Energy, matter and life at the edge
A few years ago I went to see the National Theatre of Scotland’s production of The Cheviot, the Stag and the Black, Black Oil at Dundee Rep. I had heard of John McGrath’s play, of course — it is one of those pieces of Scottish cultural furniture you assume you know without having sat in the room — but I had never seen it performed. Dundee Rep is a proper theatre, and yet the revival still carried the spirit the 7:84 company built for community stages: Gaelic song, statistics read aloud over reconstructions, audience members pulled into the action, laughter that turns sharp when you realise what you are laughing at. There is an excerpt here.
The Cheviot is a musical drama in the roughest, most purposeful sense. The 7:84 company toured it through the 1970s into village halls on Lewis, South Uist, Benbecula — places where the audience often included people who had lived through what was being described. The play is built up in three movements, and McGrath makes the connection explicit: the Cheviot sheep that replaced Highland communities during the Clearances; the stag and the Victorian passion for sport on estates that swallowed millions of acres; and the black, black oil that arrived in the 1970s with a promise of jobs and a habit of leaving very little behind. Patrick Sellar, the factor who carried out some of the most brutal Clearance evictions for the Duke of Sutherland, stands for that older history of taking the land — people cleared off so sheep could take their place. At the end the performers warn the audience that the oil corporations may prove more insensitive still. The land is yours, they say. Do not let it be taken twice.
Se firinn is ceartas a sheasas. (Scottish Gaelic for “It is truth and justice that will endure”)
I’m currently on holiday in Ardgour, and I was thinking about The Cheviot and its links to my parents’ brewery on the loch, the ferry, and Aoineadh Mòr where a village was cleared in 1824 to make way for sheep. McGrath’s “first sniff of oil” line is famous: villages that could not get a factory for a hundred workers suddenly overrun by thousands of men in labour camps, building rigs, building platforms, building a whole temporary civilisation that would move on when the field declined. In 1974 that was Aberdeen and the North-East. Exploration was already looking west, off the Butt of Lewis. The play is nearly fifty years old and more relevant than ever.
What stuck with me was the boundary. Each act asks what crosses the line, what is taken out, what is left inside, and who gets to decide. The Clearances exported people and imported sheep. The stag estates exported venison and imported wealthy shooters. The oil boom exported crude and imported money that inflated house prices in villages that had never been expensive before. The flux changed; the pattern held.
The third act brings that home. Mock interviews with oil workers and American executives — done in the flat documentary style of the 1974 television version — start as comedy and stop being funny. The numbers land: rigs, flares, yards, pipelines, wages that look generous until you price a house in Aberdeen. Then McGrath cuts back to the cleared townships, and sheep, stag and oil read as one story: something valuable leaves, something else is brought in, and the people who live on the land are told to be grateful. That montage is the argument. My parents remember Scotland before and after the oil boom; I grew up with only the tail end of it — the rigs, the money, the sense that trouble would always belong somewhere else. Re-watching The Cheviot recently, the oil act clearly reflected the behaviours of the Clearances and the sporting estates again, and to use the language of my techno-crofting posts, each act explored a different cargo crossing the boundary.
The hidden engine
A colleague from The School of Engineering (Dr Michael Merlin) who read my earlier posts on techno-crofting pointed me to Jean-Marc Jancovici and Christophe Blain’s graphic novel World Without End. Michael recognised that I had been circling the same ideas from engineering and crofting; Jancovici arrives from energy economics with a mordant wit, which I appreciate. His central thesis is simple and easy to misunderstand: energy is not one input among many in the economy. It is the condition that makes inputs arrive. Agriculture, transport, materials, computation, healthcare — all of it sits on top of abundant, high-quality energy. Take the energy away and the economy does not merely slow; it becomes a different economy, or none.
Jancovici’s provocation, which pairs well with the third act of The Cheviot, is that we are not really moving towards a renewable world. We have left one. Before coal and oil, human societies lived mostly on current solar flows: crops, wood, wind, water, muscle. Fossil fuels let us draw down ancient stocks and mistake that temporary surplus for normality. The excursion was spectacular, finite, and the return journey is thermodynamic.
The Cheviot and World Without End are odd companions — a Scottish play and a French comic about civilisational inertia — but they both refuse the comforting story that progress is a straight line, that the next taking of the land will be gentler because we are better people, that technology will lift the constraint without renaming it. McGrath names the constraint as land and labour; Jancovici as energy throughput. Techno-crofting, as I have been trying to define it in these essays, names it as flux across a boundary.
