← Azeem Azhar's Exponential View
Azeem Azhar's Exponential View
Where did all the entry-level jobs go? (With Revelio CEO Ben Zweig)
00:47:58
Junior roles in AI-exposed fields are disappearing fast. The obvious culprit is AI rapidly automating entry-level jobs. And yet, this isn't quite right. What is driving the drop is managers’ expectations about what AI will do, not the work that it's already replacing. I discussed this with Ben Zweig of Revelio Labs, which builds global workforce data from millions of individual profiles to track hiring, separations and job flows. Together, we explored the future of work and shared practical advice for new grads.
Loading summary
Transcript49 lines
- [00:00]
A
Today I want to talk about jobs. There are lots of worrying headlines about the job markets. We've heard about big layoffs at firms like Amazon and other companies going into hiring freezes like Walmart. We've heard that entry level jobs are under pressure from AI, that the amount of graduate hiring there is is going down. So what should we make of this? Is this artificial intelligence and its early effects on human work? Or is AI just a convenient excuse for bosses looking to manage their costs? To explore this, I've asked Ben Zweig of Revelio Labs to join me. He is one of the more thoughtful founders and thinkers on the topic of the labor market. And in this moment in time when the bls, the Bureau of Labor Statistics, has gone dark and become unreliable, Ben's work at Revelio Labs is providing a private sector take on what's happening in the job market at large. So, Ben, I hope you're going to be able to bring some clarity to this confusion and join me as we jointly explore what's happening to the world of work.
- [01:14]
B
Yeah, thanks for having me.
- [01:15]
A
Let's get started with what the headlines say. You know, the headlines for a decade have talked about the threat of automation to the world of work. I think it's now been 12 years since my friend Carl Fry and his collaborator Michael Osborne published that famous paper about the way in which machine learning might affect human tasks. And as we move into 2025, we are starting to hear stories. The entry level job market in particular seems to be collapsing when you look across your data at Revelio and give us a moment to understand what data Revelio looks at and gathers. What is the most accurate description of what's happening in the labor market?
- [02:03]
B
I think you're largely right that the job market is not very good right now. So hiring is at a low. Also attrition is at a low. So people are not leaving jobs as much, even despite increases in layoffs. But you know, separations from jobs are not happening to the same degree as they used to be. And that is particularly affecting young workers. But what's interesting about this is I think you kind of touched upon this Brynjolfsson paper and others out of Stanford that highlight that there is an interaction between entry level workers and AI exposure. And that's research that we've done at Revelia as well early on. But I think the researchers have gone a little deeper into it and I think it is credible that AI seems to be affecting entry level work, but notably it's not causing technological unemployment. Like was predicted by Osborne and Frey or it was predicted a hundred years ago from John Maynard Keynes. He, you know, it was in 1930, he wrote, you know, economic possibilities for our grandchildren and predicted in the a hundred years we'd all be working two day work weeks. I think it was 15 hours. So it was a 15 hour, 15 hour week. It hasn't quite been a hundred years. So maybe, maybe in 2030 we'll, we'll see that he's right, but it doesn't look like he'll be right. So we're not seeing technological unemployment, but we are seeing decreases in demand for entry level workers and that is particularly concentrated among AI exposed roles. I think you could say, oh well, you know, AI, you know, seems to be affecting the execution of tasks which you know, young people do and rather than the orchestration which favors more senior workers, people with experience in management. Because, you know, in the age of generative AI, we're essentially managing these bots to go and, to go and do things and execute on some subtasks. And I think that that is a read. Personally, I think I'm not quite sold on that yet because we're not seeing so much substitution, like we're not seeing enough substitution that you would think would result if that were the main effect. I think what we're also seeing and what's conflating this result a little bit is that there's a lot of uncertainty in the labor market, partially due to AI, partially due to policy, but I think as a result, employers have a kind of high discount rate. They're optimizing for the short term more than the long term. And young workers, generally, entry level workers, are workers that are more uncertain. You know, they require more of an upfront investment and you know, maybe they'll work out, maybe they won't. So it's a kind of high risk, high reward type of bet, whereas more experienced hires are the safer bets. So if someone's just looking to optimize for the short term, I think they'll generally favor more experienced workers.
