685. How to Survive the A.I. Shock

The AI shock and Gina Raimondo

0:00All right, let's start with a statistic and then a few questions. Here is the statistic. Roughly 60% of the jobs that people do today did not exist in 1940. Now for the questions. Five years from now, do you think you'll have the same job you have today? Or do you think that maybe your job will be replaced by AI? And if so, what's your plan? These are increasingly common

0:34and urgent questions. So I wanted to hear from someone who's looking for answers. What I'm fundamentally trying to figure out working with companies and governors is how do we implement AI without triggering giant job losses? Gina Raimondo was Commerce Secretary in the Biden administration, one of her signature projects was the CHIPS Act, which we'll hear about later. Before running the Commerce Department, she was a two-term governor of Rhode Island. And now she is leading an unusual startup called

1:07Raise Us, which is what? Raise Us is a non-profit, non-political organization dedicated to making sure that the U.S. can both lead in the technology of AI, but also make sure that workers are brought along to an AI economy. For anyone who's been around for a while, this may sound like a familiar song. Think back to what economists now call the China shock, the offshoring of millions of American manufacturing jobs,

1:38a loss that can still be felt in our economy and our politics. And now with the AI boom? I'm saying to these companies, come on, let's get ahead of it and be part of the solution and have an intentional transition. So what would this transition look like? Will there be job guarantees or wage guarantees? How about a universal basic income? We're opposed to universal basic income. And we are opposed to anything that looks and smells like

2:11a path towards socialism. I don't think job guarantees or wage guarantees are the way to go. I think that's too heavy handed. And the market should set those things. I will say, you didn't probably have me on to talk about this, but you have me on, so I'm going to talk about it. It's your microphone. I can't stop you. I am strongly of the view that we need to do a better job tempering the way capitalism is playing out in America if we want to preserve capitalism, which I think we want to do.

2:47If that makes Raimondo sound like someone who might like to be president someday, well, we can talk about that later, too. Today on Freakonomics Radio, how to survive the AI shock with Gina Raimondo.

3:12This is Freakonomics Radio, the podcast that explores the hidden side of everything, with your host, Stephen Dubner.

Economic insecurity and political instability

3:28Gina Raimondo likes to refer to herself as a recovering politician, which seems to be a popular diagnosis these days. She is a solidly centrist Democrat known for playing well with Republicans. Her co-chair at the new nonprofit Raise Us is another recovering politician, the Republican Eric Holcomb, former governor of Indiana. Here is a bit of background on Raimondo. She studied economics as an undergrad, Harvard, then got a Ph.D. in sociology, Oxford, and a law degree at Yale.

4:01Before getting into politics, she co-founded the only venture capital firm in Rhode Island. Raimondo has come to see economic instability as perhaps the biggest driver of political instability. I think everything that we hate about our politics now, more divisive, more partisan, more violent, less rational, people don't actually want to solve problems, they want to fight about them, a lot of that stems from, I believe, this worsening inequality or economic insecurity. When people

4:39are insecure and feel powerless to change things, it's only natural that you would get angry and pissed off and it bubbles over. I see that corporate profits have been rising strongly, and this has been going on for some time. This predates the AI boom. But the share of national income going to workers is at, or maybe near, but I think at, a record low. I'm curious what you see as the causes of this big gap between wages and corporate profits, and I especially want to know what you see as the consequences. Like,

5:12are we already living in the consequences, and do they continue to worsen? I think they will continue to worsen unless we do something about it. People and companies don't change out of the goodness of their heart. No one does. We all need to have the right guardrails and incentives. You can ask economists who will know more than I do about why the vast majority of the gains have gone to capital and not labor in the past 20, 30 years. I think it's probably no one answer. I think it's a combination of declining power of labor unions. Workers have less bargaining

5:49power. Second thing I think is, frankly, free money in the form of very, very low interest rates and essentially no inflation has just massively fueled private equity, hedge funds, all forms of capital investment and capital formation. The 2019 tax cuts to wealthy people and businesses, companies just are more concentrated now. And that also hurts labor and helps capital. So I think what

6:23you've seen is just lots of pressure to keep wages down. The gig economy has done nothing to improve people's economic security.

Predicting AI job losses

6:34So I look for data all over the place. And I have to say, when it comes to net job loss from AI, it's hard for me to be persuaded that anybody really knows what they're talking about. Because, you know, for every job that may be, quote, eliminated by AI, there are theoretically other jobs that could be, quote, created by AI because there are new needs and humans just keep wanting new and different stuff. That's the part that seems to get left out to me. Right now, the only data that I see that shows AI could be, quote, unquote, hurting folks in the

7:08workforce is with younger people. The data is now showing consistently that entry-level workers, college-educated folks in their 20s, are, in fact, having a rough go of it to find jobs. I believe that AI will create new jobs. There are a lot of people who are saying, there's going to be no work for humans to do because in some number of years, the AI is going to be so good, I can't imagine what human beings are going to do.

