Making Sense - Sam Harris - December 04, 2025


#447 — The Unraveling of American Power


Episode Stats

Length

20 minutes

Words per Minute

178.25171

Word Count

3,697

Sentence Count

211

Hate Speech Sentences

8


Summary

Peter Zion is a demographer and geographer. He's been on the podcast several times, and most people will have heard of him, if not here. In this episode, he joins me to talk about the Trump administration, trade, tariffs, supply chain, and China.


Transcript

00:00:00.000 Welcome to the Making Sense Podcast. This is Sam Harris. Just a note to say that if you're
00:00:11.740 hearing this, you're not currently on our subscriber feed, and we'll only be hearing
00:00:15.720 the first part of this conversation. In order to access full episodes of the Making Sense
00:00:20.060 Podcast, you'll need to subscribe at samharris.org. We don't run ads on the podcast, and therefore
00:00:26.240 it's made possible entirely through the support of our subscribers. So if you enjoy what we're
00:00:30.220 doing here, please consider becoming one. I'm here with Peter Zion. Peter, thanks for joining
00:00:39.240 me again. Hey, it's great to be back. Well, so first, you've been on the podcast several times,
00:00:43.360 and most people will have heard about you, seen you somewhere, if not here. But remind people
00:00:49.040 briefly what you do, what's your day job when you're not doing podcasts like this? My job is
00:00:53.860 to take demography and geopolitics and use it to paint a picture of the future and then show how
00:00:59.480 it matters to the people who happen to be in front of me. So most of my clients are in the world of
00:01:03.560 finance or manufacturing or agriculture, some sort of economic association or business, and I help
00:01:09.060 them navigate the crazy that is to come. Speaking of predicting the future, and actually on points that
00:01:16.940 matter, at least to me and a few of us, last time you were on the podcast, you told me that there was
00:01:22.060 no way that Trump was going to win the 2024 election. So I have to ask you what the fuck
00:01:26.080 happened right at the top here. Yeah, fair enough. Sorry, I slept that day. My general feeling was
00:01:31.920 that it was going to be America's independence that decided how things were going to go. Not people who
00:01:36.020 said they were independent, but really were Republicans and everything but name, but the
00:01:40.160 actual 10% of the electorate that splits their vote because they had decided the last seven
00:01:45.860 election. So it felt pretty safe to me. And in the post-election polls, it looks like that group
00:01:52.740 decisively voted against Trump. Everyone else shifted, however, with one exception, Washington
00:02:00.060 State. Every state in every demographic shifted substantially towards Donald Trump and allowed
00:02:06.680 him to win with the margin that he got.
00:02:08.660 Well, how would you rate his second term so far? We're going to talk about tariffs and foreign
00:02:15.000 policy and supply chain and specifics like China, but what most concerns you? Have there been any
00:02:21.620 pleasant surprises? How would you rate the last 10, 11 months?
00:02:25.080 I usually try not to be critical of presidents in their first terms because they're learning. It
00:02:30.380 takes time for the policies to kick in. The world doesn't move on a dime. Second term is different
00:02:36.040 because they should have learned from what happened in their first term. And by that measure,
00:02:39.960 it has not gone well. Most of the stated aims of the Trump administration, things like reshoring,
00:02:46.700 manufacturing, increasing its defense position, have actually taken the country in a significantly
00:02:51.000 different direction. And I have not seen an unraveling of national power on the scale since
00:02:56.640 the Soviet breakdown. So, so far, very impressed with how much damage is being done.
00:03:03.880 This is some combination of Zimbabwe-Argentine economic policy, maybe with just a dash of
00:03:12.400 French statism thrown in.
00:03:13.880 All right. So let's talk about the unraveling. What are you seeing? Because so over right of
00:03:19.880 center in Trumpistan, I can imagine many people believe that the tariffs are working. We're
00:03:26.920 on-shoring American manufacturing. We're building up our manufacturing base. Jobs are going to be
00:03:32.340 returning as a result of all this. We have China right where we want her. And it's really just a
00:03:39.180 picture of American strength. I mean, you know, Trump is the kind of guy you just can't screw
00:03:44.860 around with. Finally, we have a president like that. How are you not seeing this?
00:03:49.880 Let's start with the tariffs. When you do a flat tariff and you tax things at the border,
