Section 2.0 - Concept History and Development
In this section we start looking more deeply into self-expanding systems in general, and the seed factory concept in particular. Expansion, growth, and replication have a long history as natural processes. More specifically, they also have a long history in human culture. We take a look at their histories, leading up to their embodiment in modern technology. Then we look at how these processes and concepts can be included in our designs, and combined to form the current seed factory approach. In Section 2.1: Technology Development we look at the current state-of-the-art, in terms of existing knowledge and experience, related projects and development, open questions, and future research and development needed to implement the seed factory approach.
Expansion, growth, and replication are natural processes which existed long before humans were around to think about them. They are most evident in the natural sciences of Astronomy and Biology. We can draw on ideas from these fields to design self-expanding factories.
Expansion as a process appears to be as old as the Universe itself. Red-shifts of spectra from distant galaxies, distant supernovae brightness and light curves, and the Cosmic Microwave Background radiation imply that space itself is expanding. This seems to be the combined effect of strong repulsive forces from Cosmic Inflation during the Big Bang ~13.8 billion years ago, and Dark Energy, which continues to accelerate the expansion. If the inflation theory of cosmology is correct, the growth of large scale structures, now hundreds of millions of light years in size, can be traced to microscopic quantum fluctuations during the inflationary era.
Galaxies, black holes, and stars grow by Accretion, the addition of more matter to what is already there. Since the force of gravity is linear with mass, the ability to attract more matter is proportional to the mass already collected. This results in an exponential growth process, until all the nearby matter is consumed. Any growth process which is proportional to the current size leads to exponential increases over time, and we will see many examples of this kind growth in the course of the book.
Stars replicate in a limited sense. As faster-aging large stars reach the end of their life, they emit large amounts of gas or explode. The shock waves from these processes can create high density areas in molecular clouds. These denser areas can collapse to form a new generation of stars. Our own Sun is thought to be a third generation star whose formation may have been triggered by a supernova explosion of a nearby older star. Planets also form by a process of exponential growth from smaller bodies. A growing planet has a stronger gravitational field, so it can attract more gas, dust, and other small bodies to itself. It is not evident how astronomical processes could be used in current factory design, but we can incorporate the concepts of expansion, replication, and exponential growth.
It's estimated that life began on Earth about 3.8-4.1 billion years ago, soon after the surface cooled from the planet's formation and water accumulated. Growth and reproduction are two of the defining features of Life. Another defining feature of life is Metabolism, the use of available energy to convert raw materials into parts of the organism. Most of the energy for life ultimately comes from the Sun, and living things incorporate local materials (food) into themselves. Waste products and dead organisms are recycled by other organisms, and along with nonliving parts of the environment form an Ecosystem of related elements. We can also incorporate these ideas of energy and material sources, recycling, and ecosystems of related elements into our designs.
Small organisms like bacteria and algae consist of a single cell which divides repeatedly. Large organisms like whales and trees consist of many cells, which individually divide, then become specialized for different tasks and grouped into larger structures. In addition, large organisms can incorporate non-living matter, like the mineral parts of bones or the wood and bark of trees. The replication of a large organism involves partitioning one or a small number of cells into an Embryo or Seed, a smaller and simpler entity than the mature organism. This entity contains all the information to build the mature organism, in the form of long double-chain polymer molecules known as DNA (Deoxyribonucleic Acid), usually after mixing genetic data from two parents. The initial cells then divide and specialize until they become a mature copy of the parent organism. Humans, being large organisms, grow and replicate in this way, both at the cellular and whole organism level. The process of growth from an embryo or seed is at least 400 million years old.
We can apply the ideas of specialization during growth, replication via a smaller and simpler starter entity (i.e. a seed), and replication of information into our designs. Because so many features from biology can be included into a self-expanding factory, we can consider it a form of Artificial Life.
In Human Culture
The natural processes of expansion and replication long predate modern humans, who have only existed for about 200,000 years. As living beings we continue to participate in these processes, and they continue to operate in nature. Humans also produce Culture, which are entities that persist beyond the death of individual people. Human culture also displays growth and replication processes. Today we can purposely use such expansion and replication processes in our cultural artifacts, like factories, to meet modern needs.
- Cultural Elements
Culture includes both non-material elements, and Material Culture. Non-material culture includes things like patterns and knowledge. For example, an early Hunter-Gatherer band could persist as a pattern for a much longer time than any member lived. Language and how to use it is knowledge that can be passed down through many generations orally, without ever being written down. Material culture, like tools and cities, if maintained, can also last well past the life of any individual. The life span of an artifact, however, does not determine if it is cultural. The important feature is whether it exists independent of human biology. For example, a fast food soft drink cup may only be used for a few minutes, or sit on a shelf for years, but our internal biology does not affect these lifetimes.