7:41And that's not you saying that at all? That isn't me. Like, I gave a speech, and Reed Hastings, who's the chair of Anthropic... The non-executive chair. Non-executive chair, yeah. And I know him and I like him. He's a great entrepreneur. He ran Netflix. But he was teasing me after, and he's like, Gina, this reminds me of Jaws, like the boat's not big enough. There's not going to be work to do. Like, this AI is going to be better than everything. Maybe he's right. I don't think so. I'm with the Dubner School of Economics, which is, in some number of years, it's what you just said, right? There'll be new products, new companies,

8:17new services that AI enable. The problem is that the layoffs will come before the new jobs, and the folks who get laid off may or may not be the ones who get the new jobs. The support that we give to workers when they lose job is old-fashioned and not very responsive. And so we're going to work with governors, with employers, to create this infrastructure so that America can do both.

8:49The whole damn point is, to save America and our way of life, our economy, and our politics, our democracy, we got to be there for all Americans, even the quote-unquote losers.

Fixing the safety net and retraining

9:01I do wonder, though, why you're persuaded that government, even state governments, should be the primary engine there. And maybe you'll tell me they're not meant to be the primary engine, that this really is a collaboration between government and firms and so on. But we spoke with a few people in anticipation of this conversation with you. One of them was Michael Strain from the American Enterprise Institute. He said, you know, we've already got basic policy architecture for worker retraining and reskilling. And then there's social insurance for people who

9:32are involuntarily unemployed and a social safety net for people with low incomes. So he said, we might have to turn some of those dials and adjust the parameters, but the basic system seems to be already in place. Why do you not agree with that? I just think that's completely wrong. If that were true, then why do we have such sky-high unemployment for decades in Midwestern states that got crushed from loss of manufacturing, which led to an opioid crisis, which led to crazy, horrible politics? And there's other examples as

10:05well, other technological innovations. We actually have a terrible track record, or even the great financial crisis, frankly. If you look at Germany and other places, they did a much better job getting people back to work. The only social insurance we have is unemployment insurance. Unemployment insurance is not that effective. It's effective for a certain group of people, okay? I'll say that. It puts money in the pocket of lower wage people for a period of time so they can pay

10:38their bills. It doesn't work for people who make more than $70,000, $80,000 a year. It's incredibly rigid. If you start a business when you're on unemployment insurance, you get kicked off of unemployment insurance. If you go back to school when you're on UI, you get kicked off of UI. It needs to be fixed in so many different ways. We don't have any kind of wage insurance. There's almost no incentives for companies. Like, what happens now? A company fires people today, their stock price

11:11goes up tomorrow. There's a tax credit to help companies to do job training. It's a, think about $5,000 a year. It's rarely used because it's difficult. Last thing I'll say is, you, okay, I'm 50, I'll pick on me, not you. I'm a 55-year-old woman. You can pick on me. I don't care. No, I'm a 55-year-old woman. The prime of your life? Prime of my life. That's not what my children say. Let's say I was in the same sales job for the past

11:4430 years. What would you be selling? I'm just curious. I don't know. Software. Pretzels on, oh, software. Sorry. Yeah. Or I had a retail job. Okay, I'm now out of work. What do I do? Where do I get retrained and retooled? Probably I'll go to my community college. Chances are my community college will do a terrible job with me. Says the former governor who made community college free in her state, if I recall correctly. Correct. But making it free wasn't nearly enough. I had to get the employers to step up and create

12:15programs. We taught the people who went to community college what the companies needed, and there was a direct pipeline from school to work. My point is, we need to, in this country, tear down the wall between school and work. Are you talking about apprenticeships or something? Yes. All of the above. And in this regard, I think AI can be part of the solution. You can imagine, back to my 55-year-old laid-off worker, you could deliver career coaching or tailored

12:48education, tailored training, on-the-job training with AI in a way that is cheaper and much more effective. So for example, in Arkansas, one of our first states, we have a pilot with Sarah Huckabee Sanders in Arkansas. It's a career navigation platform. So back to unemployment insurance and the whole wraparound of your question, right now you lose your job in Arkansas, probably go collect UI, and, you know, good luck. What we're going to do, and this is the governor's leadership, she wants it to be that when you go collect UI, you're immediately given a skills

13:24assessment, which is AI delivered. Like, what are you good at, basically? So if this was you, you could say, I'm pretty good at understanding the software that I sell, and yada yada, communicating that to potential clients. Exactly. I'm good at communicating. In my actual case, it would be a challenge. I'm not sure what I'm good at. But in this hypothetical, I'm sure this woman would have actual skills. These would be assessed. Then I would be told, these are the open jobs in your state. I can click on one. That looks pretty interesting to me. Good location, good salary.

