In a recent podcast Phil suggested that bringing manufacturing home to America won’t necessarily create jobs, because most factories will be automated. They just need one man and a dog, he said. The man to turn the machine on, and the dog to make sure he doesn’t touch anything else.
That touched a nerve with Brian Hanley has spent his life refining manufacturing processes. The key ingredient suggests, is people. Elon Musk was the latest to try the lights out approach and realised it didn’t work.
Instead, if the US wants to succeed with a competitive manufacturing sector, it needs to look to post-war Japan. Workers were an integral part of the refinement and adaptive nature of manufacturing processes, in part because of the company-based (rather than industry-speciifc) union structure. Listen in to find out how Japan’s adaptive approach is what’s needed if the US is to develop a successful manufacturing sector.
Two books related to this, that Brian says should be required reading or every economist:
- Kanban Just-in Time at Toyota by Japan Management Association
- The Sayings of Shigeo Shingo by Shigeo Shingo
Hosted on Acast. See acast.com/privacy for more information.
[00:00:00] I love manufacturing, I think it's a really building solid objects that deliver value to people, I think is really, really a great thing. I think we should do more of that in this country. This is the Debunking Economics podcast with Steve Keen and Phil Dobbie. Well that was Elon Musk in his pre-wacko days, arguing something that most people would agree with, more manufacturing in America, or anywhere really, but how and what can the US learn from Japan?
[00:00:30] What has Elon learned from Japan? That's this week on the Debunking Economics podcast. Well I have no idea where we're going this week except this week we, that's Steve and I, decided we should be attacked by an American because that is very much the thing these days, isn't it? So Brian Hanley wrote to us saying,
[00:00:56] listening to the podcast you and Phil really don't understand automation or factories. This is particularly true of discrete assembly, process industrial facilities are different. So this, I take it this comes from the podcast we did a couple of weeks ago called The Trade War Has Begun. And I asked, is it too late for Donald Trump to re-industrialise America because products these days seem to have supply chains that crisscross the world?
[00:01:22] Can you really bring all those constituent parts back to one country? That's what we were discussing. That's what made Brian jump on his email. So, I mean, before we look at what we got wrong, Brian, welcome to the podcast, by the way. And is that argument correct? Is it too late to sort of try and bring all these constituent parts back to America?
[00:01:43] Well, that's probably, that's not going to happen because of the degree of specialization and the dynamic nature of manufacturing, which if I was to summarize it in a slogan, it would be innovate or die. And process industry manufacturing can be done in going in the direction of the lights out factory.
[00:02:09] And that's kind of the, the, the, the meme that's out there that we should be having AI do everything and, and that you can set up a plant to run that way. If you're manufacturing paper, you're manufacturing chemicals, you're smelting metal, that kind of thing, you can go in that direction. But even there, you know, I had a recent interaction with simple bricks.
[00:02:36] And there's now there's all kinds of different bricks. There's earth bricks, there's the classical bricks, there's, there's machines that you can buy for a few hundred dollars or a few thousand dollars that allow you to set up a brick making operation. And those are now being sold in Africa. And, and it's so everything, all of manufacturing is dynamics. So what did we get wrong then?
[00:02:59] Because we were talking about was it was our point that, you know, America could bring all of this back because it could all be automated. And what's the difference between, I think I gave the example, which was a little bit flippant, that all you need is a one man and a dog, which is, you know, the one man turns on the machine and the dog's there to make sure he doesn't mess it up. But, but I mean, is that, I mean, I mean, clearly that is too simplistic, but is that, was that your comment that when we, when we were talking about that?
[00:03:26] Yeah, that's, that, that, that's, that's what brought me to right to Steve. Yes. Because in the, in, in the automobile plant that I automated, well, I, I was putting in the integrate, the automation integration system. I designed it and developed that back when this was just not done.
[00:03:49] There were over a hundred engineering changes to the vehicles that were being implemented every day on the assembly line. This is not a static process. So all of that required human, human attendance. And it's also, manufacturing is a constant seeking of the lowest cost balance between quality that people will buy and the cost of manufacturing.
[00:04:19] And the, the, the, the example of Steve Jobs at Apple, when he, when he took over and the G3 was being manufactured at their state-of-the-art robotic manufacturing plant in Cupertino. They were building, they were, they were, they, they had a, almost a, a plant that was almost lights out that manufactured their G3 IMAX.
