For more than a decade, one of the most intriguing open-hardware projects on the internet has been the Global Village Construction Set (GVCS) from Open Source Ecology.
The basic idea is wonderfully ambitious: create an open-source collection of industrial machines that allows a relatively small community to build much of the infrastructure required for a modern standard of living.
There is a lot to admire in this idea.
But after looking at the project again many years after first encountering it, I think there is a more fundamental question we should ask:
What should a Global Village Construction Set actually contain?
Perhaps the answer isn't 50 machines.
Perhaps it isn't even 100 machines.
Perhaps the real answer is a technology tree.
The Problem With Thinking in Machines
A CNC milling machine is an impressive thing.
So is a tractor.
So is a 3D printer.
So is a cement mixer.
But none of these machines exists in isolation.
A CNC machine requires steel, bearings, shafts, motors, electrical wiring, cutting tools, lubricants, measurement equipment, fasteners, precision machining and often electronics.
And those things themselves require other industries.
A washing machine provides an even more revealing example.
A modern washing machine may contain microcontrollers, sensors, sophisticated motor drives and power electronics.
But a useful washing machine doesn't fundamentally require semiconductor technology.
A much earlier machine can be made using sheet metal, a tub, a motor, belts, bearings, switches, a pump, a heater, a thermostat and perhaps a mechanical timer.
The difference is important.
The semiconductor doesn't create the washing machine. It creates the modern electronic version of the washing machine.
That distinction should be at the heart of a realistic GVCS.
The Civilization Bootstrap Problem
Imagine something extreme.
Civilization takes a technological U-turn.
We retain our accumulated scientific knowledge, books and perhaps some surviving equipment, but the global industrial system disappears.
There are no functioning semiconductor factories. No large steel industry. No international supply chain. No container ports. No convenient global marketplace for replacement bearings.
There is a village with:
- Land
- Forests
- Water
- Sunlight
- Clay
- Stone
- Sand
- Plants
- Animals
- Some mineral deposits
- Human knowledge
What should that village build first?
A CNC machine?
Probably not.
A 3D printer?
Probably not.
A semiconductor fabrication line?
Obviously not.
The first priority is something much more fundamental:
Turn abundant natural resources into increasingly capable tools.
That gives us a ladder:
Resources → materials → processes → tools → machines → industries → advanced technology.
That is the real civilization construction set.
Start With What Nature Gives You
The first GVCS layer should therefore not contain machines at all.
It should contain resources and resource-processing knowledge.
Wood
Wood can become:
- Buildings
- Beams
- Handles
- Carts
- Wheels
- Scaffolding
- Furniture
- Boats
- Patterns
- Molds
- Tools
- Charcoal
- Paper
- Fuel
And unlike an underground mineral deposit, properly managed forests can regenerate.
Clay
Clay can become:
- Pottery
- Bricks
- Roof tiles
- Pipes
- Crucibles
- Furnace linings
- Refractory materials
- Ceramics
- Electrical insulators
Sand
Sand can eventually become:
- Glass
- Foundry material
- Concrete aggregate
- Silicon feedstock
Limestone
Limestone can become:
- Lime
- Mortar
- Plaster
- Cement-related materials
- Agricultural amendments
- Chemical feedstocks
Plants
Plants provide:
- Food
- Fibres
- Oils
- Dyes
- Resins
- Fuel
- Paper feedstock
- Chemical feedstocks
Ores
Ores provide the foundation for:
- Iron
- Copper
- Aluminium
- Other metals
The first lesson is therefore obvious:
The GVCS should begin with the material environment, not with the machine catalogue.
Capability #1: Make Things From Clay
Pottery is a perfect example of what a capability-based GVCS should look like.
A manual doesn't need to turn everyone into a master potter.
Instead, it needs to preserve the capability chain:
Clay → preparation → shaping → drying → firing → ceramic object
The practical skills can be learned through demonstration and apprenticeship.
But the GVCS should preserve:
- How to identify suitable clay
- How to process it
- How to build a basic kiln
- How to prepare fuel
- How to shape objects
- How to dry them
- How to fire them
- How to test the finished material
Then pottery unlocks much more than cooking vessels.
It produces:
Pots → storage → pipes → tiles → bricks → crucibles → refractory components → electrical insulation
Suddenly "pottery" becomes an industrial capability.
Capability #2: Make Charcoal
This might be one of the most important technologies in the entire system.
Wood → controlled heating → charcoal
Charcoal provides a high-temperature fuel that can support early metallurgy.
