Category Archives: Technology

394. When Complexity Becomes Fragility

Posted on by
Reply

Two apparently unrelated events tell us something important about the decline of the industrial era.

A technical failure in Britain’s air traffic control system caused delays and cancellations across the country. At the same time, Jaguar Land Rover announced plans to remove about 4,000 jobs over two years as it seeks £1.7 billion in savings. Its difficulties include competition from Chinese manufacturers, tariffs, high costs and the immense expense of developing new electric vehicles.

Neither event means that aviation or motor manufacturing is about to disappear. Both, however, illustrate the vulnerability of an industrial economy that has become extraordinarily complex.

A small failure with national consequences

Britain’s airspace did not run out of aircraft, pilots, fuel, airports or air traffic controllers. The disruption arose because of a failure in the flight-processing system operated by NATS.

That single system sits at the centre of an enormous network. It must receive flight plans, identify aircraft, calculate routes, avoid conflicts and pass reliable information to controllers. It connects airports, airlines, pilots, computers, communications networks and control centres.

When it falters, the consequences spread almost immediately. An aircraft that cannot leave Heathrow may be needed for a later flight from Edinburgh. Its crew may exceed permitted working hours. Passengers miss connections. Baggage is separated from its owners. Hotels fill. airline booking systems become overloaded. Disruption spreads to airports in other countries.

NATS said that the fault on 8 September was in its flight-processing system and that, although a fix had been implemented, recovery would take time. NATS statement

This is the nature of a highly interconnected system. The immediate fault may be small, but its consequences are not.

The efficiency trap

Industrial civilisation has spent two centuries increasing productivity by increasing specialisation.

An airline does not make its own aircraft, engines, computers or aviation fuel. A motor manufacturer does not produce all its own steel, glass, electronics, batteries and software. Each depends upon thousands of specialists, scattered across many countries.

This arrangement can be wonderfully efficient. Components are produced wherever skills, labour and capital can be most profitably combined. Stocks are kept low. Machinery is used intensively. Computer systems coordinate everything.

But efficiency and resilience are not the same thing.

A resilient system contains spare capacity, alternative suppliers, duplicate equipment and stocks that may never be used. These appear inefficient when accountants examine them. The pressure to reduce costs therefore removes much of the redundancy that could protect the system when something goes wrong.

The result is a tightly coupled economy. Almost everything depends upon everything else.

Jaguar Land Rover and the complexity of the modern car

The same problem can be seen at Jaguar Land Rover. The company is reportedly planning about 4,000 voluntary redundancies, nearly a tenth of its worldwide workforce, as it attempts to reduce costs. Reuters report

A modern Land Rover is no longer simply a mechanical vehicle. It is a computer-controlled system incorporating semiconductors, sensors, software, cameras, emissions equipment, communications technology and electronic safety systems. An electric version requires batteries, critical minerals, power electronics, charging equipment and an adequate electricity network.

The factory itself is only the visible centre of a much larger organism. Around it are toolmakers, component manufacturers, software companies, energy suppliers, transport businesses and specialist engineers. If one essential component is unavailable, a production line costing millions of pounds may have to stop.

JLR’s experience of a damaging cyberattack in 2025 demonstrated another aspect of this dependence. Digital systems make production quicker and more controllable, but they also create new ways in which an entire business can be interrupted. The more activities that are joined through one computer system, the more extensive the consequences when it fails.

Its present difficulties are not caused by complexity alone. Chinese competition, American tariffs, changing consumer demand and the costs of electrification all matter. Yet these pressures themselves reveal the scale of modern industrial dependence. A British factory can be placed in difficulty by decisions made in Washington, subsidies paid in Beijing, minerals mined in distant countries and software failures that may originate almost anywhere.

Complexity requires a large surplus

Complexity does not maintain itself for nothing.

Air traffic control requires secure buildings, computers, communications equipment, software, electricity, trained engineers and constant updating. Motor manufacturing requires mines, refineries, chemical works, ports, roads, electricity networks, research laboratories and global finance.

All of this depends upon the economy producing a substantial surplus beyond the immediate requirements of food, shelter, heat and basic care.

During the expanding industrial era, abundant fossil energy made that surplus possible. Coal, oil and gas enabled machines to perform work that would otherwise have required enormous amounts of human and animal labour. Increasing energy use supported longer supply chains, greater specialisation and more layers of administration.

In a shrinking economy, the difficulty is not necessarily that energy or materials cease to exist. It is that their cost consumes a growing share of what the economy produces. Less surplus remains to maintain the complicated systems built during the age of expansion.

Maintenance is then postponed. Old computer systems are patched instead of replaced. Skilled workers are lost. Spare capacity is removed. Manufacturers reduce staff and suppliers operate on narrow margins. Organisations may still appear to function normally, but their resilience gradually declines.

Failures become more frequent and recovery becomes more difficult.

Failure does not remain in one sector

We often discuss aviation, manufacturing, electricity, telecommunications, finance and public services as though they were separate activities. In reality, they form a single interdependent system.

A factory depends upon electricity. The electricity network depends upon digital control and telecommunications. Telecommunications depend upon electricity and imported equipment. Employees depend upon transport. Suppliers depend upon banks and computerised payments. All depend upon people being fed, housed and cared for.

This creates the possibility of cascading failure:

  • A computer fault disrupts flights.
  • Cancelled flights prevent workers and components reaching their destinations.
  • A factory closes temporarily because a component has not arrived.
  • Its suppliers lose income.
  • Employees reduce their spending.
  • Local shops and services lose customers.
  • Government receives less tax while demands for assistance increase.

No single event causes the decline of the industrial economy. Decline emerges from the accumulating interaction of thousands of such difficulties.

Can technology solve the problem?

The conventional answer is to add more technology. Old computers will be replaced by newer ones. Artificial intelligence will monitor networks. Vehicles will become more automated. Factories will use more robots. Supply chains will be digitally mapped.

Some of this will undoubtedly help. But every new technological layer also requires energy, hardware, communications, software, skilled maintenance and protection against cyberattack.

Technology can reduce one form of complexity while creating another. A mechanical control that can be understood and repaired locally may be replaced by software that is more efficient but depends upon distant specialists and systems. The immediate task becomes easier, while the supporting structure becomes larger.

The question is therefore not simply whether a new technology works. We must ask whether the whole system required to support it can remain affordable and dependable in a shrinking economy.

From maximum efficiency to sufficient resilience

The lesson is not that Britain should abandon aviation, advanced engineering or motor manufacturing. Some large and complicated systems will remain essential. But we may have to operate fewer of them and protect them more carefully.

Critical national systems need genuine fallback arrangements. Businesses need alternative suppliers, stocks of vital components and the ability to continue some operations when digital systems fail. This will cost more and may reduce apparent productivity. It will nevertheless make the economy less vulnerable.

At the same time, activities that do not need to be organised nationally or globally should be brought closer to the people who depend upon them.

Food production, repair, care, small-scale manufacturing and many everyday services can be organised within localities. Shorter supply chains will not eliminate failure, but they will make failures easier to understand and contain. A locality that retains practical skills, workshops, food production and human relationships possesses forms of resilience that cannot be downloaded from a distant computer.

The economy of the future may therefore contain two distinct levels.

There will remain a national core responsible for such things as air traffic control, telecommunications, intercity rail, defence and specialist medicine. Around it, a much larger part of everyday life may gradually become more local, simpler and less dependent upon continuous long-distance coordination.

The warning

The grounded aircraft and the difficulties at Jaguar Land Rover are not isolated misfortunes. They are warnings from a system approaching the limits of affordable complexity.

The industrial era taught us to admire scale, speed, specialisation and efficiency. The coming era may place greater value upon durability, repairability, spare capacity and local competence.

Complex systems will not vanish overnight. Nor should they. But as the economic surplus contracts, we will be forced to choose which complexity is essential and which can no longer be afforded.

Localism is not an attempt to recreate the past. It is a practical response to a future in which the enormous supporting structures of industrial civilisation can no longer be taken for granted. The aim will not be to preserve maximum consumption at any cost. It will be to secure a sufficient and worthwhile life with systems that ordinary people and their localities can understand, maintain and trust.

333. The Missing Pipe

Posted on by
Reply

by Ed Conway in The Material World

A short history of how Britain chose the wrong heating system a century ago, and why we are unnecessarily sweating our way through a heatwave in 2026

Jun 26, 2026

The first rule of listening to a journalist is: under no circumstances ever take business or financial advice from a journalist. With that proviso out of the way, do you mind if I pitch you a business idea?

It all comes back to something that’s rather relevant right now, with Britain and most of Northern Europe basking/melting in an unprecedentedly hot June. Records are being broken all over the place and all of a sudden everyone is talking about air conditioning.

