Afford Anything - Job Titles Don’t Mean What They Used To (And That Affects Your Pay) — with Dr. Ben Zweig (Part 2 of 2)

Episode Date: March 3, 2026

There are about 90 million unique job titles in the U.S. labor market. Ninety million. If you are trying to negotiate a raise, switch companies or launch a side hustle, that number has consequences. If titles do not line up, you cannot easily compare pay, scope or seniority. You might be doing the same work as someone with a higher title and higher salary - and never see it. That problem is the focus of Part 2 of our conversation with Dr. Ben Zweig. Zweig is the CEO of Revelio Labs, a workforce data firm that analyzes millions of job postings and online profiles. He also teaches The Future of Work at NYU Stern School of Business and holds a PhD in economics from the CUNY Graduate Center. His work focuses on how jobs are structured and how they evolve. We talk about taxonomy - the systems used to categorize work. A title acts as shorthand for a bundle of tasks. Trouble starts when the shorthand breaks down. Two people with the same title may do very different work. Two people with different titles may perform nearly identical tasks. Zweig explains how large language models can group job descriptions based on actual responsibilities rather than labels. That approach could make it easier for workers to search accurately and for companies to organize teams. The conversation shifts to management. He argues that managers spend much of their time reconfiguring roles as business needs change. Technology accelerates that reconfiguration rather than replaces it. We close with stories about bank tellers and typists. Their titles remained familiar. Their tasks transformed over time. Resource: ⁠Job Architecture: Building a Language for Workforce Intelligence⁠ by Ben Zweig Learn more about your ad choices. Visit podcastchoices.com/adchoices

There are 90 million job titles that are floating around online. 90 million. But are there actually 90 million different jobs? Because the thing is, sometimes two people with completely different titles can essentially be doing the same type of work. And other times, two people with the same title could be doing totally different work. Across companies and across industries, job titles and job descriptions don't map to each other in any standardized way.

Here's the problem. If there are 90 million job titles floating around online, how are you supposed to know what you're qualified for? How are you supposed to know whether or not you're underpaid, you're misleveled, or maybe you're applying for the wrong roles entirely? If job titles don't mean what you think they mean, that affects how you search for jobs, how you negotiate for your salary and benefits, how you position yourself. It affects, it affects how you search for jobs, it affects. when you walk into a new role, what you think you do? Like if you don't have a clear sense of how your role is structured, what tasks define it, what kinds of skills cluster together around it, you can't even adapt when those pieces start to shift as they are quite rapidly right now. Because if AI is changing the tasks inside of jobs, how do you adapt and how do you stay competitive?

If you don't know which parts of your role are scarce and which parts are becoming commoditized. And, frankly, if your title doesn't accurately reflect what you do, how do you negotiate, how do you apply for the right roles? How do you assess the field and make comparisons to what people with similar roles to you are making in other industries or at other companies? How do you bring structure to something that's this chaotic? Dr. Ben Zwegg joins us again today for part two of our conversation, which started in the

last episode. So if you haven't listened to that yet, listen to that one first and then come back to this. Dr. Ben Swag is the CEO of Revelyo Labs, a workplace data company that uses AI to analyze millions of job postings and map how work is actually structured. He also teaches a class on the future of work at NYU Stern School of Business. He holds a PhD in economics from the CUNY Graduate Center. Welcome to the Afford Anything podcast, the show that knows you can afford anything, not everything. This show covers five pillars, financial psychology, increasing your income, investing, real estate, and entrepreneurship. It's double eye fire. I'm your host, Paula Pant. Today's episode is about that first letter I increasing your income or, at a minimum, maintaining your

income in a world in which jobs, especially among knowledge workers, are at risk and rapidly changing. Again, this is part two of our two-part interview with Dr. Ben Zwegg. Please start off by listening to Part 1, which is our previous most recent episode. By the end of this episode, you will have a more clear way to think about how your role fits into the broader labor market and how to navigate your career in a world in which job titles just don't mean what they used to. Enjoy. I want to talk through the taxonomy of jobs. When people are looking for jobs or trying to figure out what their next move is, a lot of times people will go on job boards, go to Glassdoor or Indeed or ZipRecruiter and search for a thing. But there is no standardization of how a job is described.

And so the same role might be referred to as, for example, an executive assistant, a senior executive assistant, a chief of staff, a chief of staff, a. virtual assistant, an operations manager, a project manager. Yeah. Take eight different companies and they might have significantly similar job descriptions for each of those titles. Yeah, it's really a mess. We collect so many different job postings and online profiles. I think we see 90 million unique job titles, which is obscene.

There's no way a human can understand what all those are. And so many different companies have different conventions for how they use titles. And they all need to communicate to the external market to a job candidate who is searching for something. It's a problem that exists on the employer side and also a problem that manifests itself for employees who actually just want to be able to search for something and find something that's a good fit, that's a good match. There are ways to solve this problem through LOMs where it wouldn't have been possible to solve this 10 years ago. Just to give some context. In economics, we sometimes think that there's two main factors of production

in the economy. There's labor and capital. You can do some decomposition and you know that the economy is roughly two-thirds labor, one-third capital. There is a science of allocating capital. It is finance. Finance is all about making capital markets rigorous and scientific. And there is no such science for labor. Part of the reason why that is, there's many reasons why, but one primary reason is the existence of accountants. Accountants are people that categorize. There are millions and millions of people in the world whose primary job is to categorize things

for the purpose of financial analysis. So to categorize, is this sales of marketing, is this cost of good salt? They're doing this categorization manually. And they're generally accepted accounting principles. There's a financial accounting standards boards. There are standards that have been worked on for 100 years, roughly.

And we don't have millions of people categorizing job titles. And we never will. Maybe that would be nice, but we live in a time where LLMs are so good at categorizing things. It's really about finding out what's similar. And I love the way you explain that you have these different job titles where the job descriptions are the same. Because we actually do have a lot of text telling us what people do.

If we think about what a job is, carefully, we can think of it as this bundle of activities that needs to be coordinated. And those activities are right there. They're written down. We have responsibilities sections in job postings. We have bullet points on resumes, what people do. So we have billions and billions of sentences of what people do in their job in the economy. So we can create this way to find similarities between sentences.

