect guest. Alexey Musk is a professor of economics and applied AI. University of Chicago does a lot of writing on this topic. So Alex, thank you so much for coming on Atlas.
Thank you for having me. This is pretty cool.
Do you have the job of professor of economics and applied at like.
Yeah, it's worked out pretty well. It's a good time. You picked a good field. Yeah, yeah I mean I've been I've been an economist for much longer. I've been a professor of applied AI. I have been studying human behavior, human decision making for about 12 years now. Okay. Than a decade. And when ChatGPT first came out, I was kind of taken aback. And this was a few years ago now, and I was thinking after about a week of using it, I was like, this is going to be huge for the economy. And so I started talking to people who have kind of there were several people who kind of knew that it was coming and knew what the impact it was going to, it was going to have. So I started talking to those people, and I kind of quickly kind of started retooling. I smart, I started, I trained my own model, you know, I got into cool, I got into it and, you know, that's, I've been trying to play catch up ever since.
What did you see in ChatGPT specifically? Because you would have been very early at that time, a lot of people were using ChatGPT to basically as a sort of enhanced search engine tool to write poems, tell silly jokes, whatever. But you saw something that was serious for the labor market.
Yeah. I mean, once you started using it, you saw that it was able to basically not not so well in the very, very beginning. But even after a few months and like within a year, you saw that it was able to kind of do basic cognitive tasks to a decent degree. Like it wasn't like we are going to replace that person. But it was doing pretty sophisticated things. That and the jump from like where we were thinking about AI as these very, very, very targeted things like, I will play the game go or something like that to something where, whoa, it can write an essay, it can tell me about this accounting property, it can make a forecast. All of a sudden the generality of the technologies just exploded. And to me that was that was a huge deal. You know, the generality of it. I mean, I guess literally that's the G. Ryan. Yeah. But yeah. But yeah. No, I mean, absolutely. I have to say there's an aside, but like learning a little bit more about like where I was pre LMS or prejudging almost makes me even more impressed. Like the leap I don't know if like, this is a common but when you like look at like some of like what was cutting edge in 2019. Yeah. And then you look at what's cutting edge in late 2022, I'm almost more impressed than if like I hadn't known what they were up to in 2019. Like it's a huge gap in those few years. It's a huge gap. But at the same time, like there were there, there was a kind of a path towards AI and like the way that I was being worked on for a long time, which was like these very specific purpose built technology. And I think Geoffrey Hinton and other people were kind of working on their own for a long time in the wilderness of thinking like, maybe we can do something much more general than that. Maybe we can kind of come back to this idea of AGI versus these very specific tools. So then the whole term AGI, the general part of it, was kind of the reason that term came out was because in response to these very specific technologies that were being developed, which were by design general. So somebody said Shane Legg was one of the people who kind of I think coined the term, he was saying, look, let's think about the general part of intelligence and let's try to build a technology that is as general as the human mind. Let's go back to that. So like if someone makes a model that could tell the difference between written and spoken word, that's mind blowing. It's incredible breakthrough. But that's not a general technology. That's a specific scale.
What time did you have in our betting book for Joe to refer to his vibe coding? I had two minutes, 13 seconds. Should I wait a longer? No, I mean, I made it a little bit longer. Great. Let's talk for so long. No, sorry. No. Fair enough. It's a fair point. I mean, to me, like the moment when things seemed to get very serious was the release with clawed code. And at that point you went from like, okay, the model could not just tell you things, but it could actually do things for you. Was that the vibe shift that you anticipated or experienced as well?
I mean, even though many people were talking about this, that this vibe ship was going to happen, people were telegraphing it for for months and months. Look, when agents start taking off, things are going to change as far as how people how people perceive this technology. Because the thing about agents versus just like the web based browsers, they can do stuff on your computer, they can say, like you could tell it like, look, make me a spreadsheet. It will go and make you a spreadsheet using the tools that are available in your computer, not just say, okay, here is how you would make a spreadsheet, but you have to do it yourself, right? And that's a that's a paradigm shift as far as the economics of the technology.
So I set up to sort of maybe it's a straw man, but I sort of the sort of straw man that maybe we're going to knock down in this conversation. But how would you describe the sort of modal view of the impact of AI on the labor market among the economics profession, to the extent there is one?