00:03:55.100 no matter what they are, you need to break down manufacturing into two broad categories.
00:04:00.340 First, you have your simple products, things like plastics or furniture, textiles, where there's
00:04:05.860 only a half a dozen manufacturing supply chain steps. And in a high tariff environment, it's
00:04:10.560 relatively easy to relocate some of those steps within the tariff wall. So we are seeing significant
00:04:16.060 build out in the United States for things like furniture and glue and paints and all that kind of
00:04:20.500 stuff. More complex manufacturers, the hundreds or thousands of steps are very different. So here's
00:04:26.560 automotive, aerospace, electronics, computing. When there are thousands of steps and you only produce
00:04:32.280 say half of the steps in country, you have to pay the tariff every time an intermediate product crosses
00:04:37.080 the border. And the end product ends up costing significantly more. So simple math, it's easier to
00:04:43.780 take the steps that are within your tariff zone, move them outside of your tariff zone, and then you only
00:04:47.920 have to pay the tariff once. So what we're seeing right now is a steady de-industrialization of the
00:04:53.640 value-added, high-skilled labor-intensive jobs, and instead replacing them with much lower value-added
00:05:01.940 jobs and things like plastics, or maybe things that are automated because our labor force is too much.
00:05:07.000 So we've seen contraction in the manufacturing sector for the last six months. And the number that I pay
00:05:11.820 attention to the most right now is called industrial construction spending. It's a data point put out by the
00:05:16.560 Federal Reserve. And it's been negative since tariff day. So people are finishing the industrial
00:05:21.580 projects they had started before April, but no one's really launching much that's new.
00:05:28.120 So what are you expecting to see in the economy over the next 12 months as a result? I mean,
00:05:32.740 is this synonymous with inflation, what you just described?
00:05:35.860 Well, if you go back to before April, we knew we were preparing for the fall of the Chinese system.
00:05:43.620 We were preparing for the end of globalization. We knew we needed to expand our industrial plant,
00:05:48.460 probably needed to double it, and probably needed to increase the power plant by 50% to go with that.
00:05:53.520 There's no way you do that on anything less than a 40-year time frame without a lot of inflation. And I
00:05:58.380 would argue we needed to do it in the next five to six years. So inflation was already part of my
00:06:03.080 forecast. But now we have de-industrialization going on on top of that. And so if we can't turn this
00:06:10.140 around and get back to building things again, and that requires a very different federal policy,
00:06:14.760 we're going to enter a period where we're going to lose the products from abroad at the same time,
00:06:19.060 we no longer are producing very many of them here. So we have to still go through that industrial
00:06:23.440 build-out, but now against a backdrop of goods shortages. So the debate now is, do you want high
00:06:29.320 inflation plus productivity or high inflation plus a goods shortage? Neither of these are pretty.
00:06:35.940 Hmm. So obviously there's been a huge build-out with respect to data centers for AI.
00:06:41.400 And progressing.
00:06:42.260 Yeah. I mean, that's the one thing that has been a bright spot for the market. It remains to be
00:06:49.560 seen whether it's a catastrophic bubble in the end, but-
00:06:52.660 It's definitely a bubble, let's be clear. Whether it's catastrophic, yeah, who knows?
00:06:56.560 Yeah. I mean, and it can be a bubble. I mean, we should be clear, it can be a bubble without AI
00:07:01.100 itself being a bubble, right? I mean, it could be this over-investment and a build-out even if AI
00:07:05.700 is in fact the future, right?
00:07:07.940 Well, there's a lot of unknowns in that statement, so I don't want to overly come down on one side or
00:07:12.960 the other. 15 years ago, 10 years ago, people were aware of large language models, but I don't
00:07:18.060 think there was anyone in the AI space who really thought that LLMs were the thing that was going to
00:07:22.920 break open, and here it is. So expecting the model we're on right now to be what we use in 5 years or 10
00:07:29.600 years or 15 years is a bit of a stretch from my point of view. Technology doesn't go in a straight
00:07:33.420 line like that. Also, LLMs require a massive amount of processing capacity with graphics processing
00:07:40.440 units, and GPUs are the most advanced chips that humans make, and GPUs cannot be made without a
