Hello and welcome to the American Compass podcast. My name is Orin Cass. I'm the chief economist at American Compass. And today we've got a very special episode for you. It's a conversation I moderated with Teamsters President Sean O'Brien and Palantir CEO Alex Karp at a recent event we helped host in DC, a summit on labor and AI. I really enjoyed talking with them. Hard to think of two better people to bring together and get interacting on questions about technology, labor, and how the political system needs to respond. So, I think you'll really enjoy it. If you're new to the podcast, please subscribe. We come out weekly talking with the best analysts of the economy, the people shaping the future of the American right and American politics generally. I think you'll find a lot to like if you're interested in this one. As always, before we start, also have to tell you about Commonplace, our magazine. You can get the podcast there. We also publish weekly everything that matters in America. And this week's a great example. We've got a great piece from Peter Harrell on Section 122 tariffs and why they are in fact legal the way the administration is using them. We've got an essay from Dr. Leonard Sax on the politicization of parenting. And a piece from John Shu on the problems of tourism and its interaction with birthright citizenship. So check out those pieces. Please subscribe and enjoy this conversation.
Welcome everyone to my New York Drama School final assignment in which Sean O'Brien and Alex Karp encounter each other in front of a stained glass window in DC and a dialogue proceeds. I've been looking forward to this very much since Nino said we think we could actually do this conversation. There is obviously a tremendous amount of discussion within the labor world, within the technology world about AI and very little discussion between those two worlds. And so really the goal here is to give you guys a chance to talk about it. See what the areas of commonality are. What are the differences that can be resolved? What are the intractable differences we might have to think about? Obviously in many ways you are both quite heterodox and don't necessarily speak for labor overall or technology overall but I think you certainly come from those worlds, are very familiar with them, and can help us think through how those interactions might play out. So neither of you needs any introduction. Alex Karp, CEO of Palantir, Sean O'Brien, general president of the International Brotherhood of the Teamsters. Wanted to start out with one question for both of you to sort of level set, which is, the simplistic version is, is technology good for workers? If you'd like to answer straight up yes or no and defend it, you can. If your answer is it depends, on what does it depend? We'll start with you Alex.
Well, I'm delighted to be here and hugely honored to be on stage with you and thank you for helping to organize this. And in general, all I did was say yes to this but the part vignettes I've seen I think are quite important because there's really, especially in Washington DC, I've never been at a conference where anyone talks about faith of any kind. Maybe there's a reason for that, but it's... and then yeah, the labor. I would say as a way of introduction, for those of you who have not heard my various rants online, I grew up in a pretty far-left, very pro-labor family. Once when I dared criticize a union I got screamed at. How dare you? Your grandfather was in the teachers union and this is why we live well. And we had very few rules in my house but one of them was not crossing the line if someone was on strike is probably the only thing you'd get beaten for in my house. So I'm probably going to get a lot of street credit home for this. And in general, I believe, well, I really, it's always interesting to have beliefs no one believes you have, but I do believe the working person in this country has not been treated with the respect or decency, concern or empathy that he or she deserves. And most of my politics are still like my concern about open borders, concern about evisceration of education and now to AI actually go back to will this be good for the average American worker not just for moral reasons but that's how you stabilize a nation. But then to your question is it, I think you have to break down what kind of worker. I think this technology, it's also in flux but I think, and then we also just as a supposition we're talking about any time until super intelligence which I think will take a while, trained workers colloquially speaking who have some kind of vocational hands-on role I think will do very well for obvious and non-obvious reasons. Anyone who's kind of neurodivergent or viewed as crazy I think will do also very well. Everyone else it's I think going to be much much harder and that you'll see a lot of pressure on the system. And so the primary threat to in my view our broadly defined world order and American world order is people not understanding where the jobs will, where the jobs that I would say are probably underpaid, like I think workers with vocational training are systematically underpaid because some form of payment is value creation and these technologies will make you more valuable and you see this by the way like on the battlefield, the people who write the actual ways of organizing the battlefield, a very complex, very technical job many of them. You know, the best ones are not inherently the most educated. Or in factories where we're working the augmentation of labor value, I think the classic way of the way we look at value in terms of like profit maximization does not capture that a lot of the value will be created essentially by the laborer under the supposition that you have some ability to interact with the physical world. And so yeah, there's going to be a lot of pressure on highly educated non-vocational workers, especially those that don't have some kind of neurodivergent, something that you can't just replicate.