The first three posts in this series asked what a techno-croft is, what kind of community could live with one, and what “best” means under local constraints. This one asks the colder question underneath: what flows of energy and matter make such a system possible at all — what has to cross the boundary, at what rate, in what form, and for how long, for people to live well at the edge?
Earth is not closed, and neither is a croft
There is a sloppy version of thermodynamics that does the rounds in environmental arguments: the Earth is a closed system, entropy always increases, therefore we are doomed or must recycle perfectly or must return to the soil. The sloppy version is half right and wholly misleading.
Earth is open. It receives a vast flux of low-entropy energy from the Sun and radiates waste heat to space. Life maintains local order by exploiting gradients — light to chemical energy, chemical energy to heat and work, work to structure that eventually decays and is rebuilt. That qualification strengthens the edge argument. The planet is open. A techno-croft is open too.
A techno-croft is a bounded patch of infrastructure — workshop, greenhouse, brewery/fermentery, battery bank, a finite shelf of spare parts — open at every edge. Energy crosses: electricity, diesel, gas, sunlight. Materials cross: grain, yeast, steel, medicines, the occasional microcontroller. People and knowledge cross. Waste heat leaves whether you account for it or not. Nutrients leak. Recycling costs energy. No bounded human system is magically circular. The same logic applies at every scale: what matters is net delivery after the system has fed itself — whether the system is a barrel of oil, a brewery, or a croft.
Sustainability, then, is a question of boundary and time horizon: what system, in what environment, supplied by what flows, can continue for how long? Considering my parents’ brewery as an example is instructive. Fermentation is a metabolism: water, grain, hops, yeast, power in; beer, CO₂, spent grain, heat, wastewater out. Extend that metabolism with algae tanks or microbial electrosynthesis and you add more loops — without abolishing the boundary. The grain still comes on a lorry. The CO₂ still matters. When the ferry does not sail, the schedule still slips.
Living at the edge makes this accounting visible sooner. Cities hide flows behind pipes and prices; a hill-farm, a croft, an island, or a research station feels these flows in their invoices and in their empty tanks. The same gradient logic that keeps a brewery alive — useful energy in, waste heat out, net work after self-consumption — is what the oil age has been running on at planetary scale. When that surplus thins, the techno-croft feels it as price and scarcity long before textbooks rewrite their chapters.
Oil, electrons, and the pipe
I need to talk about oil carefully, because easy moralism is the enemy of good engineering.
Oil-derived liquids still carry most of the world’s transport work — road, sea, air — even as the shape of demand shifts. Road use has plateaued in several large economies; aviation and petrochemicals now lead growth. Oil remains about a third of global primary energy — still the largest single source in the Energy Institute’s 2026 statistical review — while wind and solar together remain a few per cent of the total balance. A Fermi estimate makes the hidden dependency visible: if oil supplies on the order of two hundred exajoules a year to the world economy, perhaps eighty to ninety per cent of that still ends up in wheels, propellers and turbines; the rest of the energy system — coal, gas, nuclear, renewables — still leans on diesel for mining, construction, food logistics and maintenance. If even ten to twenty per cent of that non-oil primary energy is liquid-dependent for logistics alone, that is another forty-five to ninety exajoules a year of hidden oil demand before counting the petroleum industry’s own consumption. Transport is the visible pipe; logistics is the hidden multiplier. The point is that the pipe still matters, and that it has two failure modes: it can empty slowly, and it can snap shut.
Modern civilisation runs on net work — useful energy delivered after the petroleum industry has fed itself. For more than a century oil has been unusual among primary sources: it has been able to power its own finding, lifting, refining and distribution and still leave a surplus for everything else — when it can. Coal, gas, nuclear and renewables supply vast energy, but their chains still largely move on liquid fuels. The accounting is simple to state and uncomfortable to track. Call the gross thermal energy in a barrel at the wellhead energy gross, EG. Call everything the production chain spends finding, lifting, refining, shipping and distributing that barrel energy to produce, ETP — including the steel in the rigs, the diesel in the trucks, and the submerged support system that keeps the pipe open: roads, finance, medicine, and the military logistics that secure sea lanes. What reaches the rest of the economy — the energy delivered, ED — is simply ED = EG − ETP. When ETP approaches EG, the gauge reads empty even if pumps are still turning.
