Self-Replicating Machines and the Path to Material Abundance

This article serves futurism and the move from scarcity toward abundance and post-scarcity. It asks what it would take for tools that make tools to bring the price of physical goods down toward the cost of energy and raw materials, and what people can do now. The answer offered here is that full self-replication, the old dream of a machine that copies itself completely, is the wrong near-term target. Partial self-replication plus open designs is already lowering costs, the real bottleneck is closure on the hard parts, and full replication matters most in space.

The idea in brief

A self-replicating machine is a machine that can build a working copy of itself. The mathematician John von Neumann sketched a kinematic version in the middle of the twentieth century: a constructor that picks parts from a supply and assembles a duplicate. The economic appeal is that self-copying factories could multiply exponentially, and the price of their output could fall toward the cost of energy, raw materials, and running time.

The RepRap paper in Robotica (Jones and colleagues, 2011) makes a useful distinction. An autotrophic self-replicator makes a copy of itself from raw materials without help. An assisted self-replicator has some, but not all, of the needed subsystems and relies on people or outside suppliers for the rest. The authors state that no artificial autotrophic self-reproducing kinematic machine had yet been made, and they describe RepRap as an assisted self-replicating machine.

Why full replication is hard

The kinematic model of von Neumann assumed that parts were supplied free to the machine. The NASA study discussed below notes that this hides the real problem, closure, meaning the question of whether a manufacturing system can supply itself with all of the materials, parts, and assembly techniques that it needs.

The NASA team described two layers of this problem. Parts closure means the system must make every component of its own fabrication machines. Materials closure means it must also extract and prepare every chemical element, process chemical, and alloy from ordinary rock or soil. The team expected the materials step to be the most difficult of all.

The RepRap paper notes that Freitas and Merkle, in Kinematic Self-Replicating Machines (2004), identified 137 design properties in their map of the replicator design space, each a place where a design can fall short of full autonomy.

What partial replication already achieved

RepRap is the clearest working example. It is a free, open-source 3D printer that prints many of its own plastic parts, so one machine can make a kit for another. The 2011 paper reports these figures for the early designs:

RepRap printed 48 percent of its own parts by count, ignoring fasteners. Counting nuts, bolts, and washers, fasteners made up 73 percent of the parts and the printed share dropped to 13 percent.
Some builders replaced rolling bearings with plain bearings, which raised the self-made share to 57 percent.
Materials to build the second version, Mendel, cost about 350 euros, against about 500 euros for the first version, Darwin. At the time, the lowest cost non-open-source rapid prototyping machine cost about 9,900 US dollars.
At the start of 2008 there were four RepRap machines. Two and a half years later the authors estimated about 4,500 derived machines worldwide.

The RepRap project now describes itself as the origin, directly or indirectly, of the great majority of 3D printers made in the world. It also cites a life-cycle study by Wittbrodt and colleagues finding that a household printing only 20 products a year could save between 300 and 2,000 US dollars.

The lesson is that a machine did not need to copy itself completely to start a large price drop. Open designs and a partial replication fraction were enough.

Open Source Ecology applies a similar approach to heavier equipment. Its Global Village Construction Set aims at open designs for 50 industrial machines that it considers necessary for a small, sustainable civilization with modern comforts, made at a fraction of commercial costs. The project built its first machine, a compressed earth brick press, in 2007, and reported in 2018 that the set was about one third complete.

The vitamins problem

RepRap builders use the word vitamin for any part a RepRap needs but cannot make. The project list includes fasteners, grease, microcontrollers and other chips, wiring and electronic components, circuit boards, motors, the power supply, and steel rods.

The NASA study used the same word. It reported earlier findings that 90 to 96 percent closure looked attainable in specific nonreplicating space production applications, with the remaining 4 to 10 percent supplied as vitamin parts such as microelectronics, ball bearings, and precision instruments. It also found that the gains in closure above 90 percent come at a great price, requiring from one to three orders of magnitude more raw material to be processed. The study concluded that it may be more economical to settle for 90 to 95 percent closure and import the rest.

This is the core of the abundance question. Plastic frames are easy. Chips, motors, bearings, and refined materials are hard, and raising the replication fraction on those parts is where the largest remaining cost drops would come from.

Why space is where full replication matters most

In 1980 NASA and the American Society for Engineering Education held a summer study at the University of Santa Clara, published by NASA in 1982 as Advanced Automation for Space Missions (NASA CP-2255). One mission concept was a self-replicating lunar factory. A 100-ton seed factory would land on the Moon, use only local resources, and grow into a system able to copy itself.

The study assumed a seed that produces 100 tons a year of the same materials of which it is made. Under that assumption it calculated that output could grow to about one million tons a year within ten years.

On Earth, a factory can buy a missing bearing nearby. On the Moon, every imported part must be launched from Earth, so each point of closure is worth far more. That makes full replication a space problem first.

What this means for abundance

A practical path toward lower material prices looks like this:

Raise the replication fraction of common tools step by step, as RepRap builders did by switching to plain bearings.
Keep designs open so improvements spread and many people can build and repair the machines.
Attack the vitamins one at a time, starting with the parts that are heavy, simple, or locally available, and accept imports for chips and precision parts for now.
Pair local production with recovered materials and low-cost energy, which set the price floor.

What people can do now:

Build, use, or repair an open-source 3D printer and print replacement parts instead of buying new goods.
Contribute designs, documentation, or testing to open hardware projects such as RepRap or Open Source Ecology.
Learn one fabrication skill that addresses a vitamin, such as motor winding, circuit board making, or bearing design.
Support repair rights and open standards so that machines remain fixable.

Honest limits and open questions

The figures above come from early RepRap designs and a 1980 study. Modern designs differ, and the lunar growth numbers are calculations under stated assumptions, with no factory built to test them. The slow progress of Open Source Ecology shows that open design does not remove the labor of engineering and documentation. Low material prices also depend on energy prices and on access to raw materials. Exponentially growing machines would also raise political, social, and economic questions, which the NASA study treated as open issues.

Discussion questions, essay ideas, and learning related AI prompt ideas

Which vitamin part, if it could be made locally from open designs, would lower the cost of home fabrication the most, and why?
Write an essay comparing assisted and autotrophic self-replication and arguing which matters more for abundance on Earth in the next twenty years.
Explain the NASA finding that the last few percent of closure costs far more than the first ninety percent, and discuss what that implies for factory design.
Ask an AI to list the parts of a common open-source 3D printer and sort them into printable parts and vitamins.

References

Self-replicating machines and the path to abundance

Contents