And we know that scheduling meetings, booking appointments, these are fundamentally the thing. We know those are semantically similar, conceptually similar, and LOMs are great at telling us that. Then we can see for a given job title, what is the collection of work activities that they do? If they have the same work activities, we know they're the same. So personal assistant, virtual secretary, whatever it is, if we say they have the same set of activities, then they're the same job. Full stop. One person says lawyer, another person says attorney, doesn't matter. They're the same thing. there's also the harder problem that it can solve where one person might say, oh, I'm a product manager at Indeed. And another person says, I'm a product manager at Amazon. And these may be

completely different things. The product manager at Indeed may be kind of an engineering lead. The product manager at Amazon might be doing client success work or something like that. If their collection of activities are different, even if their job titles the same, we should be able to say, these are actually fundamentally different occupations. So then a job seeker who says, I'm looking for engineering lead jobs, should actually see the job title for a product manager at Indeed.

This is something that, if done right, gives candidates a lot more clarity on what they're seeing and also allows employers to organize their own workforce and figure out what do we need more of, what do we need less of? It can also help the platforms

like Indeedon's a recruiter, because they're really in the business of facilitating discovery. So they need a common language for that. Well, and the other complicating factor is that depending not just on the industry, but also the size of the company, when we've dealt with this, when we've put up job postings, especially in a very small company, the role necessarily will involve wearing a lot of hats.

And so there has to be some language within the job description that says, all right, maybe 70% of the time, your role is going to fit this written job description. and the other 30% of the time, it's respond to things as they come up. Yeah. It is incredibly difficult to convey what it is that we need because we don't even know what it is that we need because all of these things come up all of the time. And when you are a company of four or five or six people, it's just deal with it. Totally.

This is in some way one of the most important things to think about when we think about the future of work. Yes. how do your work tasks actually get determined? Maybe they just evolve organically, maybe someone says this is what I need you to do, maybe you discover it yourself, but it's very fluid. Right.

So many people go into a job, and then a year later, they're doing something totally different. Yeah. And this is such a common experience for so many people. And it's not like I think the firms are tricking them, it's that the needs change. Right.

Sometimes we think about, oh, AI as this thing, which will reconfigure work, where we'll have these work tasks being taken away and others being introduced. And we think of technology as this secular trend which creates the need for job reconfiguration. But I think about small companies like yours and mine, where job reconfiguration is just, it's an everyday thing.

There's nothing to do with technology. Like there's different demands on the business. Maybe a client says, oh, we need this. If a client says that to me, I think, okay, how do we do this? how much they want to pay, who's got the bandwidth, is it worth deprioritizing this other thing? And I'm trying to think,

how should we reconfigure who's doing what? Or maybe someone quits, or maybe we find out someone's actually not really good at something that we thought they'd be good at or not interested in something that they thought they'd be interested in, or we automate something. There's so many different things that happen every day that causes us to think,

okay, how do we shift the borders of teams? how do we make sure that the people we have can meet this changing, evolving set of needs of the business? Right. If we had to think about what managers fundamentally need to do, I think it's about job reconfiguration. I think they need to understand the evolving needs of the business, understand the people and what they're doing, and reorient, try to continuously reorient what people do in their jobs to the needs of the business. and that's very fluid. The optimistic part of me thinks that if that's happening every day,

then the emergence of a new technology may be big, but still relatively unimpactful in the general reconfiguration of work. Because the reconfiguration of work is so much a part of everyday business for so many people.

Not to say there aren't organizations that are super rigid and bureaucratic and they have workers who are essentially like a cog on, an assembly line. Those do exist, but I'd be much more nervous about them than I would about these adaptive, nimble organizations. Would it then be the case that managers, especially as we move into a more AI-driven world, managers become even more important because of the fluidity of job roles? I think that's right. The role of middle management will see an emergence.

Wow. Yeah. That's my prediction. I also think, I mean, people love to hate on meetings, rightfully so. A lot of terrible meetings out there. But if I had to make a prediction, I think meetings will be more important as work needs to be reconfigured more often. So then managerial skills become the skill of the future for people who are listening, who are saying, what skill can I bring into the next couple of decades of my career? Managerial skills would be the skill to double down on. I think that's right. I mean, I don't know if they're easy to attain. I don't know if like management classes are any good or if there's like any good way. Or if there's like any good way. to learn managerial skills, but it's definitely something that can be learned on the job.

When I first started managing people, I was terrible at it and I hated it. And then kind of grew to like it as I got better at it, I think it can clearly be learned over time. Yeah. It can certainly be learned over time. In my own experience, I marvel at how slow of a learner I have been, at how long it has taken me to learn things that in hindsight seem obvious to give a couple of examples. One is that given the fluidity of roles, I, for a long time, I resisted giving the people on my team formal titles because it seemed like, it seemed irrelevant anyway given how fluid these roles are. So instead of giving them formal titles, I would give fun titles or funny titles. What was the funest one? So I had one who was a chief sanity officer. I have, this wasn't in my company,

but I have a friend who has somebody on her team that she calls the Excellence Ferry. Okay. Excellent fair is in charge of QA, like Quality Assurance kind of a thing. But yeah, fun or funny titles. The goal was it would evoke company culture that was sort of fun and lighthearted and that felt more like the camaraderie of friends who were all getting together to work on a project. Yeah. It was like what we wanted to evoke. In retrospect, I think it just caused a lot of role confusion.

and it took me a lot of money, a lot of lost money, before I finally I brought in a fractional COO who did an audit of our company, and she was like, you need formal titles, you need formal job descriptions, you need KPIs for every role, you need a scorecard where you have outcomes and people are evaluated relative to those outcomes.

And I was really resistant. I was like, no, no, no, that sounds like a government job. And that is not what we do here because our needs are so fluid. She explained to me that, like, no, if your needs are fluid, that means that you set procedurally every six months, the job descriptions get re-evaluated, rewritten, the scorecards get updated,

and that just becomes part of your clockwork procedures, but you need that clarity and expectation in place. Otherwise, roles are too ill-defined. Nobody knows what they're supposed to be doing. Nobody knows their scope of work. nobody knows their outcomes, and everything kind of descends into Lord of the Flies. Yeah, I struggle with this too. I mean, I also am like allergic to process in a way.