A reference barrel of conventional light crude — roughly 35° API, the sort refinery engineers use as a benchmark — carries about six gigajoules of gross energy, or on the order of 140,000 BTU per gallon. That is EG. Not all liquids in the statistics count equally. Public figures lump conventional crude with condensates, biofuels, tar-sand syncrude and refinery gains into “total liquids,” which flatters the picture. Only part of the hydrocarbon spectrum can plausibly power its own extraction and still leave a surplus — the energy window, roughly 30° to 45° API, perhaps forty per cent of the liquid spectrum. Heavy bitumen and very light condensate are chemically real; they are not interchangeable as primary fuel in a self-powered transport system. Price cannot turn a 21° API barrel into a 35° API barrel thermodynamically. Giant fields — a tiny fraction of all fields — still supply on the order of forty million barrels a day. US shale and Canadian syncrude, even at their peaks, fill only a fraction of that and often sit at the edge of or outside the window. When giants decline, volume-based substitutes do not automatically restore net work.
A barrel lifted in 1900 could deliver most of its chemical energy to the wider economy. A barrel lifted now feeds a much larger machine: deeper wells, more water pumped per barrel of oil, longer supply chains, heavier infrastructure. Marion King Hubbert — the Shell geophysicist who, in 1956, fitted US oil production to a logistic Hubbert curve and foresaw a national peak near 1970 — treated finite stocks as a rate problem, not only a reserve count. Conventional US production did crest close to his date; the later slogan peak oil names that maximum throughput, after which the easy half of a finite resource is gone even while barrels remain in the ground. Hubbert also put the thermodynamic limit plainly: when the energy cost of recovering a barrel exceeds the energy content of the barrel, production stops regardless of price. Finance can delay the reckoning but it cannot repeal thermodynamics.
Three measurable trends drive the rising bill: cumulative production (the easy fraction of each field goes first), water cut (more water lifted per barrel of oil), and depth (hotter reservoirs, more heat and disorder brought to the surface). A quick Fermi estimate on water cut: at fifty per cent water by volume, you lift roughly twice the mass per barrel of oil compared with a dry well — before you account for declining pressure or a deeper column. Industry rule of thumb: a water-oil ratio near forty to one marks economic shut-in for many fields. Depth adds temperature: a well thousands of feet down may surface fluids thirty to sixty kelvin hotter than a shallow nineteenth-century hole, and that difference enters a production-cost index linearly. Double the water mass and double the temperature difference, and the thermodynamic bill per barrel can rise by order of magnitude — which is what the historical curve shows.
I should be clear about what that index is. Chemical exergy — the work you can actually get from burning the oil — is what powers engines; the heat flowing out of a reservoir is only a proxy. The rising-cost curve I am working from treats the heat and disorder of bringing hot oil-and-water mixtures to the surface as a lower-bound gauge for how hard the production system is working, validated against a century of price and energy-cost data. I may unpack the full derivation in a later post; here I just want to highlight the engineering conclusion.
The industry may still pump record volumes — global consumption breached a hundred million barrels a day in recent years — while delivering less surplus to everything else. Track net energy rather than gross barrels. Around 2012, production cost on that gauge crossed roughly 70,000 BTU per gallon against a 140,000 gross content — half the barrel feeding the machine. From there the industry enters a Red Queen regime. The phrase comes from evolutionary biology — Leigh Van Valen’s 1973 hypothesis, itself taken from Lewis Carroll’s Through the Looking-Glass, where the Red Queen tells Alice that it “takes all the running you can do just to keep in the same place”. In nature, as prey evolve better escapes, predators must co-evolve to catch them; stop running and you face extinction. That co-evolutionary treadmill is one of the structures I work with in bioinspiration: systems that must keep adapting just to hold relative fitness. Applied to oil, it means the same thing in energy units: to keep net delivery to society from falling, the industry must raise volume faster than per-barrel surplus shrinks, or else cannibalise infrastructure, defer maintenance, and import other primary energy — gas, coal, grid electricity — to lift oil. Those other chains still lean on liquids for logistics. By the early 2020s, the model curve crossed a harder line. Chemical exergy sets a Carnot-style ceiling near 4.4 GJ of maximum theoretical work per reference barrel; after typical end-use waste, practical available work sits nearer 3.8 GJ. When production cost climbs through that band — on the order of 85,000 BTU per gallon — the average barrel stops functioning as a self-powered primary energy unit for the wider industrial world. That is the empty-gauge reading, even though engines need not stop on a particular calendar date.
