This week on the Andrew Yang podcast.
I mean, you can think of it this way. Imagine you're at the Coca-Cola company, right? And you start spending all this money on AI from Anthropic or OpenAI. There are really only two ways to recoup that investment, right? One is to cut costs, and it could be other costs, not just labor, if you can use the AI to do things more efficiently. But if you're going to find, as you say, a trillion dollars across the economy to get this money back, labor is going to be the primary target.
It is my pleasure to welcome back to the podcast futurist and New York Times best-selling author of the seminal book that's actually being re-released with some new material right now, Rise of the Robots: Technology and the Threat of a Jobless Future. You can see why I like him. Martin Ford, welcome back, Martin.
Thanks for having me, Andrew.
So, Martin, holy crap, man. You are the OG of calling AI and automation, really automation generally, because your book says, "Look, it's not just AI, it's information technologies. It's been going on for a long, long time." So, we're going to get the update from you because long-time listeners might remember Rise of the Robots in 2015, your TED Talk. And now, I don't know about you, but people come to me every day saying, "Man, you called it, AI blah blah blah." You must be getting a lot of the same stuff.
Yeah, I mean, we definitely are at a point where, you know, we seem to be much closer to this actually happening. You know, I actually wrote my first book, my self-published book, The Lights in the Tunnel, in 2009. So that's 17 years ago, and I thought then that this might be 50 years in the future, right? When I wrote Rise of the Robots in 2015, I was thinking maybe it's going to be 20 years. But now here we are. It's 11 years on, and there certainly are indications that this is beginning to unfold. And we don't know that for sure. We know we have to examine the data and see if it's really happening at this moment. But what I've always been saying is that it is inevitable. Maybe it won't happen now. Maybe the AI bubble will collapse, and it's not going to happen now. But it will happen eventually. I mean, I think it's inevitable that progress in AI and in robotics is not going to just stop, right? It's going to continue relentlessly. And eventually we're going to get to the point where really the bulk of human labor is going to be displaced. And we're going to have to figure out how we can restructure our economy and our society in order to deal with that.
You and I are on the same page completely. So, let's talk about what's going on right now, which you write about in the new material in Rise of the Robots. And I'm sure you've done some thinking even after writing the new stuff in 2025, because the stuff is happening in real time. So I talk to smart people, you talk to smart people. I'm convinced it is definitely happening now.
You know, I talk to CEOs. A CEO reached out to me out of the blue. Publicly traded company, uses AI in another industry. And he said to me, "Look, my revenue has gone up 300% and I've cut 30% of my workers. And my incentives are just to do more of that." He actually said, "The more workers I cut, the better I'm going to do. So, if this is true for me, and it's true for other CEOs, what are we going to do as a society?" And so he fast-forwarded to universal basic income, which in his case was like, let's put some markers in place. And as we kick more workers to the curb, let's increase the amount of money that flows to them. And this is a major company CEO who just reaches out to me out of the blue and says, "Hey, will you have breakfast with me so I can tell you about what I'm seeing?" And then he's like freaking out. And he's not alone. There are another, you know, host of CEOs who've told me something similar. Simultaneously, you have college kids who are running scared. Their prospects have dimmed very, very markedly on campus, particularly if they're computer science majors. And those kids would have been fielding multiple offers just two, three years ago. And now they're trying to find one job offer, and their placement rates have fallen off a cliff. So, like, I believe this is happening in real time.
And so some of the economic data is more ambiguous, which I get. The economic data tends to be a little bit of a lagging indicator. So, when people come to you and say, "Hey Martin, is this happening right now?" what do you tell them?
Yeah, I mean, I'm not as confident as you are, I suppose. I, you know, I'm open to either possibility. For sure, there are a lot of anecdotal type examples of this. In fact, I cite some of your writing in the book, right? Some of the people that you talked to in the final chapter. I have it there. And there are tons of examples like that. However, as you say, the economic data is not yet clear. The economists are definitely interested in this. They're sifting through the data, trying to figure this out. Some of the best work I think has been done by Erik Brynjolfsson at Stanford. And I cite him in the book, right? He found by looking at payroll data from the company ADP that, you know, there definitely was seemingly an impact on entry-level workers that were in fields that were closely anchored to AI. Things like customer service, computer programming, and so forth. And so far, the only really definitive evidence is mostly in entry-level jobs, which you're right, that's a tremendous problem. Even if it's only that, I mean, that's a huge problem.