Then:
Charcoal → furnace → iron → tools
And:
Better tools → better furnace → better metalworking
The system begins feeding itself.
This is what a real GVCS should seek:
Every capability should, wherever possible, help create the next capability.
Capability #3: Make Lime
Limestone looks boring.
A lime kiln looks even more boring.
But lime is one of those technologies that quietly supports civilization.
Limestone → kiln → lime
Lime can support:
- Mortar
- Plaster
- Masonry
- Construction
- Soil treatment
- Water treatment
- Various chemical processes
Again, the important thing isn't the lime itself.
It is the capability unlocked by being able to make lime locally.
Capability #4: Make Glass
Another apparently simple technology becomes surprisingly deep.
Sand + suitable minerals + heat → glass
Glass gives us:
- Windows
- Bottles
- Storage
- Laboratory vessels
- Optical components
- Thermometers
- Lenses
- Eventually sophisticated scientific instruments
A village that can make glass has crossed an important technological threshold.
Capability #5: Make Rope and Textiles
Consider:
Plant → fibre → thread → rope
And:
Plant → fibre → thread → cloth
Rope enables:
- Lifting
- Construction
- Transport
- Sailing
- Agriculture
- Machinery
- Wells
Textiles enable:
- Clothing
- Bags
- Sails
- Filters
- Belts
- Canvas
- Insulation
Again, none of these is a "machine."
Yet removing these capabilities from a civilization would cripple it.
Capability #6: Make Leather
Animal hides can become leather through processing and tanning.
Leather provides:
- Belts
- Footwear
- Gloves
- Bags
- Harnesses
- Seals
- Gaskets
- Flexible machine components
This is a good example of something that modern industrial civilization hides from us.
We don't think about the underlying capability because we buy the finished product.
A civilization rebuilding itself cannot afford that luxury.
Capability #7: Make Paper
Paper might be even more important than it initially appears.
Fibre → pulp → paper → documentation
And documentation is itself a technology.
A civilization that cannot preserve technical knowledge between generations is vulnerable to losing capabilities it has spent decades developing.
So the GVCS should include not only:
How to build a machine
but:
How to preserve knowledge about the machine.
That means paper, printing, measurement standards, diagrams, technical drawing and eventually digital archives.
The Next Great Leap: Metallurgy
Once the community can reliably produce charcoal, ceramics, lime and basic tools, metallurgy becomes the major accelerator.
The sequence might look approximately like:
Ore → furnace → crude metal → forging → tools → better tools → better furnace → better metal
Then:
- Casting
- Forging
- Sheet production
- Wire
- Fasteners
- Springs
- Bearings
- Gears
- Shafts
Now we have crossed from a craft civilization into an industrial one.
And this is where the GVCS starts becoming much more recognizable.
The Machine-Tool Revolution
Before CNC, there is something even more fundamental:
The ability to make accurate machines.
That means:
- Workbenches
- Vices
- Files
- Saws
- Drills
- Measuring tools
- Straightedges
- Squares
- Calipers
- Micrometers
- Drill presses
- Lathes
- Milling machines
- Grinders
- Shapers
- Boring machines
And this introduces another capability that deserves much greater prominence:
Metrology
You cannot have precision engineering without measurement.
A civilization needs to know:
- Is this shaft really round?
- Is this surface actually flat?
- Are these two parts interchangeable?
- Is this hole the correct diameter?
- Is this gear correctly made?
The progression is therefore:
Make → measure → correct → standardize → reproduce.
Only after this foundation exists does CNC become truly transformative.
So Where Does CNC Belong?
Definitely in the GVCS.
But not near the beginning.
A CNC machine is a multiplier of an existing industrial ecosystem.
It doesn't replace the ecosystem.
A useful hierarchy might therefore be:
Level 1 — Craft
Woodworking, pottery, weaving, rope, leather and basic construction.
Level 2 — Thermal and Material Processing
Charcoal, kilns, lime, glass, ceramics and metallurgy.
Level 3 — Mechanical Engineering
Gears, shafts, bearings, pumps, mechanical power and machine tools.
Level 4 — Electrical Engineering
Generators, motors, transformers, wiring, switches and heaters.
Level 5 — Industrial Automation
CNC, sensors, control systems and robotics.
Level 6 — Electronics
Transistors, integrated circuits, microcontrollers and power electronics.
Level 7 — Semiconductor Industry
Silicon processing, wafers, lithography, packaging and advanced electronics.
The mistake would be to assume that because Level 7 is necessary for a modern smartphone, Level 7 is necessary for civilization.
It isn't.