13:58And they say, okay, for you to be qualified for that job, you would need this kind of training or coaching. You sign up for that for free, and then you have a better shot at applying for that job. There's really nothing like that. All these systems, they don't work together. They're not that effective. They haven't been updated forever. And here's the brutal reality. Our whole system routes everyone to college, basically. And we spend an enormous amount of money on subsidizing college. So when you get out of high school, you have precious few options. You could go into the

14:33military, which is good and honorable, but only a certain number of people are suited for that or want to do that. You can go to a labor union and get a trades job, and that's great if you can, and you can figure it out, and you know that exists. Or you're pretty much routed to college. So one of our first projects, we're going to be building an advanced manufacturing apprenticeship initiative in the Midwest. What state? Starting in Wisconsin, and see if we can scale it. I have a ton of experience on this from when

15:07I was governor, mostly from learning, actually, what not to do. It's hard to do in our American system. So some of the things that we're doing are kind of, quote unquote, no regrets moves. You know what I mean? Like, let's say this AI rolls out more slowly than we think, and so the new jobs are created at about the same time and rate that the old jobs go away, and we never really have super high unemployment. Well, that's, gosh, that would be wonderful. But it would still be great

15:42if we had an apprenticeship initiative, a career navigation platform. In Maryland, we're going to be funding and working with the governor on a year of service. So if you're in your early 20s and you need a job, you can't find one, do your year of service with the state of Maryland. We'll get you a job with a local employer, probably in healthcare, because that's where their need is. You get paid for a year, you get skills for a year. Like, that's a good thing generally.

Corporate backers and skepticism

16:12I see the corporate backers of Raise Us. It's a long list, and they're really good firms. ADP, Autodesk, Blackstone, Cisco, MasterCard. But then there are the anchor funders, which are Amazon, Anthropic, Microsoft, and the OpenAI Foundation, all of whom are either making or heavily involved in using AI. And then I read the total amount of funding. It looks like you've got a commitment for about a half a billion dollars on a raise of ultimately a billion. So to me, please forgive my

16:44skepticism or maybe even cynicism at that point. But to me, that's, A, that's a drop in the bucket for these firms. And it buys you good headlines. And it feels like a little bit of reputational laundering. Like, yes, we're the firms who are making this product, in this case, software, essentially, that has already made us firms worth trillions now. And so, yeah, we'll take a nickel out of our pocket and give it to this amazing former commerce secretary who has a plan to help us transition into this new economy. But it sounds to me like the highest form of

17:19window dressing, potentially. So persuade me that it's not. Listen, your skepticism is fair. And obviously, I thought about it in putting this whole thing together. And the proof will be in the pudding. We've got to produce some results. I would say this, we can't solve these problems without them. We need the data they have. We need to know what their products are going to look like in 6, 9, 12 months. Yeah, we need the money. They gave us some money to solve these problems. Is a billion dollars a drop in the bucket? I don't know. You know,

17:55it's not nothing. But the only way this stuff works at scale is if maybe we start with four governors and come up with some good new policies that work and then 10 governors and then 20 governors, let's say things get pretty bad a few years from now and Congress does take action, which is what will be required, they'll hopefully look to what we've done and model federal legislation after what we've done. I hope we have some new ideas and some new experiments because you can't innovate at the

18:27federal government level. You have to innovate on the ground, a company at a time, a state at a time, and then prove that it works. But not only I hope Congress will look to us for good ideas, but we'll have a built-in coalition of political supporters, governors of both parties, CEOs of some important American companies. So the point isn't so much that Raise Us solves all that ails America with respect to the AI disruption. It's that it's a platform for innovation, discussion,

18:58and an opportunity for these CEOs to step up and lead. Now they may not, okay? Two years from now, you know, maybe they haven't leaned in. Maybe we haven't come up with any new ideas. Maybe they're all just implementing AI in a way that is leading to mass unemployment. But I got to try. So right now, fine, it's good for you to be skeptical. It's good for me to prove you wrong and show we can make this happen. But I got to give them some credit. I asked them to lead, and they've stepped up to lead.

19:31And how do you pitch that? Do you say to the firms, look, you need to be part of the solution because you are a big part of the problem? Yeah. I didn't quite say it like that. That's why I would not be a good politician, by the way. And we said, we're coming to you. You need to be part of the solution because we can't figure it out without you. And it's in your interest. How is it in their interest? Because let's say Dario has, you know, been very open. Dario Amadei is the CEO of Anthropic.

20:01Yes. Thank you. But I'm glad you're on first name terms with them. That makes me feel better. Well, it's kind of a famous first name. I used to drop the clothes of the dry cleaner, and I just gave my name Gina. My husband was like, do you think you're like Cher or Madonna? Huh? Anyway, Dario has said, listen, there's going to be potentially, potentially very high unemployment. So let's say that's true. Okay. Some people say there's 20, 30 million jobs that are AI vulnerable. Let's say a lot of those people lose their job. That will destroy this country.

20:35If you think our politics are dysfunctional and violent today, that will literally lead to rioting in the streets, massively increased violence. I don't want to live in that America. And by the way, it'll also lead to AI destroying regulation, which I think is bad for the whole country. So it's just not good for any company to let this play out in a way that leads to a recession and unrest. Coming up after the break, how Gina Raimondo's family history influences the

21:17way she thinks about the future. This is Freakonomics Radio. I'm Stephen Dubner. We will be right back.

Family background and the China shock

21:22Like many successful people, Gina Raimondo has a lot of drive. She attributes this to her family background. I'm from Rhode Island. That's a co-op thing about who I am. Youngest of three kids, working class Italian American family. My grandpa was an immigrant from Italy, lived in the house with us. I spent my Sunday

21:54afternoons listening to Italian radio with him, learning how to cook. And we were just like bred to work hard, which I think is the immigrant experience. Like all the sacrifices Pop made, you will honor those sacrifices by working hard and being successful. So my dad, he worked in Boulevard Watch factory, 28 years. We had a nice, happy life growing up. Never had too much, but always had enough kind of a thing. And then he lost a job. The factory closed.

