AI and Job Evolution – Ep. 297

Welcome to the Accumulating Wealth podcast. I’m Hunter Satterfield. And I’m Judson Crawford. We’re CPAs, wealth advisors, and partners at Cain Watters and Associates, a financial services firm here to help business owners navigate the decisions they face every day. 

You know, one of the questions we’re getting more and more from clients is, “is AI going to take all of our jobs?” And it’s a great question, so today we’re going to cover it. And you’re going to introduce us to the Englishman William Stanley Jevons. That’s right. We might even give some real data that’s showing that the Jevons paradox is in fact in play. Let’s go. 

 Top of the morning to you, Hunter. I don’t think that’s English. This poor man is English. Top of the mornin’. Wouldn’t that be something that Sir William Stanley Jevons said? There’s no sir. There’s no sir for this poor man. Oh, goodness. 

If you are listening to this and you are in front of your phone or computer, you should just Google who William Stanley Jevons… the picture of him. He looks like just your classic- An Irish nobleman. That could be Dickens, that could be Darwin, it could be any of them, right? Yeah, true. Oh. 

Well, yeah, and we’re getting this question a lot. You’re seeing it everywhere. Is AI going to take all of our jobs? I mean, we’re accountants. We’re supposed to be out of business, right? Totally. Yeah. But it’s actually a super fascinating topic, and so we thought we’d cover it with some real-world examples. 

In fact, there’s a lot of studies now that we are seeing, listeners, where in fact job growth is accelerating. So, we’ll talk about some of those things as well. But this is something we talked about at Annual Meeting, so if you were at Annual Meeting, you’re going to be very well informed here. If you weren’t, this is the riveting stuff you get when you come see us in Scottsdale. go.  

All right, so who is William Stanley Jevons? 1865, folks. And let me set the background very quickly. So basically, at that time, England was going through the Industrial Revolution. We were going through the end of the Market Revolution here in the States. 

And the expectations… A part of that Industrial Revolution is that you had steam power getting much more efficient. So, you had a brand-new steam engine that was called the Watt steam engine. It was much more efficient than Thomas Newcomen’s earlier design. And so, as a result, everybody thought that coal consumption was going to go down. 

But Judson, it didn’t go down. What? Isn’t that amazing? It actually went up. And so this cat, William Stanley said, “Top of the mornin’ to you,” and he was like, “Something’s weird here,” and he came up with Jevons paradox. Jevons. So what does Jevons paradox say? Well, at the core of it, it was, hey, coal became more cost-effective, so it actually led to increased use. 

So broadly speaking, listeners, as the price goes down of something, its usage goes up. Or another great example, if things get more efficient, you actually get more use up. Judson, what’s a good modern-day example of what this would look like?  

Well, I think in today’s society, we have more fuel-efficient cars, right? Which leads to more actual miles driven. Exactly. And I think that folks would think, oh, hey, look as fuel efficiency goes up, you would say, “No, we’re not driving as much.” But in fact, now it’s like, oh, we can go further, we can go explore the country, et cetera, et cetera. 

Another good example of one would be that the more- or I should say the faster network speeds get, the more we want to log on and use bandwidth, right? Because it’s like, oh, I can get things- Immediate feedback … immediate feedback. That’s right. So, at our Annual Meeting we actually went through this, and we said, “Okay, what does this look like in historical examples?” 

And there were two really good ones. The first was the cotton gin, right? So, all of us know the cotton gin, Eli Whitney, 1793. The fascinating thing about the cotton gin is that it actually allowed for faster separation of cotton fibers from seeds. So, you would think that it would actually reduce the need for labor, but in fact, it increased the need for labor because the demand for cotton skyrocketed, right? So, Judson, pre-cotton gin, we had 15,000 cotton workers in the US. Post-cotton gin, whoa, look at that number. 2 million. I think it’s more than 15,000. Yeah. 