So I definitely think there is one. There was there's a very, nice survey done by by a whole team of people, Kevin Bryan was, was one of them and Basil Halpern was, was another. And they released a survey where they, they, they asked for forecasts for from economists and AI technologists. Now, this is a self-selected group of economists. These are economists who are working on AI. Okay. So it's not the whole field. But one of the things that you got from that, that that survey was, they're very much aligned. Okay. Right. So economists, at least the ones who are actually working and thinking about that technology, they think there will be a big impact as far as capabilities. And there will be some impact on the labor market, not astronomical. And we're talking about like 2030, okay. Things like that. There's going to be substantial capability increases, but the growth is going to be pretty moderate. It's like an extra two 3%. And the really interesting thing for me from that survey was that the technologists were kind of a bit more optimistic than that. As far as both the productivity growth and, kind of some were kind of thinking that there will be much more unemployment. But for the most part, the two groups kind of agreed. I was personally surprised by that. So, and this came out, I think, last week or two weeks ago, I thought that there was going to be a lot more, daylight between the two groups.
Well, the other thing that you tend to see is people release these charts of, like, which job is most exposed to AI, and it's usually like, you know, a knowledge worker at the top or something like that. Your work is really interesting to us because you point out that a job is like much more than just the sector that you're actually working in. Tell us more about that.
So the exposure measures they came from, this literature, but mainly this, this one paper by, Daniel Rock and Pamela Miskin and coauthors that was published in science called the one of the greatest titles is GPT or GPT. GPT. You know what GPT is. But GPT in the second term is called general purpose technology. Okay. And the AI there, they they basically started mapping jobs to the exposure as being exposed to the to AI. But it's really important to understand what that number means. Yeah, that number means that A I could do 50% of a task. Right. And how many tasks are in the job that I can do 50% or more? So there's a couple things in that statement. The first 50% is not 100%. That's obvious. Right? So you still need a human in the loop. If I can do 50%, but two, it's the fact that a human job is a bunch of different tasks, right? So this is not a new point, David. A tour, has, you know, has has worked from the early 2000s with coauthors on this saying this is the task based model of jobs. They're on a similar, has the, the canonical model on this, and the idea is that when when we look at a job and we say, look, your job is exposed, let's say it's 50% exposed. It really, really matters what tasks in your job are exposed and how these tasks relate to one another. So let's say I have a job and I have a whole bunch of like, completely meaningless garbage that I'm doing, but I have a comparative advantage. And why I'm really getting paid for is like 20, 30% of the job. If I is automating the kind of like meaningless kind of rote things at my job, I can take all of that time and I can focus on the job, on the parts of the job that are my comparative advantage. What does that mean? Means I'm going to become more productive, but I'm going to get paid more, even though my job is really exposed now, what does that mean for the labor market? Now you have to think, okay, so a person is going to get so just to be clear, before we go any further, if if I'm working on a factory floor and one of my tasks is to pull a lever like that is something that could presumably be automated. But if the other part of my work is to observe, like how things are actually working on the floor and to report back to managers, that might be something that's still valuable under our sort of AI future. And if the lever part gets automated, the theory is that not only, you know, well, Tracy would be more productive and should get paid more for. Yeah, exactly. Okay. Because of the increased products. Yeah, right. This is the O-ring model of jobs. Avi Goldfarb and, Joshua Gans of this really nice paper.
Can I just ask you a quick question here, too? Like, how good are we? And by we, I guess, the economists who studied this at, like, actually being able to, like, here is a job that someone has. Write down a list of these tasks, describe...
How good are we at describing... the like, actually pretty good. I would say on that dimension we're pretty okay. That's the only database that has very, very detailed records. Like, here's a job and here's like a whole vector of things that are involved in that job.
So, and I'd say on that part, like just listing the tasks, pretty good.
The thing that I think we're less good on is how those tasks relate to one another. This is the term called complementarity.
So this is the weak links model is essentially saying like, look, if tasks are completely separable, let's say, you know, I have a, I pull a lever at my factory and I talk to people on the factory floor, and these are completely independent. If I fail to pull the lever correctly, the other part of my job is unaffected. There's other parts of the job, like cooking. For example, let's say I'm really good at 90% of the job, but like, I really screw up the seasoning.
Right. That meal tastes like garbage.