00:07:46.600 fully globalized system. So I actually look at today's build-out and like, this is great because
00:07:51.680 every data center needs power to go with it. We need the power regardless, and if we're entering a
00:07:56.520 world where globalization breaks down, we won't be able to make the GPUs, so anything that's installed
00:08:01.340 now is all we have. So while I'm not a believer in the trajectory that we're on, I have no problem
00:08:07.960 with any of the steps we're taking right now. Do you think we're in the process, however gradual,
00:08:13.880 of on-shoring our dependence on Taiwan, or is that not in the cards? No. There's 100,000 manufacturing
00:08:21.300 steps for high-end semiconductors. The FAB facility is an important one. It's an unavoidable one,
00:08:27.180 but there's another 40,000 that are like that. And once you actually have your FAB facility,
00:08:33.800 the next step down is testing and packaging, incorporating into components. We do very little
00:08:39.240 here. I'd argue that's much more important. And on the front end is the design and the logistics.
00:08:44.380 We do do that. That's the highest value-added part of the process. So this obsession with FABs is
00:08:50.560 actually taking one of the lower value-added steps that's most capital-intensive and putting
00:08:54.600 at the center of the conversation. And that's just bad reasoning from my point of view. We should be
00:09:00.060 focusing on things that are downstream, closer to the end consumer, if you're worried about national
00:09:05.120 security. Well, so I want to talk about national security and foreign policy and talk about China
00:09:09.500 and Central America, South America. But just to close the loop on this, what do you think,
00:09:15.480 if this process of what you're calling deindustrialization continues apace for
00:09:19.820 the rest of Trump's term, what are you expecting to change over the course of four years?
00:09:26.940 I mean, there's so little with this administration that goes in a straight line. So that's a really
00:09:30.980 loaded question. Key thing to remember is that Donald Trump still hasn't fleshed out the government.
00:09:35.520 He fired the top several thousand positions, filled very, very few of them, and then surrounded
00:09:40.060 himself with a cabinet that has very little technical experience in anything. So if you think back to
00:09:44.800 Obama, we had a cabinet that was all academics that had never done anything in the real world.
00:09:50.300 Now we have ideologues who have never done anything in the real world. So to expect to get good advice
00:09:56.880 going to the president, even if he was open to hearing it, is pretty thin. And we've now had
00:10:01.500 major policy changes on drug policy, on illegal narcotics, on the Red Sea, on Israel, on Ukraine,
00:10:10.560 on vaccines, on tariffs, because, and I quote, it's what I feel. We get no straight lines on that.
00:10:17.340 And so you can imagine what's going on in the business community right now. This is the reason
00:10:20.960 why industrial construction spending is falling. No one knows what policy set is going to be tomorrow
00:10:26.140 or next month or next year. And without some degree of continuity or some degree of guide,
00:10:31.020 everything is up in the air and everything is on hold. So if this sort of ambient chaos holds,
00:10:36.480 and we just hit our 600th tariff change since January 20, no one's going to put money to put
00:10:43.140 anything in the ground because they don't know if it's going to be worth their time.
00:10:46.120 So what's going on with Venezuela and Mexico? I know you focused on cartel violence. Yeah.
00:10:53.220 Trump has obviously been threatening to deal harshly with the cartels. I think it remains to be seen
00:10:59.120 what he's intending with respect to Venezuela. But how do you view our ramping up hostilities on both
00:11:06.540 those fronts? Well, let's start with the understood reality on the ground. Venezuela is a very weak
00:11:12.540 state where the state security services have been so focused on crushing political dissidents for so
00:11:18.000 long that they've lost control of large tracts of territory. And so some of the cocaine smugglers in
00:11:23.020 Colombia have now used Venezuela as a way to get their stuff to the wider world. It's a minority. It's less
00:11:27.860 than 10% of the flow. Most of it still goes north into Central America and ultimately Mexico,
00:11:33.760 especially Mexico before coming north to the border. But there is a flow that goes through
00:11:37.720 Venezuela. I'd argue most of that flow eventually goes into Europe as opposed to coming to the United
00:11:42.920 States, but it is real. And there is a degree of collaboration among members of the Maduro government