The Surprising Amount of Modern Life We Can Build Without Semiconductors
Consider a community that has reached a mature mechanical and electrical industrial base.
It could potentially produce versions of:
- Washing machines
- Dishwashers
- Refrigerators
- Fans
- Pumps
- Electric motors
- Generators
- Sewing machines
- Ovens
- Water heaters
- Agricultural machinery
- Tractors
- Machine tools
- Bicycles
- Construction equipment
These wouldn't necessarily be as efficient, compact, quiet or intelligent as their 2026 equivalents.
But they would be useful.
That distinction matters enormously.
The objective of a civilization bootstrap system should not be:
"Reproduce every product currently sold by Samsung, Bosch, Apple and Caterpillar."
It should be:
"Recover the ability to perform the underlying functions."
The Dishwasher Test
Take a dishwasher.
A modern dishwasher may contain a sophisticated electronic control board.
But its fundamental function is:
Water + heat + detergent + mechanical spraying + drainage + timing.
A primitive automatic dishwasher could therefore be constructed from:
- Metal
- Glass or ceramic
- Pump
- Motor
- Heater
- Pipes
- Valves
- Spray arms
- Switches
- Mechanical timer
No microprocessor required.
This is a powerful test for our proposed GVCS.
Instead of asking:
"Can the village build a modern dishwasher?"
we ask:
"At what point in our technological tree can the village build a useful dishwasher?"
Perhaps the answer is surprisingly early.
The Capability Tree
This suggests that the real GVCS should look more like a giant technology tree.
NATURAL RESOURCES
│
├── WOOD
│ ├── Lumber
│ ├── Charcoal
│ ├── Paper
│ ├── Pulp
│ └── Wooden machinery
│
├── CLAY
│ ├── Pottery
│ ├── Bricks
│ ├── Tiles
│ ├── Pipes
│ ├── Crucibles
│ └── Refractories
│
├── STONE / LIMESTONE
│ ├── Building stone
│ ├── Lime
│ ├── Mortar
│ └── Plaster
│
├── SAND
│ ├── Glass
│ ├── Foundry materials
│ └── Silicon feedstock
│
├── PLANTS
│ ├── Food
│ ├── Fibre
│ ├── Rope
│ ├── Textiles
│ ├── Oil
│ └── Chemicals
│
└── ORES
├── Iron
├── Copper
├── Aluminium
└── Other metals
│
▼
METALLURGY
│
▼
HAND TOOLS
│
▼
MACHINE TOOLS
│
┌─────┴─────┐
▼ ▼
MACHINERY ELECTRICAL
│ │
└─────┬─────┘
▼
INDUSTRIAL BASE
│
▼
CNC
│
▼
ELECTRONICS
│
▼
SEMICONDUCTORS
│
▼
COMPUTING
This is far more interesting to me than a list of 50 machines.
Products Become Outputs, Not the Foundation
Once we have the capability tree, we can place ordinary products on top of it.
Washing Machine
Requires:
- Steel
- Bearings
- Electric motor
- Copper wire
- Insulation
- Pump
- Seals
- Switches
- Heater
- Mechanical or electrical control
If those capabilities exist, washing-machine manufacturing becomes possible.
Refrigerator
Requires:
- Sheet metal
- Insulation
- Compressor
- Electric motor
- Refrigerant
- Heat exchangers
- Seals
- Thermostat
Again, no semiconductor fabrication is fundamentally required.
Tractor
Requires:
- Steel
- Engine
- Bearings
- Gears
- Hydraulics
- Tyres
- Electrical system
- Machine tools
Much more difficult, but still firmly inside the mechanical-industrial tree.
Computer
Now the situation changes dramatically.
We need:
- Semiconductor devices
- Precision electronics
- Memory
- Processors
- PCB manufacturing
- Displays or equivalent output
- Storage
- Software
The computer therefore sits much higher in the dependency tree.
The Difference Between Open Design and Open Capability
There is an important distinction between:
Open-source product design
and
Open-source civilization infrastructure.
A project can publish the CAD files for a CNC machine.
That is useful.
But suppose those files require:
- Imported precision bearings
- Imported servo motors
- Imported electronics
- Imported cutting tools
- Imported steel
The design is open.
The capability isn't.
That's not necessarily a criticism of open-source hardware. Open-source hardware remains extremely valuable even when it depends on global supply chains.
But it isn't quite the same thing as a Global Village Construction Set.
A true civilization-oriented system should explicitly show:
Which dependencies remain external?
Three Levels of Openness
We could therefore classify every capability into three levels.
Level A — Open Design
The plans are available.