Sean, how do you think about the effect of technological progress both historically and what we're looking at now for workers? When does it work? When doesn't?
Well, I think if you look at our brand, the International Brotherhood of Teamsters have been around since 1903. I'm a proud truck driver. So that last speaker that wants to retire us and send us a check, just send it big and send it on time. But I'm a fourth generation Teamster. And if you look at our brand, two horses, right? So goods and services were provided to this country with horse and buggies. And if you don't think in 1903 when the teamsters started and then through the evolution the combustible engine came in we weren't scared. We were scared as hell. We were scared as hell we were going to lose jobs as a result of that technology back then. I think when we look at technology and efficiencies I think a couple things. There's always the fear of the unknown because we have never had a seat at the table when this technology, whatever it may have been or whatever it may be, we've never had a seat at the table to discuss the effects and impacts. Never had a seat at the table to give our opinions on what would work, what wouldn't work. And I think as a whole when we're representing workers and we're representing the American worker, not just union workers, it's important to realize the consequences of this technology of these efficiencies of the AI. And I think there's an opportunity to embrace technology, to embrace efficiencies, protect jobs, more importantly, get people home safe at the end of their shift. But I think the biggest struggle is collaborating between those two worlds. So you have the AI gurus who use four or five clap words, right? And then you have people like me who were brought up blue-collar, didn't get the opportunity to go to college, had to go right to work, but we're looked down upon because we don't have the education. Or maybe we're looked down upon because they don't think we're sophisticated enough to understand it. And I think the fear of the unknown is what bothers us or troubles us. I think truly if people collaborate, both worlds, people sit down, have a seat at that table instead of working people being on the menu, we can help develop that menu because we're at the table. And I think like anything else and you know I'm taking a different approach and you know we're doing a lot of things differently in the teamsters union not just for the teamsters but for the American worker is we want that seat at the table. We want to embrace technology that's going to make work that much better but at the same time we've got to protect jobs and create jobs as a result of this technology. And Alex and I have had discussion prior to this whole event. And he eloquently explained to me one time the white collar jobs that are going to be affected and wiped out through technology. I think for the first time those white collar workers that are going to be vastly affected that was talked about earlier, they realize they're an American worker. They're not a white collar worker. They're an American worker. So the other fear of the unknown is what is going to happen to these folks when their jobs are eliminated as a result of AI. Where do they go? What is the retraining? What is the guard rails that are going to be put up to make sure that people are retrained and there's opportunity to live a middle class lifestyle when efficiencies and technologies are implemented. So I would say this, we always get portrayed or I always get portrayed as a hindrance to technology. We don't want to be a hindrance. We want to embrace it. We want to help design it. We want to help implement it and we want to create opportunity for people.
So, by the way, the only thing I would say is if people in AI or in technology don't figure out how to make sure what you're saying happens, I don't think you're on the menu. I think we're on the menu.
I don't want to say that the federal government will send us a peanut tomorrow.