I've got a daughter that she's just finishing up her first year of college, so she's not going to graduate for another three years. I don't know what the world is going to really look like then. But my intuition would be that it probably begins with entry-level workers who are doing more routine, predictable work. But over time, it's going to scale up to impact even more experienced workers. And then the question is, how long will that take, you know, before we really see that impact? I think the economists are going to debate this. I mean, they debate everything, right? So, by the time the economists tell us definitively that this is really unfolding, I think all the rest of us will already know, right? It'll be obvious. But yes, I'm very concerned that this might well be the moment in time. But the point I would make is, even if it doesn't turn out to be that it's going to happen right now, it definitely will happen. Probably not too far in the future.
Yeah, so we have seen a real shift in the language that a bunch of CEOs are using. So, even two or three months ago, Dario Amodei was saying 50% of entry-level white-collar jobs are going to be gone. The CTO of Microsoft said something similar. There are a lot of folks saying, "Look," I think Mark Benioff said like maybe we're not going to hire, you know, junior people anymore. And then, about three weeks ago, everyone started saying job fears are overblown. The CEO of Goldman Sachs wrote an op-ed in the New York Times saying, "Hey, there'll always be new stuff for the smart people to do," and like it went, "we're not going to cut jobs." And that one interested me in part because someone at Goldman Sachs said to me over dinner a number of months ago that he's on the team that's supposed to look at figuring out how many workers they can automate.
Honestly, and he had like a very, very significant number. You know, he was like 40, 50%. And so hearing someone inside say that and then seeing something very, very different from the CEO in the New York Times. So there's been this massive concerted effort, in my view, for a lot of the leaders to try and assuage the job concerns. And I think they're assuaging the job concerns in part because people have gotten so mad at them. The approval rating for AI is 26% in one poll. Very famously, a bunch of college seniors booed commencement speakers that used the syllables AI together.
Like anytime it came up that they would get booed. And several of these companies are trying to go public. So, you know, they have this massive interest in at least making it to the public gate so they can, you know, be valued at a trillion dollars or something along those lines. And so I've seen like a real shift in the language. I personally think that they just got together and were like, "Look guys, people are starting to despise us, and like really despise us. Protest the data centers. Like, we've got to stop scaring people." Now for me, my interest is not necessarily scaring people. My interest is just trying to tell people what I actually think is happening. And it's one thing I enjoy about you because you're, in my opinion, like a disinterested observer in the sense that, you know, you also just try and call things like you see them in real time, but also in your case trying to project into the future. And so, what you just said was like, "Look, even if it's not now, it's going to happen whatever like several years from now or whatever the time frame is." But do you get incoming in a particular way where people are like, "Hey Martin, you're scaring people. Stop it," or anything along those lines?
I mean, I haven't heard that too much. I mean, I've always been perceived as kind of negative since I, you know, I've done a lot of speaking engagements, speaking to groups, and usually the response to that has been quite positive. But obviously, when I started doing that back in 2015, the level of concern was not so high, right? Which is something that was kind of interesting and futuristic. Now the fear is absolutely really there. But I think people recognize that we, you know, we need to talk about this. We can't just sweep this under the rug, and that's going to be the worst possible outcome if we're just in denial, and then this unfolds, and we don't have any plan, right?
I mean, really my intent in talking about this is to get other people, smart people, you know, policy wonks, people like that, to really engage with this, because I don't pretend to be smart enough to figure out all the answers here. It's going to take a lot of work. I mean, we can talk in general about a UBI and things like that, but actually implementing that is going to take tremendous work, and it's a tremendous political challenge. I mean, so we really need people more engaged with this. And, you know, that's sort of my hope by publishing this update to the book, that I can get this out there, and really get a lot more people engaged and taking this seriously, so that if it does unfold perhaps rapidly, we've got some plans that we can begin to execute.