Level B — Open Manufacture
The community can manufacture the object using locally available industrial inputs.
Level C — Open Material Chain
The community can manufacture those industrial inputs itself.
This distinction is crucial.
A locally manufactured washing machine assembled from imported motors might be Level B.
A village capable of making the motor, copper wire, steel, bearings and insulation locally is approaching Level C.
A village capable of producing the machines that make those components is approaching industrial self-reproduction.
The Real Target Should Be Self-Reproduction
And here we can recover one of the most powerful ideas from the original GVCS.
Open Source Ecology has long envisioned the machines as a product ecology, rather than simply unrelated machines. The concept includes progressively moving from externally sourced components toward greater local production of components and materials.
That is exactly the right direction.
But the system needs to be expanded.
Instead of:
50 machines
we should think:
Resources → processes → materials → components → machines → products → better machines.
The end goal is not a collection of open machines.
It is a self-expanding industrial ecosystem.
A Realistic GVCS Might Have 200 Capabilities but Fewer Machines
This sounds paradoxical.
But it isn't.
One kiln could support:
- Pottery
- Bricks
- Tiles
- Lime-related processes
- Glass-related processes
- Metallurgy
- Refractory materials
One forge could support:
- Tools
- Agricultural implements
- Fasteners
- Machine components
- Repair work
One lathe could support:
- Shafts
- Bushings
- Pulleys
- Machine repairs
- Replacement parts
One electrical workshop could support:
- Motors
- Generators
- Wiring
- Transformers
- Appliances
The machine count is less important than the capability graph.
A New GVCS Could Have Six Layers
Layer 1 — Resources
Wood, clay, stone, sand, ores, plants, water and biomass.
Layer 2 — Processes
Pottery, charcoal, tanning, weaving, glass, lime, forging, casting and pressing.
Layer 3 — Tools
Hand tools, kilns, looms, presses, pumps and saws.
Layer 4 — Machines
Lathes, mills, generators, motors, tractors, appliances and other industrial machines.
Layer 5 — Industries
Metallurgy, ceramics, chemicals, electrical engineering and machine manufacturing.
Layer 6 — Advanced Technology
CNC, robotics, electronics, semiconductors and computing.
And then there is a seventh layer that may ultimately be the most important:
Layer 7 — Knowledge
The GVCS must preserve:
- Drawings
- Measurements
- Process parameters
- Material specifications
- Repair procedures
- Testing methods
- Failures
- Alternative designs
- Local substitutes
- Manufacturing knowledge
- Educational material
Because the ultimate product of the system isn't a tractor.
It is accumulated capability.
The "Thingy" Principle
This also solves an issue that initially seems impossible.
How do we put something like pottery into an engineering manual when pottery requires hands-on skill?
We don't pretend the manual can replace apprenticeship.
Instead, we put the capability specification into the manual.
For every capability:
- Inputs: What do I need?
- Process: What must I know how to do?
- Tools: What equipment helps?
- Outputs: What can I make?
- Quality: How do I know it worked?
- Failures: What commonly goes wrong?
- Unlocks: What does this capability make possible?
- Next level: What more advanced capability can replace it?
This could apply to hundreds of things.
Capability Card: Pottery
Inputs: Clay, water, fuel.
Tools: Basic hand tools, moulds or wheel, kiln.
Process: Clay preparation → forming → drying → firing.
Outputs: Ceramic vessels, tiles, pipes, bricks, crucibles and refractory components.
Quality tests: Strength, porosity, dimensional stability and thermal resistance.
Unlocks: Food storage, construction, metallurgy, chemical processing and electrical insulation.
Next level: Controlled ceramic processing.
Capability Card: Charcoal
Inputs: Wood and controlled oxygen supply.
Tools: Kiln or charcoal pit.
Process: Controlled thermal decomposition.
Outputs: Charcoal and useful heat.
Unlocks: High-temperature metallurgy.
Next level: Controlled industrial carbon production.
Capability Card: Rope
Inputs: Plant fibres.
Tools: Fibre-processing tools, spinning and twisting equipment.
Outputs: Cordage, rope and nets.
Unlocks: Lifting, transport, construction and mechanical transmission.
Capability Card: Lime
Inputs: Limestone and fuel.
Tools: Lime kiln.
Outputs: Quicklime and hydrated lime.
Unlocks: Mortar, plaster, construction and chemical processes.
Now multiply that by 100 or 200.
That becomes a civilization manual.
AI Could Make This Version Dramatically More Achievable
This is where modern AI becomes interesting.