Why, many are asking, do we not have more air conditioning? Look at a country-by-country comparison and – to the extent that such data exists – the UK usually comes out at or near the bottom of high income nations.

Percentage of households equipped with AC in selected countries, 2018. IEA data. UK figure from English Housing Survey

The short answer of course is that, historically, Britain hasn’t had much need for AC. Compared with countries like the US and Japan we have far fewer “cooling degree days”, and hence much less demand for AC.

But there is a slightly longer and much more interesting answer, which is that this isn’t just a story about AC, and the lack thereof. It’s a story about something much deeper: about the way our homes are built (specifically the plumbing), about net zero, and the government’s heat pump strategy.

But before we get to all that, back to the explanation about why Britain still has so little AC, despite the rise in average temperatures in recent years. It mostly comes back to two things:

  1. Decades ago we committed ourselves to heating systems that don’t lend themselves to cooling, and…
  2. Even if you want an air conditioner today it’s surprisingly hard to find one, and nigh on impossible to find a portable one that actually works particularly well.

To take number 1 first (and I covered some of this in a Times column a few years ago) at some point about a century ago, we plumped for what are known in the trade as “wet” or hydronic systems, with boilers that heat up hot water and send it round our homes.

There are many good things to be said about wet systems: they work well in cold climates, and that hot water can be used for bathing and washing too. But most of the rest of the world doesn’t have radiators like ours. They use systems that blow air around the house, either through ducts and vents or a few wall-based units.

Air-based systems (HVAC as they’re sometimes called) aren’t perfect. The air they puff out can be stuffy and dry, though you can add filters and humidifiers if that bothers you. And you usually need a separate unit to heat your water. Even so, they have one enormously attractive advantage: connect them to a heat pump and they can both heat and cool your home.

And this is where the AC debate collides with the heat pump debate. Heat pumps, as you will probably already know, are a sort of electric alternative to a gas or oil boiler. In thermodynamic terms, they are a rather amazing technology, as you’ll see if you ponder the helpful diagram below. Whereas what a gas boiler does is to take the energy embedded in a molecule of methane and burn it, using the heat to warm the water running through your pipes, what a heat pump does is radically different.

Instead, it essentially captures the heat in the air around us and uses electricity to multiply that heat up so that it becomes warm enough to warm a radiator or a hot water tank (or to puff it out in the form of air).

Now, there are some problems – the chief one being that heat pumps have typically not been able to get that eventual heat up to quite the same levels as a gas boiler, though the latest generation of heat pumps are a lot better on this front. But the real marvel is that – even leaving aside the net zero point which is that they are a lower carbon way of heating our homes – they are so, so much more efficient at turning the energy we put into them into the heat in our homes.

I say “heat pumps” but I’m actually oversimplifying, because in practice, there are two distinct categories of heat pumps. There are heat pumps that heat up water (air-to-water) and heat pumps that heat up air (air-to-air). The point of my column a few years ago was we were being encouraged by the government to buy air-to-water pumps, the assumption being that we should hang on to our radiators.

You can understand why they chose to prioritise “wet” systems. The last thing they want to do is tell people they need to rip out their entire plumbing system. The merit of an air-to-water system is that you simply replace your boiler and you’re away. But the problem with an air-to-water heat pump is twofold. First, it’s actually less efficient than an air-to-air system which puffs out that multiplied heat in the form of air.

Second, it’s, well, not much of an upgrade. For the actual consumer buying them, the long-and-short of it is that they would have to pay rather a lot for a system that would cost about the same to run (at current electricity prices) and results in slightly less hot radiators. Which, to my mind at least, helps explain why so many households have been so reluctant to switch. It is far easier to encourage people to spend lots of money on their home if they feel they’re actually improving their living standards.

The argument in my column was that the government should really start offering grants not just to these water-based systems but also to air-to-air pumps. All of which brings us back to this heatwave and to AC, because, you see, air-to-air pumps are, essentially, air conditioners. In the winter they provide hot air; in the summer they can blow out hot air. Really, we should be calling them “temperature pumps”, since “heat” is only half of what they do. Anyway, bafflingly (to me at least) as of a few years ago the government wasn’t offering any public support for anyone wanting to buy one.

The good news is that since that column the government has changed course and now offers some grant money for those wanting to buy an air-to-air system (I like to think the two things are connected but they almost certainly are not). The grants are far lower (£2,500 vs £7,500 for an air-to-water system) so there’s still an incentive to stick with the old “wet” radiator system. Moreover, the official advice fails to mention that if you opt for an air-to-air system you also get air conditioning thrown in for free. It’s almost as if they don’t want people to know about this amazing technology!

Anyway, once properly installed, you will end up with one big box with a fan and pump outside your house (like the picture at the top) and another unit inside the house delivering hot or cold air inside (like the one below). The whole thing will set you back a few thousand pounds but, all being well, soon enough you will save back that money in heating bills.

But for those who aren’t yet making the leap and installing a permanent system, the reality of having air conditioning in your home in the UK is decidedly unsatisfactory. Because the only way to do this is to buy a portable air conditioner.

Portable air conditioners are inherently far less efficient than the split models I described above. But for those who don’t yet want to rip up their walls and install a proper air-to-air heat pump, they are about the best thing there is for dealing with heat waves like the ones we’ve had this summer.

However, bafflingly, nearly every portable AC available for purchase in the UK today is about two or three times more ineffectual than it needs to be. There are a couple of reasons for this. The first is that they almost all use an old-school technology to cool the air. Once upon a time, the compressors inside old air conditioners/heat pumps (the bit that helps actually change the temperature) were simple on-off units. Either they were cooling at full pelt or they were doing nothing. These days, most proper heat pumps and air conditioners have what are known as inverter compressors, which can ramp up or down the cooling depending on how much is needed.

But here’s the thing: nearly all portable AC units available in the UK are based on the old on-off technology. The upshot is a) they are far more power hungry than they need to be and b) they are very, very noisy, especially when the compressor whirrs into action at full speed. For some reason, it’s very, very hard to find a portable inverter AC unit in the UK (which is not the case, for instance, in the US).1

The second problem with portable AC units is that – and I’m not making this up – they arrive without the right pipes. When you buy a portable AC in the UK, it will, almost without exception, come with a big hose at the back, which you need to stick out of a window.

That hose is an exhaust, essentially sending all the hot air outside. But strictly speaking, you really ought to have two hoses – one to send the hot out and, critically, another to suck in air from outside. The full story is to be found in this New Scientist piece but the long and short of it is that basically every portable air conditioner on the UK market today is doomed to be about two or three times less effective at cooling a room than it could be. Not because of anything wrong with the machine – literally because they don’t throw in the right plastic hoses to attach to it.

And here’s where we get to the business plan. In theory, every single one of these AC units could, through the addition of a few simple pipes and a 3D printed adaptor, be made twice as effective. Rooms could be cooled at twice the speed – for a few plastic parts that cost only 20-30 pounds. And yet, as far as I can work out, the only way you as a consumer can actually get one of these kits is to 3D print it yourself.

I can’t help but feel the market has failed on all sorts of fronts here. Millions of Britons want cooling in the summer. The demand for cooling will only grow in the coming years. But for those who don’t want a permanent air-to-air heat pump, the only options available during a heatwave, save for fans or sweating their way through it, are far, far worse than they need to be. It strikes me there’s an enormous business opportunity here.

Or maybe, well, I’m just a journalist. As I say, never take advice from folks like us. And perhaps I’ve missed something – if so, please comment below! Either way, the overarching point is that sweltering in misery during heatwaves is not an inevitability. There is technology out there to help you through it. The only problem is that it’s far harder to get hold of it than it really ought to be.

The only one I’ve encountered is the brand new Meaco Cirro (though NB you have to opt for the two more powerful models to actually get the inverter. And as far as I can make out the Cirro only has a single hose so loses out on all the efficiency gains you’d get from a double-hosed unit…

T

330. Small Modular Nuclear Reactors are a Dead End

Posted on by
Reply

MuseLetter #398 / May 2026 by Richard Heinberg

Download Printable PDF Version

The nuclear power industry is currently promoting designs for small modular reactors (SMRs) that will supposedly be cheaper, safer, and faster to build than older nuclear power plants. Bill Gates and Amazon are investing in the technology. Moreover, some environmentalists, including Mark Lynas and Bill McKibben, support SMRs in the hope that they can lower carbon emissions. And, according to polls, far more Americans now approve of the development of nuclear energy than was the case just a decade or two ago.