[00:04:44] And first thing Steve did was he announced that was shutting down and he was moving everything to China. He, he, he, he, Apple made Foxconn what it is because Foxconn was willing to turn on a dime, not present any restrictions to what the product could be.
[00:05:04] Whereas the product that Apple was, was, was, was, was, was restricted by the, the high investment in, in robots and specific ways of doing things that they, that they had, had, had put in, in Cupertino. So he wanted to, he wanted to get it, get into a situation where there was no need to design to our manufacturing system.
[00:05:33] And that's what a human intensive system does. So you're saying that they were, by having labor doing the jobs that would be done by a machine, they can do anything. Whereas a machine, you've got all that capital investment and you are making a product as defined by the design of the machines, which is not particularly flexible. Yes. And when you're dealing with robots, you also have to, have to realize that you're dealing with IT.
[00:06:03] You're dealing with, it's software intensive. And if you've ever dealt with IT departments, you know that IT departments always have a backlog. And so the IT department can become your bottleneck when you're, when you're using robots. People think of robots also as big, solid machines that, you know, don't wear out much.
[00:06:24] But the fact of the matter is that they're big machines that put very high wear stresses on small points on all, on all the joints and the, and the actuators. And because they do repetitive motions, those wear points are, are very constant, tend to get very concentrated. Right. Like runner's knees. Yeah.
[00:06:48] Like runner's knees, except that they don't, runner's knees have stem cells that regenerate them and robots don't. So it sounds like you're saying then that, you know, labor's always good. Low cost labor is always going to win against automation. Generally speaking, that's probably true.
[00:07:10] I sent you a couple of books, the images that can ban just in time at Toyota and Shigeo Shingo's zero quality control. We'll add links on the notes to the podcast. Elon bought this idea of the lights out factory, by the way. Elon tried to do it in, in Hayward. And there was one point where he had, the assembly line actually got moved out into the parking lot so they could continue to make production.
[00:07:40] And he eventually, you know, he, after beating his head against a wall, as he tends to do, he figured out that this was, you know, this was not going to work. And these books that I, that I, that I, the Kanban just in time at Toyota and Shigeo Shingo's zero quality control are two books that I think should be read by every economist.
[00:08:02] Because they, they lay out the, both the dynamic production of, of manufacturing and how, how it, how it's done and the necessity for, for continuous innovation very well, as well as what the methods are. And those methods came about in interaction in Japan with, with their unions.
[00:08:29] And the, and the Japanese unions are unique, are unique in the world and that they, their, their laws were given to them by the allies after the war. And those laws are the strongest union laws on the planet. So they actually protect the laws, the laws protect workers. Yes. Okay. Which of course America ceased doing about 50 years ago. Yes. But that's written into the Japanese constitution effectively.
[00:08:54] Um, I don't know, I don't know the details of how those laws are, but they're, they're, they're very strong and they adapted to them. It wasn't without kicking and screaming in the fifties. The, the code of Bushido, there, there, there's, there's this idea among, uh, many people here that, you know, Japanese management has always been touchy feely. They've always been, you know, very caring about the workers and all that sort of thing. And nothing could be further from the truth.
[00:09:24] Prior to World War II, there, people died in manufacturing plants in Japan as, as they did in the U.S. in the thirties. After the war, there were, there was a, there was a lot of union conflict in the steel industry, as is, was the case here in the U.S. Steel workers are, you know, big, strong guys and they, they're, they're, they're more likely to, to tangle.
[00:09:49] So to give you a sense for how this ended up there, there's, there's an account I read of an American who went to Japan, started a company. And a few years later, he had his first strike, which he didn't really expect. He got mad about it. And he would, he got very upset about being taken hostage in his office. He ended up falling downstairs and breaking his arm.
[00:10:11] He went to the police and the judge ruled against him because he had interfered with protected union activity. This ended up creating a situation where there was no necessity for unions to cooperate, collaborate with each other between, not just between industries, but within an industry between, between companies.
[00:10:38] And that created the foundation for the, for the Japanese company union. So you had a union at the company level. So that's it. So, so you saying, I mean, I might be stepping too far ahead here, but are you saying that created competition? Yes. Because I could see that if you've got a. Or, or it made productivity and changing manufacturing easier. Yeah. For the corporation. Because everyone in that company is working to try and improve their lot because that secures their employment. So they'd, they'd be looking for a more efficient company. Yeah.
