Rob, Hello, I'm Rob Hirschfeld, CEO and co founder of RackN and your host for the cloud 2030 podcast. In this episode, we continue to discuss whether or not AI export controls work, but we take a really interesting twist here, because what we talk about is manufacturing. What we talk about is innovation, and it's not whether or not you can control AI chips, but what does it actually take to build innovative product? That's where we really have challenges on export and controls. There's military manufacturing, goods, things like that. That's part of what this, AI embargo, or chip embargo, is about. And yet, in our discussion, we really talk about how challenging it is to actually build truly innovative manufacturing and what the barriers are. This goes deep, and I know you will enjoy it.
You want the topic for the day? Sure this is, this is your topic? Actually, yeah, it's based on our expert controls. Conversation is spurring innovation for manufacturing it and power, based on our thought about export controls. So I will remind you we were talking about, does control it? So I went to a boy this is a unwinding a bit. I went to a open for debate session talking about us controlling chips exported to China, which is very much in the news at the moment, on rare earths and things like that. And so the question becomes, that we were getting to and picking up here is so can you, you know, do the limits work, and what does innovation actually look like? So if you want to improve innovation in a country, what, what should you do? I think that's where we were going, although take it wherever we need. Was it the was it the mother of invention idea that you know, when you're under constraints, you find other approaches to it. And that's, you know, it kind of spurs, taking new approaches to it, which was kind of the story that, you know, deep seek, you know, got from, from the I think there's an element of that, but I also think there's an element, and I think this is where Joanne was going. It's like, I you don't need the chips to crush everybody, right? The idea of advanced manufacturing with with advanced robots and, you know, running, you know that,
excuse me, guys. Let me just find out if this is important, you're not. It's a call, of course, that we've been waiting for. I'll
put you on mute in the background. So what was, yeah, I may have missed the session, or I'm forgetting it. What was the, what was the point that Joanne was making that software was efficient or no, that you don't need, right? There's, there's the chips are about end up being limiting your your ability to train models. But you don't need to train models to create new robotic process, manufacturing or right? A lot of, a lot of the the downstream innovations aren't limited by your chip capabilities, right? You're, you know, you're training a model, yay. Deep seek. Help train a model faster, refine a model faster. But you know, a lot of where the China innovation is, is, is actually in, you know, better robots, reducing costs for drones, you know, improve drone navigation, right? All. So all of this stuff that's actually much more important in from a from a global commerce, a global competitiveness perspective. I mean, we're suddenly learning just how vulnerable the, you know, the rest of the world is to China saying, yeah, oh, you can't, you can't have these rare earth metals that are necessary in magnets to do high performance magnets. Oops, that seems like a major myth. Yeah. I mean the Yeah. And I, you know, I totally get, you know, why? Because, for the last 20 years, China has been saying, hey, we'll pollute and destroy the environment to do the refining and and meant, you know, manufacture. Of these, these materials. And, you know, I've been saying for 20 years that we've out. We're out. We're not just outsourcing labor, we're outsourcing pollution to third world, you know, to not no longer third world, but to other, other countries, you know, and I that's, fortunately, that's not quite where the conversation comes from or has goes with the there's a pretty significant benefit of us having done that, I'm sorry, of us having done, oh, having, having, having, having outsourced our 30 production to other countries, right, low wage production, right? We've, you know, there's right, you know, it turning around and saying, you know, hey, we're going to manufacture what is mythium, or whatever the you know, some of these, these metals, or even lithium in the US means that we're going to have to or deal with our trash or, you know, recycled plastic or whatever, and we've been shipping that outside of the borders for a long time. Oh, yeah. And, you know, there, we haven't put a appropriate value on our ability to do this, that type of of outs, you know, outsource, yeah, and, you know, we, you know, what that comes down to is that, you know, it, we're, we're putting