- [04:45]
A
That is such a fair summary and I'm in danger of agreeing with you too much. So you're going to allow me to, to, to push back on a few of those observations. But let's unpick some of the, the points that you, you raised. You talked about Eric Brynjolfsson's paper. You've also talked about this idea of managers becoming more risk averse. And if you're more risk averse, taking on board a millennial who might need Their kombucha and early break on a Friday afternoon to take go for a meditative stroll is higher risk than whipping the corporate wage slave in their 50s for another, you know, few hours in the week. You know, that that's a, that's a decent picture. But the other thing that you drew up was just how complicated and multifactor labor markets are. And one of the things that strikes me about labor markets is that there are so many forces at play. From interest rates to the effect of the pandemic, to political uncertainty impacting how long bosses are willing to make investments. For that, it's really difficult to unpick those causes. And the starting point for that needs to be the data that you can look at and that you can make sense of. So we would typically go off and use BLS data or similar data like that in the United Kingdom or, you know, you know, Yale University has their budget lab and they run surveys. What is the, what is the way that you look at to make sense of what workers are doing and what hirers are doing? So give me a sense of the sense, the sensing that you're doing the data that you're gathering.
- [06:24]
B
We started Revelia Labs with the goal of producing high quality data that can be used for any downstream, you know, analyses. And right now we are not seeing government data. You know, the BLS hasn't put out their employment data in, you know, over two months. The government's been shut down and there's a lot of turmoil there. UK public employment data is not fantastic either. I think what we're really trying to do is kind of start from very micro level data. So actual individuals, you know, trying to understand like the, the actual firm dynamics at any given company and then aggregate that up globally.
- [07:00]
A
I mean, how do you do that? What are you going to show us? Are you going to show me? Because you've crawled LinkedIn and you've figured out how many customer service reps there are at United Airlines compared to field engineers. And that gives us a number. Or is it not as granular as that? Or is it more comprehensive than that?
- [07:16]
B
That's exactly right. Yeah. We start at the micro level data, you know, individual profiles and resumes and stuff like that. And then we can see headcounts, start dates, end dates. We can see inflows, outflows. So we can see the stocks and the flows of people. Of course there are biases in the data. There's, you know, there's sampling bias and that can be adjusted for. There's lags in reporting and that can be adjusted for, I mean, not easily, but, you know, it's, it's possible to, you know, that's why we have all these data scientists and data engineers kind of working hard. But yeah, we need to adjust for issues with the data. We have to classify it to the proper occupations and industries and, you know, all the, all these different dimensions of employment. Once the data is properly debiased and categorized, then it becomes possible to see actual stocks and flows of employment in the whole world. So I think one real limitation to government statistics is that they're done domestically. So, you know, the US has their, you know, it has the bls and you know, every, every country's got sort of their version of it and they do things differently and have different assumptions and it's hard to standardize between those things. So if we are looking to analyze some policy or even just like technological adoption that happens unevenly throughout the globe, like comparing countries to each other seems really important. Public statistics are really limited for analyzing things like policy and technological change.
- [08:35]
A
So let's get a sense of, then go back to what we really think is happening. You talked about the Brynjolfsson paper and that Brynjolfsson paper was something that came out in the summer. It was in late August, early September, with one of Eric's postdocs. And they looked at a lot of hiring data over a couple of years, two or three years, and they found that if you look to AI exposed occupations, computer science or programming and customer service, the level of hiring for entry level jobs was much lower than in occupations that were not exposed. But in those same occupations, mid and senior level hiring or employment levels had not declined. In some cases it had increased. And the argument was, and I think you put it earlier in one of your answers, it's challenging to hire somebody young because while they're cheaper, they need training. They may not have decided that what they want to do is be a COBOL developer or a CSS jockey and so they might flake out after six or nine months. So again, why not just, you know, whip the 45 year old for a few more years? So that was Eric's argument. And the paper was called the Canaries in the Coal Mine. It did really well, understandably. I thought it was a great piece of work. The buts, and I think this is why it becomes complex, this discussion about what are we actually seeing. So the buts from my side were, number one, it felt really early for ChatGPT and Gen AI, which are not even three years old to affect hiring. I mean, I've worked with big companies. They, they do not move with the degree of alacrity that you might expect. So that was, that was my first challenge to Eric. The second challenge was simply the, the idea that sometimes, especially if you're having to sale uncertainty, you may have overhired in previous years during COVID you may have cost concerns. It's much easier to not hire a 24 year old than it is to fire a 44 year old. So as a manager, you like water, you'll take the path of least resistance. I suppose the third argument was really what are we seeing here? Because a few weeks later the Yale Budget lab or Yale Economics Lab came out and said, well, we've looked at, you know, widespread data and we can't see these effects at all. Now of course, Eric and Barat's effects as they described in Canaries in a coal mine was highly, highly plausible. Right? It fits lots of mental models we might have had. But equally there are mental models and arguments that I've put forward that might be as credible. So when you look at that through your lens, how would you disentangle what was really going on?
- [11:14]
B
We did the same analysis at Revelio and we found actually before, before either of these papers came out and we, we found that this phenomenon that Eric and co authors did document is actually happening. You know, we had different measures of AI exposure but found the same, the same pattern.