What actually happened after the model crossed is the point. Pumps kept turning. Headline liquids stayed above a hundred million barrels a day. Fields better than average still delivered; mixed liquids accounting counted barrels outside the energy window; debt, strategic stocks, OPEC discipline, gas-fired extraction and grid-powered refining kept molecules moving. Borrowed work is the honest name for that subsidy stack. Price is a poor proxy for net energy under those conditions: Brent can sit in the mid-sixty dollars while the thermodynamic cost of the average barrel keeps climbing, and it can spike above a hundred and twenty-five dollars when a chokepoint closes, even though the wells did not suddenly deepen. Order-of-magnitude estimates put the share of gross energy reaching the non-oil economy near sixty per cent in 1900 at the production boundary; by the mid-2010s, after counting production cost and the further waste in real transport use, practical net fractions fall into single figures. The gauge went low; the engine kept turning on borrowed work — at rising brittleness.
Cumulative production of conventional crude follows a logistic curve with a thermodynamic asymptote on the order of two thousand gigabarrels for the self-powered fraction. We are deep on that curve. Volume-based reserve-to-production ratios count barrels; they miss net work. A trillion barrels of ultra-heavy bitumen in the ground fails as a reserve in the thermodynamic sense if each barrel costs more energy to surface than it returns.
Think of the curve as a fuel gauge. When the Corran Ferry is cancelled then one has to reschedule, borrow from neighbours, eat what is in the freezer. When Hormuz constricts you pay more, defer trips, feel the slack you never built. Civilisation does the same at scale.
That is the chronic stress: depletion as rising cost. Then there are acute stresses, which make the boundary visible in headlines.
In April 2020, global oil demand fell by roughly twenty per cent in a single month — the fastest collapse in modern data. Nothing geological happened. No giant field expired. Planes stopped. Cars sat idle. The boundary closed from the consumption side. In Ardgour the ferry schedules wobbled, tourists vanished, and Dad’s brewery felt the shock long before any spreadsheet caught up. Price crashed while the thermodynamic cost of lifting oil did not: wells did not become shallower because planes stopped flying. Price and ETP decoupled — exactly what you expect when markets clear fear and demand, not reservoir depth.
In the spring of 2026 the boundary closed from the supply side. Conflict in the Gulf interrupted the Strait of Hormuz — roughly a fifth of global oil and gas transits that chokepoint, on the order of twenty million barrels a day. Vessel traffic through the region fell by more than ninety per cent at the worst. Brent crude, which had been in the mid-sixties a barrel in January, peaked above a hundred and twenty-five in April. Diesel and jet fuel spiked harder than crude; rationing and hoarding appeared in import-dependent Asian economies. A partial ceasefire reopened the strait briefly in June; fighting resumed in July. Ardgour sits on a price signal that moves when the Gulf convulses — ferry fuel, heating oil, the diesel in the lorry that brings groceries from Inverness. A rough Fermi estimate for a four-person croft burning two to four thousand litres of diesel a year: a fifty-dollar product spike passed through in full is on the order of £1,000–2,500 a year before ferry surcharges and food freight.
Partial bypass pipelines — Saudi Arabia’s link to the Red Sea, the UAE’s Fujairah route, Egypt’s SUMED — might move on the order of ten million barrels a day between them. Hormuz normally moves about twice that. Geography will not rebalance like a spreadsheet. Producers without alternative export paths simply wait, or discount, or flare. Insurance premiums and rerouting around the Cape add real cost without lifting a single extra barrel from the reservoir — risk premium and routing on top of lifting cost. The gauge and the invoice are related instruments, not identical ones.
Chronic and acute hit the same question. Chronic stress is depletion as rising cost: Energy To Produce (ETP) climbing year by year, net delivery (Energy Delivered, ED) shrinking per barrel, the slow leak in surplus work, masked by borrowed work until it is not. Acute stress is flux interrupted — COVID from the demand side, Hormuz from the supply side — on the scale of days or weeks, when price and fear can outrun any smooth depletion curve. A techno-croft must budget for both: the long gradient and the short shock.