Yeah. There's pros and cons to that. Sometimes there's confusion, and that's what you get. But also, I get that we don't want things to feel like you're working in a government job. And I wonder, this excellence fairy, like when she went to the proverbial cocktail party and someone said, hey, what do you do? Yeah. What do you think she said?

Do you think she said, oh, I'm an excellence fairy? Or do you think she, like, outlined the stuff she did? Or do you think she just chose a different title? Yeah, that's a great question. I don't know. Yeah. When people ask what I do, I don't really know what to say because it's very varied. Sometimes I say I'm an entrepreneur, but sometimes I say I'm an economist, which is more of an

identity than an occupation for me. Like, I did a PhD in economics, so I feel like in my soul I'm an economist. But it doesn't really describe my day to day. spend my time running a company and doing all the things that entrepreneurs do and the fact that I have a background in economics is like not so relevant to that. I wonder if we kind of have these identities and the occupations where occupation is and our job title, our occupation is really shorthand for the collection of things we do. Like if someone says, oh, I'm a QA analyst, then we kind of know what they do. And you're like, okay, you like testing frameworks, you can do this and that.

So we maybe have some sense of the stuff they do. And if someone says, I'm an entrepreneur, then someone hears it. And they say, okay, I kind of know the collection of things you do. If someone says you're a teacher, like, you know what they do. I think we can also kind of maintain this other identity that is like, oh, well, I have my professional brand and professional, I don't know, I guess just professional identity that is a little bit distinct to the day to day. and maybe when people ask what do you do maybe there's a part of them that's asking like how do you spend your day

because that's a valuable question to get to know someone but maybe a part of that question is like how do you think about yourself professionally? I don't know it's like a tough question to answer yeah yeah I mean and so back to what does the excellence fairy say at the proverbial cocktail party I would imagine most people would not lead with that title because the title is not informative

so I think you would have to going back to a job is a bundle of tasks, you would have to at that point just describe the bundle of tasks that you do. Technically, officially, my title is Excellence Ferry, but what I read, yeah, what that means is that I do A, B, C, D, E. Right.

Yeah. But that, that, like, takes a while. Yeah, you do need some sort of, like, shorthand. Right. I mean, I think that's in some ways, like, why we have taxonomies generally. We have taxonomies of products, and taxonomies of the animal kingdom.

Like all over, we have the shorthand, which is like a mental, it's just a mental shortcut. I mentioned I have two girls in preschool. And so much of raising a kid, I think, is about introducing categories. I've been working on taxonomies for longer than I've had kids. So sometimes I'm just like, oh, wow, we are instilling this idea of taxonomies at such a young age.

Right. We are just like, here's farm animals, here's shape. here's colors, here's noise. Like, we were trying to get them to a point where they think in terms of categories. And that way, you know, when we say, oh, there's a clock, we can glance at it in a millisecond. Like, it could be a clock we've never seen before, but we know how it behaves. We know the direction, things go. We know the properties that sometimes things are fast, things are slow.

And that is something that we are also much better at than machines. it's very expensive. It's very computationally expensive for an AI system to recognize a clock. Our brains are wired in a way that's very cheap for us. It's like very computationally inexpensive. I think really because we have these categories, sometimes I think that we are just taxonomical animals. Right. We really need these heuristics to do kind of mental classification and association in cheaply. And yes, we are of classification animals, absolutely. Even the notion of an animal.

Yeah, exactly. Right, I mean, we're not birds, you know? I think a lot about the platypus because that is the animal that defies Linnaean classification, right? It's an egg-laying mammal. Yeah. Based on the rigidity of classification systems, like that should not exist. Mammals, by definition, do not lay eggs. unless you're a platypus in which case you do.

And so you have occasionally, when you have these rigid taxonomical structures, you have the occasional outlier like the platypus that defies those boundaries. And I don't know if there's ever been a good understanding about what to do. I feel like society has collectively just said, we're not going to talk about the platypus. Next question, please. Yeah. Yeah. The way we categorize things should.

really be based on what's useful rather than what's like technically correct. I'm sure we all have that annoying friend who tells us that a tomato is really a fruit. Yep. Yeah, yeah. I like to say my favorite fruit salad is guacamole because avocado and tomato. Yeah, exactly. But like practically, we treat it like a vegetable. Yeah. And we categorize it like a vegetable. And that is the more convenient categorization for us. There's this really fun book. It's called Why Fish Don't Exist. and it's about the first president of Stanford University who has this crazy story. He probably murdered the wife of the founder of Stanford.

Wow. It was nuts. And he was a taxonomist. So he would categorize fish and try to find a new kind of fish and label it and name it and all that. There's been these advances in computational evolution that can sequence the genome of all these different animals

and they found that fish aren't actually a category in that they don't descend from a common ancestor. There are fish that we can't. consider fish that are descended from reptiles or amphibians or mammals or birds or mollets or whatever and we consider them fish now because that just is what happens when you live in an ocean for 200 million years you kind of become a fish it's not actually like the evolutionarily correct way to think about how to categorize animals we shouldn't stop using the category of fish because it means something and you can explain it to a kid and you can understand it pretty easily

and we know they behave, we know they have fins, we have all these properties that we associate with fish. In this way of categorizing things, we always have things that are at the borders. A platypus may actually be practically at the border of a lot of different things. That's kind of always the case. So even in occupations,

we have marketing and sales, and it's useful to think of those, but then we have people that are kind of like between marketing and sales. Yeah. And they're this like platypus-type role, there's some cost to discretization.

By creating categories, we are discretizing, which means we are throwing away the like within category variants. We can say dogs, and we know there's like big dogs, little dogs, and that is like a broad category.

And we know that we're losing a lot of nuance when we just say dogs. And same thing when we say mammals. Like there's weird mantas, like there's platypuses and there's humans and there's whale and they're all really different. That's why we have hierarchy in our taxonomies.

So then we have these like categories where maybe platypus is less weird, but still quite different from like primates. Yeah, I mean, because you think about the classification of dog. You have everything from a chihuahua to a German Shepherd or a Great Dane falls under this very broad category known as dog. You don't have that same level of variance in cats

in the domestic house cat. Yeah. A cat is a small carnivorous mammal that generally adheres to a certain dimensions, a certain body weight, certain dimensions, it doesn't really stray outside of this very narrow band. Yeah. And that is not the case with dogs.