- [11:28]
A
So, so what was that pattern that Revelio found?
- [11:31]
B
Like specifically, it was that the difference in AI exposed roles and non AI exposed roles was more negative for entry level workers than it was for more senior workers. So the idea is that, you know, AI exposure is more harmful to more junior employees than it is to more senior employees.
- [11:52]
A
And what does AI exposure mean? I mean, who's got an AI exposed job? Does a graphic designer have an AI exposed job? Does a paralegal have an AI exposed jobs? What, what is the measure by which you assess AI exposure and how do you control for factors relating to this industry or the sub industry or the firm itself?
- [12:12]
B
The most important thing to understand is that exposure is not something that it really doesn't happen at the job level. It happens at the work activity level. You can think of jobs as collections of work activities, collections of tasks. And you know, AI does not automate jobs wholesale. No technology really automates jobs wholesale, but automates components of jobs. There really is no job that has zero AI exposure. Every job is somewhat exposed, but some jobs are more exposed. So, you know, you mentioned, like paralegals and stuff like that. There's a higher prevalence of tasks that have high AI exposure and a lower prevalence of tasks that are more immune to AI, specifically generative AI. So, so this is not a commentary on like, you know, technology at large. This is really this specific technology. I believe the pattern, but I also believe that it's, it's not so plausible that these companies are just substituting labor for AI that quickly. I mean, you know, you said you've worked at large organizations. I've worked at large organizations. I used to run workforce analytics at IBM. And I know that like, nothing gets done the way they want it to get done. I don't know when they'll start actually adopting AI technology, but, you know, it's not going to be overnight. And I don't really believe that firms are as sensitive to this technology in how they can. They can substitute labor. Which leads me to believe more that they're just taking safer bets and they're discounting the future more. And I think what we're actually seeing is less of a commentary on what AI is doing and it's more about what the expectation of AI is doing. There is some thought that, you know, oh, AI is very powerful. It demos very well. You know, you can do all, you know, you dazzle people and someday it will have a big effect on workflows in a large organization. Maybe not today, but since we expect that that might happen, maybe it's better to take safer bets while we wait and see how this thing plays out.
- [14:08]
A
We have this ad that goes out around Christmas, which is a dog is for life and not just for Christmas, because lots of kids get bought puppies and they have to keep them for a long time, right? So a young grad hire is for three years, not just for three weeks. And your argument here is that even if AI can't do what people say it might be able to do, the fact that managers believe it might be able to do that has them looking at these long, long term decisions, which is hiring a person, making a commitment to them, being willing to work with them for two, three, four years. They're putting that off. And this actually, Ben, is super fascinating because a mismatch between expectations and reality can be thought of as exuberance. And it's that mismatch, when we think of stock markets and stock prices is the exuberance that leads to bubbles. It leads to the idea of ultimately disappointment and the wrong allocation of capital and resources. So what I'm hearing from you and Sort of drawing a thread out is if there are these expectations about what AI can do, encouraging companies to take decisions about 2027 and 2028, which is what the hiring of a grad in Q4, 2025 is actually about, and those expectations are not met. That is actually a kind of a fundamental problem, not just for the graduate, but actually also for the company. How do we test and identify that it really is the expectation mismatch hypothesis that you've laid out here and not something else.
- [15:51]
B
It is still very early and you know, you mentioned the title of that paper was Canaries in the Coal Mine. Really just because, you know, we're looking at very early signals. But I think as more data comes out, I think what I would want to look at is, you know, what actually happens to adoption within firms. So are firms adopting this technology in their workflows? I think we'd be able to see that from, you know, what people say on their resumes and profiles. Like, you know, what, what do they say in their bullet points? You know, like on their, on their resume. What are the work activities that show up even in job postings? I think we'd be able to see it.
- [16:21]
A
By that you mean that I'm a large company, I'm hiring and I say must have advanced prompts and context engineering on at least two different LLMs as a, as a requirement. And that is an indicator that you attract.
- [16:34]
B
I think you could see it in qualifications and requirements. I think you could also see it in the sections on responsibilities. So if they say, you know, your job, you're being hired to do these, you know, six responsibilities and they, and they outline, you know, responsibilities that are really about, you know, interacting with AI systems. We would reasonably conclude that, that this actually is a responsibility that they're giving to, to people in the actual job. And, and it actually is a core part of their workflow. So I think if we start seeing more of that, I would say we have a lot more evidence that there is some adoption going on. Which, by the way, adoption doesn't necessarily mean that this will lead to labor substitution. Like you could have adoption of AI, that that augments certain workers and creates more employment opportunities.