[00:11:06] They realized that they could destroy their companies very easily. They, they, they, they, they, they knew that management had to come to the table to, to, to work with them. So the organization of Japanese unions actually suited more flexible manufacturing than was possible in America. Yeah, that, that, yes, it ended up doing that. And it also falls in line with Deming's work, which of course the Japanese took to, which the Americans either neglected or rejected.
[00:11:35] Well, Deming was, Deming was infamous for having a quarter to a third of his seminars stomp out furious when he told them they were the problem by definition. Management was the problem, not the unions, but management. And in Japan, this message was, there was also an interaction with, with Japanese culture. Japanese culture is a more group centered culture. It's not so individualistic.
[00:12:03] And, but this message that Deming delivered, you know, it fell on ears that were culturally open to that message, that you're the problem. And statistical quality control got implemented. This created, this created the foundation for how you could have Kanban, JIT and the Andon system that made zero quality control possible.
[00:12:33] So explain those, you just gave me two big words there. So explain what those systems are. Okay. Kanban is a system for pulling inventory into the plant, kind of like a grocery store, like customers do at a grocery store. And the idea is to minimize inventory of everything. Because the basic insight there is that all of your inventory is dead money.
[00:13:00] And it's just sitting there on your floor in your plant, like diamonds scattered around, waiting to be lost. So this is like the just in time manufacturing approach. Yes. You know, everything from screws to assemblies come into the plant. And once they're in your plant, there's an opportunity for them to be destroyed, stolen, damaged, you know, just lost, whatever.
[00:13:26] So that system came about from the factory floor. And eventually that was extended out to the dealerships where they set up a pull through system all the way through. And if you're willing to wait for your car, you can do that. Although they also ended up moving the plants closer to the people. So this is so the workers are providing this evidence and advice.
[00:13:51] One of the keys here is that the deal of being in the union was so good that if you're in the union, you essentially can't be fired for no cause. You can't be laid off very easily. And so all of the engineers, all of the accountants, all of those rank and file people, they were part of the union.
[00:14:15] And that meant that the unions knew that they knew as much about running the company and what was going on as anybody in management did, which, among other things, meant they were able to contribute. And so the idea, so they would, the interplay between people and automation, I guess they'd have every incentive to say, well, it makes sense to automate this bit because that is just going to make us more efficient as a company. We're not going to lose jobs. We're just going to keep jobs.
[00:14:43] And so they would have this inherent interest in automating where it made sense. So you get this hybrid about maintaining that flexibility with people, but automating where it made sense. And you talked about having to stay ahead of the game. So they'd be there saying, well, OK, we need to change the automation. Yes. And one of the keys to that was the Andon system. Andon just means lamp.
[00:15:05] And it started out as a light switch that workers would turn on, a group of workers would turn on at their station when they didn't have enough to do. Can you imagine that in America or the UK? No. Just keep quiet about it would be the best approach. Yes, keep quiet. Yeah. Long lunch. Yes, exactly. Let's take a break here.
[00:15:28] The unions worked hard on the elimination of headcount because they knew they had to get costs down in order to compete better. And they also knew that management couldn't take the easy way out and just fire them or lay them off.
[00:15:45] So one of the effects of that was that the groups would eliminate one headcount at their station or their area of expertise on the plant. And the person that they would eliminate would be the best worker, the one who had contributed most to making it more efficient.
[00:16:10] And the reason for that was that what management then had to do was find new places for them. And those people went into R&D and into setting up new product lines, et cetera.
[00:16:25] And this created a pressure from the bottom that was behind the endless incremental improvements to products and changes, in some cases, changes of the entire kind of work. You know, there were steel workers retrained to be software people, et cetera, et cetera. You know, some of those people succeeded, some of them didn't, and they'd move them around to find a place for them.
[00:16:49] And this foundation evolved to what they called automation with a human touch, meaning it was centered around the work group in the plant in stations. So I can understand why you got upset about my one man and a dog analogy, because it's the complete opposite of what really works in reality, which is a great example of what's happening in Japan.
[00:17:14] And I want to talk more about that and whether you think there'll be a bit of a resurgence from Japan now the economy is perhaps starting to kick in again, but also interested in about how Elon Musk changed his approach as well. We're going to take a quick break and we'll come back and look at all of that. Brian Hanley is our guest with me and Steve this week. Back in a second.