Thank you listeners. Please make sure to like and subscribe, and if you have any questions send them to
[email protected] and we'll do our best to answer them. Thank you.
Hello listeners. You like money? I like money. Let's help everyone save some money. So I started Noble Mobile, the first cell phone plan that pays you to use it less, a data dividend. That's right. It's for real. Go to noblemobile.com/yang and find out how you can get bonus cash just for signing up. Get months for free, the whole shebang. It's awesome. Go to noblemobile.com/yang to find out more.
Yeah, I'm totally with you. I mean, if we wind up kicking thousands and thousands, maybe hundreds of thousands or millions of workers to the curb, there has to be some kind of a response, you would think. And it is in many ways a political challenge that, as everyone knows, I've set myself to trying to address. I ran for president on universal basic income in 2020. You know, I've been working on democracy reform efforts. I think the two-party system right now is not designed to actually respond to the challenges of today because their incentives are not actually to solve problems. So I'm with you on all fronts. I'm also with you in calling for a universal basic income, which now I think is totally common sense. You know, when I walk the streets, the vast majority of people are like, "We need it, we need it." And I was just at a policy conference in Michigan, and I was the scary one, Martin, where they brought me in and said, "Hey, let's talk about AI," and then I scared people. And my logic is this, and this logic I think is very hard to argue with.
The hyperscalers, the mega tech companies, are slated to spend 725 billion dollars on AI infrastructure, data centers and the like in 2026. And they spent hundreds of billions last year, and they have some scheduled for next year, so you're talking about a trillion-dollar ballpark. You spend a trillion dollars, and so one of the comps I make is like, look, a really modern NFL stadium might cost you two billion. So you're looking at setting up 10 NFL stadiums per state in like a two to three year period. I mean, that's like the scale of the investment we're talking about. And so how do you get that money back if you are throwing down that way? And when people think about it, it's like, okay, sure, as consumers we might pay the AI company something. Sure, as companies we might pay the AI company something. But a lot of that money has to come out of cost savings from cutting workers. Like, there's no other way to make the math work. And we can talk about whether this math makes sense at all in terms of like maybe the valuations are just out of whack, or the capital expense is excessive relative to, you know, the commercial opportunity in a given time frame. But logically, if you're going to spend hundreds of billions of dollars, you're going to have to kick a lot of workers to the curb to make the math work, is what I said to the folks in Michigan. And they thought about it, and they were like, "Yeah, like how else do you make those numbers work?"
Right, no, I think that's exactly right, and that's what I said in the book, is I think that, you know, the big plump target for getting a return on this investment has got to be cutting labor costs, right? I mean, you can think of it this way. Imagine you're at the Coca-Cola company, right? And you start spending all this money on AI from Anthropic or OpenAI. There really only two ways to recoup that investment, right? One is to cut costs, and it could be other costs, not just labor, if you can use the AI to do things more efficiently, but certainly labor is the biggest target there, right? I mean, if you're going to find, as you say, a trillion dollars across the economy to get this money back, labor is going to be the primary target.
The other way you could have a return on the investment, of course, is to use all this AI so you can sell more Coca-Cola, right? You could boost your revenues. I mean, and it's got to be some combination of those two things, but I really think that the labor costs are, you know, the low-hanging fruit here. That's where companies are going to have to ultimately go if they're going to get a return on this investment. And, you know, just in the last few days, there are all these reports of companies that have been spending huge amounts on tokens, basically, you know, AI services from Anthropic or OpenAI, and they're not seeing that return, right? There was just a report from Axios, I think, that said some company out there actually spent half a billion dollars on services from Anthropic and, you know, really didn't get anything back. So, I think we're getting to the point where they're going to have to start evaluating, you know, how can we get a return on this investment? And that's increasingly going to put pressure on them to look at labor-saving approaches.