- [17:21]
A
I mean, we should, we should touch on that point as well. Right? Because of course, Excel showed up 35, 40 years ago and it's created more work than it's, it's taken away. What would you look out for? And where have you been able to see places where AI is enabling more work through lifting productivity rather than going into this Task substitution modality we've been discussing.
- [17:44]
B
I think. Here's how I would kind of frame up this, this question. So, so we mentioned, you know, jobs are collections of tasks and activities. As some tasks get automated, let's say, you know, you have 20 different tasks or activities or responsibilities in your job, and let's say, you know, six of those get automated. Then you're left with 14. Do you, do you double down on those 14 and get extra good at those things, or do you take on more? That's a big question. Like, how quickly can jobs kind of reconfigure? So I think that depends heavily on the organization. So there, there's, you know, while we're on the topic of good academic papers, there's this wonderful paper called Adaptive Organizations where it goes through this model of firms as being either adaptive or procedural. So in adaptive organizations, I mean, I say this like as a founder of a startup where, you know, people's jobs change all the time. So, you know, and this has nothing to do with technology. Usually sometimes, you know, we get a client request and, you know, we say, oh, now there's extra work to do. Who could do it? Who, who's interested? Who wants to take this on? And, you know, we figure it out. Or someone leaves and we have to reallocate their work, or someone says they're not interested in something or, or we automate something and we, like, don't have to do that anymore. And we kind of, we're constantly reconfiguring what people do in their jobs. And I think it's a very common pattern for someone to enter a job thinking they're going to do one set of things. And then as the job evolves, they find themselves doing totally different things. And that is the mark of an adaptive organization. An organization that can be fluid, that can think about jobs fluidly, where people, you know, do what the business needs. And business needs change all the time.
- [19:24]
A
So if you're listening to this and wondering, when are they going to tell me what I need to do to get a job in this tough market? Just hang in there. Because at the end, Ben and I give some really specific advice. This point about the adaptive organization is a really, again, it's one that we need to unpack. So when I go off and talk to C Suite execs, I've had a discussion with a chief human resources officer with, in a firm with 47,000 employees, so of a reasonable size, and she was saying, see jobs as bundles of tasks, you know, units of work that need to be done rather than as single, lumpy activities. And we're now going through a process of unpacking what each job is into a series of tasks. And they're having to do this from both the top down, a centralized team, but also enabling individual managers to look at their teams and teams of teams and, and say, what are the tasks that comprise which jobs? It's a big effort. And of course, economists have been talking a long time about jobs as bundles of tasks and tasks having certain skills and activities attached to them. It's super theoretical. It is super theoretical. I mean, I have managed teams of my biggest teams, hundreds of people. We never thought in those terms. So perhaps let's break this down with some really simple step by step. Give me a real example of converting, you know, a specific job that people listening might recognize into its set of tasks and help us understand how would an HR manager do that and how many of them are doing that.
- [20:59]
B
Organizations generally don't have good taxonomies of work activities and tasks. And that's something that, you know, we're trying to do. You know, I wrote this book about job architecture, which really is about, like, you know, helping create taxonomies of what people do in their jobs and how to think about jobs as bundles of tasks that, that change. So I think having the right data behind this and the right taxonomies and categorizations is really critical if we are going to manage this purposefully. Now, of course we can manage it, you know, reactively and, and, you know, that's not always terrible for a small company. For a big company, I think that becomes much harder. But I, I mean, I'll give you a personal example. For my company, we have a team of economists and, you know, we have 12 economists. You know, they all have PhDs and economics and that, and that's their, their background, you know, and the term economist is really, it's more of an identity than an occupation because what they actually do is whatever needs to be done. You know, they write newsletters and we produce content and they do, you know, we work with journalists and they do media relations. And we just started producing public labor statistics. And that was just a few months ago. And now they are involved in creating public labor statistics. That was not in the job description. And sometimes we talk to a client and they want to, they want to, you know, understand the data and talk to, you know, an expert, and they get on client calls and they are the trusted expert of, you know, what, what we need. And sometimes those responsibilities shift. Sometimes they do more client work. Sometimes they're producing reports or newsletters.
- [22:29]
A
They sound very productive. Are you sure they're economists?
- [22:33]
B
Yeah, yeah.
- [22:34]
A
Not like economists I've met to be fair.
- [22:36]
B
But yes, you know, we call them economists because that's their training. But you know, what we actually have them doing is, you know, some mixture of client success and media relations and data engineering and data science. Their title is really, is really more a function of the talent pool, less a function of their occupation.