22:24How old was he? He was 56. I was a little kid. My brother and sister were just about to go to college. One was in college. Did that fundamentally change your family's fortunes? Oh my gosh, hugely and completely. My brother and sister are, as I just said, a little older than I am. So they, when that happened, I was a kid and they were off to college and out of the house. I grew up in a different house than they did because certainly there was way more economic insecurity,

22:55way more. And just my dad was just always grumpy. Like he felt bad. It was pride, dignity. It was brutal. And do your kids work as hard as your father did? No, no, no. Does it concern you? Or do you say, well, you don't have to work that hard anymore? It concerns me. Yeah. Maybe I'm just a hard ass. So you're calling your kids lazy on my air? No, I didn't go that far, but they are a little less tough and they've had an easier life.

23:30But I could imagine some people saying, well, that's wonderful. That shows that civilization and human progress are really trucking along and we should be grateful for that. But what are you missing when you don't have that work drive? I think people are genuinely happier when they have a little less and have to work hard to earn and achieve. One thing that makes an American an American is we are strivers. We like to get ahead. We like to take risk. We're entrepreneurial. America's unique. If

24:05you walk into a restaurant, a bar, someplace, you meet somebody for the first time, you say, hi, my name's Gina. What's your name? What do you do? And everybody else in the world complains about that. Yeah. They think we're weirdos. We are our jobs, our careers. We judge ourselves. Are we worthy based on are we successful? I bet you, your friends, your family, you feel good about yourself if you have a good job and you're getting ahead. And that's why I think America is like the land of the career and the work and the worker. And a job is way more than a paycheck. Yeah, you need

24:38the paycheck, but it's dignity, pride, purpose, identity. So do you kind of see your dad's face or think of him when you think about what's maybe going to happen if there is massive job replacement because of AI? Yes, absolutely. It's why I took the job as secretary. It's motivated so much of my career, frankly. I mean, I leaned into manufacturing so significantly as secretary because I saw how

25:08good it was for us when we had it and how bad it was when we didn't. He always used to say, I needed a bridge to another chapter of work. He wasn't ready to end, right? Like he wasn't 65. And by the way, they needed the money. They needed purpose and a job. Do you know where his job went? What country? China. So he was part of the early, what we now call the China shock. Is that right? Late 80s. Yeah. So with your adult reckoning now and having been in government for a long time,

25:40describe the China shock. I mean, the fact that we call it the China shock is kind of funny because it wasn't supposed to be a shock. This was, if I have it right, the Clinton administration thinking based on previous administrations interactions with China, that it was a really good idea for China to be admitted into the WTO, the World Trade Organization, and become a trading partner and a labor partner. So assess that decision and then the downstream consequences, knowing what you know now. You know, this is a classic case of it looked good on paper and the economic theory and it

26:15didn't play out. The theory was if we draw closer to China, we'll draw them to a market economy and away from communism. And also the classic theory of trade, comparative advantages. China has comparative advantage in manufacturing, lots of cheap labor, let them do what they're good at. The products will be cheaper. That'll help Americans. We'll buy stuff that's cheaper. Let us do what we're good at, which is services. And by the way, a lot of that happened. A lot of that did happen. Yeah.

26:46China did lift millions and millions of people out of poverty through manufacturing and America became more productive, higher wage, bigger economy, and a global juggernaut in services, software, and all kinds of services. So all of that actually worked. The problem was the unintended consequence of the millions of Americans who were just like precipitously put

27:17out of work from their manufacturing jobs. And here is the really important point that I worry a lot about with AI. There were winners, right? All the people who got a better, higher paying job in the service economy and quote unquote losers. But the losers were concentrated in particular geographies and non-college educated men. And when you hit harder, quote unquote, the losers, man, it can reverberate. You destroy communities. I mean, we're still paying for it in

27:47our messed up politics today.

Learning from past trade mistakes

27:49So you recently gave a TED Talk discussing the China shock. And in a conversation you and I had not long ago, you said, we screwed that up. Good idea. Terrible implementation. Number one, we started too late. Number two, it was massively bureaucratic. Number three, it's completely divorced from employers. What makes you think this time can be different? Let's assume that there is a massive job replacement due to AI or massive job loss, or at least shifting due to AI. And at this point,

28:21it is still an assumption, really. But let's assume that that is true. What makes you think that this time can be different if you're following essentially the same playbook, which is to say, we will be alert, we will retrain, and we'll figure out how to redistribute these gains a little bit better? What makes you think it can work this time? First of all, it is a big gamble, right? I'm not promising success, but I'm going to will it into existence, truly, because it has to be. But secondly, gosh, let's learn from the past.