This year, the world has been plunged into a global energy crisis: with the closure of the Strait of Hormuz, nearly a fifth of world oil shipments have been held up, with economic impacts likely to reverberate for months or years. World leaders are suddenly desperate for energy alternatives, and are turning to solar, coal, and nuclear. At the same time, electricity demand for data centers is exploding, and builders of those centers hope to use SMRs to power artificial intelligence (AI).

In short, it looks like a great moment for the nuclear industry.

Yet Indigenous peoples, technology critics, and old-school environmentalists still oppose nukes—even in new, highly touted forms. I agree with their critiques. In this article, we’ll look at the current nuclear revival and see why it may end up being a zombie attack.

Nuclear Renaissance?

Before looking at SMRs specifically, it’s helpful to understand the status of the nuclear industry in more general terms. The industry’s potential resurgence comes after three decades in the doldrums following the Chernobyl catastrophe in 1986. Today, roughly 440 nuclear power plants, spread across 30 countries and with a combined net capacity of around 400 gigawatts (GW), provide about 10 percent of the world’s electricity. The US, which has the largest number of plants of any country (96), is seeing a slow phase-out of old reactors (average age 44 years), but has commissioned three new ones during the last decade. China is now operating 60 reactors, with up to 40 others under construction. India is likewise hoping to grow its nuclear industry rapidly and is experimenting with fast breeder reactors. Globally, the International Energy Agency (IEA) forecasts total nuclear power capacity to grow to over 700 GW by 2050, and small modular reactors are expected to make up a significant share of this growth. A year ago, the Trump administration unveiled an ambitious nuclear strategy that includes a goal to quadruple the United States’ nuclear capacity by 2050, with SMRs playing a key role.

The principal drivers of renewed interest in nuclear power are climate change (globally), the Trump administration (in the US), tech companies’ voracious demand for electricity, and Asian nations’ hunger for more industrial power. Most nations want to limit their carbon emissions, and the main low-carbon alternatives to fossil fuels are solar, wind, hydro, and nuclear. Solar and wind are intermittent (“variable”) sources, requiring energy storage to align electricity supply with demand. Hydro has limited potential for growth. That leaves nuclear power, which has the advantage of being reliable and steady, and has possibilities for expansion.

If it’s helpful to understand why the industry is growing again, it’s just as important to know the reasons for its long period of dormancy:

  • Cost: Nuclear power plants are complex and expensive, employing technology that’s internationally regulated due to concerns about the proliferation of nuclear weapons. Despite over 80 years of the industry’s development, nuclear plants still take a long time to build and are often plagued with cost overruns.
  • Fuel: Uranium, the fuel for nearly all existing nuclear power plants, is a depleting nonrenewable resource, and supplies are running short. Uranium mining is a dirty, expensive process, and mine closures, mostly due to resource depletion, are expected to lead to fuel shortfalls by 2035. While geologists have identified more uranium resources, opening new mines will entail further environmental destruction and harm to human communities, of which the uranium mining industry already has a grim history.
  • Waste: Despite decades of research, the global nuclear industry still has found no good place to put the 300,000 tons of nuclear waste—as well as 480,000 tons of depleted uranium in the US alone—that it has produced in the last 80+ years.
  • Safety: While nuclear accidents are relatively rare, they can be devastating and expensive when they occur. The Fukushima disaster of 2011 resulted in direct cleanup costs of up to $180 billion as of 2016, but the damage still has not been completely contained, and indirect costs to human health have been estimated at half a trillion dollars. Further, nuclear power technology is still tied to the threat of nuclear weapons proliferation.
  • Water issues: Nearly all nuclear power plants use water as a coolant and are highly vulnerable to droughts and floods. Droughts reduce the availability of water for cooling, while floods (nuclear plants are generally built next to rivers, lakes, and other bodies of water) damage safety infrastructure and risk contaminating water sources.

If the nuclear industry can overcome its historic obstacles, a door is open. According to the industry, small modular reactors are the main way forward.

SMRs: Promise or Hype?

The main arguments for SMRs are that they would be cheaper and faster to build than conventional power plants; that they would be safer; and, being smaller, that they could be installed to power remote towns or data centers. The idea is to build components in a centralized factory and then assemble those components at power generation sites.

“Small” is defined as 300 megawatts of electrical power or less. While most existing nuclear plants are in the one-gigawatt (1,000 MW) range, some proposed SMRs are 20 megawatts or less; these are called “micro” reactors.

For the most part, SMRs are still at the design stage. China has one SMR under construction. In the United States, TerraPower, founded by Microsoft’s Bill Gates, has received a permit to build a 345-megawatt (not exactly “small,” but close) sodium-cooled reactor in Kemmerer, Wyoming.

Clearly, it is possible to get funding and approval for these new-generation power plants. The big question is, can SMRs deliver on their promises to overcome the historic drawbacks of conventional nuclear power?

  • Cost: SMRs will only be cheaper to build if large numbers are ordered; the first prototypes may be even more costly than conventional plants. Meanwhile, construction costs per MW of capacity will likely be higher, and operating costs are largely unknown until real-world data can be collected. The cost of electricity from SMRs is therefore also yet to be determined, but preliminary estimates put it much higher than solar or wind.
  • Fuel: Most proposed SMRs use uranium, but some designs on the drawing boards would use depleted uranium or thorium as fuels (see below). For now, however, the uranium fuel constraint looming over the nuclear industry remains in place. SMRs also won’t use their fuel more efficiently than conventional reactors, despite some claims to the contrary.
  • Uranium from Seawater: The supply limits of uranium could be greatly expanded by harvesting it from seawater, where the potential resource is enormous—albeit at a concentration of about 3.3 parts per billion. The total oceanic uranium resource is estimated at 4.5 billion tons, over 500 times all identified land-based uranium resources. However, extracting the uranium will take a lot of energy: the best existing technology using absorbent materials will offer an energy return on energy invested (ERoEI) of about 4:1, which is lower than the ERoEI for solar, wind, hydro, fossil fuels, or conventional uranium mining.
  • Waste: Some proposed SMR designs would be breeder reactors that could get rid of depleted uranium or even nuclear waste by using them as fuels—but this technology has faced significant challenges (see below). Otherwise, SMRs will do nothing to solve, and may actually worsen, the nuclear waste dilemma.
  • Safety: SMRs are designed to be safer than conventional nuclear plants, using passive, gravity-driven cooling systems that don’t require electricity or human intervention to shut down. However, their overall safety is controversial. There is still no real-world data to support the industry’s promises. And having lots of smaller nuclear plants dotted across the landscape could make it easier for nuclear materials to end up in the hands of bad actors. The resilience of SMRs in the face of more frequent and more severe natural disasters is also controversial; a 2021 study concluded that storms, droughts, and higher ambient temperatures linked to climate change are likely to pose operational risks to all nuclear power plants.

The biggest remaining advantages of SMRs are the speed with which they could be deployed once the manufacturing infrastructure is in place, and the prospect of providing non-grid-tied dedicated power sources for data centers.

What about further technological advances?

When confronted with the limits of one technology, nuclear advocates often shift the conversation to another. However, close examination usually shows that each technological “solution” has its own problems:

  • Fast breeder reactors: If nuclear fuel is scarce, why not develop fast breeders, which produce more nuclear fuel than they consume? Currently, Russia operates two fast breeders and India’s first one reached criticality in late April. China has a fast breeder reactor for research. The US, France, and Japan operated breeders in the past but have shut down research along these lines due to high capital and operational costs, safety risks related to sodium coolant, and nuclear proliferation concerns.
  • Alternative cooling systems: Water-cooled reactors (a category that includes nearly all existing commercial nuclear plants) pose risks of loss-of-coolant accidents due to pipe breaks, high-pressure operation failures, age-related component deterioration, and earthquakes or other natural disasters. The industry’s solution: use sodium or helium as a coolant. Unfortunately, sodium is highly chemically reactive and ignites upon contact with air and reacts explosively with water, while helium is a depleting non-renewable resource that is becoming economically scarce at a rapid rate.
  • Thorium reactors: If uranium is scarce and might lead to weapons proliferation, why not use more-abundant thorium? China already has an experimental two-megawatt thorium reactor in the Gobi Desert. However, thorium reactors have steep development costs and produce a highly radioactive byproduct, uranium-232, which decays into isotopes that emit penetrating gamma rays, making fuel handling and maintenance more hazardous and costly. Also, thorium reactors require a “driver” fuel: thorium-232 is fertile, not fissile, meaning it needs a different radioactive fuel (like uranium or plutonium) to initiate the chain reaction. Therefore, proliferation concerns remain.