In the language of my last essay, oil is often the global optimum for portability and energy density — the best artefact if you ignore repair, supply lines and geology. The local optimum at the edge is messier: solar and wind where the resource exists, storage and demand-shifting, biomass where appropriate, and a few honest fossil imports where they still do irreplaceable work — ferry diesel, backup generator, the tractor on a wet November day — counted openly. Renewables are flows rather than stocks. They need collectors, grids, inverters, maintenance, and a mining chain that still largely runs on diesel. Building them restructures civilisation around different gradients — and those gradients still sit inside the open-system accounting that began this section: net work after the machine feeds itself, waste heat leaving whether you notice or not.
Nuclear belongs in that restructuring as firm electricity. Jancovici’s point lands harder once you separate electrons from liquids. A reactor displaces coal and gas on the grid; it can power workshops, heat pumps, data links and power-hungry loops such as microbial electrosynthesis when the wind is flat. It will not, by itself, put diesel in the ferry or the tractor. The local-optimum filter still applies. At croft scale you almost never run a village reactor; you draw grid electrons from a mix that may include nuclear and hydro upstream — Torness on the east coast is closer to Scotland’s energy metabolism than a speculative small modular reactor on a sea loch — plus local solar, wind and storage for the hours the line is thin. Uranium, enrichment, specialised components and long waste stewardship sit in the honest-import bucket: you will not smelt those in a shed. Construction and maintenance still lean on liquid logistics. Nuclear shrinks the electricity half of the boundary problem; the liquid half remains until mobility and machines change.
I sort imports into four buckets, mentally, when I think about a techno-croft boundary:
- Flows to localise — heat, water, some food, some power.
- Stocks to hold — fuel buffer, spare parts, medicine cabinet, seed bank.
- Imports to accept honestly — advanced chips, specialist drugs, materials you cannot make on site, and — at national scale — fissile fuel and enrichment services behind firm grid power.
- Dependencies to reduce because they are brittle — single-source parts, just-in-time food with no pantry, anything that fails the week the strait closes.
This is bookkeeping.
Minimum viable metabolism
If techno-crofting is metabolism at the scale of a community, then minimum viable metabolism is the set of flows without which people cannot stay: the place empties, or life there shrinks to crisis management rather than ordinary living.
The list is broader than survivalist handbooks admit. Calories, water, heat, and oxygen come first. Then electricity for light and tools. Waste processing. Energy storage. Spare parts. Medicines. Labour and attention— someone who understands the water system, someone who can nurse a sick child, someone who will notice the pump bearing before it seizes (this Actionable Intelligence is the foundation of a new venture that I’m advising: Crystallised AI).
Minimum viable metabolism sits below flourishing. Flourishing includes repair, health, care, governance, boredom, ritual, beauty, competence and meaning. Beer is optional biology. Fermentation, craft, sociability and shared work are deeply human — which is why the brewery keeps returning in these essays as proof that metabolism is also culture.
A post-oil world, in the sense I find useful, means lower net throughput, more visible flux, more slack — designing as if the ferry, the strait, and the marginal barrel can all fail on different timescales, while oil molecules may still appear where they earn their keep. COVID and Hormuz were unplanned stress tests. They asked, crudely, what still worked when mobility or supply convulsed. The answers were uneven: Zoom for the professional classes, empty shelves for others, diesel price for everyone who lives past the last bus stop. Read as a design lesson rather than only as a disaster, COVID also showed that a large fraction of liquid demand was optional mobility. Planes stopped and cars sat idle because people stopped going. Hormuz imposes the same cut through price. A techno-croft that waits for either shock is late. The planned version is to reduce travel on purpose — fewer Fort William runs, fewer Edinburgh hops treated as routine, more shared trips, more work that crosses as bits rather than as bodies — while thickening the place enough that staying is ordinary life.
That is where the earlier essay on techno-croft communities comes back as energy accounting. “Rooted in place, wired to the world” is an energy rule: if income and collaboration can travel as fibre or satellite, diesel need not. Export bits; import fewer passenger-kilometres. Rituals of weather, work and maintenance — the autumn week when roofs and turbines get checked before the gales, the shared harvest, the winter supper in the hall — function as trip-avoidance infrastructure: reasons to be here, skills kept local, neighbours you can borrow from when the ferry is cancelled. Cohesion is slack. Isolation is a hidden liquid bill, because lonely places send people away for work, care, entertainment and meaning, and every departure burns fuel.