[00:17:33] We're going to take a quick break.
[00:18:09] The Debunking Economics Podcast with Steve Keen and Phil Dobby. So Brian Hanley is with us. We're talking about assembly processes. What works best? We've heard in the first part, it's not just machines. It's got to include humans as well. And Japan seem to have got that sorted out pretty well. America less so. Before we go any further, though, Brian, we haven't introduced you. We just sort of kicked straight into it. Tell us a bit about yourself and why you know about all this stuff.
[00:18:37] I know about this stuff because back in the 80s and part of the 90s, I did factory automation work. I was one of those bright, young, fire-breathing software guys who had some engineering background.
[00:18:57] And I was thrown into the deep end for a project at Ford Motor Company to automate their plant. And the result of that was that we came up with this system that was an information utility.
[00:19:21] And it could be programmed to deliver whatever build information somebody wanted on the assembly line from their station or, you know, from a computer terminal. And it could be hot-plugged, which was very innovative then. Hot-plugged meaning? Meaning you could, as everybody does today, you know, thinks is normal.
[00:19:51] You could plug your computer in. You could plug a terminal in, and it would just work. And you could unplug it. It would just work. It wouldn't cause any problems. And you could say, okay, this limit switch signal, you're going to deliver this information, or, you know, this happens over here.
[00:20:11] We also solved a problem that had been worked on for decades at Ford with rescheduling of vehicles in the plant. Because different plant areas had what they called body banks, which allowed you to reset, reorder the vehicles on the assembly line. But you only had, like, from 5 to 12 lanes to resort them in. It's, you know, it's a classic queuing problem.
[00:20:41] And a young man that worked for me named Larry Henry, who's a black kid from Detroit, he figured out a way to do it. And he did the Kobayashi Moru method, you know, Captain Kirk winning that. He changed the game. He said, well, let's go back, put this all the way back into the build sequence before we even start. And we work it all out beforehand. And that worked.
[00:21:11] But if you build machines that are capable of delivering product X, can it, you know, can they, I mean, if you've got that light side approach, or you've got a minimum of people working in them, can those machines be recalibrated to make product Y, which might be, you know, 50% different? Can you take stuff out? That may be. That may be possible. And, you know, again, the issue becomes, which is more easy to deal with?
[00:21:41] Is it more easy to deal with people who understand what needs to be built? or is it more easy to deal with a software department that may or may not understand what needs to be built on the assembly line and may or may not. To give you a sense for what can happen with people who are not well-versed with robots, there was an incident at the plant where I was. I didn't see it.
[00:22:09] I wasn't there at the time where somebody went into a robot enclosure that started up while he was in there. And they have lockout switches to turn everything off so that nothing, you turn off the power so that nothing can happen in there. And didn't do that, thought it was just fine. And the robot cut him in half.
[00:22:37] Because somebody up in the front office decided they were going to run a test on that robot while that guy was in there and they didn't know. Yeah, I mean, there are quite a few serious incidents during the early robotics introduction period. Well, imagine how much it worked. Well, maybe it happened less with artificial intelligence, which is a conversation we can get to. Because does a computer with artificial intelligence act more like humans and can they evolve the process themselves?
[00:23:06] But before we get to that world, what did Elon Musk do then? So he went from, you said he went down this approach of trying to do this lights out approach. He ended up backing off and he's de facto doing the human-centered approach. I can't say a lot of details about it because I wasn't in there doing the plant work. I read about it.
[00:23:30] And I remember at the time thinking, yeah, well, I could have helped him, but oh, well. But he did spend a lot of time there himself, didn't he? So he was pretty hands-on. So this realization. He's a very hands-on manager and his management style is very effective for physics problems. Yeah. He's single-handedly managing to destroy the value of his company, of course. So he's undoing all the good that he created.