Garbage. Right? So you haven't succeeded in your task. You haven't succeeded on that. So when the tasks are interrelated, screwing up on one or two tasks means you did not complete your job. And it's basically kind of almost the zero-one sort of relationship. So the extent of that complementarity, of how these tasks are related, will determine the extent to which automation is going to affect the labor market, and we don't have good numbers on that.
So this is really interesting. We're good at writing down the list of the tasks. We are not good at writing down the sort of like deep relational links to the task and how they fit together.
Exactly, exactly. So that's something we need data on. The other part that we really need much more data on, and I recently was quoted as saying we need almost like a, you know, Manhattan Project-level effort on this, is this term from economics called consumer elasticity of demand. And basically, sorry, elasticity of consumer demand. And that basically means how much will people buy more of something when the price changes.
So let's say a person becomes a lot more productive.
And they, for the same sort of resources, they can make a lot more of the product, their wage rises. What does that mean for the labor market? If they become more productive given the same kind of inputs, their wage rises, but also the firm is probably going to be paying less money to produce the same output. If it's a competitive industry, the prices are going to go down. If the consumers don't respond by buying a lot more of the product, the firm is going to fire a bunch of people because they can do more with less. But if when prices come down, people buy way more of the product, then they might hire more of the same people. And in many sectors we've seen kind of the second thing play out.
So people are arguing that software is actually one of those sectors. So there's been a bunch of talk, kind of looking historically at like, what does productivity mean for the technology sector? It usually means a lot more consumer demand. And so there's this really active debate now about what are coding agents actually going to do to software engineers. And some people are arguing, look, we have seen historically pretty elastic demand, and so we're going to potentially see a lot more hiring in that sector. And many people are saying this, but other people are saying, wait, maybe it's not as elastic as we think, and people are going to become so productive that we're really going to see, about, downsizing.
That was kind of the argument that Jerry Seinfeld was making in our 'Defending Software' episode.
Should we talk about that more? You know, people are worried, right, about... Yeah, a white-collar wipeout. I'm worried. Like, what would... So maybe the question should be, what would have to be true about either the nature of AI capabilities or the relationship between tasks and jobs? What would have to be true such that this scenario could unfold, a wipeout?
Yeah. Two things. Well, let me talk about three things.
One is just full automation.
The models are so good that they just automate all of the tasks. That's like a very simple scenario to think about because obviously people are going to get fired.
If it's fully automated. The other one is the one we've just been talking about where people become much more productive, but consumer demand is not elastic enough to absorb that extra production. So you're going to have much fewer people doing a lot more stuff. So again, you're going to have a lot of unemployment. The third thing is related, but is basically how many jobs each person has will determine the incentives of the company to actually invest in the automation technology. So let's talk about like the one-task job. Let's say a person is just pulling the lever, and let's say right now that doesn't even look exposed, right? We look at the exposure graph, it doesn't look exposed. But let's say we're kind of getting close, and it just needs a bit more money to get to the automation switch. Well, the company has a lot higher incentive to invest that money if they know that if they invest that money, hey, they can get rid of that person completely, whereas they have less incentive when, you know, let me invest in automating the lever pull. If I know that I can't fire the person because he also does a lot of stuff. So we have to think about the incentives of the firms to automate in the first place. These are large projects to do the automation. It's not like, oh, OpenAI releases a model, all of the companies adopt it overnight, we see it in, you know, a week later we see the outcome. There's a lot of organizational kind of going back and forth. A lot of systems need to be changed, all of this sort of thing. And so companies need to know, like, look, if I spend the money on it, I'm actually going to save money as a result.
So studying the archetypal guy pulling one lever aside, what are the real-world jobs in your framework that are actually most exposed to AI risk?
The one-dimensional work? I guess I hate to say one-dimensional because every job is multi-dimensional, but if I had to make a guess where economists and other people should be kind of worried, I'd say stuff like truck driving. Yeah. And stuff like warehouse workers. Like if you Google, you know, warehouses built in China or something like that, these warehouses look nothing like what we think about warehouses. They're completely, completely automated. They have robots like crawling on the walls. And they're just, there's no human in the loop at all in these warehouses.
And so... Oh, the warehouse gets automated and then the warehouse gets automated.
So part of that automation is going to be kind of loading that truck.
And then the truck gets loaded through automation. And then that truck drives from A to B for this.