28:55We're still a young country and celebrating 250 years. We've seen now a few of these economic transitions. Let's learn from it and make it better this time. So for example, what you just talked about, it was completely reactionary. We didn't start to think about helping communities that were hurt from loss of manufacturing until they were already reeling, which is part of why I'm

29:26starting Raise Us Now. Let's get prepared. Second thing is, it really was like a federal government one-size-fits-all approach, which it just often doesn't work. So what was there back then? Was there federal money going to states for retraining programs? Yes. The biggest initiative was called TAA, Trade Adjustment Assistance. And it was a giant pot of federal money given out to states that were disproportionately affected by the loss of

29:59manufacturing jobs. By the way, some of it did work, okay? For example, there was wage insurance. You lose a job, you take another job at a lower salary, the government tops up your pay to get you where you were. For a period of time, not forever, I assume? Yes. Not forever, for a period of time. It works best for people who are further into their career. And the whole point of it is, don't allow yourself to be unemployed for a year or more. Get back into the workforce, even if the job isn't at the same rate of pay. Lots of data to show that was actually pretty

30:33effective in certain communities. Where we really made a mistake with TAA is that the person who lost their job had to certify to their state that they lost their job because of the trade opening up with China. Ah, how do you do that? It's so, so hard. It's so hard. So the average person was waiting for benefits for almost a year. And loads of people gave up. I know nothing about how legislation like this is made. You do, plainly. How does something like

31:08that work its way in? That sounds to me, as you describe it, kind of like a poison pill that someone put in there because they didn't want all that federal money to go out. They want to make it harder for displaced workers to get their money. That's an interesting comment, and I hadn't ever thought of it that way. Honestly, those things are often put in there, maybe by the poison pill folks, but also maybe by like, quote unquote, good government people. Don't waste the money. You don't want people getting this money for whom it wasn't intended. And maybe to prevent fraud if you have to certify

31:42that your actual job is now being done elsewhere. Correct. There's lots of proposals out there now saying companies need to publish all the jobs they're getting rid of because of AI. If somebody's going to lose their job because of AI, they need to tell the person they're losing their job because of AI. Personally, I oppose all of those because it's just impossible. It just weighs programs down with bureaucracy. But here's the real point. Companies didn't do anything. The whole thing was

32:16government top-down led to help folks who lost their job. What I'm trying to do here is to go to companies, including these tech companies that are making this AI, and saying, this isn't going to work out well for you if you don't get in the boat now and help us try to find some solutions. It might feel good today to have a big riff of 10,000 people, but that won't end well for this

32:47country or you. It remains to be seen whether we'll have any impact on the way companies implement AI. But as you say, I lived it, right? We put our blinders on. We sent all these jobs to China. Super excited about the increased profit and efficiency. And then we ruined the day we did it. The Chips Act, the thing that I spent so many hours getting passed in the U.S. Congress. Same thing, Stephen. Everyone said, oh, it's a lot cheaper to design chips in America and make them

33:18in Taiwan. Let's put our blinders on and just make everything in Taiwan. And then we wake up one day, we don't make any leading-edge chips in America. And we say, oh, wow, every leading-edge chip the U.S. military uses is made in Taiwan. Holy cow. So we spent hundreds of billions of dollars of taxpayer money to reshore the industry. It happens every time. How stupid can we be? Just put your blinders on and focus on the efficiency and the productivity. So I'm saying to these companies,

33:50come on, let's get ahead of it and be part of the solution and have an intentional transition.

The Chips Act and semiconductor deals

33:57Let's talk a little bit more about the Chips Act, which you really helped create and shepherd and implement when you were Commerce Secretary under President Biden. I gather that the Chips Act is a fundamental precursor to raise us and the conversation we're having today. Is that right? Yes. It was bipartisan, for one. Yes. Talk to me a little bit about what you feel it's accomplished and especially what you feel it's accomplished even during the second Trump administration. My sense is that because it benefits a lot of manufacturers in Republican states, that there are very few, if any, Republicans

34:33that I've seen in Congress who are interested in taking Chips down, correct? They have continued to implement Chips essentially on the same course as we had it, for which I'm grateful. And I think it's a national security imperative. Also, you made a really important point. It was bipartisan. We work really hard to have a lot of Republican votes. And I did that intentionally because even though I'm gone and the president's gone, all the Republicans

35:03that are still in Congress, it's their baby too. So they're still there pushing on it. But was the underlying Republican sentiment in favor because there would be jobs and factories and firms in their states or was it more open than that? It was much, much more about national security. Some of the biggest, most vocal, like Roger Wicker, who's from Mississippi, was hugely supportive of it. He's the chair of the Armed Services Committee. There's not a chip factory in Mississippi. I don't think there will be.

35:34But he's like, listen, the United States is, this is scary if we have to rely upon Taiwan for all our chips. You were essentially the chief or at least a chief dealmaker in this. You were going to negotiate with these semiconductor firms like Intel and TSMC and Micron and so on, yeah? I did, yeah. Can you just describe that? It sounds fascinating and important. And I'm just curious what those conversations were like. Was it an easy sell or no? It was not. I'm laughing because the bill passed. Everyone in the industry was euphoric because they

36:07knew they were going to get money. But when I sat down and said, OK, what do we have to accomplish with this amount of money? We wanted to make a certain amount of chips in America, a certain amount of memory chips, yada, yada. We realized, oh, it's actually not that much money. So every company came in with their ask. TSMC made an ask. We need this many billions, Micron, all of them. And I told all of them, like, guys, you'll be lucky to get half of that. The implication being that you need to then pony up the rest to build the factory here and employ

36:40our people. Correct. Exactly. So the first meeting was always a little rough. And what kind of leverage did you have if they just said, ooh, that sounds like a lot of trouble and a lot of money. No, thank you. I had that problem, especially with TSMC. TSMC is an extraordinary company. It's an unbelievably well-run, efficient company, the best in the world at what they do. And they would say, like, Secretary Gina, I mean, we're happy in Taiwan. We've got dozens of factories. We know what we're doing.