- [22:51]
A
That's a good example because you've, you've unpicked the different things they do. But what that means, most people probably not working in firms that are as dynamic and adaptable as a technology startup like your own. They'll be working in small enterprises where the, perhaps the HR manager doubles as the finance controller or as the head of legal or they're working in these enormous organizations where you know, change management walks up, works up hierarchy, up to hierarchy. I'm really looking at this notion of how do you turn the theory which is a job isn't a job, it's a bundle of tasks that a person does. And that bundle might be a little bit dynamic or a bit more static. And it's the understanding of that bundle of tasks that helps us make most sense of how you develop that individual in the shape of a series of automation and augmentation technologies. You know that that is the theory and there's endless white papers, entire parts of the Amazon have been cut down to print out the white papers from think tanks about this over the last decade. But let's turn that theory into like really practical example driven reality.
- [24:07]
B
At these very large organizations there's a lot more process and if you have a, an organization that is procedural, you may have a job within that organization where someone's job is very narrow and they, they are a specialist in, you know, one or two subtasks, you know, one or two tasks where they, they are hired to do kind of like singular things really well. You know, you can think of them as like on some sort of assembly line. They do the same thing day in, day out. That's work that I think is, is getting less and less common. You know, work that, that is like super, super specialized. But it still happens in large organizations. And if that is how we think about work and workers, then it really is possible for that work to get displaced when those tasks get automatable. So that is concerning and I think when there's a lot of rigidity around what people do in their job, you know, either from Occupational licensing or just organizational like standards or just, you know, wherever rigidity comes from in these like big bureaucratic organizations that, that can inhibit our flexibility to respond to technological shocks. But not just technological shocks, also just needs of the business. You know, people get cut because, you know, businesses need different things as they, as they evolve. I am concerned for, for these big organizations. I, I do think that their work can become more adaptive. So I think in the, you know, in the 60s, you know, Peter Jucker wrote about how, you know, workers are going from these like Taylorious workers where, where they're like super specialized in one thing to, to knowledge workers where their job is actually more about connectivity and coordination and like sort of navigating, you know, complexity. And in that, in that world where you're navigating complexity, I think you do have more diversity of tasks and the, and the work is more about orchestration among people that execute subtasks. So I think we are becoming more like managers and knowledge workers.
- [26:10]
A
I mean Drucker was a brilliant, brilliant character, right, who, who foresaw a lot of this. There still feels to me like there is a series of complexities and nuances that are hard to capture. You know, you've built a taxonomy of tasks in this, the job architecture, which is your, you know, the theory encapsulated in your, in your book. That idea that well, we can construct a taxonomy and we can perhaps look at an organization and figure out the propensity for particular tasks to feature in a job bundle. And therefore if Alice has this job and Bob has that job, we can get a sense of what bundles of tasks they look at. But of course that's a correlation and that correlation might be high. But these people are real people with real experiences, real things that they do that might not show up in your ganogram. There's a whole bunch of relationality. So you know, Alice is one of these hyper productive economists. But what you don't realize is she's phenomenal interpersonally. And the head of sales always likes taking her along because she can open that account there and then. And where is that captured? And it's not captured in the job architecture, it's not captured in the statistical top down process. And it's that nuance, you know, maybe it's tacit knowledge as Polanyi might have talked about. It's also relationality. How do you operationalize this given those complexities? And what is your advice to companies thinking about that? Because someone in a big company might say, I actually can't go off and do a nuanced long form interview with every one of my 85,000 staff or 225,000 staff. So where do we go now?
- [27:53]
B
I think it's a very big challenge. You mentioned all these white papers that were in on how we kind of, you know, reconfigure the workforce. In all these white papers, you, you mentioned there's plans for how to reallocate people toward the needs of the business. And I think there's some good ideas and some bad ideas there. One thing that I think is an overrated idea bordering on bad idea is re skilling and reallocating people to where the needs are highest. I think that has been such a trendy idea. And there's all these like internal talent marketplaces. And you know, people talk, they think about strategic workforce planning as being heavily involved in reallocating people. I think that has not worked particularly well, mostly because principal agent problems within the organization. So if you say, you know, hey, here's a, here's a part of the business where we don't need as many people and there's someone there who's got these skills and interests that align to other needs in the organization. So we recommend, you know, you take this person and move them to that team. Because of all this tacit knowledge that you mentioned, the manager is probably going to say, you know, no way in hell are you taking, you know, my best person. And I think for better or for worse, you know, managers don't really care about whether they lose someone to outside the organization or whether they lose someone to a different part of the organization. Especially in these big companies. That's not what they're incentivized to think about. And you know, politics are political. You know, they're going to pull whatever levers or strings they, they have to, you know, block this. And that'll probably work. And you know, the more that works, then if they don't put up a fight that sends a negative signal. Now you're like, oh, you're actually going to let go of this person now. I don't want them. So, you know, the market begins to unravel. So I think a lot of programs that were initiated to kind of reallocate people from, you know, low demand areas to high demand areas have just not worked because they haven't worked within the organizational constraints. And so I'm not optimistic about that. I think actually the data part of it is not the hardest part. I think it is possible to understand the work activities of jobs. I think it is Possible to understand the skills and other attributes of people and think about those skills as inputs to completing the work activities and have like a really nice framework for how to think about, you know, where we can maximize output and how we allocate labor to maximize output, but what levers we can pull. That's where it's super challenging. I think one approach that I think has been underused is actually shifting the borders between organizations and say, you know, hey, this is like an in demand group that has a lot of responsibilities and has a lot of things that it needs to do and not enough people. And here's a part of the organization that has too many people and isn't responsible for as much. And I think designating responsibilities more flexibly could be an interesting approach that I think some organizations are implementing, but not as many as I'd like to.