Well, so I actually think what's happened to white collar workers is obviously a fascinating, incredibly important discussion. I actually want to stay focused a little bit on what's going to happen to blue collar workers because I think from a sort of skill bias technological change perspective you could certainly make the case oh this is the shift that's kind of to their benefit. But first of all there are certainly many who could also be very affected. Truck drivers being an obvious category. And secondly, our economy and public policy has proven itself very capable of looking at a situation when blue-collar workers are very much in demand and declaring a labor shortage and bringing in temporary labor or promoting offshoring or doing a whole bunch of things that have not necessarily actually delivered the outcomes we might want. Even where the basic economics suggest we should be seeing something better. And so, you know, Alex, I think you've been outspoken on this point that it's incumbent on the technologists. It's incumbent on the folks who are going to be making decisions about a lot of this stuff to make sure that workers have a seat at the table instead of being on the menu or else everyone's going to be on the menu. But what does that look like in practice? What are the levers that you think we have to pull that are actually going to make sure this does work out?
Let me answer a slightly different version. It's like a lot of times we assume in philosophy or in policy that it's a lot of it actually comes down to is what you're saying actually in accordance or in discordant with the underlying technological economic developments. And if you just broad swath look at Europe and America there are certain tendencies that are somehow extremely discordant with the AI revolution. So the AI revolution does make people with actual interesting and physically placed skills more valuable or can, but it also means we cannot have policies of importing labor like that is just insane. Like there is no, so like some of these things that look moral like immigration policy. Who should we, how do you justify importing labor into economies that will become much more efficient where the kind of people I think you represent can be trained and become more powerful or more economically powerful and then you have wide swaths of people who, and this is where I'm not sure, a lot of the people in policy positions are also in jobs that are training that you know are going to be particularly susceptible to replacement and then that's a very very hard thing to deal with. And so you know just having a baseline of we are going to go to a place where you know certain kinds of things we're going to need more of, other things are going to be rationalized, that infects a lot of lives and then what are the policies around it including taxation, training, upscaling, how do people, like if you just say like if a labor is now four times more valuable how do they capture that? Like the classic way we remunerate people in a company is essentially a pyramid where a lot of the people at the top make a lot of money and kind of the people at the bottom have had to fight through organizing labor or however to make sure they get a ever smaller portion of the pie. But what happens when like if you're using large language models in our products, those people are actually the people at the front line that are going to take over certain kind of decision-making and actually can do it much more efficiently and so that they're not replaceable and also can begin for the first time like one of the big issues you have in America with manufacturing especially manufacturing of complex devices is we really couldn't manufacture certain things in this country that you know you could in Taiwan, you could in Germany, you could in Japan, you could in Korea because they were very technically and culturally specific. But now you can and you can do it with like American workers under like the conditions we have. And so those policy shifts have to work backwards from the AI revolution, not forwards from some economic theory that in my view is just no longer applicable. I mean last and to say kind of the most obvious thing I mean this revolution is actually happening in America. Every part of the stack is American. So that's also hugely unique. I can't think of any other if you like do military technology or any other thing. It's very hard to think of a technology that is this culturally this country specific. So like there's a lot of levers, but I would say to your point like it's just there has to be like we're part of how this is going to work if only also because like the fissures in this society that are so scary are political and if we don't have kind of people of good faith who may not agree on every issue but want to make it work, we're going to blow this thing up.
So, let me push back on one thing before we come to sort of what does that mean for labor's role? The description that you're giving of the frontline worker is four times more valuable now or could be. Yeah, could be makes it implicit to jump to and therefore could be compensated at four times the level. That does not necessarily happen. And in fact there are several others who could instead declare themselves to be more valuable and capture that. One would be whatever middle manager is left who deploys the system and says he is actually the one who is four times more valuable. One would be the seller of the system who says it is my system and I am going to essentially charge up to the profit, you know, essentially capture as much of the value upstream. So is your model and way of thinking about this...