I have been sketching Fermi numbers for a four-person techno-croft — I will return to a fuller pass in a later post — and the orders of magnitude are instructive even when wrong. A remote household might burn two to four thousand litres of diesel a year across vehicle, tractor, generator and ferry-dependent travel: on the order of thirty to sixty barrels, felt hard when the price doubles. At roughly six gigajoules per barrel, that is eight to twenty-five gigajoules of liquid fuel a year for mobility alone — before heating oil, before the diesel embedded in every delivery. Split those litres mentally into livelihood and habit. The tractor and the essential ferry crossing are hard. The weekly long drive that could have been a call, a shared van, or a local job is soft. Halving discretionary travel does more for net energy at the croft boundary than inventing a new fuel for the same mileage. A doubling from seventy to a hundred and forty dollars a barrel on the global market lands on the ferry fare, the heating oil invoice, the cost of bringing barley to the brewery — and on every trip you never questioned.
Electricity is a separate line. Four adults in a modest but wired croft might draw four to eight megawatt-hoursa year for lights, tools, communications and a small workshop — an order of magnitude one to two kilowattsaverage, spiking higher when someone is welding or running a freeze-dryer. Solar and storage can cover much of that where the resource exists; firm grid power — including nuclear and hydro upstream — covers winter weeks that weather alone will not. The live question is maintenance and what the electrons enable.
Protein is harder. Four adults might need roughly two hundred grams of protein each per day — call it three hundred kilograms a year. Algae or single-cell protein at pilot efficiencies might plausibly be produced on site at the cost of tens of square metres of cultivation and kilowatts of light and stirring, but only if someone maintains the system when a seal fails at dusk in February. Minimum viable metabolism is an engineering target and a social one: who carries the cognitive load when three subsystems need attention at once? Who keeps the skills alive when the person who understood the reactor moves away? Community cohesion answers the second question as much as training manuals do.
Which loops are worth closing?
Perfect circularity is a diagrammer’s fantasy. Every loop costs energy, complexity, materials and attention to close. Techno-crofting means identifying the critical loops and closing those well.
My working tiers:
Must close locally — heat resilience through insulation and storage; water security; the ability to repair critical infrastructure; a basic food and protein buffer that survives a missed delivery.
Should reduce dependence — packaging, some feedstocks, some electronics turnover, some fertiliser and nutrient imports where on-site cycling is feasible without building a cathedral of pipes; discretionary travel that community life can replace.
Accept honest imports — advanced semiconductors, specialist medicines, high-performance alloys, some fuels and tools you will not smelt in a shed; and, at the scale of the grid behind the croft, firm power whose fuel and waste you will not manage locally.
The useful question is which outside-world dependencies are thermodynamically, materially or politically too expensive to leave invisible?
Biosphere 2 is the cautionary tale I return to when circular-economy rhetoric gets too smooth. The Arizona enclosure of the early 1990s failed on closure arithmetic: oxygen levels crashed because soil microbes respired more than the designers accounted for; CO₂ seesawed; pests and politics multiplied. The project was useful anyway. It demonstrated that cycle diagrams lie gracefully on slides and badly in steel and glass. Closure is hard, measurable, and maintained by human intervention.
In a recent chat Dr Michael Merlin put the repair question wider than the pump itself: can you make or adapt the part, maintain the tool that makes the part, and keep the skills alive locally when the person who understood the subsystem leaves? Biosphere 2 is where that question lands at scale. The enclosure needed constant human tending — soil work, air management, political negotiation — because the loops were never as closed as the brochure promised. A techno-croft that closes heat and water locally but imports every microcontroller without a spare and a schematic repeats the same error at household scale.
The task is to see which old-world dependencies we have been paying for invisibly, and stop pretending they are free.
Space as stress test
NASA’s CELSS work, ESA’s MELiSSA programme, the Soviet BIOS-3 experiments, and Biosphere 2 all pursue variants of the same harsh lesson: how little can you import if you intend to stay alive? Ardgour is not Mars, and these programmes are not templates for Highland living. They are stress tests that remove the fiction of endless resupply.
Space agencies count kilowatts for life support, square metres for crops, litres of water in loop, hours of labour for maintenance. The arithmetic is unforgiving. A habitat crew cannot assume a cargo ship every fortnight. Neither can a croft assume a lorry every morning — different reasons, same pressure: how long the store lasts when resupply fails.