[00:23:58] So how has this all changed your view, Steve, in terms of the ability for America to come back as a manufacturing nation, for example? One of the main things that I was aware of before Brian raised this issue was the issue of machine tools experts. And I think it was Tim Cook who made a comment at one stage that if you wanted to have a meeting with all the machine tools experts in America, he could hold it in his boardroom. If you wanted to do the same thing in China,
[00:24:28] he'd need a football stadium. And what that means is you don't have the skilled staff who know how to operate machine tools and make machine tools, that it's a necessary prerequisite for establishing a manufacturing system. So I'm aware of that issue. And is that the major bottleneck? Does America's main bottleneck come out of the fact that it doesn't have the skilled engineering
[00:24:54] and machine tools staff to be able to re-industrialize in the first place? And to adapt those machines to future demand. I would tend to agree with that. That's something that Steve Jobs said to Barack Obama. He said, you know, I can't get the number of process engineers that I need at any price. Part of that, it was a deliberate strategy by the Chinese government
[00:25:22] when they overproduced engineers and they overproduced engineers who understand these things. And the engineers that I've spoken to in the U.S. who know Shigeo Shingo, Zero Quality Control, Kanban, etc., just like the back of their hand, and they just assumed everybody else did, they were all Chinese. All of them.
[00:25:48] And it was kind of a surprise when I said, well, no, this isn't general knowledge. I have yet to meet an American who has read these two books. Wow. And yet you regard them as essential reading for anybody trying to establish manufacturing. Absolutely essential reading. Particularly Shigeo Shingo, to give you an idea of the kinds of things that he was able to do using his process methods.
[00:26:17] He was able to cut the amount of time it took Japanese shipyards to build a container freight ship from four months to three months and then to two months. He was able to take the time required to change a 1,000 ton press die
[00:26:42] for an auto plant from eight hours down to three minutes. Oh, my God. Okay. And there's a third book that I didn't include that's called SMED, Single Minute Exchange of Die. And that process is about iteratively changing the way you do this so that you pipeline and you have all the parts ready.
[00:27:10] And these kinds of techniques have to some degree, but probably should get more attention from military people because if you can do the equivalent to that for a military, you're going to win. So I wonder whether, though, what's happened is that companies that are making goods which change over time to consumer demand, and iPhones are a great example of that,
[00:27:38] and the competition that they face, rather than saying, well, okay, we need to have a very adaptive process for how we make these. They just deal with lots of different suppliers. And if there's a component that needs changing, they just change the company that provides it. I mean, is that the word? Yes, it is. That's harder than it sounds, isn't it, Brian? I would have thought that, yeah, that's the problem. In fact, that's one reason why Musk has been successful in re-engineering in Tesla and SpaceX.
[00:28:08] He's doing everything in-house. It's much easier to do in-house. And yet Apple, you know, the components of an iPhone apparently are coming from all over the place. So it may be complicated, but they're doing that. And I'm wondering whether they're doing that because if they think, well, we need to change a component, and it's not the right one, we just buy another one from someone else. Yes, you can to a degree, but a good friend of mine, he's an audio engineer, and he's worked all his life on audio systems.
[00:28:36] And they are constantly, of course, part of the problem that he has is that he's worked for smaller firms. And so smaller firms don't define the market, and they have to take what companies produce. And if a company decides they're going to stop producing a particular chip, well, you know, they have to then re-engineer their product around that. And that's a bunch of work. You know, it can take months.
[00:29:04] He spent a year or so re-engineering a product to solve those problems. During COVID, it became a crisis for many companies. So how have things evolved then in Japan? I wanted to briefly cover the progression from Deming, which was statistical quality control, to the next step that Japan took, which was total quality control,
[00:29:33] where they measured all over the place, and they also extended their measurement back. Toyota developed this method they called the five whys, which means if there's a defect in the product from the machinery or from the process, ask yourself, why is that? Why did that happen?
[00:30:03] And then, okay, maybe it's a drill bit wears. Okay, well, why did that drill bit wear? And are these drill bits... So they extended their quality control process. Like a fault tree. Yes, exactly. Exactly. And they extended their measurement processes back to push them back into their suppliers and into their intake processes.
[00:30:31] And that got defects way, way down to low single digits. But then they found, well, it doesn't... They're not going to zero. And that was where Shigeru Shingo's insight that quality control is not a value-added process came in.
[00:30:54] And he started a process where the andon, the lamp, was used to stop the assembly line if there was a defect. You stop the assembly line and everybody works on figuring out why did this happen and how can we prevent it from happening again? For example. A simple example is, let's say a particular worker has to install five screws.
[00:31:24] Well, if they don't always install five screws correctly, how can you do that? Okay, we're going to present those screws to them in a little bin that has five screws in it every time. And how can they make sure that happens? And that creates a feedback loop for quality control at the station to the person who's doing the job. Therefore, they can't make a mistake because it's built into the manufacturing process to have a check against that mistake.