No, no, we really can disaggregate the productivity and this naturally flows to the workers that it's about the, that it's about something about AI or that look, there's a subtle question you're asking which probably would require a different panel. There's always a question of Pareto optimality so I'm not saying that Pareto optimal, mean like if they create four x you get compensated four x, depends on if you have you repping you or... So there's always that. But I would say as a general trend, one of the political and cultural sociological problems we have is like look at how panicked many of the people are in the tech industry who aren't like AI proximate. It's because these technologies expose parasitic and weak products. So you have a strong product if you actually know what you're doing and someone augments that, you have a weak product if you're taking credit for it. And it was much easier to hide that weak product. And what I think you're going to find is the middle management who are like doing things that can be done in our product with large language models or maybe other products or different kinds of models. They have a weak product. Like it's like a societal version of this would be it is like a lot of these things are pentesting what is the underlying value and it is very very hard to hide. Like I think a lot of what is happening with skepticism towards government is does anyone here think government's efficient? Like yeah, so but you can't really hide how inefficient it is and that's so like what I think one of the things I, you know, when our private conversation one of the things I was saying I won't say what you just, like you're going to find that I believe I could negotiate a deal where workers get paid substantially more, people in the middle have to be retrained to be workers and the person at the top who's probably a neurodivergent freak is getting paid more than probably you might want but the people at the bottom are making a lot of money and I think that's how every structure is going to look in this country and one of our advantages as Americans and it was we can do this. I think it's going to be much harder for other nations to do this.
So to what extent should we expect this to happen fairly naturally? To what extent does labor play?
Well, labor has to be at the table like I stated earlier because think about all the efficiencies and technologies over the last 30 or 40 years that have been implemented. You've seen such a wage disparity between the upper level management and that manager that's taking credit for work he's not actually doing or performing and there's been such a wage discrepancy that's why you've seen over the last 5 years, I think part of it which was helpful, you always look for a positive out of negative, you look at the pandemic right in our world and for once we were looked at as essential, we provided goods and services to this country probably in the worst uncertain terms with total disregard for our members safe, healthy, and the safety of their families as well. So, we actually demonstrated our value. So, we were able to demand what we were worth. Now, to Alex's point, you know, if we don't have a seat at the table, if we don't collaborate and coordinate and actually get this thing right, you're going to see two classes of people. You're going to see the filthy rich and the filthy poor. And there's going to be nothing in between. And the one thing we always focus on, and I try and work bipartisan when policies are being made, is that the one thing that nobody in this room can disagree with is when you have strong jobs, you have strong communities. And when you have strong communities, you have good education systems. And if we do not collaborate, we've got an opportunity right now that we've never had before where labor comes together with tech, labor comes together with architects of AI, and we can actually have a coordinated, concerted effort to get this thing right, I think everybody wins. And you know, the problem is if we don't, that bottom tranche of blue-collar workers are going to be stuck where they're at. They're never going to be able to climb that ladder. They're going to be stuck in a community that's, you know, disintegrating day after day. And you're going to have probably 10% of the world over or 10% of the nation over the next 10 or 20 years that are going to be born into wealth generationally as a result of all the savings. And so how is that ever going to right itself?
Yeah. So I want to push you a little bit further on what it means to get workers a seat at the table because on one hand we have examples like the teamsters that do that very effectively. On the other hand, we're at 6% private sector union representation. That will not get up to 30, 40, 50% as fast as AI is coming. I let me know if you disagree, but I certainly that would be my expectation. How do we make sure workers have a seat at the table, are represented? Is it more unionization very quickly? Are there other mechanisms out there we could be using? Do we need to create new ones?
Well, I think we've got legislation that's bipartisan that we're hopeful that everybody will adopt, which is the Faster Labor Contracts Act, which will help. And this isn't about increasing union density. It's about protecting American workers and American jobs. If employers fail to treat their employees with dignity and respect, that's when they come to us. If employers are doing the right thing, paying their people what they're worth, providing good benefits, providing a future, a career path to the height of their talent, then they don't need us. But the reality of it is at the end of the day, the seat at the table, I mean, is having conversations like this. Do you ever think 10 years ago that you'd have a general president of a labor union sitting at an American Compass event with the guru of AI? Absolutely not. Right. I knew nothing about AI or American Compass, so definitely not.
Yeah, you were still watching like Barney or one of those shows, right?