Interesting. Because you know, obviously a lot of people in freight will say the way you make that argument is very different. Then they'll say, well, yeah, driving a truck is much more than the driving part, right? So it's like, okay, you can have a Waymo truck, but who's going to deliver it? Who's going to deliver? It is actually a big deal. Like if somebody stops it on the road, a Waymo truck, they could just stop it on the road and rob the truck. Right. That's one element. But to your point, you know, if like one of the tasks that a truck driver has to do is that coordination once they've gotten to the warehouse. But if the warehouse is already automated, that is a heavy thing, then that no longer is as important, perhaps, for that to be a human task.
Exactly. And think about the incentives of the company to invest in this technology. It's huge. These are very, you know, these are some of the only jobs, truck driving, where, you know, you don't need a college degree to earn a lot of money. And so there's a big incentive on the company.
So okay, I get that. But on the other hand, even going back ten years, I think if you went to Davos, there were probably people saying, truck, I'm worried about the future of truck driving because AVs have been around as like a thing since before I just, yeah. So in terms of like post-chatting jobs, etc., that would be concerned with like, what do you see out there or what are you looking at?
I mean, I think everybody's looking at software engineering. I think you have to think about like where the technology works best now is verifiable tasks, right, where you have a lot of data where you can say this is good or bad, not in a supervised learning sense, but in general it needs to be verified. That's why like math there in research, math has been like the big kind of boom. As far as what are people talking about on the internet as being automated, math is verifiable. Yeah. You know, a proof is either right or wrong. Once you do the proof, it's much easier to check if it's right or wrong rather than construct the proof. And so jobs that have large components where we have a large data bank of data to train the models in a way where the output is verifiable are going to be potentially more exposed in the sense where you can automate more tasks within the job. Now, the thing that we haven't talked about yet is new tasks, right?
So we're talking about a very static sort of economy where there's the lever, there's me walking around, and if I'm automating these things, that's the end of my job. But you could imagine a scenario where you automate a part of a job, and all of a sudden this person is freed up, or this task was actually a complement to a task that wasn't even, you know, imagined by the organization that this person is now doing that's not automated. So that's something that I think people should be looking at especially. And this is data that actually AI companies have, is what new things are people doing.
Like, say more about that because this gets to the, you know, like what new jobs could we actually see from this question, which I never see a satisfactory answer to.
So if they do have that data, they have, okay, all the data sets, but they have data about like, okay, so this is a software engineer and you know, a year ago, these are the sort of tasks that this person was working on through our system. These are the sort of queries and things like that. And you could see like some of these queries being automated fully by the agents. Now they're asking potentially different questions, are, can we classify these as different tasks that are not fully automated, where the AI system is actually a complement to those tasks? So this is not like a perfect picture, okay. But this is data.
But so it's not really like a new job per se, but it is freeing up the software engineers to like ask about different things or shift in focus that they in yellow obviously, you know, vibe coding and...
Yeah. Right. Exactly. Finally we're freed up from the drudgery of our day-to-day life to work on that.
But no, but like, this gets to sort of, you know, the big question is, like you mentioned, one scenario is just that, like the technology can do all the tech, right? How seriously do you take that possibility? Because then it's game over, right? Like it's like, okay, it does all the tasks and then it's going to keep getting better. And if I can learn to do a new task, well then if it can do all the tasks, there's probably not, then maybe I'll learn something new. But to learn that task, how seriously should we take out, take this possibility that the models are on some time frame, on track to just be able to do all the tasks?
So a lot of parts of that question, one, physical versus versus just kind of digital, right? So I think there's a scenario where it can do everything kind of sort of these sort of cognitive, nonphysical tasks, whereas the physical world is completely, you know, these robots.
Let's just talk like email jobs or computer jobs, okay. Let's talk about computer time.
So I think I take that scenario pretty seriously. Okay. I think if we get, I haven't seen any data to suggest that the models are slowing down as far as their capabilities. You know, Methos was released yesterday or two days ago or something like that. And if you, we don't have great data on this, but if you look at like where it is on the kind of line of capabilities, it's just on track and it on track is very, very fast.
Yeah. Right. So the developments are happening very fast.
So as far as like email jobs, I think there is a scenario where pretty much everything is automated. And then you have to ask, are people going to be moving to the physical jobs or will there be new jobs that we haven't thought about before? So, you know, if you look back in the 1940s, like, I think more than half of the jobs that we have now didn't exist in 1940.
Yeah. And so what did the new jobs look like?
I mean, I have a theory, please. It's very similar to the one that you didn't like.