the cost, the long term costs, on somebody else, often around either because we, we kick, we, if it's all domestic, we kick the can down the down the street and continue to do that, or, to your point, send it to some place where, hey, you know, we'll take it on, you know, as a bargain, and we'll we, we don't have the same kind of controls and and rules and regulations? Sure, yeah, yeah, the quick was the, is the question, or was the kind of the pillar of this innovation that its impact on innovation? So it's, how do we, you know, how do we actually have a reasonable control. What is it? What would it take for one, you know, what is it going to take for China to circumvent, I think we're answering this to circumvent the controls. But what would it take for us to revert it? What would it look like? I mean, here's, you know, so, so I'm like, oh, I want to set up an advanced manufacturing facility using three printers and robots, right? Okay, yeah, hey, I, you know, I you know, and I think this is a reasonable thing. I'm like, okay, look, downtown Baltimore has a ton of empty buildings, power infrastructure and people, you know, I could build a, you know, agile manufacturing facility, you know, on, on, you know, yeah, yeah, great, makes sense. I'm gonna have to go procure printers, right? Gonna have to procure robots. I'm gonna have to configure them. I'm gonna have to train people to then do the work, to supervise the machines. You know, if I did it, I would be building a high tech thing. It would be right, all robots, and, you know, robots tending right. The question is, what's the lead time to get all of that? What's the what's the real cost? And then going forward, what are your What are your recurring costs? And at the end of the day, can you make that a business that has a good market? You know is, are you? You know, are you? Have you priced yourself by doing it that way? Have you priced yourself out of the market, for example? Well, and this is an interesting thing, because if I was going to build the machines, I'm assuming the machines aren't tariffs aside. Tariffs aside aren't any more expensive for me to acquire. We're talking about innovative, the non, the downstream innovation from AIS, right, improved printing. AI, arm, you know, AI, integrated, like, like, modern manufacturing is not people and pierce been doing a really nice series about this, talking about, like, you know, they're talking about, just heard a piece. They're talking about building trusses, which used to be people with saws and hammers, and they've it's becoming increasingly automated. So, you know, when you build a trust now it's assembled much more by robots. Well, prefab, but even the assembly is now, you know, to spend, yeah, yeah. Um. Which stuns me, because it's, it's like, that's a, you know, multi million dollar and here's, here's where I'm getting back to, that's a multi million dollar machine. And I think Asian, the Asian, Chinese factories that they're building now are not, hey, it's tons of workers coming in from the countryside anymore. They're now building them with robots. And they're much more flexible, they're much faster, they can handle much smaller piece parts, you know, and and China has an advantage in they have the expertise, and they've been building the it's easy to permit, and they've been building the facilities much more aggressively, the way, the way it seems. There's a skill, there's a skill, there's a skill set that there's a population with skills at all the levels that you can probably, that you can draw on, as opposed to, you know, starting day one with training,
and there's and there's so there's, there's the Chinese, but then there's also the Europeans, because the Europeans are very adept at using robotics for building materials like trusses and pre prefab that are completely made on production lines that are lights out where the robot arms do the, you know, the laser measuring and the position and the, you know, the arms come down, and they do the cutting or or whatever, or the finishing parts. So I know of, I think there's three or four companies in Germany, and there's probably two or three in the Netherlands that specialize in this kind of stuff. Excuse me, and what they've done is they've added visual inspection capabilities to CNCS and other kinds of equipment that would used in the manufacturer, whatever those are, along with the robotic arms and automated conveyors and everything else, here's what they didn't add, though. They added the AI for all the visual inspection and everything else. But those machines cannot call and order their own spare parts. They can't call home. They can't they can't tell you when you have a other than the sense some sensors that may have may be available on those machines to do root cause on their own. They can't say, I'm broken in x part. They can just say, I'm stopping.