The Soviet BIOS-3 experiments in Krasnoyarsk are worth a footnote because they are so blunt. A small crew lived on algae and wheat grown under intense lamps, recycling air and water in a sealed chamber for months. It worked, after a fashion, because the experiment was small, heavily instrumented, and staffed by people who treated maintenance as the main job. The transferable lesson is that closure is labour-intensive before it is heroic.
In my first essay I used Mars as a comparison for constraint. I stand by that. The useful transfer is discipline.
Abundance, agriculture, and the watts underneath
ARIA — the UK’s Advanced Research and Invention Agency — has been scoping visions of manufacturing abundance and lights-out agriculture: automated greenhouses, robotic farming, factories that run with few people on site. I watch these programmes with genuine interest and a firm filter. The techno-croft test asks whether a technology improves the local optimum once materials, repair, energy, ownership and failure are counted — the same filter I argued in the essay on global and local optima.
Abundance without net energy accounting is just another sealed industrial stack. A lights-out greenhouse that produces salad at scale but cannot be understood, repaired or powered by the community it serves is a dependency with better marketing.
The more interesting edge for me remains microbial electrosynthesis — the thread from my first post. At pilot scale, electroautotrophic microbes can reduce CO₂ to simple molecules such as acetate using electricity; downstream fermenters can upgrade those molecules to protein and other products. The requirement is watts, square metres, operators who understand the biology, and an honest answer about where the electricity comes from when the wind is low and the Strait of Hormuz is closed. Firm power — including nuclear on the wider grid — makes that answer less weather-hostage without making the kilowatt-hours free. A back-of-envelope bound: if producing a kilogram of microbial protein needs on the order of ten kilowatt-hours of electricity at plausible pilot efficiencies, three hundred kilograms a year is three megawatt-hours — a significant fraction of the household’s total draw. MES routes electrical gradients into chemical food — one possible loop worth closing partially, if the boundary arithmetic works.
ARIA-scale abundance and croft-scale MES meet at the same question: what energy input enables how much useful output, and who maintains the converter? Hype answers with renderings. Techno-crofting answers with a meter and a maintenance log.
Material dialectics
Energy and material systems are ownership systems as well as technical ones — who controls the machine, who carries risk when it fails, who is allowed to modify what crosses the boundary.
Marx took the dialectic into material life: production, labour, supply, waste. I do not need a treatise on communism to note the engineering truth: a technology can look optimal because its contradictions have been exported — into mines, shipping routes, maintenance contracts, debt, care work, or ecological damage elsewhere. A diesel tractor you cannot repair because the controller is encumbered by an end-user licence agreement is a political arrangement on wheels.
The next question in this series is what legal and technical permissions cross the boundary with the energy and materials. FabLabs and open-source tooling. Convivial machines in Illich’s sense. Right to repair. Community fabrication versus consumption. I touched on Neil Gershenfeld’s FabLab vision in my essay on techno-croft communities — email the design, mill the part locally — but the permissions matter as much as the bandwidth. A tractor you cannot patch is a boundary you do not control.
That is for the next post. Here I want to leave the thermodynamic frame on the table.
Living inside visible limits
Techno-crofting begins with a simple accounting exercise: what comes in, what goes out, what degrades, what can be repaired, and who knows how to keep the flows moving? Ask that honestly and the live problem is how to live well inside visible limits — with enough energy, enough material, enough slack, and enough shared skill to keep going when abundance is no longer invisible.
McGrath warned the audience at the end of The Cheviot that the oil companies might outdo Patrick Sellar. Jancovici warns that we mistook a stock for a permanent river. COVID showed that demand can vanish in a month. Hormuz showed that supply can vanish in a week. The chronic curve and the acute shock converge on the same croft question my parents answer every time the Corran Ferry is late: what still crosses the loch, and what must wait?
What flows must cross the boundary, at what rate, in what form, and for how long, for people to live well at the edge? Thermodynamics answers the net work in each barrel. Geography answers the pipe. Firm electrons — nuclear included — shrink what electricity must ask of weather and diesel. Community answers how much mobility you demand in the first place. A techno-croft designs for all three.
This is a live essay, as always. I expect to revise it as the argument sharpens — and as the world keeps testing our assumptions faster than I can write.
If you wish to discuss any of this, or to engage on a project or a talk, you can reach me at a.a.stokes@ed.ac.uk.