Yeah, it is weird that there's a lot of variance in size, but you could make the case that there's less variance in other things. So I don't have a dog, but people that have dogs I think see a lot of similarities in themselves and other dog owners and they can relate to each other, even if their dogs are totally different types of dogs, because there's some commonalities. They're friendly and they have this positive energy and they want to please you and there's all these things that identify them. And there's other dimensions that don't identify them well, like size. Whereas cats, they are identified by some dimensions.

But they have a great deal of personality variance. Yeah, I mean, I'll take your word for it. I don't really know anything. I like, I don't see, yeah. They have a huge degree of personality variance. Even two cats raised in the same household at the same time, massive personality variants. Same with humans. I mean, humans don't have a lot of size variants.

We are relatively all the same size, especially compared to dogs. But yeah, lots of personality variants. Yeah. And we categorize ourselves in all sorts of ways. There's all sorts of like demographic categories. we categorize ourselves based on ethnicity and things like that that are also kind of arbitrary. Exactly. Yeah. I'm Nepalese. That's a definition based around a set of national borders that didn't even form until the late 1700s. Right. So what would that make me pre-1700s? I don't know. I don't know. Yeah. I mean, maybe back then.

You know, revered to Big Five personality traits or something. I think in some ways it's a good example of like how deep we want to. to go based on like how familiar we are with things. So with humans, like we have such a nuanced understanding of different types of humans, you know, by saying someone's Nepalese, that's informative because it tells us about their background, how they might think about things, like their upbringing probably carries information about socioeconomic status. There's information that we can convey, but that is relevant to us, but probably not relevant to a dog who's like looking at humans. Like they probably don't care. And in some way, it's a little bit similar with jobs where if you're in the finance part of an organization,

you probably only care about like what's sales and marketing, what's R&D, what's this and that, broad categories of occupations. But if you're in like talent acquisition, you really want to know like, oh, what's the difference between like DevOps manager and an infrastructure development, like things that are like really precise. Right. So having this level of hierarchy, how much detail, do we want to get? Like, how deeply do we really care to understand something? In the world of jobs, we have this situation in which there's so much variance in job titles,

in job descriptions, in qualifications needed, in the fluidity of what that role becomes, in what types of products or services that role supports, in the size of companies. The more we talk through it, it's a wonder that a job title or description means anything at all, Yeah. Just given how much room for interpretation and fluidity and variance there is in the similar jobs. Sometimes when we build these taxonomies, we think how many unique occupations are there in the world? How many unique activities? If we take the collection of things people do in their jobs, how many are there in the world?

Like how many distinct things do people do? It is finite. So for job titles, let's say, at about 15,000, you start getting like real synonyms. there's probably roughly 15,000 occupations. If you go deeper, you're really splitting true synonyms. And in work activities, probably like 60,000, which sounds like a big number. But sometimes I think about that and I'm like,

oh, wow, that's actually like not that big. It's a number that we can kind of wrap our head around. And like just the collection of all the things in the world in every industry and every geography, the whole set of things that people do in their job, 60,000 things. Like that's the whole economy. That's the whole labor market.

It just seems not that big. Like sometimes I think about that. I'm like, okay, we can understand this. Like, that's not more than we can process. We can certainly, like, store that amount of data. We can, like, have a catalog that has 60,000 elements. It's way smaller than, like, a library. And, yeah, maybe this isn't so hard after all.

Maybe understanding jobs, activities, skills. Like, maybe these things are actually quite doable. Maybe that does sound like a lot. Yeah. And that 60,000 is always changing. Again, going back to how what 60% of the jobs that currently exist didn't exist back in the early 1900s. I don't know what the rate at which they're changing. And if that rate, I would imagine, although I don't know this, that rate has sped up quite a bit just in the last decade or two.

I would imagine that too. It's very hard to get data on what work activities people did going back a long time. So we have good data on the occupational breakdown, which is why we do these things at the occupation level and find there are lots of occupations that didn't exist. We have social media managers and AI engineers and all that. We don't have elevator operators and switchboard operators and things like that. The better question that we really can't answer is,

what are people doing today that they weren't doing 100 years ago? I love to think of this example of bank tellers. There was this big article, I think, in the Wall Street Journal, when the ATM came out, that bank tellers were just going to be completely devastate demolished, completely obsolete. Now we have something like 10 times more bank tellers than we did when that article was published.

So we didn't see a reduction in the quantity of bank tellers. There's a very common narrative that we hear in economics, which I think is wrong, that says that this is an example of something called Jevin's Paradox. So Jevins Paradox is that when the price of an input falls, we can sometimes actually spend more on that input. There are more use cases for it, more demand for it. Yeah, exactly.

So if the price of gas to drive our car falls by half, maybe we'll drive more than twice the amount and actually spend more money on gas, something like that. It's like possible that's the case. The story goes that the price of depositing and withdrawing money became so much lower. That led to a scenario where instead of having a bank in the center of town, now you have a chase on every corner. So we just expanded the just banking as a phenomenon. The reason why I don't find that to be a satisfactory answer is that it assumes that bank telling is bank telling. And I think that's just not the case. So right now, people that are bank tellers are not primarily depositing and withdrawing cash.

They're primarily doing customer support and relationship management and helping people navigate products and like different types of credit cards and things like that. And it's really much more of a customer service, customer success type role. and it's really just a different job. It could have happened where some executive at Chase Manhattan at the time had made a call that we're going to stop calling them bank tellers, we're going to start calling them relationship managers,

and we'd be telling a very different story right now. The fact that we call it the same thing is just kind of unimportant. Six of one, half a dozen of the other. Yeah, yeah, it's a different job. Whether we use the same title or not is, you know, I almost wish we didn't. But sometimes we do.

You know, sometimes jobs evolve gradually. I get this example where the role of statistician is totally different than it was 50 years ago. We still have the title statistician. Maybe we shouldn't. It's kind of irrelevant what it's called in a way. When the role fundamentally changes so much that it is no longer recognizable to a person who had been practicing that same role, let's say, 40 years in the past. It's like a tree falling in an empty force. It's still the same job.