AI doesn't need to magically invent 200 machines.
It can help organize humanity's existing knowledge into a capability graph.
Imagine feeding an AI:
- Historical engineering manuals
- Agricultural manuals
- Traditional crafts
- Machine-shop handbooks
- Metallurgy texts
- Chemistry references
- Open-source hardware projects
- Old industrial manuals
- Repair manuals
- Archaeological knowledge
- Modern engineering data
Then asking:
"What is the minimum prerequisite chain for producing a working electric motor using locally available resources?"
The AI could produce a dependency graph.
Then:
"What capabilities are missing?"
Then:
"What is the simplest version of each missing capability?"
Then:
"Can an existing machine produce the tools needed to build the next machine?"
That is exactly the kind of problem AI is unusually well suited to.
The Ultimate Metric
I would therefore abandon a simple:
"GVCS completion percentage."
Instead I'd measure something like:
Civilization Capability Coverage
For example:
- Food production — percentage of required capabilities available
- Basic construction — percentage of required capabilities available
- Mechanical manufacturing — percentage of required capabilities available
- Electrical generation — percentage of required capabilities available
- Chemical industry — percentage of required capabilities available
- Precision manufacturing — percentage of required capabilities available
- Electronics — percentage of required capabilities available
- Semiconductor manufacturing — percentage of required capabilities available
The numbers would only be meaningful if rigorously defined, of course.
But conceptually, this is much more informative than saying:
"37 out of 50 machines are complete."
Because the question becomes:
What can this civilization actually do?
The 1,000-Year Test
This is where the thought experiment becomes particularly useful.
If civilization had a thousand years to rebuild itself, I wouldn't want it to begin by trying to recreate a 2026 factory.
I'd want it to build a ladder:
Woodworking
↓
Kilns
↓
Charcoal
↓
Ceramics
↓
Lime
↓
Glass
↓
Metallurgy
↓
Iron tools
↓
Machine tools
↓
Mechanical power
↓
Electricity
↓
Motors
↓
Industrial machinery
↓
Precision manufacturing
↓
CNC
↓
Electronics
↓
Semiconductors
↓
Computers
↓
Advanced automation
At each stage, the civilization becomes better at producing the next stage.
That is technological compounding.
And that is what a Global Village Construction Set should really be about.
The Real Goal: Not Recreating Civilization, But Making Civilization Reproducible
There is a subtle but profound difference.
The original GVCS asks us to imagine a collection of machines capable of supporting a small modern civilization.
The improved concept asks something deeper:
Could a small community start with ordinary natural resources and progressively recreate the technological capabilities of civilization without requiring the entire existing global industrial system?
If the answer is yes, then the system has achieved something extraordinary.
It doesn't need to start by manufacturing a smartphone.
It needs to start by making a pot.
Then a brick.
Then a kiln.
Then charcoal.
Then iron.
Then a better tool.
Then a lathe.
Then a motor.
Then a generator.
Then a washing machine.
Then a CNC machine.
Then, perhaps eventually, a semiconductor fab.
The first pot and the eventual semiconductor are part of the same technological tree.
That is the GVCS I would like to see.
Not simply an open-source collection of machines.
Not a collection of CAD files.
Not a catalogue of futuristic diagrams.
But a public, open, continuously improvable map of human technological capability — from local clay to civilization-scale industry.
And perhaps the most important feature would be this:
Every generation should leave the next generation with more capabilities than it inherited.
That is a much more realistic definition of Global Village Construction.
A Possible Structure for the Project
If this were developed as a real open-source project, I would organize it into six major repositories or layers:
| Layer | Contents |
|---|---|
| 01 — Resources | Wood, clay, stone, sand, ores, plants, water, biomass |
| 02 — Processes | Pottery, charcoal, tanning, weaving, glass, lime, forging, casting |
| 03 — Tools | Hand tools, kilns, looms, presses, pumps, saws |
| 04 — Machines | Lathes, mills, generators, motors, tractors, appliances |
| 05 — Industries | Metallurgy, ceramics, chemicals, electrical, machine manufacturing |
| 06 — Advanced | CNC, robotics, electronics, semiconductors, computing |
Every entry would contain:
Prerequisites → procedure → tools → inputs → outputs → quality tests → failure modes → substitutes → downstream capabilities → more advanced version.
That, in my view, is where the original Global Village Construction Set idea could evolve into something considerably more ambitious — while paradoxically becoming more grounded and less futuristic.
The goal isn't to give a village a 3D printer.
The goal is to give a village a path from a lump of clay to the ability to build its own 3D printer.