This is today's technology that that that's in place
with today's technology. Yeah. So like, you know, I think you've heard me tell the story before. I got a call from a manufacturer that makes machines that weigh and sort fruit. They happen to be the largest one in the world, and they sell to all the distribution channels that go for you know, when you go to the grocery store and you buy strawberries in a box or whatever, that's not how they get them from the grower. They get them in huge quantities, and then they have machines that you put the fruit in. It sorts it, it weighs it, and it packages it all at the same time. Well, this particular one, they had built all the artificial intelligence for visual inspection and sorting, but they couldn't. They didn't do anything using AI to impact mechanicals or report better data, or call home for spares, or anything like that. The
infrastructure around the operation is still, is still relatively okay,
yeah, very fragmented. And so bringing things together, and it's it's also the integration of automation with AI gets you where you want to go. And in their case, they had 18 factories around the world, all making these giant monkeys machines to sort berries and fruit and, you know, any kind of produce, type of thing in small quantities and nuts. And what they were finding was that they needed more more capability around AI that was not visually oriented, because there's only, you know, there's a finite level of I don't want to put com, I don't want to say compliance, but quality is determined in certain ways on certain types of fruit. So you can take 1000 pictures of strawberries and. Say this one's good and this one's not good, and you can also do regular visual inspection, but they wanted to have more capability built in that AI would assess, well, did the produce I get from this grower that I'm now sorting result in a lower yield of actual product that I can sell, and more waste or less waste, etc. Those were the kinds of things that they wanted to automate, but when they built their equipment with all the visual inspection tools, they never brought that part into consideration. And there's an impact between the visual inspection stuff and other forms of AI. Do they interfere? If you're only generating images, do you need to add a different level of data set to it? How do you do that? How do you go back and reinvent the wheel for a machine that might be the size of a room, right? That's geared to opticals.
Is that, I mean, but is that something that's going to come as we get this? And this, I think, is part of what I'm thinking about, like, if you're building a next generation factory, you're building it with, you know, as you I'm assuming more modern, more modular components, more they're lighter weight, or potentially custom managed, like, no, no, no, not there yet.
Well, it's not so much we're not there yet. It's that certain things just can't be built that way. I'm not saying, I'm not saying you can make everything on 3d printers, but there are certain like presses, right? They use metals. And you know, the it's the size of the plane, it's the size of the raw material that's going to go through it. I mean, if you've ever watched any video on how they actually manufacture cans, and I'm not trying to be political here. I trying to be political here, the roles of the tin. The tin is so fine that if it's not put in properly, if it goes off even a millimeter in either direction, you have a huge amount of waste and a huge amount of cost you cannot it's not like you can crunch like aluminum foil and sort of recycle it to reuse it with that kind of stuff. It literally fractures, and so it doesn't make for a good upcycling kind of capability. But that's one area with presses with certain the weight of the engine that drives the machine is too heavy for lightweight fabrication of the machine, or the printing of the machine like there's tolerances, there's limits. There's all that
kind of, no, you can't, you can't do it's one of the reasons why you can't just use a inexpensive robot arm to do a lot of the this work, right? You know, simple like, Hey, I'm picking and placing, or I'm moving something from one to another, but you actually have to have, you know, tools. And that's just like one of the things that the US lost. And I have a, I have a friend who's, you know, who does steel manufacturing, and he's like, I, you know, where I can it's so fascinating, because he's like, Yeah, I need a tool to punch the metal into the into shape, and bend the metal into shape and things like that. That's a tool. And inside of that. So you have a machine, and then inside the machine, you actually have the dye that it's going to do that punch or shape and, you know, and those are wearing parts, and you have to have somebody who can maintain them and clean them and refine them and all that. And those are, those are skills that, you know, he doesn't have to do it enough to maintain that person. It's a highly, you know, it's a specialized skill, so he needs to outsource the dye somewhere. And they're, you know, huge things. So he's either buying extra dyes so he can ship them out of the country to have them maintained or serviced, or ship them somewhere else. And now you've got a supply chain where free travel across the border or ability to ship is essential to him being able to manufacture wherever he's going to manufacture, which he'd rather do next to the diamond. The die people for this reason, because sometimes there's an emergency, right?