Yeah, exactly. It's an amoeba splitting into two. Like, is it still the same amoeba? I don't know. Yeah. I mean, maybe there is some convenience to keeping the same set of job titles, which is fine. I think that's like a good reason to keep job titles. But it raises a very interesting point, which is that then casts this question mark on the conversation of quote unquote, will job survive.

What does it mean for a job to survive if a job title iterates the scope of work and like iterates and iterates and iterates gradually? enough that it evolves to something that it didn't used to be, and that new something requires a totally different set of qualifications, experience, skills. What does it mean for that job to have survived? Yeah, I think that's right. I mean, so much of this conversation, I think, ignores that within job transformation, which is, I think, the most important part of all of this. I mean, there was this famous essay by John Maynard Keynes in 1930, which is called Economic Possibilities for our grandchildren. He makes a prediction that 100 years from now, and we're right around the corner 200 years from 1930, he said that we'll be working 15-hour weeks. So that productivity will grow so

much that we will run out of things to do and we'll just live a life of leisure. That's basically the prediction every generation with every technology. Sometimes I think about that essay and I think oh, he was dead wrong. We're not working 15-hour weeks. He was actually right on our productivity growth. Like productivity has grown actually faster than he predicted. But we have this insatiable appetite. And now we want to consume things that we didn't consume before. And we have this growing need for what we think should be done. And I think, okay, we're not going to run out of things to do. There's lots of things that we want in the world that don't exist in the world. So that kind of makes me feel like, oh, he was wrong because he didn't account for our growing

appetites. Then sometimes I go back and forth in this. Sometimes I think, actually, maybe he was right. Maybe we actually do live a life of leisure, but that's happening within our job. Like right now, you and I are working and like, I'm having fun, you know, so, you know, this isn't like backbreaking labor. You know, we're not in the sweatshop. And that is the reality of a lot of people's jobs now. Like jobs have gotten more social and there's more a little more fun. There's a little more leisure. and that is the transformation of work that does give us a little bit, I mean, people are pursuing jobs based on what they want to be doing more than what just pays the most more and more. That's something we have lots of evidence for that like newer generations put a higher priority

on things other than compensation, work-life balance and management and culture and all that. And that is consistent with this idea that we actually actually, maybe are pursuing more leisure, but so much of these phenomena of job displacement, of automation, like, it's happening within jobs. It's not happening between jobs. It's so important to think about what's happening within each job. How is each job transforming? The advancement and the proliferation of knowledge work, I mean, that that is a good point that there is a degree of leisure within the domain of knowledge work in particular, because to be able to sit in a climate-controlled environment with adequate air conditioning and adequate heat and

push buttons with your fingers all day. Yeah. That is certainly quite leisurely in comparison to a lot of the very dangerous jobs that millions and millions of people have had and millions still do have in other industries, which then bringing it back to AI makes it so jarring the notion that this is the first thing that has really disrupted knowledge work. Yeah. for sure. I mean, it does feel weird. I forgot who said it, but someone had some famous line that I want AI to do my laundry and dishes so I can write and make art. I don't want AI to write and make art so I can do laundry and do dishes. Yeah. It is kind of threatening the things we like about our jobs, the thoughtfulness, the deep thinking, the analytical component. I get that

that's scary. The question is, what is the remainder? Is what's left over even more fulfilling, or is it less fulfilling? Like, if we really were being forced into more jobs that require dexterity and things like that, I don't think we all want to be busing tables and doing laundry, but maybe we do want to be overseeing more ambitious, exciting projects and coordinating more abstract activities that create more value. Maybe that is actually exciting in a way that AI can unlock our workflows to be more exciting and more human in a way. Because there are frontiers that we haven't even approached.

We're not building cities underwater yet. Yeah. It strikes me we still take trains just like we did in the 1800s. There are so many facets of life that can be innovated and iterated and new frontiers that we haven't even approached. that could be developed. Totally. There's so many pockets of our economy that have had really low productivity growth. Housing, you can argue, has had even negative productivity growth. That's creating all sorts of problems for affordability and walkability, you know, things that are desirable. Walkability and density of cities and how do you balance density with people. I mean, people

inherently do enjoy having space, but also enjoy living in the community that comes with living in density and how do you balance those two? We need more productivity. And it's worth kind of remembering that like the future is already here. It's just not normally distributed. We sometimes talk about like our own jobs as being, we can do more interesting, more enlightened work. But there's so many parts of the world where that's not the case where they are still kind

of working in sweatshops. And that kind of productivity, like taking these innovations to other places can be really useful, also bringing them here. We don't have bullet trains that they have in. Korea or Japan. I took the Metro North today. It's fine, super annoying. Yeah. But it's not a bullet train. Yeah. And it does feel like we as a society, I mean, me personally, probably all of us, do feel a little bit impatient with how quickly we're like making progress or not making progress in certain pockets of the economy. It's like we are hungry for more. Yeah. Whenever I take the trains

around New York, like when I take the Metro North train or the New Jersey Transit line, where I go to Penn Station, I do feel a little bit like I'm living in the Gilded Age. Yeah. I feel like it's the late 1800s, early 1900s. I don't feel like I'm living in 2026. Yeah, we want more. And we'll get there. I mean, one is that we'll get there through technology,

but also through combining labor with technology. In my company, I'm sure you feel the same way. I would love to have 30 more people. I know exactly what I would do. Yeah. If someone can free up more of their time, I've got a long list of things I want people to work on that grows faster than our ability to work on them. Even within a company, there's just so much more that we'd like to deploy labor to do.

Yeah, yeah, that is true. And I think almost every small business owner feels that. Looking at the jobs report, looking at labor, you know, labor stats from the past year, what we've seen is job growth has been slow. And most of that growth has accrued to the largest. companies. Looking at the labor reports, the hardest hit companies have been the smallest ones. And yet, as you and I have just talked about, both of us lead small companies. And both of us are like brimming with ideas for what I know exactly what I would do with. I don't know what I would

do with 30 people. I definitely know what I would do with like 10 more people. Yeah. I could definitely put 10 more people full time to work tomorrow. And I would be thrilled if I had the hiring budget for that. Yeah. And the minute those 10 more people were deployed, give me six months and I would know what to do with the next 10. You're right. That employment has been slow, like way too slow for the economy at large, especially so for small firms. It's hard to pin down exactly what's driving that. I suspect it has more to do with capital markets than with labor markets. It is harder to get financing for a small business, for a startup today than it was back in 2021.