Currently, there is little real-world data regarding these “new” nuclear technologies, even though all have been discussed or experimented with for decades. The nuclear industry hasn’t actually solved its many dilemmas, and the current nuclear renaissance isn’t being driven by novel solutions so much as by the rapid worsening of society’s energy-related problems, primarily climate change: world leaders are now so desperate for reliable low-carbon energy sources that they are willing to overlook substantial risks, if only the nuclear industry will put a shiny gloss on its latest iteration of products. And leaders of the tech industry, keenly aware of the soaring electricity demand from AI, are even more desperate for ways to power the exponential growth of their companies without risking a backlash from the rest of society, which may suffer from higher electricity prices or shortages.

If not SMRs, then what?

Nuclear power is a product of high-tech modern industrialism. The proponents of nuclear power assume—and nuclear reactors rely on—global supply chains, uninterrupted grid power, reliable water resources, and functioning political systems. The future that’s unfolding around us is a polycrisis in which supply chains, grid power, water, weather, and politics-as-usual are all threatened. In these unfolding circumstances, the only solutions that make sense are ones that are small-scale, local, low-risk, and nature-based.

What to do about carbon emissions? Yes, we need to replace fossil fuels with low-carbon energy sources—but these should be as low-tech as possible, and we should aim to reduce overall energy usage.

What to do about AI data centers? That’s easy: don’t build them. We are rushing headlong into an AI-managed future without an adequate understanding of what AI is, does, or is likely to do in the future. Besides, AI appears to be perhaps the biggest investment bubble in history.

Most political and economic leaders have taken the attitude that we must go to any possible lengths to save industrial modernity. But industrial modernity is the essence of our problem: it is a crisis-generating machine—and one that, prior to its inevitable self-destruction, is creating enormous wealth for a small minority of people, while entrapping everyone else in dreary systems of employment, payment, debt, dependency, and distraction that leave little time for reflection on the futility of it all.

Moreover, SMRs will do nothing to solve our immediate global energy crisis. The oil shortages that are already sweeping over the world in the wake of the US-Iran war cannot, in most cases, be offset with electricity—at least not right away. While electrification is a good interim energy strategy for gradually winding down modernity with minimal casualties, it’s one that will take time, and some things will be hard or impossible to meaningfully electrify—including heavy manufacturing and air travel. Meanwhile, the world needs gasoline, diesel, and jet fuel now; SMRs will take decades to deploy.

The opinion you hold about SMRs will have a lot to do with your general attitude toward technology. If you think humanity’s fate and future rest with high tech (including AI and advanced rockets to enable colonization of other planets), then you’re almost guaranteed to believe that SMRs will help us get there. But if you think, as I do, that the global polycrisis is an inevitable outgrowth of industrialism and its consequences (resource depletion, pollution, and overpopulation), then you’re likely to view SMRs as a pointless and dangerous waste of resources.

Once we see why industrial modernity is unsustainable, the most important question becomes: what is a viable exit strategy? On our way out the door of modernity and back toward simplicity, we need to minimize the creation of new problems and re-learn nature’s elegant solutions. When our priorities are thus reoriented, nuclear power makes no sense.

Featured Image: Adobe Stock

328. Today’s Data Centres – Tomorrow’s Local Resource Crisis

Posted on by
Reply

Artificial intelligence and cloud computing are driving a rapid expansion of data centres. They are often presented as symbols of progress, bringing investment and employment. Yet beneath the headlines lies a question that deserves much more attention.

Economist Tim Morgan recently observed:

“There’s another aspect of this that’s worrying. If somebody builds a data centre in your locality, what happens to your cost of electricity and water, when they’ve got effectively bottomless pockets?”

This question goes to the heart of the localist argument.

Electricity and water are not unlimited resources. Every locality has finite generating capacity, finite distribution networks and, increasingly, finite water supplies. A large data centre may consume as much electricity as a small town and millions of litres of water each day for cooling. When such a development arrives, it becomes a powerful new competitor for essential resources.

The owners of these facilities are often among the wealthiest corporations in the world. They can afford to pay prices that ordinary households, farms and small businesses cannot. Even if they negotiate long-term contracts, the extra demand they create still requires investment in new infrastructure. Ultimately those costs are often spread across everyone else.

The result is that local people may find themselves paying higher prices for electricity and water, while having little influence over decisions that affect their daily lives.

This illustrates one of the weaknesses of an economy organised around perpetual growth. New developments are assessed mainly by the value of the investment and the contribution to national output. Much less attention is given to the effect on the resilience of the locality itself.

Localism asks a different question.

Instead of asking whether a project increases national GDP, it asks whether it strengthens or weakens the ability of the locality to provide for its own people. Does it leave enough affordable electricity for homes, workshops and local industries? Does it protect water supplies for farming, food production and daily life? Does it improve the long-term security of the community?

In an age of growing resource constraints, these questions become increasingly important.

A shrinking economy makes the issue even more significant. As energy becomes more expensive and investment capital becomes scarcer, every kilowatt of electricity and every litre of water become more valuable. Local communities cannot assume that additional supplies will always be available.

This suggests that essential resources should increasingly be regarded as strategic assets belonging first to the locality. Major industrial users should demonstrate not only that they can pay for these resources, but also that their use does not reduce the resilience and prosperity of the surrounding community.

Future planning may therefore need to move beyond traditional economic assessments. Before approving major developments, localities may need to ask whether they can genuinely afford to allocate scarce electricity and water to activities whose principal benefits flow elsewhere.

The debate is therefore not about opposing technology. Data centres undoubtedly have a role in modern society. The question is one of priorities.

When resources become constrained, should communities compete with global corporations for the essentials of life, or should those essentials first secure the well-being of the people who live there?

That is a question localism is uniquely equipped to answer.

304. AI Data Centers Are Not the Railroads of Today

Posted on by
Reply

The AI boom shares all the risk profiles of previous speculative manias but lacks society-wide benefits while generating fast-metastasizing negative consequences and costs.

Charles Hugh Smith

Jun 01, 2026

The idea that the current bubble in AI data centers is an echo of the railroad-construction bubble of the 1870s is appealing–but only half-right. The completion of the first transcontinental railroad in late 1869 sparked a speculative mania of raising capital to build railroads, which were seen as “can’t lose” investments in a technology that lowered transport costs from $1 to ten cents.

But not all routes had the potential to become profitable, and the resulting collapse of the railroad bubble devastated the developed-world economies, triggering a deep economic downturn from 1873 to 1879 that was called “The Great Depression” at the time (or “The Long Depression”).

The term for speculative frenzies channeling vast sums into investments with difficult-to-assess risk profiles is mal-investment, and mal-investment on a large scale triggers financial panics and economic depressions in a well-understood feedback loop.

Money invested in digging a mine that doesn’t yield any gold can’t be recovered. That capital is gone. There is an opportunity cost to every investment: that capital could have been invested in something else that was more productive than the speculative bet on something with unclear risks and payback.

As the scale of losses become apparent, credit tightens and the pool of capital available shrinks. Short-term loans that can’t be rolled over into longer duration loans trigger bankruptcies which quickly lead to bank runs (financial panics) and layoffs as businesses close. This decline in wages, revenues and the velocity of money is self-reinforcing, and the recovery process–being both financial and psychological–takes years.

The parallels with the AI speculative investment mania are obvious. Just as any railroad was viewed as guaranteed to be immensely profitable because railroads generated enormous efficiencies that reduced costs, all AI is guaranteed to be immensely profitable because AI generates enormous efficiencies that reduced costs. But in the real world, use cases for specific railroads and AI applications are stretched along a spectrum which isn’t visible in the early stages of a speculative boom.

Individual use cases don’t automatically guarantee an entire class of use cases will be successful. That one railroad–or application of AI–profitably reduced costs does not necessarily extend to all railroads or AI applications.

Nobody wants to wait around for the long process of sorting which use cases are actually beneficial and which are mal-investments, as the big money is made by making big bets in the early days. Human greed is a remarkable force, especially when combined with self-serving hype and the euphoria of the herd running.

In the current confluence of greed, hype and euphoria, the possibility that the inevitable aftermath of vast mal-investment is a Great Depression doesn’t exactly resonate. AI isn’t a railroad, it’s the most amazing force in the Universe, etc. This is Wetware 1.0 in action: the psychology of speculative frenzies doesn’t change, and so here we are–again.

Those are the parallels of the railroad mania of the 1870s and the current AI mania. But that’s only half the story. Railroads did dramatically lower costs, turning unprofitable ventures into profitable ventures not by reducing production costs but by reducing transport costs, which prior to railroads might equal production costs.