[00:31:54] Yes, they still can, but they have to be drunk or something. They've got to be creative to make mistakes rather than just, you know. And he called that pocayoke, which means mistake-proofing. And that, again, gives you a dramatic advantage over companies that can't do that. So what you're saying, in a sense, is that a manufacturing system involving labor, where the labor is acknowledged as part of the process and their insights are listened to by management,
[00:32:23] that is more flexible than a highly automated plant, where the software engineers have to get the design right. And to redesign, you've got to rebuild machines rather than retraining individuals. Right. And that redesign, I mean, so trying to fix faults is part of it, but also, obviously, human taste evolves and what we want to consume. I wonder if you don't have that process, and I wonder whether this is a fault in America and Europe and other places,
[00:32:50] that companies that have automated just keep on pushing the same stuff, and so long as consumers keep buying it. But ultimately, they lose out on an opportunity there. So I'll give you an example. America has a particular way of making cars that's not particularly attractive in Europe. So America doesn't sell a lot of cars to Europe, because I feel like, you know, American cars haven't evolved a great deal. Americans are driving very similar cars, and Europeans have a different taste.
[00:33:20] So if there was a way to evolve the manufacturing process so that you could easily provide cars that are going to appeal to European taste, you might ship more cars, and you might, you know, go a little way towards that balance of trade problem. Yeah, although, as Steve knows, the balance of trade problem also has to do with the fact that, the rather intractable fact that the U.S. is the reserve currency. I know. There's a very minor way to solving that. But anyway, maybe that distracts my point.
[00:33:48] But I mean, my point was, if you've got an automated process, and it's fairly rigid, you are just going to keep on. Well, it's almost like sweating the asset, isn't it? We're just going to keep on making this and hope people buy it, because we can't adapt. But that means that you're open to international competition, particularly as you're saying from Japan. Well, yeah, yeah. But American auto plants do adapt. I mean, they're adapting. They're just adapting to a different kind of customer.
[00:34:16] American roads are bigger, wider, longer, straighter. European roads are narrower, curved, shorter. You know, things are closer together. You need to be more nimble. You don't need the steering wheel. The roads are all straight. You can, yeah, you can actually take your hands off. I've done this before and gone, you know, to see how far I could go and gone, you know, a mile or so and still been in the lane.
[00:34:45] Hope your insurance company isn't listening to this. Good point to take a quick break. Back in just a second. We're going to go.
[00:35:28] Yes, today we are talking about manufacturing processes. And are we saying then that the US, you know, Donald Trump's aspirations to bring back manufacturing? Well, I mean, first of all, it's a curious aim anyway, isn't it really? Because you've almost got full employment. So I'm not quite sure. And he's getting rid of as many people as possible. So it would have to come from automation. But you're saying, well, that's a problem.
[00:35:55] You can't just automate because you will need people. So America's got two problems with Donald Trump's strategy. One is, are there going to be enough people? And secondly, are those people going to be able to work in this hybrid environment with robots so that you've got and probably not because you've because of, you know, how we started out because of the labor structure and the union structure in Japan, which is company based rather than industry based.
[00:36:25] Yeah, that's a big factor. I'm thinking, you know, I was down at SpaceX for a couple of months or not down at SpaceX, but I was down near there. And they have these, they popped up these training centers for manufacturing for people who want to go work for SpaceX. And so the education system will adapt if there's demand.
[00:36:52] I think the education system is going to be a bottleneck for the Americans as well. Yeah. Yeah, it's going to be, it's, it's, it's going to be something of a bottleneck. Yeah. So where, so where does America go in all of this thing? Is it, will it see more manufacturing coming, coming back? That's a really big question.
[00:37:11] And I think there will be some degree of it, but again, you know, the, the, the fundamental structural problem of the, of being the reserve currency and having your major export be dollars is really hard to fight. This is what, I was actually talking about this on a different podcast today. And the point is that Trump is trying to hang on to a, what he calls a strong dollar and have a strong manufacturing sector. And the thing is they're, they're contradictory.
[00:37:41] The strong dollar means your manufacturing has a cost disadvantage. And there'd be the best form of a strong dollar would be a dollar that's, you know, constrained to America, not used as an international currency. But that's the last thing he's going to agree to. So you've got these conflicting objectives coming out of a rather conflicted White House. Well, he'd say his tariffs will take care of that, Steve. I mean, tariffs, tariffs are obviously going to fix absolutely everything.