Venture capital as an industry has transformed very rapidly. First of all, it's been very unsuccessful as a business model in the last decade or two, but also the nature of venture capital has been much more concentrated in specific bets

that attract a lot of risk capital and then they've been becoming more private equity-esque with the rest of their capital. So, you know, the big Silicon Valley VC firms, in the last decade, basically invested, like the first five to ten years ago, we're basically just investing in crypto. And now they're just investing in AI.

I mean, I'm exaggerating a little bit, but I think it's like true enough that they're really frothy valuations, tons of money going toward AI startups, still crypto startups now. But SaaS companies, hardware companies, science, And they are not getting funded.

They're being funded on multiples of like earnings rather than like multiples of revenue, which is like weird. You know, we think of that as like private equity behavior. Right. Which is very risk off and trying to, you know, pressure companies to focus more on profitability, a little less on growth on the margin. So the excitement around AI is probably funneling some capital away from growth of small business businesses.

So it makes sense. the extent that the VC model prioritized growth rather than profitability because the priority was forward-looking earnings, right? The priority was, we're going to place evaluation on you based on what we think you might make five years into the future. Or 20 years into the future, yeah. And now that we're entering this AI world, forward earnings, forward revenue is so unpredictable. The value of a company is so unpredictable that the VC model can, you know, can no longer place a valuation today on what a company might be worth two years in the future even, because who knows what the world's going to look like two years from now?

Yeah. So because the future is so unpredictable, you can't discount that back to today, and therefore you can't really place a good valuation on companies today. And so then you just have to value a company based on a revenue multiplier. Yeah, which is the odd-fashioned way. I think that's exactly right. Right. Investors, companies, I mean, we are in a high discount rate environment. We are discounting the future a lot relative to the present. We are prioritizing for the present. Some of that has to do with interest rates and the macro economy. I think a lot of it has to do with technology, like you said, I think a lot of it has to do with policy as well. There's uncertainty around like what is the world going to look like. We have a lot more uncertainty around the geopolitical situation, around what the technological frontier is, around what. migration patterns. I mean, so many things that really affect the composition of our economy,

they're so uncertain right now. So I think it's creating an environment where it's very hard to invest in the future. So I think it's very rational. I don't want to blame VCs. I mean, maybe a little bit. Yeah, no, it is. Yeah. It's completely rational that the valuation. And I think that it's actually probably a good thing. When you're making a lot of speculative bets, placing these speculative bets on companies, not based on their current revenue, but based on expectations for future revenue, which is by definition speculative, you get a lot of busts. Whereas if you're, if you are placing valuations on companies based on current earnings or current earnings growth, when you're placing a valuation on a company based on actual numbers

that they have achieved, the valuations come down, but the valuations are also, rooted in something that is much more fundamental and real. In your situation, if you wanted to hire 10 people, you could accept a bad valuation on your business and raise some money and hire 10 people. But that would be very expensive. But if we were in a different environment, you would borrow money cheaply and hire 10 people if you knew that they would be deployed to product events. Is it then the case that because of the changing nature of the VC model and the changing nature of how we are placing valuations on small businesses, that's why we've seen so much of the labor growth accrue to large companies?

I think so. I'm sure there's other parts to the story, but that's one that I think is certainly happening. There's other like frameworks to use to think through that. Maybe there are more efficiencies in large organizations. When I teach the future of work, I sometimes think, why don't firms grow to infinity? Why don't we have like infinite specialization or the opposite?

it. Like, why don't we just have a bunch of independent freelancers like running the economy? What dictates the optimal size of firms? And I think ultimately it comes down to coordination costs. There's some costs to coordinating a small team that's lower than coordinating a big team. And there's some returns to specialization. So a big company can have specialists and they can really benefit from that specialization in a way a small company can. I think it's this balance of coordination costs and specialization. And the relative cost and benefits, of those change due to technology and regulation and things like that. So I wonder if the technology we see now is reducing the coordination costs in large organizations relative to the coordination costs within small organizations. So a small company, a company of five people, like they talk

to each other. Like their coordination costs are the same as they've always been. But in a large company, they coordinate using technology. And as technology gets better, we get Slack and summarization and this and that and Zoom and whatever. Maybe coordinating across lots of people is becoming cheaper. This goes back to what we were talking about at the top of the show, the difference between execution and orchestration. Because that coordination cost, that really speaks to orchestration. And if the cost of coordination and therefore, of that orchestration comes down, that would accrue to large companies because in a small company, if you've got a team of five people, AI can help with execution, with task execution.

Yeah. But the orchestration of that in a team of five, a company of five is not going to change. Totally. The execution benefits accrue to small and large companies alike. That really shouldn't dramatically affect the different distribution of company sizes. I don't know if we have a lot of evidence driving the change in the returns to large companies versus small companies. It's worth thinking through. Do you think this pattern will remain, this pattern of large companies being so far the winners in the AI landscape? Like, will that pattern continue to persist?

Or will eventually small companies, which can have more niche specializations, will they be able to kind of innovate the way out of this? Yeah. Or specialize the way out of this? Yeah, I would put my bets on small firms relative to large firms because here's one way to think about who's well positioned to take advantage of new technology. Like we spoke about job reconfiguration and how managers really need to reconfigure jobs. I think a lot of big companies had been successful because they have very clearly streamlined processes and a lot of efficiency from like fine-tuning processes with a lot of structure. you know, if you think about a big bureaucracy, like with an assembly line, like everyone's got a very specific job, I think it's harder for them to reconfigure the tasks in people's jobs. So I think that they will be not so well equipped to be able to restructure what people do as easily, whereas small firms are adaptive in nature.

We are constantly reconfiguring what people do. And when a new technology comes out, it's easy to use it. There's no bureaucracy to run through. I mean, I still hear of big companies that are still struggling to figure out their privacy policy around using chat EBT for their work. And they just have to use some like gated version that's super clunky and you can't export data. You know, they like can't get it together. And I think that's going to hurt them a lot. I'm also very concerned about artificial rigidities in what people do.

so like occupational licensing and professional trade organizations, like if you are a phlebotomist, you're allowed to like draw blood. And if someone asks you a question that you know the answer to, you're like legally not allowed to answer that question. You have like very hard constraints on what activities you're allowed to do, what activities are not allowed to do.