I was out trying to learn how to drive a truck.
No, that's it's a very interesting point. How far? All right. Was it the Wiggles? I was actually Lamb Chop's Play-Along was my preference, but that was a little bit earlier. The point you're making though about how quickly things are changing I think is exactly right and so that I think creates a lot of space for opportunity. I want to ask you guys one more question about the technology which is that there is it seems to me a huge trust problem right now between to some extent how workers perceive the deployment of technology in the workplace to a significant extent how Americans perceive the effects of recent waves of technological innovation coming out of Silicon Valley. I guess actually start with you Sean and then come back over to you Alex. How would you characterize that? Is AI sort of like is AI mostly a vague this seems scary could take my job or is there a more concrete sort of bill of particulars that is what has frustrated people so much in recent years?
Well, I'll try and answer that in two parts. AI is terrifying to our members because as I stated earlier, the fear of the unknown. You know, one thing you'll hear me state often is, please just state your intentions, tell us what's coming our way, and let's figure out how we'll deal with it. Over the years history, whenever technology was implemented and/or talked about it was just implemented not thought about the ramifications and the consequences of what this technology whatever it may have been whether it was the combustible engine whether it was robots whether it was any type of technology and I think the problem is on our side of it right now labor American worker we need answers to questions we are a society now I think where we're a lot more sophisticated. There's a lot more information to be had, easy, accessible, good and bad. We just need these gurus of technology, these AI innovators to actually state their true intentions of what they're looking to do with this. And I know I've had conversations with Alex privately. He's one of the only few, if not the only one that is actually I've ever heard say, 'Look, AI is here. It's great. It's great on the battlefield. It's great protecting this country. It's great protecting our nation. It's going to be great. But we also got to protect the people that are going to be affected by AI.' So when I say there's a lot of skepticism, there is. There's a lot of uncertainties just as history repeats itself. But I think there's a lot of opportunity as well not only to create jobs but to create credibility amongst opposite interests.
Yeah. And so you know Alex I think to Sean's point there are some folks who take this issue very seriously in the technology community. Some who in my perception take it maybe less seriously. How do you feel like the tech community is actually engaging with the trust issue? Is it there is no issue?
Could it be worse? I mean like it's like we're going to take your job. We're not going to give you anything. We're going to get very wealthy. We don't inherently talk in a way you could one could easily understand what we're saying. In many cases, we're talking our playbook, which is actually masking a longer timeline on development or conversely, the fact that there are some really hard things to talk about is this is definitely not a technology where everyone wins. In fact, that on the battlefield, the chief advantage of AI for America is implemented correctly and certainly with hiccups, we should be able to establish ourselves as a military force for which there's no equal. And or at least prevent any other nation from being able to do that, which I would imagine would be bad. I believe very deeply would be bad for everyone in this room. And like that's something like nobody should...
The people who believe they want that are somehow living in a world where that could never happen or misinformed about just the way people see that. I don't think our adversaries are amoral at all. They just have a different moral sense of what it means to be a worker, what it means to have or not have a first, second, or fourth amendment, or in the theological sense, the radical nature in which our rights are protected in direct tethering to a higher being is even radical in Western Europe. Those are literally our God-given rights. And then the moral anchoring of that, especially vis-a-vis a technology that is moving to what you could call super intelligence, meaning how do we anchor rights above that technology and fight for them. So then it's obviously something that should augur to our benefit, especially in the military and adjacent located places. There are a lot of people who would disagree with what I just said, and so I think that's legitimate. You can argue that. I would say on the labor and then privacy, a lot of the critiques come down to, well, what will happen to me and my children? I can't think of any technology that is at least been where I've interacted with it, meaning the last 25 years and in a weird way before that, that had this kind of aspiration and actual ability to have a global impact on labor. There are all sorts of disparities in this culture. We are a society divided by class, I would say by gender often. Many Americans would say by race. I am much more of a classist. But in any case, obviously disproportionately affected depending on what you are in that stack. But every single one of these things, AI is going to have a different and disproportionate impact on. And then we have the reality that there is no class of people that has as little support probably as tech leaders in this country. And there is an issue of how much wealth can one very small group of people have without that pissing off people or giving them a sense this is not a fair, maybe potentially rigged game. So I think that I'm not going to even delve into the way in which these technologies, they feel somehow like again, theological aspirations with no tethering to a major religion, and in the context of contracts, I would say Christianity. So this is just a cluster.