Well, it's like, you know, and I know that I've said this before, the whole supply chain during COVID went from just in time to just in case. And we're still, we're now back to just in case, like I just was creating a module, software and agent, or the the beginnings of the design for it, to see whether or not you could manage the fluctuations in. Tariffs or the fluctuations in the global supply chain better by bulk buying inventory at certain times based on things like commodity markets, foreign exchange, shipping costs, landing costs, and just the Well, today, the tariff is 25% tomorrow, it's going to be 50 and anticipating that. And could you better manage your supply chain, therefore your production and all of your cash flow, by doing it that way? So I started trying to tap certain kinds of data sources that are public domain, you know, in the spot markets, commodity markets, forex market, and put all that together to see if I could actually make it work. Because there's a lot of companies who keep asking me for it like, you know, is there a way to do this? And I know that their procurement people know what their either their three PLS, or other trading partners. Training partners are doing to try and help them. But if it makes sense to bulk buy and store, as long as there's no issue of useful life with that good why wouldn't you do? Yeah,
there's, there's, there's, you know, shelf rot. But of course, were you I'm trying to understand, were you actually looking at commodities markets and models that people use to make their edging decisions and their and their their purchasing decisions and in advance? Were you trying to take that kind of thinking and apply it to this particular problem, or were you actually pulling some of that into the model, for example, foreign exchange and predictions regarding, you know, the Change in exchange rates, things like that, was yes, the latter, as opposed to the former, or both, both.
I tried one way. Then I tried just pulling data out of freely available, you know, sources that I could find where it would allow me to download a bunch of stuff, not in real time, because by the time I downloaded it was out of date. But I just wanted to see how the data was to see and to bring in and model kind of do the math for if the commodity is down, or I see a forecast where I can envision that the commodity is going to drop in price at the same time as exchange rates are changing. At the same time as you know shipping costs are dropping based on certain routes and whatever the commodity is I'm buying. Because if, if it's think about copper right. The price of copper goes up and down on a commodity market, and wire has been fluctuating dramatically, not only because of China, but other countries as well, and the tariffs. So if I can find a way that the model can balance, when is the best time for me to buy copper wire that I'm then going to use in the manufacturer of something else, like connect, you know, ether internet, or any other kind of cable, or whatever electrical cable, then I know that I should buy more copper and hold it somewhere, not pay The tariff on it, until I actually move it across my floor in my warehouse from, you know, one side to the other side, because of the way that's designated, in an opportunity zone where I can reduce the cost, etc, etc. And it became like this, you know, tremendous algorithm. And I'm like, okay, so I threw it to, I actually threw it into perplexity first to see what it was going to tell me. And then I dropped it into oh three. And like I said, this was, you know, this is my first tape at this of trying to figure out what the what the variables are that I should actually consider in this to get to the point of being able to forecast. So I then did it with some plastic resins. Because, you know, the toy manufacturers in the US are being killed, right? Because there's 80 what is it? 85 or 90% of all the toys that are produced are produced in China, right? Nobody can afford to bring them in, even the previous tariff, let
alone the new low. 30% there the
new low. Right, right, right. So, you know, I had heard a piece from the, you know, the guys at Mattel and toyco and somebody else who makes educational toys, and they said, you're gonna have shortages on, you know, long before Christmas, because we can't bring this stuff in. We can't afford the tariffs without raising our prices, and parents are already cash strapped, and everybody's feeling the pinch. So they're fascinating.
So retailers would rather have empty shelves than raise the prices. Is what you're saying.
The market won't bear the price increases. That's the other part of the equation that I'm trying to figure out.