The differences between railroads and LLM / generative AI are significant. While many railroads went bankrupt when the bubble burst, those that actually served expanding markets were eventually put to use as the tracks were still useful many years after being laid. A new locomotive type might enter service decades later, but the tracks remained useful and valuable for decades–with proper maintenance. The rails were not obsoleted every few years, nor did the the entire rail lines have to be replaced every few years.

AI is not permanent. It is constantly being obsoleted. A new class of lower-power consumption chips could obsolete the current class of AI chips, requiring a mass replacement of the entire processing foundation of AI. Innovations in software could reduce the processing demands, turning existing data centers into expenses rather than profit generators. AI software that users download onto their own computers negates the need for “renting” data centers (i.e. buying processing power with tokens) by generating models from the user’s own data. These are just a few potential forces undermining the utility, lifespan and profitability of the current build-out of data centers.

While the cost structure of railroads were relatively straightforward, the costs of AI are complex and difficult to assess as initial costs are not total ownership costs, as maintenance expenses are still unfolding and future costs of resources and energy are trending higher.

While the cost reduction and efficiency benefits of depending on AI are as yet unclear, the costs of sorting “good AI” from “bad AI” are already mounting as real-world expenses. The market continues to underestimate the AI slop problem and what it means for enterprise adoption and spending.

Create enough hallucinated legal arguments, flawed engineering calculations and backdoor-ridden code, and the slop vats fill faster than our capacity to tell good work from bad, writes Tim Harford. How can we tell good AI from bad? (Financial Times)

Cedar Owl recently published a comprehensive overview of the Total Costs of Ownership of AI / Robotics and concluded they may exceed the costs of human employees. Will the cost of an AI Robot be higher than the salary of a Human Employee? AI Robot vs. Human Worker Total Cost of Ownership (cedarowl.substack.com)
“AI didn’t remove cost–it changed where the cost lives.”

As for profitable use cases, it’s too soon to tell. Individual cases don’t necessarily scale to the entire sector or economy. The hype is AI is scalable and applicable everywhere, but this isn’t what real-world experience is finding.

Unlike railroads, whose cost-reduction benefits were immediate and measurable, the sum total of AI benefits is not just unclear but potentially negative. The negative effects of AI slop and malicious applications are already visible but the full consequences of their expansion cannot yet be determined.

Recent polls reveal a profound skepticism in the younger generations whose lives will be most impacted by AI. Gen Z Is Using A.I., but Doesn’t Feel Great About It. Only 15 percent said they saw A.I. as a net benefit.

The structural limits of AI are equally visible but the full consequences of these multi-factor limitations cannot yet be determined. A recent article in Scientific American summarized one key limitation: the illusion that AI is “thinking,” “understanding” and “reasoning”: AI and human intelligence are drastically different–here’s how:
“They are extraordinarily powerful tools when used as what they are: engines of linguistic automation, not engines of understanding. They excel at drafting, summarizing, recombining and exploring ideas. But when we ask them to judge, we unintentionally redefine judgment–shifting it from a relation between a mind and the world to one between a prompt and a probability distribution.”

There are many other structural limitations whose nature limits “quick fixes.” “To grow skills, people need to go through hardship. They need to develop the muscle to think through problems,” he said. “How would someone question if AI is accurate if they don’t have critical thinking?”

“This is the contradiction that has many AI boosters talking out of both sides of their mouths: The use of coding agents is actively diminishing the very skills needed to effectively manage the coding agents.” (via Manoj S.)

CEOs are quietly realizing the AI replacement plan has a problem. Two problems, actually.
“One: the token costs for running AI agents are now exceeding what they were paying the employees they fired.

Two: when the tokens run out, the AI stops. Just stops. No continuity. No workaround. Just a spinning wheel where your workforce used to be.”

AI coding frontloads one form of productivity by backloading the entire system with higher maintenance costs down the line. These costs are not visible in the initial phase, and by the time they’re piling up, it’s too late to reverse these structural costs.

The sums invested in AI data centers–and committed to planned data centers–are on a large enough scale that even the most robust economy is vulnerable to disruption when the revenues needed to justify these extraordinary sums fail to materialize and the total operational costs and costs of ownership become measurable.

Matt Stoller offered an apt analogy of AI data center capital investments:

But in a sense, the entire AI narrative is a bit like selling huge amounts of picks and shovels as everyone rushes to the mines, and then betting there will be gold when they all start digging. Much of the stock market is made up of investor speculation that pick and shovel companies are about to hit the motherlode. But we don’t actually know how much gold there is, or even if there is any gold at all. So far, every powerful and rich person has insisted that there’s so much gold we can’t imagine it all, and anyone who thinks otherwise is a Luddite Marxist loser.”

Perhaps most importantly, once we subtract the hype, there is no evidence-based answer to the question: will our society / the public benefit from AI? Or are all the proposed benefits of reducing costs and generating innovations concentrated in the hands of AI’s owners and corporate users?

Cui bono–to whose benefit? What’s being touted as beneficial to all–equivalent to railroads–is at this point only beneficial to owners and monopolistic-cartel corporations, the very asymmetry that is fast undermining the foundations of our social and economic systems.

Put another way: is AI actually solving the core problems undermining our society and economy–systemic asymmetries of costs, wealth, power, agency and opportunity–or is AI adding new problems–brain rot, dependence on black box systems owned by a handful of tech corporations, AI slop, deepfakes, and a tsunami of malicious AI?

For all these structural reasons, AI data centers are not the railroads of today. The AI boom shares all the risk profiles of previous speculative manias but lacks society-wide benefits while generating fast-metastasizing negative consequences and costs.

292. AI, Income Tax and the Return to Rural Localism

Posted on by
Reply

A recent article in The Telegraph raises a profound question about the future of society. If artificial intelligence removes large numbers of white-collar jobs, what happens to the tax system that depends upon human employment, and what happens to the millions of people who no longer fit into the high-technology economy?

The issue is no longer theoretical. Technology firms, economists and even AI company executives are now openly discussing the possibility of mass redundancy amongst professional and office workers. Predictions range from the disappearance of entry-level jobs to severe reductions in software, administration, finance and design work.

The modern state is heavily dependent upon income tax and national insurance contributions from human labour. If AI systems increasingly perform the work once carried out by accountants, clerks, designers, legal assistants, administrators and software engineers, then the tax base itself begins to erode. This concern is now being openly acknowledged by OpenAI and others.

The political response so far has largely centred on ideas such as taxing AI companies, taxing data centres, introducing universal basic income, or retraining workers for an AI economy. But these responses still assume that the industrial-consumer system itself will continue largely unchanged. That assumption may prove false.

If AI concentrates wealth and productive power into a relatively small number of corporations and highly skilled technical elites, then a growing proportion of the population may simply become economically marginal to the formal system. The danger is not merely unemployment. It is the gradual separation of society into two economies.

The first economy would be the AI economy. Highly automated, urban, capital-intensive and dependent upon large concentrations of computing power, electricity, finance and advanced infrastructure. A relatively small number of people may prosper greatly within it.

The second economy may increasingly become a human economy. People excluded from the high-productivity AI system may drift toward forms of life where human labour still retains value because it meets immediate local needs rather than competing with global machine intelligence.

This is where rural localism may emerge, not as an ideological movement, but as a practical adaptation.

In such localities, people may no longer expect secure careers within national corporations. Instead, value may come from food production, repair work, woodland management, small-scale construction, care of older people, local transport, water systems, energy generation and simple manufacturing. Many of these activities are difficult to automate economically at small scale, especially in dispersed rural areas.

Paradoxically, those who do not benefit from AI may rediscover forms of resilience that highly automated urban populations lack.

A rural locality producing some of its own food, fuel and essential goods may prove more stable than a city population dependent upon welfare transfers funded by increasingly fragile corporate taxation. The ability to grow food, repair tools, manage woodland, harvest water and maintain simple infrastructure may become economically important once again.

There is also the question of housing. If large sections of office employment disappear, demand for expensive urban property may weaken significantly. Already there are warnings that AI-related job losses could destabilise housing markets. This may gradually push some people away from metropolitan centres toward cheaper rural areas where survival costs can be lowered through partial self-reliance.

However, rural localism will not be easy. Land prices remain extremely high. Planning systems still assume economic growth and commuter lifestyles. Rural infrastructure has been weakened over decades. Villages often lack workshops, smallholdings, local rail services, markets and affordable housing. Most importantly, modern populations have lost many practical skills.

Nevertheless, if AI continues replacing cognitive and administrative labour while energy and living costs continue rising, the logic of localism may strengthen naturally.

The future may therefore divide into two worlds existing side by side. One world dominated by AI, automation and concentrated wealth. The other based increasingly upon locality, practical labour, shared resources and reduced dependence upon the formal money economy.