Another great example from the Market Revolution would be the Erie Canal, right? So the Erie Canal, for those up in the Northeast know, basically connected the Atlantic Ocean over to the Great Lakes. That allowed us to get shipping freight all the way out to the Midwest, and that was basically – before that, Chicago didn’t exist because we couldn’t get stuff out there. It was so expensive to get freight over land. In fact, prior to the Erie Canal, it cost $100 a ton to get freight over land, and it dropped all the way to $10 a ton via the canal. 

So, when we were looking at the stats on this, there was actually no measurable freight volume before the Erie Canal, like it wasn’t a measurable statistic. The best thing we could find is in 1825, which is several years after, that we were moving 217,000 tons per year. Just 30 years later, 2.3 million, so it 11x just 30 years later. 

So, I think the takeaway here, Judson, is that technological shifts they don’t steal labor, right? They just reallocate labor. Yeah, but Hunter, this is AI. It’s going to replace everything, and we’re just going to sit here. We’re not going to do anything. It’s like the movie… That’s right. What’s the Disney movie? We will all be jobless. The Pixar movie? Oh, Up. Not Up. WALL-E, right, where we’re in our chairs in space, and we don’t do anything. Gotcha. You have daughters. How do you not know this? Yeah, I saw it a long time ago. I don’t remember it. Yeah. I kind of remember what he looks like. Yeah. That’s about it. 

Yeah, I mean, I think the big takeaway is what the Jevons Paradox teaches us is that you shouldn’t confuse task automation with job automation, right? And again, we don’t necessarily know where this is going. All we can do is follow this and say, “Okay, well, for basically the last 200 years, Jevons Paradox has rung true, and so it presumably rings true here in some respect,” right? Prove it. Oh, okay. 

Yeah. Well, look, I think we’re starting to see statistics that have started to drop on this, right? And what’s really interesting about it you have to go a couple different places to find this stuff. One really great one is just new business creation. 

So, you would think that when AI comes in, people do – they just sit on their hands and they don’t do anything. But in fact what we’re seeing is there’s almost a direct correlation between AI adoption rate and new business creation in those same sectors. So, there’s a really good stat about the number one sector thus far in business applications since December of 2022, so what are we, like three and a half years, is professional and business services. 

Which everybody’s basically said that’s the one that’s going to be replaced, right? Incidentally, it is the second highest of all on AI adoption rate. So highest on new business formation and second highest overall in AI adoption rate. What other ones, Judson? Let’s see. We’ve got the information sector, which is the number one adopter. Damn, 42% AI adoption rate thus far. Yep. 

And yeah, it’s what? It’s second. Second overall in business applications. I think this is another good one- healthcare and social assistance. So, it’s third highest in new business and it’s top five overall in AI adoption rate. So, I mean, again, I think you go and you look at probably the least one on there, which would be like transportation or mining, basically no adoption rate from an AI perspective, and no new businesses either, right? 

I mean, I think the reality is, and we’ve talked about it on the podcast before, we’re already seeing like, “Oh my gosh, what could we do with this and expand our business, not reduce it?” Yep. Well, I think the other one that’s interesting because it does align so much with what we do in professional and business services in general, is the education services, right? 

It’s not the highest, but high adoption rate and high growth, right? Because we can’t have… this purely, it can make it more efficient, it can do everything, it cannot actually teach a person everything. Absolutely. No, you’re exactly right. 

The Balaji, who’s this sort of world thinker, he said AI is a middle to middle. It’s not top to bottom, right? You still have to have the interface of people on each side, and I think that’s the whole purpose of Jevons paradox, is that you actually still have to have people involved, whether it’s feeding the model or taking the results of the model out. But again, folks, that’s real information from the Census Bureau as well as information from the government on business formation. 

So, another one that I thought was interesting, Judson, is radiology, and I actually have a client who is a radiologist, and this is a conversation we were having years ago. Like, dude, AI could easily take this out, but the data shows differently, right? 