There are these regulated industries that I think will also not be able to adapt to technological change. change and I think that'll hurt them. I'm concerned for the future of those organizations that are typically larger, whereas the smaller upstarts that are trying something a different way can adapt very flexibly. That makes a lot of sense because the more you have that occupational rigidity, the less that you are able to, we just talked about the fluidity of roles,

the less you are able to adapt to the fluidity of roles which is necessary in a rapidly changing environment. And if you can't, it's adapt or die. If you can't adapt, you die. Yeah. And our company is someone, one of our economists just wrote a newsletter about consulting firms and how consulting firms are shifting their workforce dramatically. They're taking on lots of AI roles. They're like trying to embed AI and everything they do, like very aggressively. And right now these like AI roles, like roles that didn't exist before AI, people whose job it is to use AI technologies and implement it, they now exceed the number of entry-level consultants in consulting firms. I don't know.

When I saw that result, I was like, wow, this is crazy because we think of consulting firms as like the launching pad for ambitious young people. Yeah, yeah, it's like the Harvard to McKinsey pipeline. Yeah, yeah, exactly. People talk about Palantir as this example of the Stanford computer science undergrad is the new Harvard MBA. I think we are seeing some of that. And consulting firms have been allowed to adapt.

And I think of law firms as kind of a good parallel to consulting firms because it's a similar model. It's a similar like up or out type of model where you have the partners who are the experts really selling business. I think there's just a lot of similarities in their operating model. But law firms are bound by professional associations. There is the bar association. And consulting firms are not. we don't see this happening at law firms, but we do see it happening with consulting firms. I think that probably is due to the Bar Association having imposed rigidities on what lawyers

are allowed to do, how they can do it, and I think that's worrisome. How should an individual in either of those professions deal with that impending risk? Like, how should an individual who is currently a practicing attorney be thinking about that? For practicing attorneys, I think what the professional association is doing is sort of protecting the incumbents. So I'd be much more nervous if you were about to look for a job in law. If you're in law, if you're like a, what is it, L3L? Yeah, yeah, something like that. Yeah, yeah, the 30-year law student.

Yeah, so I think they've been having a hard time, even the last couple decades. I think the majority of law school graduates do not end up practicing law, which is crazy. But yeah, so it's been a tough market for lawyers generally. I think it'll get even tougher because there are, number one, like constraints on new entrants and constraints on how productive they could be, how quickly they can implement new technology or just like use new technologies, but also they're challenged by free and cheap substitutes for legal work.

You know, if someone wants to like write up a contract, they can just do that using AI right now. Maybe there's a case to be made for whether that's a good idea or not. I personally use AI for a lot of contracts, but even just with this book, I wrote this book with Wiley and they sent a contract over. I just asked Chachapit, like, what's standard, what's not standard? Like, is this a good term?

What should I push back on? And I did this negotiation with essentially an AI lawyer. And I think it was fine. It was good. And who knows, maybe it'll come back to bite me. But, you know, it's better than paying a couple thousand bucks for a lawyer to look it over. I think there is competition just from practitioners having less demand for lawyers. Especially for the kind of lower level tasks.

Yeah, exactly. You know, procedural lower level tasks, drafting up a lease for the rental of a single family home, that type of a thing. I mean, most lawyers are just going to copy and paste that anyway. Exactly. Or use the template from the state. It reminds me of a joke. A couple of friends and I, we got these invitations. The invitation said, oh, dress code is business casual.

My friend looked at me and she was like, what's business casual? And my other friend was like, eh, it's when you don't read the contract too closely. Yeah. Nice. So, yeah, I think the new one is like, yeah, when you use AI to read the contract. Yeah, yeah, this is time and a place. That's business casual. Yeah.

We talked earlier about somebody who's between the ages of 25 to 30 and they're starting their career. I want to close this out with someone who's old enough that they are established, they have a network, let's say between the ages of 35 to 55, right? I guess that's a wide range, but like you're old enough that you're established, you're an incumbent, you're old enough that you don't necessarily want to retrain in a different occupation. Yeah. At your age and your current stage of life, you don't want to go back to school to obtain. some high barrier to entry credential, but you're also not ready to retire. So that 35 to 55 age range, what should anyone in that bucket know as we think through the future of work? I would almost segment it into two types of people there. Some people in that age group are

working at big companies. Like you said, more and more. If you're playing a role in a big company, you're probably something like a knowledge worker. You're probably connecting things. answering emails, involved in a variety of different business functions. And your job probably does transform actively all the time. My advice to that person would be to try to pay attention to how your job is transforming. Even just like every once in a while, just look back to what you were doing three months ago and just think, oh, how did my job transform? What led to that?

Was that because my manager wanted me to do different things? Is that because I just took on new responsibilities? I mean, there's this phenomenon of job crafting. It's this idea that people can sort of shape their own job based on what they like, what they think is more productive, what they find more meaningful. And that probably happens a lot, even when we don't explicitly think about it.

So I would encourage people to just be purposeful about thinking through what you do in your job. Are there parts that you like and don't like? How would you like to transform? that and how should you interact with your manager to align that to the objectives of the business. If you can find a way to have a more fulfilling job and also be delivering productive output, great, go for it. Reconfigure your job such that it does that. There's other people in this age range who are in more narrow jobs that are not knowledge work. So let's say, you know,

someone's an altrucian or something that is, I mean, maybe that's not a good example, because that's probably not so likely to get automated, but something where your occupation might be vulnerable to automation. Most jobs don't get automated wholesale, but there have been times when that did happen. Switch word operator typists, things like that. That is concerning. So will we see AI displacement, like labor displacement?

How responsive can suppliers of labor be? I think probably quite responsive. So in this example of switchboard operators, there was really very little unemployment that resulted from switchboard operators. Most people who were operating switchboards went into secretarial work. They were able to reorient what they did very quickly. And I think because the skills and tendencies overlapped, even though the work activities were different,

if you do feel like your job is sort of at the cusp of what might be automated, I think I would just encourage them to take an inventory of their skills, their interests, what they're good at, what they would do if they didn't have to worry about money, which can be a good starting point to thinking about like, where is my energy? But sometimes what does require some adaptation, going back to the Luddites, don't smash the machines. Don't be like John Henry. It's much more productive to think of how you can start. sort of reorient your capabilities towards some other productive ends.