In that sense, rural localism may become less a lifestyle choice and more a refuge for those left outside the AI system.

This piece was written by ChatGPT, prompted by Barry.

265. AI could be the end of the digital wave, not the next big thing

Posted on by
Reply

thenextwavefutures•5h ago

the next wave

Carlota PerezAINicolas Colin

View originalSubscribed

I have deliberately tried not to write too much about AI, because the signal gets swamped by the noise. But I think the picture is becoming clearer now. This week on The Next Wave, I’m going to re-publish versions of posts originally on my newsletter, Just Two: one from last summer, and one that goes live this week.—-

Just by way of a thought experiment: what if the current surge in the bunch of technologies that goes under the label of ‘AI’ isn’t the beginning of a whole new technology surge, but is actually the final stage of the digital surge that started in the 1970s and accelerated at the turn of the century?

I’ve been wondering this for a while in a vague kind of a way because I haven’t been able to see the business model that supports the huge investment in AI in the USA. (I’ve written about this before on here.)

This is a long way in to couple of pieces by Nicolas Colin, the strategy and innovation blogger, who has been wondering the same thing, but a lot more coherently. He calls this ‘late cycle investment theory’.

Like me, he is a fan of the work of the academic Carlota Perez, who built on the work of Christopher Freeman to develop a model of how technology and finance interacted to create new long surges of investment, starting with canals and cotton, that run for 50-60 years. (She calls them ‘surges’ because unlike ‘waves’ each technology embeds itself in the society and its infrastructure.)

The two most recent surges are a cars/oil surge, which started in 1908, and the Information and Communications Technology, which started in 1971.

(Source: Carlota Perez)

I’m not going into all of the theory of the Perez model here—it’s online if you want to do that, and I have written about it elsewhere—but the relevant point for the present discussion is that it follows an S-curve, and the first half is slow going, as new infrastructure is ‘installed’, and some of it is below the radar. The internet was a closed academic network for most of the first part of its S-curve.

From infrastructure to ‘deployment’

Halfway through, after a lot of infrastructure has been built out, and usually following a financial crash in which some of the investors in that infrastructure lose their shirts, ‘deployment’ companies take over, with actual customers and business models, and have an accelerated period of growth, before they hit market limits and turn into ordinary businesses. And the investors who have made large returns from that period of growth start looking elsewhere—for the technologies that will make the next surge.

The reason I like Perez’s version is that her model has had a lot of explanatory power over the last 25 years as I have watched the evolution of the tech sector.

That’s a long way into Colin’s argument, and let me quote from his first article directly:

Seen through a late-cycle lens, today’s markets show signs that we’ve entered the maturity phase of the computing and networks revolution. The theory, therefore, leads to specific, testable predictions about where capital should go and which strategies will outperform.

Three indicators

He points to three indicators from the tech sector that support this observation that we’re in the ‘late cycle’:

  1. The startup funding collapse of 2022 wasn’t just a correction—it may be structural. As investor Jerry Neumann argued in his landmark Productive Uncertainty, startups rely on uncertainty as a competitive edge. When good ideas become obvious to everyone—including well-funded incumbents—the startup model faces real strain.
  1. Then came AI, revealing new dynamics. ChatGPT’s breakthrough didn’t come from a garage startup but from OpenAI, backed by Microsoft’s vast computing power. Google, Meta, and Amazon responded with billions. This pattern—big tech deploying huge capital against well-understood problems—fits the late-cycle theory exactly.
  2. Most tellingly, platform saturation now looks almost complete. Digital transformation has reached most sectors where computing and networks can plausibly work. What remains—healthcare delivery, education, construction,  government services—may reflect the paradigm’s natural limits, not untapped markets. [His emphasis]

Optimising the existing system

In the second article —some behind a paywall—he looks specifically at the way AI is being deployed, and I’m going to quote/paraphrase quickly the visible bits of this here.

Late-cycle investment theory suggests AI is the efficiency breakthrough of the computing and networks era, not the start of a new one. Just as lean production refined mass production in the 1970s without replacing it, AI optimises the existing paradigm rather than creating a new one.

(Data centre. Photo by penguincakes/flickr, CC BY-NC-SA 2.0)

Colin’s done a lot of analysis here, and he’s assembled quite a lot of evidence which he shares. I’m not going to spend a lot of time on this, because I’m more interested in the bigger strategic questions that get raised if he is right.

What a new technology surge looks like

But it’s worth summarising some of the observations. First, that at the start of a new technology surge, you don’t know it’s happening. You understand the decisive moment afterwards, the moment at which an innovation transformed the cost structure (the Spinning Jenny, Watt’s condensing engine, the Ford production line, the microprocessor). But with AI, the moment was very visible, to the point of being choreographed.

Second, the amount of capital investment is off the scale. At the early stage of a surge, investment tends to be patchy and not fully understood—the sector exists but it is not completely legible yet.

And third, Colin suggests that AI allows computing to reach sectors that have in some ways resisted it:

Like lean production, which extended mass production’s dominance for decades through efficiency gains, AI doesn’t mark computing’s end but its maturation. The technology spreads to previously untouchable sectors, creating the illusion of radical novelty whilst actually representing computing and networks’ final conquest of the physical economy.

Late deployment

It’s worth pausing here. Although Perez dates the end of each of her surges from the date of the innovation that makes the next surge, possible, there’s a kind of ‘late deployment’ stage in the old surge while the new one is still in its early stages of development.

Late deployment: So although the ICT surge dates from 1971, much of the final innovation in the cars/oil surge also dates from then. In the UK at that time, there’s still a huge roadbuilding programme of motorways and ring-roads, and these then made possible the emergence of long-distance logistics, big-box out of town retailing, and edge of town business parks. Colin’s arguing that AI is the equivalent of bigger roads and big box retail—different, but more about embedding the technology more deeply than the kind of transformational change that eventually causes a new and distinctive form of abundance.

There’s also social pushback—in the UK the campaigns against big ringroad schemes started in the late 1960s and early 1970s. And perhaps we’re seeing some of that about AI. The U.S. map of local pushback against data centres from Data Center Watch covers the whole of the country, in red states and blue. People seem to hate Google’s inserting of AI tools into its search results, and hate even more that it is all but impossible to turn it off. This doesn’t speak to an exciting technology that is being embraced by its users. A note by Ted Gioia on his music blog says that:

Most people won’t pay for AI voluntarily—just 8% according to a recent survey. So [tech companies] need to bundle it with some other essential product.

Or as Ed Zitron noted recently of Notion:

Notion bumped its Business Plan from $15 to $20-a-month per user thanks to its new “AI features,” which I imagine sucked for previous business subscribers who didn’t want “AI agents” or any of that crap but did want things like Single Sign On and Premium Integrations. The result? Profit margins dropped by 10%. Great job everybody!

Normal returns

This matters for a couple of reasons. In the first place, late stage post-deployment technologies do produce returns on investment, but they’re normal returns, not increasing returns.

But in the second place it sheds a different light on what amounts to a ‘business model war’ going on between China and the United States at the moment through their different approaches to AI.

I think we know plenty about the American model. It is fuelled by a transhumanist ideology that is just this side of The Rapture, as Sam Altman of OpenAI reminds people every week of the year.

The Chinese model of AI

As the Exponential View newsletter explained on Sunday, quoting the policy organisation RAND, the Chinese model is completely different:

In Washington, the AI policy discourse is sometimes framed as a ‘race to AGI.’ In contrast, in Beijing, the AI discourse is less abstract and focuses on economic and industrial applications that can support Beijing’s overall economic objectives.

Azeem Azhar of EV added some gloss:

Chinese teams… publish leaner open-source architectures and partner with specialists in areas such as healthcare analytics (Yidu Tech) and adaptive learning (Squirrel AI).

This is partly driven by constraints: China has far less computing power than the US, and needs to build lean. This also means that its model is far more exportable. But the important point here is that if AI is a late-stage technology and not the next large surge of innovation, the Chinese model matches the moment. Perhaps we shouldn’t be surprised: unlike most countries, a third of the full members of China’s Central Committee are technocrats.

—

238. They Were Never Going to Scale

Posted on by
Reply

Embracing Appropriate Technology in a Contracting World

Ludovic Viger

That one simple sentence cuts through decades of hype like a knife: THEY WERE NEVER GOING TO SCALE.

It was the quiet warning from the Appropriate Technology (AT) movement—the one the mainstream green energy narrative did quadruple backflips to ignore. Vast solar farms, towering wind turbines, gigafactories churning out batteries—they promised endless clean power on an industrial scale.

But physics, geology, and economics had other plans. As we hit 2026, the verdict is in: globalist green tech was never going to replace fossil fuels at the level required to maintain an infinite-growth economy.