Well, and that's another reason why...
So glad you threw the first f-bomb, because you were all expecting me to do it, right?
You didn't put that on that gambling app, did you?
Well, sorry. I wanted to pick up on the trust point because the cluster I think is exactly right and the question is how does that resolve itself? Does that resolve itself? It seems to me the only two resolutions are either something happens that leads the tech community to take seriously the need to earn trust and be trustworthy or they face a populist defenestration. Which happens first or is there another outcome?
You know, you're assuming people know. Like in the weird thing at Palantir now, it's the craziest thing ever. We're deployed on a lot of the world's most important things and then I get all these calls and in most cases the problem is you assume very smart people can do very simple things and you're unfortunately completely incorrect. So you're asking, I think there's going to be increasing and increasing political pressure on the system. I think populist movements, by which I mean forget populist because in some ways I'm sympathetic with both certain forms of left and right-wing, but people like you want to represent your members and there are granular and quantifiable ways that these technologies can be proven. And if you're at the table with your members, you can evaluate it. Getting as many people in the tent that want granular improvements that can be measured is, I think, the most important thing, with the widest tent possible that includes everybody who actually wants measurable improvement. Because on the other side of that tent, you'd have people of all backgrounds, all parties that actually don't want that, and this is the perfect technological development to just throw gas on everything and burn down everything.
Well, I think part of it is too, you make a great point. You're always going to get the people that don't feel included, that no matter what good this technology could bring, no matter what efficiencies, no matter what opportunities it could provide, you're going to have people, if they're not included, trying to burn it down. So I think it's important to your point that the more people that we can get under that tent, and look, no one has to agree on a perfect science, there's going to be, to your point, there's going to be sacrifices, there's going to be winners, there's going to be losers, but at the end of the day, if you develop a technology, you develop a plan that's going to generate opportunity, generate growth, and close a wage disparity, I think it's a home run.
That's a perfect segue to the last thing I want to do. I warned you guys we were going to do this. We're going to do a little bit of interest-based bargaining. For those who don't know, regular bargaining, you just pass your offers across the table. Interest-based bargaining, we're actually going to identify some underlying interests, find some ways to address them. To my left, representing labor, Sean O'Brien. Slightly more to my right, representing capital, Alex Karp. You alluded earlier to the idea that you could envision a sort of a grand bargain. And so I want to identify what are the elements of that from each of your perspectives. What are the couple of core interests that a bargain would actually have to accommodate and make sure are well addressed?
I mean, there's lots of insane regulation around everything. So I'm sure everything I'm about to say is not feasible. But the way I would negotiate a bargain with workers, organized workers, non-organized, I would go to them and say, 'Well, help me get rid of all the bozos in your business, or let's figure out a way where basically I'd make you capital partners.' Like, I think we can reform this business and the way we do our software license, I'll guarantee you a license, meaning your current salary, and it'll go up 25, 30% a year, and some portion of the actual growth in the business. It could be the same as what we make, it could be more. And in return, when people run around calling me evil because I've helped take out terrorists, you'll remind them that, well, you know, we're a friend of labor.
Well, all right. So that's Palantir specifically, but go one notch wider. If you're thinking about the industry, this collection of businesses, capital interests that want to introduce AI quickly, are investing trillions of dollars in capital at this point, need to generate a profit on that investment. What are the must-haves for any arrangement with workers in that deployment?