37:11And by the way, the PRC would probably rather us stay there, correct? Oh, my God, of course. And actually, maybe even the Taiwanese government because... Leverage against China? Yeah. If you think about it, it's leverage. It's the economy. They call it the silicon shield. In Taipei, because China needs it just as much as the rest of the world. So it's pretty good bet China will never blow that up. So what did you do? Did you just tell them your dad's bull of a story and they somehow were won over? Never underestimate. Just sheer determination, willpower, sticking with the problem. But the

37:47real unlock was that I was able to get the biggest customers of TSMC, which were American companies, Apple, NVIDIA. I think between Apple and NVIDIA, it's like more than 30% of TSMC's revenue. And they did it. You know, Jensen, the CEO of NVIDIA. Jensen Wang. Tim Cook, the CEO of Apple. Lisa Su, the CEO of AMD. All TSMC's biggest customers leaned on TSMC,

38:20frankly, to say, we're too vulnerable. If all of our chips come from the same factory in Taiwan, we need you to manufacture in America.

38:34Tim Cook, as you may know, recently completed his term as CEO of Apple. His successor, John Ternus, took over on September 1st. I'm Stephen Dubner. This is Freakonomics Radio. And we finish up our conversation with Gina Raimondo right after this break.

Measuring success at Raise Us

39:01Raise Us, the new operation that Gina Raimondo is running, will have its initial rollout in four states, Utah, Maryland, Arkansas, and Connecticut. For those of you keeping score at home, which seems to be everyone these days, two of those are red states and two are blue. One goal of Raise Us is a reform of job training. Raimondo and I spoke about job training back in 2018, the first time she was on Freakonomics Radio. It's episode 313. If you want to listen, it's called How to Be a Modern Democrat and Win. At the time, she was governor of Rhode Island.

39:35I would talk to guys in the building trades who had been out of work for years, losing their houses, losing their marriages. So I said, we have to completely turn upside down the way we do workforce development. We had a model, which I often refer to as train and pray. Train people and pray they get a job. And it turns out that's not effective. So now I asked Raimondo, with train and pray off the table, what will success look like for Raise Us? Success is, a couple of years from now, we are judged by one thing. Do people get jobs,

40:09keep jobs, and transition more quickly from one job to another if they get laid off? Two years from now, we'll have an apprenticeship program. We'll have some changes in unemployment insurance. We'll have this year of service. Like, is it working? We spoke with Raj Chetty, the Harvard economist, who I believe is on your advisory board as well. And he raised a point that in order to measure this success well, you have to figure out your time horizon, for one. What kind of data infrastructure would you need, whether it's from firms or elsewhere? You need access to earning records and so on?

40:42So today, for example, we had a meeting working with some employers asking that question. In other words, what would it take for you to redeploy these people instead of lay them off? And what are the jobs that they could go into? And how much of an incentive do you need? These are the discussions that we're going through right now. We have a data lab of folks who are going to work with us to collect the data. Raj Chetty is like the best in the world at this. I mean,

41:13he's a top flight academic. I'm a little bit more simple minded. Some of it can be dollars in, jobs out. Like when I was governor, we spent $5,000 of the state's money per person to get them a job through a training program. So if you take someone who's unemployed and get them a job reliably, that's a win. You talked about the AI firms in particular having an interest for a variety of reasons. I wonder how you think about whether employers have a moral obligation to protect workers going

41:52forward. My sense is that we believe the government does have a moral obligation to protect its citizens, including employees and so on. But when you think about the leverage that employers now have in the economy, we mentioned earlier that shareholder profits are at an all time high. How do you think about that question, the moral obligation of a firm to protect workers, at least to think about them as, you know, maybe workers first, but also citizens second? Something's happened in the past 50 years in this country where shareholder value has become

42:27singularly important and focused on. The whole industry of asset management certainly didn't exist the way it does now 50 years ago. I do think employers have an obligation to their employees to be a good employer as well as to their customers and their shareholders. You know, often I hear CEOs say, well, if I take my company hat off and put my America hat on, I would do X, Y, Z. And I understand

43:00that, you know, they're running a company, they're not running the country for sure. And they fundamentally have to have an efficient, profitable company. I think that we'd be a much stronger country if CEOs led more on these critical issues. A lot of them want to, but they're afraid to. Politics today is the worst. And I have some sympathy for that.