- [30:45]
A
The flip side of flexibility is instability. And instability is connected to people's own psychological status, their sense of psychological safety. It is connected to the degree to which a job adds to their dignity or takes away from that dignity. There's already issues with the notion of the organogram and, you know, a bunch of people moving from one to, to another. A lot of what we've discussed feels like it doesn't add to the core components of a job that connect to, well, being, that look at people as more than just staccatovite worker bees to be, you know, optimized within the great machine. What is that balance? And what have you seen in working with organizations about how well this job architecture. Hey, Ben, you're not Ben, you're actually employee number X who has skills in this bundle of tasks and we're going to change that bundle, right? That, that to me already doesn't feel like it's a nice thing to be done to.
- [31:49]
B
One, I have two, two kind of answers here. One is that I think there really is some, some very real tension between adaptation and stability. If an organization needs to adapt, maybe, you know, some instability is just the price to pay. So I think that that may happen and maybe that's just, that's just the trade off that has to be navigated. But I do think there are ways to be purposeful about giving people, you know, autonomy and let them feel like there's integrity in their job.
- [32:18]
A
Can you give an example of one of those ways that gives integrity and purposefulness for someone as they go through this, something that you've seen in the field.
- [32:27]
B
So one, one is this concept of job crafting when work reconfigures sometimes that's top down. Sometimes, you know, the manager says hey, you know, you, you're doing too much of this, you need to do more of that. And that feels kind of micromanaging but more and more it's actually coming from the employee, from the worker. And they say hey, you know, like I think this can be done a different way or I think, you know, I'd like to use this technology versus that technology. And you know, I, I am more interested in doing this, less interested in doing that. So I think that should happen more because that is a way to reconfigure work that puts a lot of the power and responsibility on the worker. I think the role of the manager is less about delegation or should be less about delegation and more about communicating the needs and changes of the organization well and be kind of a liaison between the concerns of the business and the skills and interests and capabilities of the workers of their team. The people in the organization, the workers, the people actually executing on tasks like really, really should be given a lot of discretion about how to reconfigure their work. And I think that, that ultimately respects their intelligence.
- [33:34]
A
What you're describing is credible. Ben can see why, how and why that would be more effective. It does strike me that it, that'll work for some types of corporate cultures. I mean what, what you have described. I'm thinking about tech companies and the big tech companies. I'm thinking they could do this falling out of bed in the morning before their coffee because that is the inherent way in which these companies have been architected. I'm not sure if Dunder Mifflin, you know, the paper manufacturer sales company has the right cultural wherewithal and how many companies in America actually have that, that kind of cultural capacity and have the, the depth in the management and leadership talent to facilitate that. So I'm going to leave that as a question, not for you or I to answer. I'll leave it with listeners to think about because it's a, it's a big challenge. The thing, other thing we haven't talked about though is the fundamental instability of the technology. So you talked about a really important point which again I want people to hold on to this notion that maybe it's managers expectations of what AI might do in the next six months, 12 months, 24 months, that is slowing down the pace of hiring. But there is another piece that's happening with AI which is undeniable and that is, it is getting more and more capable. That's happening across three vectors, vector number One is its generalizability, its ability to be applied in lots of different domains. In the most sophisticated things I do with GPT5 G5 Pro are within my domain. And then I see scientists doing it in their domain and I see other experts in their domain, that generalizability is broadening. The second is the complexity of a task that an AI system can do end to end is increasing. You know, deep research is the first example of that. You know, 10, 20, 50 minutes, hours worth of human work. But we're seeing through the coding, evaluations we run our own as well. This notion of doubling of task length then the AI system can reliably complete is extending. The third area is the guardrails and the scaffolding that is being built around these tools that may not be AI itself, but the agentic frameworks, agentic policy management, operations, exception handling, error recovery, these are all getting more and more mature. And those, those three things are shifting independently of each other, often by different groups. So the question of can an AI, whatever it is, do this task is changing on certainly one aspect on its own is doubling every six or seven months. So again, if you recognize that, I hope you do. What have you, what is, how is that instability shown up in the data that you have been able to, to analyze and, and get your hands on?