Yeah, I mean, they're getting hit already, independent of their own costs. If they have to do anything that UPS their prices in even inventorying stuff early on, it's very, it's very, this was, this was the lawsuit with the wine importer, who's sort of the point of the spear on the US tariffs. He's like, look, here's my, my problem isn't, you know, yay. Terrorists are not tariffs, he said. But my, my, my problem is, is that I have to pay the tariff when the when the wine hits the sure country right, and I have to pay it, not when it's sold, right? So he's paying tariff on products that may or may not move and right, or you just don't know when it's going to move right, when it's going to move. So he's got to float the tariff until he sells the product. And he doesn't, you know, there's no indication that he can pass the price of that tariff on in the in the pricing. So it's, yeah, it's a really, from a business perspective, it's a very unsettling calculation to pay, pay money out on product that you haven't sold, although, I guess, well, yeah, go ahead, straight
in manufacturing. Think large manufacturing where you have massive quantities of inventory, whether you're depleting your you know, you've significantly depleted the inventory or not. Put the tariff issue aside for a moment. That inventory cost now, the cost of your scrap, the cost of your waste, the cost of your emissions, and your cost of goods in manufacturing overall has just jumped, and you're working with very, very slim margins, because the latest inventory is going to be that much more expensive than what you had. How are you going to reconcile that with contracts like think about the big tech companies like Dell and whatever that outsource to EMS the like Sanmina Jabil, flex, all those guys, right? All of those companies that actually build those boxes, they're on fixed programs, and they compete based on fractions of cents, right? Because it's high volume, right? And now suddenly, you right? I mean, they all always, most of them set up third party supply chain arms that were arm's length from the organization, but they would go and buy, like, a year's supply of dims or sims or whatever, like jelly bean parts that are used in the manufacturer. When those run out, that whole program is at risk. And we're talking billions of dollars. We're not talking millions of dollars on those programs, plus the lead times, no can deliver, no can get goods into the country if you can't even import the raw materials, forgetting about the tariffs, because you can afford that extra hike. But you can't get the raw materials. You can't deliver on time. Time is contractually obligated for your payment, as is delivering a se finished good to a three PL, a third party logistics provider for a merging transit model. So my sub assembly and your assembly and Rich's assembly all get put together by the third party logistics provider and then shipped to the customer. You miss one of those deadlines, you are you're dead. Yeah, you're really Sol. So it's affecting everybody and and how
would you even but I mean, even with all of that, even if you had the data, the the there's so much unpredictability, yes, and you know what's going to what is going to happen with the tariffs, you know, are they going to change up, down, sideways, and when that we've gotten. History to speak of in that. So we're living in, you know, pure chaos on that, on that side. So what you have to do then is to start to play, you know, kind of Bayesian statistics and actually doing Monte Carlo simulations to try to figure out, all right, assuming, making whatever assumptions you can make, this is when there's going to be a change, and the direction of the change, the magnitude of the change, and then trying to model what that, in turn does not just to our economies. Here, but foreign exchange and, yeah, I mean, the the the interactions there, the buildings, of that, of that model, just a, it's a it's a bear,
it is a bear. But I had to try. I'm not suggesting that I'm going to figure it all out. But that's exactly what I was doing, rich running Monte Carlo simulations. And the only, the only thing that you didn't mention is that there is a lag, sometimes, sometimes between the announcement of a change of a tariff and the actual time it gets changed. And little bit of that, little bit of give, give and get, helps you with
it. So, yes, what it well it, what it does is it says, I can't give you a long range model, but what I can do is say fairly quickly, as soon as something is announced, what are what you know, what is my best tactical approach, which is, go to the bank, get as much money as you can buy this good inventory area, you Know, warehouse it wherever, because you need to do that, or, you know, some other variation on the strategy, yeah, yes,
and that's exactly what I was playing with. And I'll give you a little insight that I learned if you if you get the news of something happening the first five minutes thereafter, you see more fluctuation in foreign exchange than you do at any other time over the next 24 hours. So if you can lock in your price on the foreign exchange then and keep that as constant across the Monte Carlo model, you're good, you're golden. If you wait more than five minutes, maybe up to 10, you have a little bit more variation, but it's very short time windows, just like you know, the spot price of a commodity itself on Chicago markets or worldwide. It's a free, very short time window that you can play with. So I started to be able to put a little bit of a guardrail to see where I could actually capitalize on it. And I used copper, Canadian dollar, US dollar, and sources where copper is predominant, like in South America, etc, where it's coming from. The tariffs the time from mining to assay to finish out of the mine, that was one set of variables, transportation to a port, port onto a ship from a ship. And so if I tell you that the equation alone as a Monte Carlo was probably three quarters of a page of a Word document, and then I dumped it into an LLM and said, Help me out here, guys. And I literally it broke. It just said, I can't, oh sure.