Electoral systems and political incentives

43:24How do you feel about quarterly earnings reports? I read that the SEC is proposing stepping way back on the requirement for quarterlies. Yeah, they are, which I think is a great idea. We've become too obsessed with it. And no CEO I know likes these quarterly earnings. They don't like having to be slavishly focused on their share price. It makes it hard for them to make long-term decisions and long-term investments. Listen, we're in such a precarious place as a country. The political violence and dysfunction that

43:59Congress can't solve a damn problem. They don't function. We can't put out a budget. We're a global embarrassment on so many topics. I think if you're a CEO, you are an American, you have children, you care about the future. By the way, why are our companies so successful? Because we have rule of law, democracy, a court system, free-flowing capital markets. None of these companies would be successful

44:30as they are without those things. And how do you think all those things, those foundational attributes of our dynamic economy, how do you think they're doing under the second Trump administration? Strained. Like, very strained. We have to get back into the business of realizing that having good infrastructure, a good rule-based democracy, sensible politics is good for everyone. What's sensible politics look like? And let's even take a pretty big step back because

45:02I'm curious to know what you think of our electoral system generally, changes that you'd like to see. And, you know, I think the problems that I've heard you talk about before, I think we'd agree on that the primary system leads to extremism from both parties, and there are litmus tests that are kind of unrealistic and gerrymandering, I know we've talked about as a big problem. But I'm curious whether you think the electoral system is both stable and working.

45:30Listen, nothing is perfect. But even in the time I've been in politics, like not very long, I got in in 2010, wasn't perfect then. The primary was the same thing that you just said. There was too much money in politics. Your first race was for treasurer of Rhode Island. Yes, treasurer of Rhode Island. And you won your first time? Mm-hmm. I did. You've never lost an election, have you? No, I never have. I've had really hard elections every time, but I've been lucky not to lose. But here I ran. This is not that long ago. There was a vibrant local press corps. I did pension

46:07reform. I think I had six reporters crawling all over me every day. That's gone. It's not just gone in Rhode Island. It's gone everywhere. There's not really much local press anymore. And therefore, accountability in government is a different issue. Correct. Accountability, a news source that's fact-based about your community that the average person can read and trust. When we were doing pension reform, every Rhode Islander had an opinion. They knew what was going on. It was on every radio, every television, every Providence Journal.

46:40That's all pretty much gone now. What would need to happen for you to run for president in 2028? And if you did run, what would need to happen for you to win? Well, I would have to be somewhat convinced that there were a path to victory for someone like me. Someone like me meaning what? You know, practical, have made a lot of tough decisions, a record as a bipartisan. Like, I'm a person of compromise and problem solving. The algorithm doesn't like that at all.

47:12Yeah. The algorithm hates that. And you know what the irony is, Stephen? To raise the money you need to raise, you have to do a lot of it online, obviously, in small dollar. And, you know, that was thought to be better to raise loads of money in tiny increments. And it is. That was the Obama model. It worked really well. Exactly. The Obama model. And by the way, that is exciting and preferable to some high roller writing a big check. But here's the dirty secret. The way to raise the money online is to be more and more outrageous, more and more against compromise. And it unfortunately

47:50hardens people's positions in a way that I think makes it good for today's elections in a primary and very hard to govern. Let's go back to raise us in the little bit of time we've got left. One critique I've heard is that it's not really a jobs plan, but more like a reputational buyout for building future data centers. The financial journalist John Markman wrote, the $500 million that you've raised is not an apology for jobs already lost. It is a down payment on the political permission to keep

48:22building. What do you think of that critique? I think it's silly. If we accomplish nothing, then OK, maybe that's fair. I don't think any of these data center companies think that giving us 20 million bucks is going to get rid of all the challenges they're having in local communities with data centers. So I don't think that that's why they're doing it, actually. I think they're doing it because this is going to be a problem for the country if you don't get it right, and they want to be part of a solution, and we want them because we can't figure it out without them. But I do want to say this. It would be a bad thing

48:57for the country if we over-regulated AI and lost ground in the global AI competition. It would be a bad thing if Congress woke up tomorrow and said, no more data centers in America. Like, you can't put the toothpaste back in the bottle. AI is the next frontier of technology. Either we're going to lead the world or somebody else will. I mean, I want my kids to grow up in a world with AI, with the benefits of AI. We just have to cushion the downside in the transition, like we talked about,

49:32from here to an AI economy. But I don't think it's a win to kill data centers. So your plan with Raise Us is compared to some other plans in response to AI. Yours is extremely mild compared to some. There's one policy proposal from Bernie Sanders, which was, I think, to essentially confiscate half the stock value of big AI firms and put it into a public fund. Okay. So first of all, that's not very much money. Wait, you're saying that half the stock value of the top AI firms is not very much money?

50:05I mean, today they're worth a lot. We'll see where that goes. Who's an AI company? We're just going to go after two companies? How are we defining an AI company? I mean, is it just going to be anthropic and open AI? Are we also going to go after Google? It's an impossible thing to implement. But also, why just them? This is another thing. Okay. So open AI would just pick on one company for a second or Google, whatever. XYZ company invents a technology and then they sell it to 50 U.S. firms and those U.S. firms implement it in such a way that they fire half their employees.