00:11:48.800 in facilitating those flows. I don't know if calling the cartel is the right term. That implies a degree
00:11:55.340 of overarching control that probably doesn't exist, but it is a problem. So anything you want to do
00:12:01.800 against Venezuela on a state security point of view, there's an argument to be made there.
00:12:06.360 However, the Trump administration has yet to brief Congress about any of the details.
00:12:10.440 So none of the intel that supposedly identified the ships has ever been shared with Congress. And
00:12:17.060 what I'm hearing from the military is that there really wasn't any ever. And that's before you
00:12:22.160 consider the legal implications of some of the specific decisions. We had both Trump and Hagseth
00:12:28.120 brag about performing war crimes the week of Thanksgiving, because if you take out an enemy
00:12:35.380 combatant and they've been completely disarmed of any sort of offensive capability, and then you shoot
00:12:41.860 them, that's a war crime. Every treaty the United States has ever signed on international law
00:12:46.580 codifies that. And so now they're backtracking. It's like, well, maybe that's not what we meant.
00:12:51.680 In addition, we now have the USS Ford down there, which is our top super carrier, badass piece of
00:12:57.460 military equipment. And it was only on the Monday after Thanksgiving that Trump convened a national
00:13:02.860 security meeting with the Joint Chiefs, Secretary of Defense, Secretary of State, to begin the discussions
00:13:09.260 about what the goals in Venezuela are. So to this point, the policy has been, let's shake some sabers and
00:13:18.320 see if Maduro will just leave. Now they're starting to have the early discussions about what the policy
00:13:24.620 might be. Very much, of course, on the wrong side of the card.
00:13:29.780 Yeah, it's remarkable that the, these kind of extrajudicial killings of people on boats in the
00:13:36.260 Caribbean is, I mean, it's being kind of roundly condemned, but on its face, it seems like it's,
00:13:43.900 has to be in violation of some law or past practice because, I mean, we don't even kill
00:13:50.780 drug smugglers when we catch them and convict them of drug smuggling. I mean, that's not a capital
00:13:56.220 offense. And yet now we're simply annihilating people who are alleged to be smuggling drugs. And we
00:14:03.440 have been shown no evidence that that's even true. And as you just pointed out, there's this recent
00:14:08.220 case of this really patent war crime where we've bombed the boat. Now people are scrambling in the
00:14:15.940 water and we bomb them again just to make sure everyone is good and dead. I mean, that is a,
00:14:20.380 just a, unambiguously a war crime. But I mean, the whole idea that we can be just killing people
00:14:26.580 who are suspected of smuggling drugs. Well, has that been a past practice? Is that in line with
00:14:33.400 any of our laws? It's certainly not officially. I'm never going to say that it wasn't done because
00:14:37.080 there's a lot of that happens with covert action in Latin America that, you know, maybe is distasteful,
00:14:41.140 but it's not technically illegal by U.S. laws. And the, the, really the limiting factor here is the
00:14:47.240 War Powers Act. So that was something that was put in place by Congress bipartisan in the aftermath of
00:14:52.420 Vietnam, uh, passed by a two thirds vote, president vetoed it. It overcame the veto. It was like the
00:14:58.480 biggest overcome of the veto that we've ever had in modern history. Every administration since then,
00:15:04.280 including the Trump administration, has said that the War Powers Act is unconstitutional. It's an
00:15:09.300 unconstitutional check on presidential authority, but it has never come up to the courts. And Congress
00:15:16.260 since then has never called the president to account. So either the president has presented
00:15:22.520 his case after 30, 60, 90 days to justify, and then Congress votes to authorize the military action.
00:15:29.540 That's for example, how Iraqi freedom happened, or the president is done within 30 days and folds the
00:15:35.480 situation down. This is the first time that the War Powers Act has been directly challenged over that
00:15:41.400 30-day threshold. But Congress has done nothing. And unless and until Congress actually decides to
00:15:48.600 act, technically it's legal. Doesn't mean it's right, doesn't mean it'll work.
00:15:53.620 Congress has been sitting on its hands on, uh, virtually every question. What about Mexico? What's
00:15:59.220 happening with our posture, uh, vis-a-vis the cartels? And, and, uh, what are the implications of