The Wall of Physical Reality

The math simply doesn’t track. To swap our current energy base for industrial renewables, we require a “Great Mining” of rare earths and minerals that don’t multiply on demand. As we move to lower-grade ores in politically unstable regions, the Energy Return on Investment (EROI)—the actual “profit” of energy we get back after spending energy to build the tech—is plummeting.

Add to this the 35% spikes in battery costs due to recent trade tariffs, and the “cheap energy” dream has curdled into a protectionist nightmare.

The Liberation of Limits

This isn’t defeatist. It’s liberating.

THEY WERE NEVER GOING TO SCALE frees us from the myth that salvation lies in bigger, faster, more centralized tech controlled by distant corporations. It reminds us that chasing “scale” often chains us to fragile global networks and bureaucracies that fail the moment a shipping lane is blocked or a mineral cartel raises prices.

Instead, we can turn to technology that is small, adaptable, and human-scaled. This is the philosophy of E.F. Schumacher’s Small Is Beautiful. It’s not about rejecting progress; it’s about choosing progress that serves people and place, not abstract GDP metrics.

What is “Appropriate Technology”?

Schumacher argued for “intermediate technology”—tools more effective than traditional methods but far simpler and cheaper than high-tech industrial solutions.

The AT Manifesto:

  • Repairable: Fixed locally with local tools.
  • Capital-Saving: Prioritizes human ingenuity over massive debt.
  • Resilient: Works when the global “Just-in-Time” supply chain breaks.
  • Nonviolent: Low impact on the Earth’s finite resources.

Think biogas digesters from farm waste, hand-built solar cookers, or passive solar designs that heat homes without a single circuit board. These are technologies with a human face.

Share

Contraction vs. Collapse

We are entering an era of economic contraction. Resource peaks and debt burdens mean endless growth is off the table. But contraction isn’t collapse if we prepare.

In a world of “Efficiency,” systems are optimized to be lean, which makes them incredibly fragile. One disruption cascades into a total shutdown. Appropriate Technology embraces “Redundancy” and “Simplicity.”

The Complexity Trap

Joseph Tainter’s The Collapse of Complex Societies warns that societies eventually invest in so many layers of bureaucracy and tech that they reach “diminishing returns.” We spend more energy maintaining the system than we get out of it.

AT counters this by simplifying:

  • A community-owned windmill solves a need without a billion-dollar investment.
  • A backyard biogas setup turns waste into fuel without a global pipeline.
  • A hand-cranked grain mill works when the grid is down.

When the giants stumble, the small and the simple remain standing.

The Cooperative Engine

How do we deploy this tech without falling back into corporate traps? Through Cooperatives.

Worker-owned models align technology with local needs rather than shareholder returns. In contraction, co-ops buffer shocks: when markets falter, members prioritize livelihoods over dividends.

  • L’Atelier Paysan (France): This cooperative designs open-source farm tools with farmers, ensuring tech remains a tool for the worker, not a shackle.
  • The Mondragon Model: In the Basque region, large-scale cooperatives prove that we can have sophisticated industry while keeping benefits circulating locally.
  • Energy Collectives: Across the globe, neighborhoods are installing “micro-grids”—rooftop solar kits and micro-hydro owned collectively—to power homes without feeding the corporate beast.

The Script Has Flipped

The mainstream ignored this wisdom because it didn’t offer a way to get rich quick. But in 2026, getting “resilient” is the new getting “rich.”

THEY WERE NEVER GOING TO SCALE isn’t a counsel of despair. It’s an invitation to reclaim your agency. Free from scalability myths, we escape dependence on fragile megastructures. We build what we can maintain, and we own what we use.

How to start:

  1. Audit your tech: What do you own that you can actually fix?
  2. Join a “Library of Things”: Tool-sharing co-ops reduce the need for everyone to own a resource-heavy machine.
  3. Invest in Low-Tech: Support local energy collectives or experiment with passive solar heating.

Small is still beautiful—and in a contracting world, it might be our only way forward.

202. A New Take on the Future

Posted on by
Reply

Gail Tverberg’s latest post in Our Finite World is worth reading.

Her conclusion sums up this new take.

I do not expect that there will be a formal World War III. Instead, I think the United States is already in a cold war against practically every other country because there cannot be enough goods and services to go around. The US can’t go into a formal war against China because it provides parts of the supply chains for many essential goods the US uses today. Even Europe is a competitor for essential goods. For example, the less oil Europe uses, the more oil will be available for other countries.

While new technologies such as artificial intelligence and energy recovery may eventually alleviate our energy problems, it is unlikely that such approaches will solve our problem in the near term. As a result, governments are likely to be less able to keep their promises. Historically, families or “villages” of extended kin have provided safety nets, rather than government programs. Perhaps now is a good time to be thinking about how we can move in this direction, as well.

Which sounds like localism to me!

195. Data centres in the USA are creating water pollution

Posted on by
3

We have glibly talked about the huge amount of energy consumed by data centres and whether the UK National Grid will be able to cope.

But do we know what we are talking about?

This is a photograph of an Amazon data centre in Morrow County, Oregon, USA.

Data centres need water to cool the servers, most of which need to be kept at 70 to 80 degrees to run effectively. The volume of water required is not the problem; The used water ejected from the data centres can end up, via water courses, polluting the fields used for growing food, which has been found to be contaminated.

Here is a map of planned data centres in the UK.

What the map shows (and what is worth noting)

These are particularly AI-ready ones backed by the government’s push and private investment.

The map shows a spread across the UK, including major clusters near London, in the North East, North West, and other parts.

  • The bulk of new data-centre proposals are concentrated around London and the South East, but there are also major planned sites across northern England, Wales and Scotland (reflecting a push beyond just the capital region). Data Centre Magazine+2Blackridge Research+2
  • Some of the largest upcoming developments include:
    • Elsham Tech Park (North Lincolnshire) — a massive planned data-centre campus. Blackridge Research
    • Cambois Data Center (Northumberland), near Blyth — a 720 MW facility under planning. Blackridge Research+1
    • East Havering Data Centre Campus (London Borough of Havering) — a large new centre in or near the London green belt. Barbour ABI+2Blackridge Research+2
    • Humber Tech Park (near the Humber Estuary / North Lincolnshire) — another significant planned campus. Blackridge Research+1
  • According to recent analysis, more than half of the proposed new data-centre facilities will be in London and its surrounding counties, though a substantial number are dispersed across the UK — including Wales, Scotland, Manchester, and other locations. Data Centre Magazine+1
  • The map also reflects a shift in policy and regulation: the government is now formally designating data centres as eligible for “infrastructure-scale” consenting under the revised rules for “nationally significant infrastructure projects (NSIPs)”. This means major new facilities will have an easier path through planning, which is likely to accelerate build-out across many of these mapped sites. Pinsent Masons+1

⚠️ What the map doesn’t guarantee

  • The map shows proposed or planned data-centre sites — not all will necessarily be built. Some may stall, be altered, or cancelled depending on planning outcomes, grid/water supply issues, or local opposition.
  • Timing is uncertain — while many sites appear on the map now, actual construction and operation may be several years away.
  • Impact on local infrastructure (power grid, water supply, transport) is real. The government recognises this and has introduced support mechanisms (including incentives for locating data centres where grid capacity for renewables is stronger) through initiatives like AI Growth Zones (AIGZ).

157. The End of the Industrial Era and the Rise of Local Makers

Posted on by
Reply

The twentieth century was the age of great industries in the UK. . Steel, shipbuilding, textiles, cars, chemicals and electronics were produced on a vast scale, with workforces counted in the tens of thousands. Entire cities were shaped around a single factory or plant. The rhythm of life followed the whistle of the works. But that world is passing.

Globalisation, automation, and the limits of growth have hollowed out the giant industries. Many can no longer compete, either with low-cost overseas production or with new methods that require far fewer workers. Where once a steelworks employed 20,000, today the same output can be achieved with a few hundred. The vast labour armies are gone, and the communities that depended on them are left adrift.

This decline marks more than an economic change. It signals the fading of the industrial era itself, when economies were measured by the mass production of heavy goods. The future lies not in ever bigger factories, but in smaller, more flexible systems closer to where people live.

Localism offers a different path. Across the country, small manufacturers are emerging to meet needs that global supply chains once served. A workshop making timber window frames for local housing repair; a community bakery supplying fresh bread within walking distance; a small brewery producing distinctive ales for local pubs. These are modest in scale, but rooted in place. They keep money circulating in the local economy and provide jobs that cannot be outsourced to the other side of the world.