50:41Well, it doesn't seem to me that Google or open AI is purely to blame there. My whole point with Raise Us is it's in everyone's interest for the United States to lead the global AI competition. I have one thing I want to put on your radar, which you may or may not find interesting, but

Capitalism, socialism, and historical models

51:02it goes to your question of my wildest dreams. Capitalism and democracy have proven to be wonderful forces to allow for individual liberty, creativity, and prosperity. It's like the Reese's peanut butter cups of- Deliciousness. Economic systems, yes. And it works. Right now, the way capitalism is playing out in this country is causing a lot of people so much economic insecurity and anxiety that they're saying, what's wrong with socialism?

51:37So I'll ask you, what's wrong with socialism? Well, it hasn't worked anywhere. It'll take away our freedom. It'll take away our liberty. It'll take away our creativity. I mean, it's not a place where I want to live. I want to live in a place where anybody, any immigrant, anybody can have a good idea, get well-educated, take some risk, and be successful. So I fall in the camp of make some changes to capitalism to save capitalism. One of

52:09the things that I think has to happen is, like you said, more business leadership. You don't have to take your business hat off to do something that's good for the country, because something that's good for the country is going to be ultimately good for your business. The model that I'm using for Rezos was something that was created in the 1940s. It's called the Committee for Economic Development. It was created as the Second World War was winding down, and businessmen, they were all men, were very

52:41worried that as we transitioned to a peacetime economy from a wartime economy, and millions of GIs were coming home, and there were more GIs than jobs, that we were going to have a high recession. So a group of businessmen, CEOs of Studebaker, Goldman Sachs, etc., got together and said, we're not going to let that happen. It was entirely private sector driven. They created like 10,000 or more local committees and

53:14communities all over America. They leaned on companies to create more jobs, and they helped them figure out how to create more jobs. And their theory was having money in the pockets of the average American helps business. They had just gotten through the Great Depression, and they didn't think the country could handle another Great Depression. So they had what I'll call a planned, intentional transition to a peacetime economy, and it kind of worked. Yeah, it kind of worked.

53:44So this is what I think we need now. Interesting. This was literally partnership, or at least an association of firms trying to figure out how to grow the pie, yeah? Correct, correct, correct. Some of the examples are so interesting. They would purchase an abandoned manufacturing facility and use it as an incubator for startup manufacturing companies. They would look at the business plan of a local company and help it figure out how to find some growth so they would add employees. This is why I think we need these AI companies at

54:18the table. These people are smart enough to create these models. They have to help us figure out how to use these models for growth. And it will feel better if it's collaborative versus forced, I assume is the idea. Completely. It's always so fun to talk to you. If we didn't love to elect politicians in this country who shout and tell crazy stories and pound their chests all the time, I'd say you'd have a great chance at winning something big like president. So who knows? I'll write you a letter of recommendation if you ever

54:49need it. Oh, could you please? I'm going to need a few hundred million copies.

54:54That, again, was Gina Raimondo. My thanks to her. Thanks also to some friends of the show who helped us prepare for this conversation, including Michael Strain, Raj Chetty, Claudia Golden, Kyla Scanlon, and Scott Lincecum. And you remember those questions I asked at the top of this episode? Well, I would love to hear what you think about the future of AI and your work. Our email is Freakonomics Radio at Freakonomics.com. Also, it has come to my attention that many of you who

55:26listen to Freakonomics Radio have never read Freakonomics, which is fine, even awesome. But in case you are curious, we recently published a 20th anniversary edition of Freakonomics with a new forward, and you can get it wherever you get your books. Meanwhile, coming up next time on

Drafting NFL quarterbacks

55:43Freakonomics Radio, at the start of every NFL season, I find myself asking the same question. How can it be in this age of advanced analytics that it's still so hard to predict which college quarterbacks will thrive in the NFL and which ones will fail? Listen, quarterback evaluation is probably one of the most studied and still most uncertain problems in all of professional sport. In trying to answer this question, I spoke with a Hall of Fame quarterback, a professional scout, a pioneering sports psychologist, and a Silicon Valley entrepreneur who wants to fundamentally

56:16change how athletic potential is evaluated. I also took a special kind of IQ test. Each time you touch a box, it will change to look like another one of the possible shapes below. Oh, man. No way. Is that what it's supposed to do? To begin. Good luck. How to draft an NFL quarterback that's coming up next time on the show. Until then, take care of yourself. And if you can, someone else, too. Freakonomics Radio is produced by Renbud Radio. You can find our entire archive on any podcast app.

56:49It's also at Freakonomics.com, where we publish transcripts and show notes. This episode was produced by Augusta Chapman, with help from Mandy Gorenstein. It was edited by Ellen Frankman and mixed by Jake Loomis, with help from Jeremy Johnston. The Freakonomics Radio network staff also includes Dalvin Abawaji, Eleanor Osborne, Elsa Hernandez, Gabriel Roth, Elaria Montenacourt, Pete Madden, and Tao Jacobs. Our theme song is Mr. Fortune by the Hitchhikers, and our composer is Luis Guerra. As always, thanks for listening.

57:20Let me say a quick thing, and then... You can say a long thing. You're a governor, damn it. The Freakonomics Radio network. The hidden side of everything.