Actually, we should close on the story of switchboard operators and typists, because I think there's some hope there. Both of those job categories became obsolete. Yeah. But it actually then led to a proliferation of more advanced, more sophisticated jobs. Your mom was a typist? Yeah, my mom started her career as a typist. She was like, trained on a specific typewriter, the IBM, whatever, and we had to get trained. And she was very proud of being able to do however many words per minute. That was her job. Now, what's interesting about being a typist is that you're not just typing. Like, you're also figuring out how to message things, how to translate information and make things concise. It's not really just

hearing words and like typing them out. As computers became more widespread and typing got easier and easier, she kind of gradually made her way into more secretarial work and then started doing more more secretarial things and was in that occupation for a while. And then eventually was doing secretarial work at large companies at law firms. She also worked at IBM for a little while, did more, more secretary. I'll work and eventually was involved in managing a lot of information and files and

categorizing things, which is kind of what people do as a secretary. And eventually just became the expert on filing documents related to, I think it was like subsidiaries of a company and like the documents for incorporating the subsidiaries of something and became this like manager of a subsidiary database for a company. I don't think she ever thought of herself this way. When she told me about this, I was like, oh, so you're like a database administrator. And she was like, I guess, you know, like she never thought of herself as a database administrator. But like, that kind of is what she did.

And that is a role that you can hire for today. But she always thought of herself like some hybrid of like secretary and paralegal and this hodgepodge job of like understanding and categorizing information. These things do happen gradually. The activities didn't change as much as the job titles changed, which is kind of a counter example to the bank tellers, where the activities changed a lot and the title stayed the same. So, yeah, I mean, you kind of have both sides of the same coin. And I think that there's quite a bit of optimism there for people who are worried about job displacement because both the bank teller example and the typist to database administrator counter example both illustrate the changing nature of. jobs, regardless of what the title is, the changing nature and then the very fluid nature of work.

Yeah. I mean, we spoke about like how responsive is labor supply to automation. Can workers find new things to do? Another issue is how responsive is labor demand to automation, to technology. So do firms actually adopt technology? And when they adopt technology, is it fast or is it slow? And I think for better or for worse, it's a bad thing, but firms don't adopt technology all at once. They're slow, especially big companies, are really slow in implementing change. On the one hand, that's not good for those companies, but on the other hand, it does give us time to adapt. Workers can find new things to do. They have time because firms are not going to implement new technology just like that.

Yeah, we have a chance to kind of reconfigure work by ourselves. through talking to our managers, by understanding the changing needs of the business. All of these things happen organically. And yeah, I think that is kind of a hopeful story. Beautiful. Well, thank you for spending this time with us. Where can people find you? Yeah, thank you.

Learn more. LinkedIn? I post a lot about labor market insights and data on LinkedIn. That's probably the easiest. Wonderful. Well, thank you. Thank you so much. Thank you.

Yeah, thanks for having me. Thank you to Dr. Benzwegg. What are three key takeaways that we got from this conversation? Key takeaway number one, job titles are chaos, and that affects how much you make, because there are 90 million unique job titles floating around. That's messy and chaotic, but also it creates information asymmetry because if titles are inconsistent, then you can't easily compare roles. You can't easily figure out whether or not you're underpaid.

So understanding your actual tasks and how those tasks map to the market, that gives you leverage when you go in to negotiate for your pay. I think we see 90 million unique job titles, which is obscene. There's no way a human can understand what all those are. And so many different companies have different conventions for how they use titles. And they all need to communicate to the external market, to a job candidate who is so. searching for something. It's a problem that exists on the employer side and also a problem that manifests itself for employees who actually just want to be able to search for something and find something that's a good fit, that's a good match. That is the first key takeaway. Key takeaway number two,

managers become more valuable in an AI world because if AI automates execution, then the skill that is scarce and valuable is orchestration, management. And so the ability to reconfigure roles, to realign teams, to rethink workflows, those orchestration types of tasks increase in value. And that means if you want job security, promotions, raises, be a great orchestrator. If we had to think about what managers fundamentally need to do, I think it's about job reconfiguration. I think they need to understand the evolving needs of the business, understand the people and what they're doing, and reorient, try to continuously reorient what people do in their jobs to the needs of the business. And that's very fluid. The optimistic

part of me thinks that if that's happening every day, then the emergence of a new technology may be big, but still relatively unimpactful. in the general reconfiguration of work, because the reconfiguration of work is so much a part of everyday business for so many people. Finally, key takeaway number three, jobs don't disappear, but they transform.

Automation very rarely just wipes out a role overnight. Instead, it reshapes what happens inside of that role. It reshapes the tasks that are demanded of that role. And if you're adaptable, then that's good. news because it means that your upside comes from evolving with the work rather than jumping ship. So much of these phenomena of job displacement, of automation, like it's happening within jobs. It's not happening between jobs. It's so important to think about what's happening within each job. How is each job transforming? Firms don't adopt technology all at once. They're slow, especially

big companies, are really slow in implementing change. On the one hand, that's not good for those companies, but on the other hand, it does give us time to adapt. Those are three key takeaways from this conversation with Dr. Ben Zwegg. He is the CEO of Ravlio Labs and the author of a book called Job Architecture, as well as an adjunct professor at NYU Stern School of Business teaching the future of work. Thank you for being an afforder, part of the Afford- Anything community. If you enjoyed today's episode, please do three things. First, share this with your friends, family, neighbors, colleagues,

cousins, babysitters, teachers, dog walkers, postal workers, share it with your barista, with the guy at FedEx, with a person in your kids' carpool, share it with your former professors, share it with your favorite economist, and your least favorite economist, share it with all the people you know, because that is the single most important way that you spread these ideas and you spread the message of great financial health. Number two, please subscribe to our newsletter. It's completely free and worth every penny. afford anything.com slash newsletter. Number three, please open your favorite podcast playing app

and leave us up to a five-star review. Please write a few words while you're doing so. Let us know what you enjoy about the show. I read every single one of these, and these are incredibly important at helping us book fantastic, insightful guests. Thank you again for being an afforder. My name is Paula Pant.

This is the Afford Anything podcast, and I'll meet you in the next episode.