Technology supports this shift. Small-scale digital tools, such as 3D printers or computer-controlled cutters, allow local workshops to design and manufacture customised goods without the capital of a giant plant. In towns and villages, shared workshops – sometimes called makerspaces – give craftspeople and engineers access to equipment that once only large companies could afford.

We see it in Totnes, where local food processing and brewing firms supply a growing share of the town’s needs. In Sheffield, once the heart of heavy steel, small workshops now produce specialist blades and tools of a quality that mass producers cannot match. In rural areas, small sawmills and furniture makers are reviving traditional skills while adapting them to modern demand.

This is not nostalgia, but evolution. As big industry contracts under the weight of global pressures, small-scale making rises to fill the gaps. The end of the industrial era does not mean the end of production. It means production becomes embedded in communities again, on a scale that fits within the limits of resources and the needs of people.

The giant factories that defined the past will not return. The future belongs to the small workshops, the local makers, and the networks of exchange that sustain them. In their modest way, they point to a new economy – one that is lighter, more resilient, and more human.

148. AI: False Saviour of a Hollowed-Out Economy

Posted on by
Reply
by Charles Hugh Smith: August 27, 2025

What nobody seems to notice is all the incentives for deploying AI are perverse.

The real story of the US economy isn’t about AI, it’s about an economy that’s run out of rope. AI is being hyped not just by promoters reaping billions of dollars in stock market gains, it’s being hyped by the entire status quo because it’s understood to be the last chance of saving an economy doomed by the consequences of decades of artifice.

The real story of the US economy is that decades of “financial innovations” finally caught up with us in 2008, when the subprime mortgage scam–a classic example of “financial innovations” being the cover story for greed and fraud running amok–pulled a block from the global financial Jenga Tower that nearly collapsed the entire rickety, rotten structure.

Our political leadership had a choice: clean house or save the scam. They chose to save the scam, and that required not just institutionalizing moral hazard (transferring the risks of fraud and leveraged speculation from the gamblers to the public / Federal Reserve) but pursuing policies–zero interest rate policy (ZIRP), quantitative easing, increasing the money supply, and so on–that had only one possible outcome:

An economy permanently dependent on inflating asset-bubbles that enriched the top 10% while the bottom 90% who depend on earned income fell behind.

The desired goal of permanent asset-bubbles is the “wealth effect,” the cover story for transferring all the gains into the hands of the top 10%, who can then go on a spending spree which ‘trickles down” to the bottom 90%, who are now a neofeudal class of workers serving the top 10% who account for 50% of all consumer spending and collect 90% of the unearned income and capital gains.

This arrangement is inherently unstable, as “financial innovations” suffer from diminishing returns. Eventually the debt-serfs can no longer borrow more or service the debt they already have, and every bubble being bigger than the previous bubble guarantees the next implosion will be larger and more devastating than the previous bubble-pop.

So what does a system that’s run out of rope do? Seek a savior. The rope has frayed, and the rocks are far below. The impact is going to be life-changing, and not for the better.

The choice remains: clean house, end the bubble-dependent frauds and scams, or find a way to inflate yet another credit-asset bubble. Clean house and lose all our bubble-wealth? You’re joking. The solution is to blow an even bigger bubble. Hey, it’s worked great for 17 years.

Never mind that the precarity of the bottom 90% is accelerating as both the state and Corporate America have offloaded risks onto households and workers; they have OnlyFans, 24% interest credit cards, zero-day-expiration options and side hustles to get by. Never mind that for many Americans, basic services are on the same level as impoverished developing-world economies. What matters is maintaining the wealth of the few at the expense of the many, by any means available.

Enter the savior of our asset-bubble-dependent elites: AI. AI is going to change the world, we’ll all be watched over by machines of loving grace, profits and capital gains will be in the trillions of dollars, yowza, because we’ll fire half of you and give you enough Universal Basic Income (UBI) to scrape by, and some of you can join our security teams protecting us from the impoverished rabble.

There’s just one teeny little problem: AI is a false savior. It doesn’t work as advertised, it has multiple inherent limits that can’t be overcome by scaling up processors, and the dystopian consequences of even this first wave are already uncontrollably destructive:

AI psychosis and addiction to AI chatbots is making users even lonelier and more isolated than they were before embracing AI, AI Slop is overwhelming legitimate content, AI agents are just good enough to degrade already pathetically deficient corporate services, LLM models are Swiss-cheese security risks, and 95% of all corporate AI projects founder.

What nobody seems to notice is all the incentives for deploying AI are perverse. Those seeking nickels from Big Tech platform engagement / views win big by flooding the web with AI slop, scammers and fraudsters now have much more powerful tools to deceive and defraud (deepfake videos and voiceovers), and rather than watch over us with loving grace, the AI systems are scraping their own inaccuracies and hallucinations and deceptively presenting this slop as accurate.

The status quo is counting on AI to be the savior of a hollowed-out economy, but it’s a false savior. The frenzy has inflated another credit-asset bubble as planned, but another bubble that enriches the few isn’t going to fix what’s broken. Rather, unleashing AI tools in a system of perverse incentives is accelerating the decay and collapse of the entire system by replacing authentic value with illusions of value–what I call Ultra-Processed Life.

I hate to be the bearer of unwelcome news, but enriching the few at the expense of the many is the problem, not the solution. So the AI bubble mints more billionaires, well that’s swell, but the process of inflating bubbles that enrich the few is what’s destabilizing our economy and society.

This chart presents the consequences of 17 years of bubble-inflation “wealth effect”: wealth for the few and “effects” for the many:



Asset bubbles have been good to the top .01%:



And it’s been good for the top 10%, too: $107 trillion in net worth, and of course, “I earned every penny of it.” The bottom 50% with $4 trillion–well, better launch your OnlyFans site, even though there are already millions of other desperate people hoping to scrape up a few bucks by selling themselves online.



Contrary to the hype, AI isn’t the savior of the bubble economy–it’s the hype-heavy straw that breaks it for good.

Before we bow down and worship a false savior, it’s probably a good idea to learn about the limits of our AI savior and the self-serving hype, and the banquet of consequences being laid for true believers.

MIT report: 95% of generative AI pilots at companies are failing

LLMs + Coding Agents = Security Nightmare

AI Is a Mass-Delusion Event

The potential of generative AI for personalized persuasion at scale

What If A.I. Doesn’t Get Much Better Than This?

The Real Demon Inside ChatGPT

ChatGPT May Be Eroding Critical Thinking Skills, According to a New MIT Study

AI Industry Nervous About Small Detail: They’re Not Making Any Real Money

Which jobs can be replaced with AI? Jobs that have already be degraded to the point of uselessness.

I’ve written 17 essays on AI this year: here are four:

AI: Over-Promise + Under-Perform = Disillusionment and Blowback

Maybe AI Isn’t Going to Replace You at Work After All

Good News! AI Can Do More BS Work

AI Is a Mirror in Which We See Our Own Reflection


Check out my new book Ultra-Processed Life and my updated Books and Films.

Become a $3/month patron of my work via patreon.com

Subscribe to my Substack for free


My recent books:

Disclosure: As an Amazon Associate I earn from qualifying purchases originated via links to Amazon products on this site.

Ultra-Processed Life print $16, (Kindle $7.95, Hardcover $20 (129 pages, 2025) audiobook     Read the Introduction and first chapter for free (PDF)

The Mythology of Progress, Anti-Progress and a Mythology for the 21st Century print $16, (Kindle $6.95, audiobook, Hardcover $24 (215 pages, 2024) Read the Introduction and first chapter for free (PDF)

Self-Reliance in the 21st Century print $15, (Kindle $6.95, audiobook $13.08 (96 pages, 2022) Read the first chapter for free (PDF)

When You Can’t Go On: Burnout, Reckoning and Renewal $15 print, $6.95 Kindle ebook; audiobook Read the first section for free (PDF)

Global Crisis, National Renewal: A (Revolutionary) Grand Strategy for the United States (Kindle $6.95, print $16, audiobook) Read Chapter One for free (PDF).

A Hacker’s Teleology: Sharing the Wealth of Our Shrinking Planet (Kindle $6.95, print $15, audiobook $17.46) Read the first section for free (PDF).

Will You Be Richer or Poorer?: Profit, Power, and AI in a Traumatized World
(Kindle $3.95, print $12, audiobook) Read the first section for free (PDF).

The Adventures of the Consulting Philosopher: The Disappearance of Drake (Novel) $3.95 Kindle, $12 print); read the first chapters for free (PDF)

Money and Work Unchained $6.95 Kindle, $15 print) Read the first section for free


Become a $3/month patron of my work via patreon.com.

Subscribe to my Substack for free