What were you using to if you don't mind my asking, what were you using to model that? I mean, were you actually trying to build a working model where you could change, you know, you had the dials and leverage to change it, yeah, were you putting it into a, like, something like a, you know, a sophisticated spreadsheet kind of approach. What? What were you doing to make that work? Yeah, okay.
And then I put it into a spreadsheet. And then I was looking for a better tool. So I started playing around with some of the AI tools for math and leverage, trying to leverage that. And then I said, Okay, if I was going to use like, you know, Claude or perplexity, or you. Like something more sophisticated than just GPT. Not that it's not sophisticated. What kind of tool recommendation Could I get? And I found something that I haven't actually opened yet, which seems to be open source, and it's it's for complex math. And I'm like, okay, but I was primarily doing it with a piece of paper and a, you
know, used to, they used to be more of them around, but, and they still exist. But there are some pretty interesting and sophisticated systems dynamics models where you you're basically, you know, putting, you know, be then flows and gating, gates on them and and so forth. And in it, you can use them to introduce cues and delays and stuff like that, as well as where, you know, warehouses. I mean, they the forester MIT was pretty famous for having a big system dynamics modeling program. And there's some that are still out there, that there's some, there are a couple that I remember seeing not too long ago that are still, you know, desktop models, which, given the power of our desktops today, they do, they can do some pretty, pretty good stuff. And they're, they tend to be also visual. To get, yeah, they also seem, they're, they're also visual. So you can play around with kind of putting, kind of the major, major actions and and so forth. In them, I'd almost be willing to bet you that those are going to be more informative than whatever you're going to get out of you know, well, immediately, unless you can feed that into the LLM and have them, have the LLM, kind of lay it all out there. Yeah. Well,
what, what I was trying to use the LLM for was, if it could give me a historical trend that I could run it against pre tariff, that that would be a good way to try and do a comparative between the two. Now, if you ask me, why did I do this?
Why did you do this? Why did you do this?
Because, well, two things, one, we're sitting on pins and needles, waiting to see if a deal is going to close or not. And two, because I thought I need a cash cow, and I need it quick. And everybody keeps asking me about, you know, tariffs is a hot topic, especially on this side of the border. And I was like, Yes, I can figure this out. And the more I thought about it, the more I went, you know, commodities traders would know this, this sh 80. Forex traders would know this. There's got to be a way of pulling feeds together to do it quickly and to get a baseline out of the llms to start with. And then I could actually have a tool, and I could go look, instead of releasing the first you know, the platform and the software as it is, here's your here's your teaser model, power of management, when to buy your next set of inventory goodness, because People really want to know that stuff. So that's, that's what set me down the be practical, figure it out. The
thing that's stunning to me is, you know, you know, we started this conversation on, what does it take to actually build, sort of this innovative manufacturing supply chain pieces. And the answer keeps coming back to it's, it's so incredibly complex. I'm not sure anybody would want to I mean, people want to buy goods, but the margins and the complexity of what we're talking about is is really, really high. And it's one of those funny what you're describing to me, there's an element of, I've already got a very complex problem, right? Having the menu, you know, in China, the manufacturers are stacked up next to each other, which is why they cluster manufacturers like this, because the services and the supplies and the inventory are complex enough that locality does help, and then you just drop a bomb in the middle of it, and you know, all of a sudden people are like, I want more localized supply chains,
right? But it's not just that they want the localized supply chains. It's that they no longer. Topic. It's that they're trying to wrap their heads around the fact that they're, they no longer exist in that locality,
right? No, it's this, is this is the thing that's silly is that it's not a matter of, I mean, I've heard this over and over again. It's not a matter of wishing manufacturing back, right? You have to. And I think this is this is this, to me, is where my interesting question comes in. And you stepped out for a second, but we were like, all right, Rob wants to build printing, robotic based manufacturing in his hometown. Great. But the effort to do that, to actually procure the raw materials and specialists and material, all right, there's, there's so much ancillary stuff, besides just, you know, willing a, you know, new manufacturer into existence that that it's not as simple as, I mean, it's the start. It would be a starting point, but you would have to, you know, have all sorts of things that are, you know, ancillary to it, say, I'm reading the