- [36:27]
B
It's very tough to see that in the data. I mean, the team at Anthropic, they have a team of economists as well. They had published like what tasks are people using Claude for? So I think that's probably the best data we have on how people are using this technology and how that's evolving. I think you're right. As these systems become more kind of agentic, you know, it's less about doing a very precise task and maybe doing a collection of three or four things and stitch together, there's some coordination of work tasks, you know, call it a process or something that they are then able to do. That's kind of, you know, what we're seeing in the data. You know, theoretically you can think of a job as, on the one hand, a collection of tasks, but on the other hand, you're not just executing tasks, you're also orchestrating between those tasks. You're also deciding what comes first, what comes second, how do these things fit into each other. You're playing the role of an orchestrator, of a coordinator. And that is still happening, probably happening more than ever, because now there are some tasks or processes that interact with other things and maybe interact with people, maybe other systems, maybe other Vendors or whatever it is and that orchestration is becoming much, much more important. I think that is something that I would bet is probably a little bit related to this, you know, idea of favoring more senior employees.
- [37:44]
A
Yeah, this notion that orchestration is a fancy word for being the pointy haired boss in the Dilbert cartoons. Right. That's what managers do. We orchestrate resources within the company and across suppliers and subcontractors. And yeah, the old word was just management.
- [38:00]
B
Yeah, no, I, I, I could see a world where, where even entry level workers are kind of middle managers in a sense. They have more orchestration of systems to do and that, that is kind of how we think about middle management. You know, they're, they're sort of piecing things together to ultimately deliver, you know, something more more abstract and more generalized to whoever they're delivering that to.
- [38:20]
A
I mean, I'm really glad that you've got here because that allows us to go back to the beginning of our discussion, which is that the canaries in the coal mine, the idea that entry level jobs, particularly if they're AI exposed, are coming under pressure. What's driving that is to some extent the risk lens that managers are taking from hiring the uncertain rather than doubling down on certain existing workers. So let's now look at those entry level, the entry level workers, people who are graduating, they are now better off than they were earlier in our conversation, which is they know why they're not being hired. They're not being hired because the hirer wants someone who has already got orchestration skills, AKA is already a manager and the hire is risk averse. Right now, given that, let's get practical. What would need to change for those entry level workers to now be suitable to be hired? We know what the symptom is, we've identified it collectively, the two of us over the last 30 or 40 minutes. What do they need to do?
- [39:27]
B
I teach a course at NYU called the Future of Work. I've been teaching it for five, six years and I was just thinking what should I tell them? What should I teach them? How should I evaluate them? How should I make sure that they're well positioned for the workforce? My thoughts are changing on this every day. I think what I'm kind of coming around to is this idea that managing projects kind of end to end is going to be really important. So some type of education, like I think our education system needs to involve more orchestration and coordination. I think about the MBA program. You know, my dad did an MBA in like the 60s and I think at that time, you know, you get an MBA and you, you know, that's a path to like, run a division somewhere. And, you know, in the past few decades, you finish an MBA and you're like a level two analyst. You're not running anything. You're. You're just. You're a worker that has like a little bit more of a credential. Maybe now business education can shift, you know, back to where it was in the 60s, where it could be more about actual management. Like maybe management as a science deserves more attention if we're all going to be managing AI systems. So I think that is important in the very, very short term in terms of, like, looking for a job. I don't know that there are great answers for someone who's like, on the job market today. I think they're in a very tough spot.
- [40:41]
A
I mean, great and not great, right? Sort of great that you're being frank. Not great, because that seems to be the situation. I'll posit a couple of thoughts. So one is this idea that through education or experience, people need to develop the capabilities that demonstrate orchestration. And perhaps some of those you can do even if you're interning. Because a large part of orchestration is being a good manager, is asking good questions. Another part is asking bad questions. Another part is framing problems and thinking where resources might emerge. So maybe there is a part which is to unbundle. We've been talking about bundling and rebundling skill task bundles, but unbundling the tasks and capabilities that go into this idea of orchestration. And some part of it may be prompt. How do you prompt? How do you context engineer? And there may be other dynamics. The bit that still seems gnarlier. How do you solve for that novelty risk anyway? So one of the reasons we had discussed that managers were slowing down their new hiring was that Jason from outside the firm, who's 24 and just finished a grad degree, is not known to the firm. And Jason can get all his orchestration skills and he does it through Khan Academy and Execed and whatever else and demonstrates his micro credentials, but he's still outside the firm. So are there any mechanisms, are there any things that as an economist one could come up with that allows that signaling to de risk for the manager?