[00:38:08] But I mean, it's so, but this, this hybrid way of working seems to be the message that America hasn't gotten onto yet. But if he, I mean, if he did, I mean, is there hope? I mean, I'm looking at Oliver Wyman, which is a management consultancy firm. These are going back a while, 2004 to 2014.
[00:38:28] The annual growth added, compound annual growth added by the manufacturing segment in the United States over that decade was 3% compared to 2.8% in Germany, 2.1% in the UK, minus 0.9% for Japan. But they weren't great years for Japan, obviously. So manufacturing was contributing a great deal to the U.S. economy back then, actually more than the growth in the economy as a whole. So, I mean, that wasn't so long ago.
[00:38:57] What was going right there that hasn't been going right since? That's a good, that's a good question, Conorblade. The one thing that does come to mind is probably, probably energy, energy costs. But America has a big advantage on that, obviously, as well, which is, you know, a factor we've talked about in, you know, the fact that energy is half the price basically in the U.S. than it is in Europe. How can Europe compete with the United States?
[00:39:22] It's the only way it can be, it can be more innovative, but that seems less likely to happen in Europe than anywhere else in the world. And energy costs is really hard to fight, too. So we've got an advantage in energy costs. So finally, then, this idea that, you know, which obviously in the tech sector is rife, this idea that we're getting components from all over the place, that your iPhone, I don't know how many different components, how many different companies are involved in that,
[00:39:49] seems to be what you're saying is, well, that's not an efficient way to work. The efficient way to work is like the Tesla approach, to try and do as much of it in a house as possible, because then you've got, you don't have those supply chain issues that we obviously saw during COVID, and you've got the flexibility to adapt as you need to, and you're not going to be left high and dry. As long as your supply chain holds up, then that strategy works great.
[00:40:15] And it works for small companies that are producing a new product, and they can't afford to, you know, to do any vertically integrated manufacturing system. That works for them. But then they have to, in turn, be able, you know, like my friend who works in the audio industry,
[00:40:40] that has to be able to be re-engineered on a regular basis in order to keep up with what the supply chain is going to give you today or tomorrow or the next week. And I guess the other thing as well, is if you've got a bunch of workers and management and growbots all working together to try and refine a product, I mean, that's not actually defending you against something else that comes along that could potentially obliterate you, you know,
[00:41:10] because all you're doing is thinking within your small sphere about how you improve what you do and how you're responding to demand. It's not looking at, you know, the true innovative product that could change consumer behavior. Yeah, that's true. And the other side is artificial intelligence, just, you know, as a point to leave us on. Could those robots start to say, well, okay, we can adapt.
[00:41:33] So we don't need humans because, you know, because we'll do the adaptive thinking that they were doing. Yeah, what we have in artificial intelligence today is primarily LLMs, large language models. And those are, those have hit the asymptotic limits of what they can do. I would suggest looking up Gary Marcus and what he's had to say about it.
[00:42:03] I worked with earlier AI. There's been, there's, this could be a discussion all of its own. Well, maybe we leave it for that then. Do you know what I have to say? I love it when we have a guest on because Steve is unusually quiet. He's, he's, but. I had this terrible handle in the guest, let's talk. Isn't that a crazy idea? But Lee, it's been great having you on.
[00:42:31] Appreciate spending this time with us. And look, let's talk again sometime soon. Yeah, maybe we'll, we'll talk about artificial intelligence. But Brian, good to have you. Thank you. Nice to be here. Interesting stuff, wasn't it? And that is the Debunking Economics podcast for this week. Join me and Steve back again next week. Thanks for listening. The Debunking Economics podcast. Disney's Hercules.
[00:43:00] Nur noch bis August in Hamburg. Hercules! Jetzt bis zu 40% mit unseren Abschiedstickets sparen. Nur für kurze Zeit unter musicals.de. If you've enjoyed listening to Debunking Economics, even if you haven't, you might also enjoy the Y Curve. Each week, Roger Hearing and I talk to a guest about a topic that is very much in the news that week. It's lively. It's fun. It's informative. What more could you want?
[00:43:29] So search the Y Curve in your favourite podcast app or go to ycurve.com to listen. The Debunking Economics podcast.