Yeah, the message I know, yeah, you know, it's Think of it this way. And we have this in one of our slide decks. A typical plant floor has nine different translators and 10 different pieces of messaging software to communicate the data off the variety of equipment. That's one line in one factory, yeah. Multiply that by, you know, where I used to be, 3 million square feet of manufacturing space in one location. Then multiply that times 28 that's a multinational and each one of those facilities, each one of the lines is different. Each one of the plants is different, and you have just to communicate. You know, my temperature is this, my vibration is that, my whatever, whatever. And they're all in different file formats. They all have to go through translation. Or you're you're doing real time feeds. They still have to be normalized. They still have to be to to an extent, tagged just for that level. Then you have control, you know, SPCs control SCADA systems, control systems. You have MES systems on the manufacturing institution that don't agree with the ERP or anything else there's
but wait, hold on a second, when we're almost, yeah, sorry, no. The thing that's weird, though, is it feels like those are popping up out of, you know, out of nowhere in China,
because, because there is so much more limitation on the choice of equipment that they have. It's all made there. So there's a certain there's a greater level of standardization, because if it's uniform to one country, then you're not going to argue with the the powers that be in the government that say you have to standardize on X protocol or y protocol, right? This is what we did, also,
government support, government support and bank, bank support, loan support, heavily, heavily subsidized. So I go down the street, I get a loan. The government says, I buy the Chinese robot with the Chinese assembly line, with the Chinese control system, and then, you know, and all of that stuff lines up. Sort of fits together, because, in part, because they're doing it over and over and over again. And wow, all right,
the fastest. They don't have 100 years of legacy to deal with,
right? They'll just build, they can build a they can build a new plant or new concrete, put it up.
Very fast, very
fast, right? And, and there's, is actually less of a there's a less of, there's there's a reduced issue of the problem of, I'm going to say kind of legacy or failure in these cases, these things go up, go into, into operation, if they're successful, great. If they're not, they're very quick to cut their losses. And they, you know, the powers that be, in particular, the government basically takes, takes on a lot of the, you know, the liability of some of these. Places. So if you're in, if you're in good with, you know, a source of finance that's, you know, connected, yeah, that's, it's fascinating, yeah.
But the last point is, there's a trade off on the rapidity and agility for which they move, and that is quality right product quality is still well below par in a number of ways and in a number of situations. For their domestic use, it's perfectly acceptable by Western standard. It may not be, and so that's another issue, but to stand up a factory today or or wait to to have manufacturing come back into the US minimum five years before you're going to see a real result, which is why I think I'm in the camp, like a lot of other people that go, why the hell are you doing this?
If you want to do it, pick, pick the thing that you want to do it in, and enable that, yeah, like chip manufacturing or solar manufacturing, or something like, or you could, you could actually, you could actually, with conscious effort, create a real opportunity in, in a, in a forward looking narrow car, electric cars, all right,
as always, a fantastic discussion about what it really takes to innovate, what the data AI implications are, what the manufacturing and supply chains are. We really put a global lens on innovation, and I love these discussions for that. I hope you're enjoying it too. If you do, let us know, I'd love to hear what you think about the podcast. And you are always invited to join in our round tables and look at our schedule at the 2030 dot cloud. See you there. Thank you for listening to the cloud 2030 podcast. It is sponsored by RackN, where we are really working to build a community of people who are using and thinking about infrastructure differently, because that's what RackN does. We write software that helps put operators back in control of distributed infrastructure, really thinking about how things should be run and building software that makes that possible. If this is interesting to you, please try out the software. We would love to get your opinion and hear how you think this could transform infrastructure more broadly, or just keep enjoying the podcast and coming to the discussions and laying out your thoughts and how you see the future unfolding. It's all part of building a better infrastructure, operations community. Thank you. Applause.