 Yeah, absolutely. And what we see is that really since 20 years ago, that the trajectory of the number of radiologists has not slowed down one bit, right? The income of radiologists have lagged a little bit, but in recent years with the onset of AI, income has actually caught back up with that long-term trajectory, and we’re seeing that radiologists are making now, again, more than $500,000 per year, right? 

Yeah. I think that this is probably the best, especially since so many of our listeners are going to be in the medical field. I think this is the best example of Jevons Paradox at work. So now we can read X-rays much more quickly and much more efficiently. So again, theoretically, you would need less radiologists, but what you actually can do, because you can do it much more quickly, is just read more of them. 

That’s right. Right? The same radiologist can actually read more of them. If that person is reading more X-rays, they are making more money because it is a per X-ray. Even if that unit cost comes down, their overall productivity has gone up. So, I mean, it’s perfectly aligned with what Jevons Paradox looks like. Absolutely. 

Okay, two more, listeners. Before we get to the summary of what we think you ought to do here. So, I thought this one was great. This came out from U.S. Bureau of Labor Statistics just recently, and basically what it looked at was it actually looked at the unemployment rate for young people. Now, why would it look at young people, Judson? 

Because they’re young. No. Because they’re entry level Yeah, they’re entry level, they’re very high level of adoption. We see that here at the firm. Like, our young folks are the ones that are trying all the things on AI. They’re experimenting, they’re playing with it, whereas us old folks are struggling a little bit more. Takes time. It does. It takes time for us, you know? 

But what was fascinating is the BLS data showed that the unemployment rate for 16-year-olds and then 20 to 24-year-olds, those are two ones, and they almost are always in correlation. For the last 35 years, they are in direct correlation. So basically, what that means is your 16 to 18-year-old unemployment rate’s damn near the exact same as 20 to 24. 

In this case, over the last six months, all of a sudden, they’ve started to converge. The 16, 17, 18-year-old rate is actually rising slightly on unemployment. The 20 to 24 is plummeting, and that’s completely counterintuitive to what we would think because you would think all those entry-level jobs are being taken out because of AI. 

But in fact, what’s happening is that everybody’s saying, “I need the 20 to 24-year-olds not only to fill those entry-level jobs, but to start because I just started a new business, and they’re the ones that understand AI adoption,” and et cetera, et cetera. So I mean, again, I think that supports the narrative quite nicely. 

The final one before we get to some takeaways, listeners, is there’s been this huge shift since probably 2015, 2016 and some of you I know listening do this. You basically are outsourcing some of your work to the Philippines for call centers. 

And we see that too- and it’s other places, not just the Philippines, but we see that too even, Judson, when we make a call, right? Like to a call center, we’ll get somebody that’s over there. And that’s been a big shift since 2016. What’s fascinating is that there are now almost two million workers that are doing work in the Philippines on call centers and it’s basically doubled in the last eight years. 

And again, that’s counterintuitive because you would say, “Oh, now a lot of this work can be done by AI.” But in fact, what’s happening is that the lower cost per interaction does not mean fewer interactions. It means more customers are being served, right? So that is, again, that’s Jevons Paradox at its core. 

So the takeaway from all this is that the data, Judson, is supporting that the same things that happened to the Market Revolution or the Industrial Revolution in England or the dot com era when we had fiber optics or the internet mobile cloud era when we saw an explosion of jobs even though the internet came around and we theoretically wouldn’t necessarily see that, the same stuff is happening again. 

Absolutely. And I think that, again, it may be lessening to a degree, but I think that the last thing that you listeners, our clients can do is be afraid of this, right? Or think that it doesn’t have a place in your business. Does it mean that your job will likely change? Yes. Does it mean that some of your staff members’ jobs will change? Absolutely, right? And some of those are hard but ultimately what you have to take that for is keeping up and also the efficiency that you can gain and what else you can do. 

Yeah, and I think a lot of the jobs are going to change for the better, right? Like you and I, when we were starting out here at Cain Watters, there was a ton of data entry. Some of these tools we’re now seeing… it’s a button. You click, and then it automatically inputs all the data. Well, what is that going to do? I mean, are we all of a sudden going to just go fire all of our advisors? No. It gives our advisors the opportunity to be more intentional, more one-on-one with clients, whatever else it is. 