00:16:05.380 that? Uh, this, this is actually more of a Biden administration question than a Trump administration
00:16:09.340 question. Um, under the Biden administration, um, a significant amount of progress was made at
00:16:14.720 cracking down on the Sinaloa cartel, which for most of the last 15 years has been the most powerful
00:16:19.800 organized crime group, not just in Mexico or in the United States, but in the world. And there was a
00:16:24.940 general agreement that it was the most effective of the drug smuggling organizations because El Chapo ran
00:16:30.740 it as business. But El Chapo was eventually arrested and has now been put away. And combine that with the
00:16:38.600 effects of the last five years prosecution of his old organization. Um, his son's basically running
00:16:44.140 the place like, um, Uday and Cuse, if you remember those Iraqi assets. And has it been two years? Is
00:16:50.260 it only a year? Relatively recently, one of Los Chapitos, that's one of El Chapo's sons convinced the
00:16:56.940 guy who was the chief accountant of the organization to get in a plane and fly to El Paso to look at some
00:17:03.140 real estate and the DEA was waiting. So we basically had a shattering of the organization
00:17:08.920 top and there is no longer one Sinaloa cartel. There's like a dozen and they're fighting one
00:17:13.900 another. That sounds great. But the second largest drug trafficking organization in Mexico,
00:17:21.160 in the world is called the Holisco new generation cartel. And they are an order of magnitude more
00:17:25.980 violent. And now they are the most powerful entity and they are taking over a lot of former Sinaloa
00:17:30.680 territory. So by taking out the big bad, we've helped inadvertently to create a new big bad that
00:17:36.440 by many measures is a lot worse because Sinaloa, their kind of guiding idea was you don't shit
00:17:43.400 where you sleep. You bribe local law enforcement. So they're on your side. Holisco, their idea is that
00:17:49.760 the shit is the point. And so the first thing they do when they go into a new territory is shoot the
00:17:53.900 mayor, shoot the police chief, randomly shoot a bunch of people in town. So everyone knows who's in
00:17:57.640 charge. And so nobody stands against them. Different model, a lot more violent, not as
00:18:02.500 effective in the long run, but it generates a lot more activity that we are now seeing among other
00:18:09.880 things in the murder rate in Mexico. So has the Trump administration done anything new on that
00:18:15.620 front? I know he's said that he wants to send in SEAL Team 6 to deal with the cartels. Have we done
00:18:21.460 anything that's actually changed our posture with respect to them? Not hugely from my point of view.
00:18:26.300 And I don't see that as a reason to love or hate the Trump administration. The core problem is
00:18:31.160 demand drip. And as long as Americans like their cocaine, you know, this is going to happen.
00:18:35.360 And we've discovered that if you take out the kingpins in the case of the Sinaloa,
00:18:39.300 you create a more fractious environment where there's a lot more competition.
00:18:43.200 That's, for example, one of the reasons why fentanyl has risen. Cocaine requires a supply chain
00:18:48.020 that involves a half a dozen different countries. And it takes about four man hours between the
00:18:52.460 the fertilizing, the growing, the processing, and the smuggling for every dose of cocaine.
00:18:57.620 Well, fentanyl, you can buy a bunch of precursors on the open market, make a batch of a million doses
00:19:03.820 in your garage in a week. And it takes four man seconds to get the fentanyl in. So we've,
00:19:12.960 by cracking down on cocaine to agree, we've inadvertently created a new product set.
00:19:17.300 And taking out cartel leaders will have absolutely no effect on that. And one of the things we've
00:19:23.280 ironically seen with the border is by stopping the flow of illegals, it's made it very, very,
00:19:29.700 very easy for all the drug smugglers to just ship stuff in containers. So again, I don't blame Trump
00:19:35.960 for that switch. This is one of those things that as long as people want their narcotics,
00:19:41.520 there is going to be an economic pipeline of some form for it.
00:19:44.440 What do you know about the recent pardoning of the former president of Honduras who was a
00:19:50.320 convicted drug smuggler?
00:19:52.640 If you'd like to continue listening to this conversation, you'll need to subscribe at
00:19:56.600 samharris.org. Once you do, you'll get access to all full-length episodes of the Making Sense
00:20:01.880 podcast. The Making Sense podcast is ad-free and relies entirely on listener support. And you can
00:20:08.300 subscribe now at samharris.org.
00:20:14.440 Thank you.