- [42:10]
B
It's a tough part of labor economics because, you know, signals have been eroding over time. So, you know, higher ed was always like a great signal and you get, you know, a diploma. And that was always a good signal and still is that's part of education and part of education is actual, actual education actually increasing your, your skills. And that part is more and more suspect. You know, I, I think both of those are eroding a little bit because education is, you know, formal education is not as good in administering credentials, is not the best or certainly the, not, not the cheapest way to actually learn things. And I think, you know, higher ed in the US at least has been more about athletics and stuff like that. Like, like that that has been an issue. So I, I think, you know, deriving some sort of signal is very tough for employers. There's a whole field of talent intelligence where people try to do sourcing kind of more creatively and intelligently. I think there are some out of the box ways to find some signals. Maybe, maybe there's projects that people can do and they can talk about. There's assignments that now a lot of organizations are giving to their applicants. There's kind of like AI interviewing and stuff like that that's opening up that the hope is that that could be a more informative version of a resume that gets at someone's fluency in a topic.
- [43:26]
A
It does feel to me that we understanding that what the concerns of the manager are is the first step and you know, we identified one which was you need to be a good orchestrator, you need to be a good manager even as a new hire. So there are things that you can go off and do and you know, credentialize or prove. The second is about de risking yourself as a novel resource in the organization. And I'm going to unbundle this in two ways. So one way would be that as a novel resource I may not know how long you're willing to stay somewhere. You know, the last thing I want to do is hire someone who's going to stay for six months. So is there something that the candidate can do to prove a long term interest in a particular direction? And then the second thing that I need to know is that work is tough and, and I need to know that you've got the grit to stick through the ups and downs, stick through being told what to do. And are there ways in which you can evidence that the second feels it's a little bit easier to evidence because you can say, well, I went off and did a seven week route march just with two bags of M&Ms. And I came out the other side. Then there might be ways of evidence in the first. Do you think if we could do that, if graduates could do that, that would help address the Concern that this very, very scared hiring market is exhibiting?
- [44:53]
B
Yeah, I think so. There's this wonderful paper that shows that the value of an MBA program is much higher for women than it is for men. And the, you know, part of the reason why is that, you know, one factor that determines the gender wage gap among professionals is this concern that women may, you know, women tend to leave the labor force at higher rates than men in their, you know, in their 20s and 30s. If a woman gets an MBA, they're spending, you know, $200,000 on, on their professional career and it's very costly. And you know, if someone is making that upfront investment that is a credible signal that they do not intend to leave the labor market, the labor force in the foreseeable future. So the idea is that signals really have to be costly. A big tuition bill is one way to, to show that, that you can bear that cost or something that just takes a lot of time or is painful or, you know, is, is something unusual that, you know, imposes a cost on you and has real trade offs. We probably should look for signals that actually can't be given out for free. Like anyone could get some certification from whatever, you know, voodoo online college. But if those aren't costly, they're not going to send a strong signal.
- [46:02]
A
Let's turn this into that specific piece of advice. So there'll be lots of people listing out of the sort of tens and hundreds of thousands who will have kids who are in that final year of college, in their undergrad. There's a bit of pressure, but there is time. There is time few months before you really hit, formally hit the hiring market. If there is one thing that Junior should do that mom and dad should encourage Junior to do, so that when they get out there in the fall of 2026 to look for a job, they're best placed. What is that one thing?
- [46:37]
B
Yeah, so, so one, one is really related to this technology that I think managing projects end to end is really important. Just doing, doing something that goes over the finish line, you know, finishing things, doing things that require coordination across a lot of different subtasks, I think is a strong signal and will get stronger. So I, I think, I think just taking on an ambitious project and just grinding it out and doing it, I think that that is something that I think is not easy but like will pay dividends. The other is more generic advice, kind of evergreen advice that it's, it's about networking and you know, every, every industry is, is a small world and you know, get to know people get to know how they talk, how they think, listen to podcasts on certain domains and figure out, like, how are people in this space that I want to be in? How do they think? How do they talk? And, like, just engage with them. And that is not going anywhere. Like, that's always good advice, I think.
- [47:28]
A
You know, economists often talk about payoffs. Ben, we've got a payoff out of this conversation with your last two pieces of advice. Ben Zweig, thank you so much for the time today.
- [47:38]
B
Thank you.
- [47:39]
A
Thanks for listening all the way to the end. If you want to know when the next conversation is released, just hit subscribe wherever you're listening. That's all for now, and I'll catch you next time.