So, I mean, I think again, it’s going to be make for happier employees. It’s going to make for happier clients, but you can’t get there, I think number one is to Judson’s point. I think number two is this: You can’t get there if you don’t try it, right? 

Listeners, I mean, if you are out at a dental show that you might go to, or if you’re not in the dental industry, if you’re not out doing some CE on what this looks like for you and your business, you are going to fall behind because everybody else is out there doing it right now. That’s right. 

And if you’re in the middle or the latter half of your career, don’t let that stop you. Again, your competitors are using those things. But I think the other thing to think about, and we’ve talked about this before, is to make sure you even know what the capabilities are that are in your current software and systems that you’re using in your practice. 

Because a lot of times, like for instance, let’s take a practice software. The practice software has been investing in AI tools that you could have in your office right now that you may not be using to make you more efficient, right? And if you don’t know how to capture those things, you are missing out. 

 Yeah. If you are not partnered with the right vendors in your life that are already integrating and doing this hard work for you in your software tools or your supply ordering or whatever else it is, they will either very quickly or there’s somebody out there that is, right? And so, I think that’s a big takeaway. Like, you have to do it, but then you also have to lean on your relationships that you have to say, “How are you doing it as well?” 

I think one of the things that for you and I is most scary, and this is probably best summarized by our good partner, Dan Wicker. He’s a great, great partner. We love this man. He is getting older for sure. Yeah. If you’ve seen him recently… When Copilot was downloaded to his computer – I’m telling the story and I don’t care, guys. When Copilot was downloaded to his computer he actually asked Copilot, he says, “Hey, Copilot, is it actually possible to teach old dogs new tricks?” 

And Copilot responded with this great, funny bounce back, but basically it was like, “Yes.” And you know what, Dan, and you and me and others, like, it’s taking us longer but we’re doing it, and it’s working brilliantly. So even if you’re on the older side, get in there, ask your younger employees, “Hey, how are you using it?” Ask your friends, “How are you using it?” 

But the takeaway is AI’s not taking all of our jobs. Don’t confuse task automation with job automation. Get out there and work your butts off, and it’ll end up being a really good thing for you. 

Yeah, I agree. And one thing to loop back just a little bit that it reminded me of is, and this is even going back a little bit before the onslaught of all of the AI technology but when there were software that were helping our clients with billing and that became more and more popular, the biggest reason that many clients didn’t want to do that is because they said, “Well, I’ve had so-and-so doing my billing for all this time and I don’t want to replace her job. I don’t want to lose this person. She’s been with me for all this time.”  

Never once did I think we see people that all of a sudden just had somebody sitting in the front of their office and not doing anything, right? Absolutely. It freed them up to better their business in other ways, right? 

And probably – we may be talking about in future episodes, just because you have a software or something that’s helping doesn’t mean it’s foolproof, right? Which is something that I think that we may want to explore in future episodes is like, can we use this? Can we rely upon this for everything, right? Because we’re talking about how – today we’re talking about how this can be used to make us more efficient and everything else like that. Maybe we should talk about what it looks like to use this in certain ways, like in financial planning or tax advice. 

 Yeah. No, that’ll be a great future episode. We need to mark that one down and maybe record it next week. How about that? It’d be weird. Top of the morning to you. Top of the morning to you, Englishman. 

The best way to keep up with this is to subscribe to this podcast. That way, you’ll never have to miss an episode, and you can go back and listen to all the others. If you’re enjoying it so far, leave us a review. Have a question, comment, or suggestion for a future episode? Drop us a line at cainwatters.com/wealth. We really do answer these. And if you want to learn more about what we do when we’re not recording these episodes, visit cainwatters.com to see how we’re helping over 3,500 clients reach their long-term financial goals. 

We actually do know the difference between England and Ireland, just in case you’re wondering. Good day, mate. Put another shrimp on the barbie.