All kinds of cultural elements can exhibit growth and replication. A hunter-gatherer band can grow to the point it divides into two bands. A language can grow by the addition of new words, and by new people adopting it. New languages can form when people live far enough from each other that they rarely communicate. Shifts in vocabulary and usage then lead to enough changes that there are two distinct languages instead of one. Widespread distribution of media like books and video have tended to stabilize language content, and worldwide communications means nobody is too far away any more. Although the trend is now for small languages to disappear in favor of the more common ones, the remaining languages continue to grow by adding new words and speakers. Cities can grow by physically adding to an existing city, and historically they reproduced by setting up colonies in new locations.
The oldest elements of culture that showed growth and replication predate modern humans themselves. They include Band Society, and early types of technology in the form of hunting, foraging, and tool-making skills. Species of the genus Homo, of which modern humans are the only surviving member, expanded repeatedly across the world over the past few million years. Successful social groups therefore presumably divided when their population became larger than a given area could support. Skills could grow by accumulation of details and technique and by specialization. Each person only has a limited memory and time to learn. So specialization allows more total skills to be passed on. Cultural replication included making new copies of tools from raw materials, but not generally using an existing tool to copy itself. Instead, a collection of tools was brought to bear on making copies of its members.
In a broad sense, Technology includes all the knowledge, skills, and artifacts used towards practical ends. So a tool like a hammer is a technological artifact, but knowing how to make and use hammers is a part of technological knowledge. The oldest technologies of all are foraging and nest building using nothing more than hands and feet. Other species of the taxonomic family of Great Apes (the Hominidae) do this. Presumably so did our hominid ancestors millions of years ago. We can speculate about using sticks and unmodified rocks to knock down fruit or kill prey, but those kinds of tools don't leave lasting evidence of their existence. The oldest identifiable tools are split and flaked stones from Kenya that are 3.3 million years old. These types of tools were not used to make more tools of the same kind. If you wanted to copy them, you needed to start fresh with new rocks. The skills and knowledge, however, could be copied by observation and imitation, even if language did not exist yet. The original stone tools could be used to make secondary tools. For example hand choppers and scrapers made from split rocks can be used to cut and peel branches, which can then be used as clubs, or handles for stone axes.
The earliest use of tools to make more tools may date from using Digging Sticks and Hammerstones to excavate and flake stones, which in turn can cut and sharpen more digging sticks. Over time, hand tools such as these started to be used to make better tools and artifacts, a process which still continues today. Eventually, specialized groups of toolmakers developed, such as blacksmiths, who were able to copy their own tools plus make tools for other trades. We can make a distinction between the toolmakers - those who can make tools, and nearly everyone else - who use tools made by others. Modern society is filled with tools, from kitchen utensils to smartphones. The vast selection tools and machines found throughout civilization can all trace their existence to the self-making toolmakers.
Early blacksmiths could copy tools for themselves and others, but their work was not mechanized and certainly not automated. Animal, water, and wind power later augmented human muscles, and increased total production. More recently steam and electrical power further improved production rates. As the scale of production and the knowledge involved grew, toolmakers also became specialized. The mining industry, steel mills, foundries, and machinery builders still retain the collective ability to make more tools and machines of the type they themselves use. So as a group they are still able to replicate themselves, and make tools and machines for others.
Agriculture is the part of technology concerned with purposely growing living things. That living things can grow and make copies of themselves from seeds or by giving birth was evident to people at least as long as agriculture has been used, which is about 12,000 years, and was probably known before that. Understanding the detailed mechanisms only began a few hundred years ago, including replication of DNA, division of cells, and all the supporting biochemical elements. Biological seeds usually contain the plant embryo, which does the actual growing, a supply of nutrients to support the initial growth, and a protective coating. But long before understanding the details of how they worked, people knew seeds grew into full sized plants, which in turn produced more seeds. This is known as the Biological Life Cycle.
Because each plant generally produces many seeds, their number can grow exponentially until they are limited by competition or available resources. So any natural landscape which can support them tends to be covered in plants. Total oceanic Primary Production approximately equals that on land, but is more thinly distributed because the water absorbs sunlight before it can be used. It also consists of more microscopic organisms on a relative basis due to lack of solid anchorage in deep waters. Once people learned to exploit the supply of nutrients, in particular plants and animals, and started raising them intentionally, the natural ability of living things to grow exponentially increased the human food supply dramatically. In turn, human population also grew exponentially. The inventiveness per person of modern humans has probably not changed much in the past 200,000 years. But with vastly more people today, and better ways to pass on knowledge, the accumulation of new knowledge and culture has also grown exponentially.
- Starter Sets
Starter sets have occurred throughout human history, especially when settling new areas. At first, these were a hand-carried set of tools to be used in a new location to hunt, chop, and grind. The need to carry them limited their total mass. Once more complex levels of technology had developed, colonizing a new location via ship or land travel included bringing along a larger starter set of plant seeds, animals, tools, and an inventory of finished items like nails, which could not be made immediately. Once they arrived, people set about clearing, planting, and building, eventually reaching the point they could make their own tools. Until then, they traded with their original location for items they could not yet make. The contents of a starter set has varied by era, but it was well understood that you needed one to survive and flourish in a new location. In the modern world, the idea of starting fresh at a new location has become less common. Improved transportation and cheap mass production are used instead to deliver finished items when needed. An idea can still be useful, though, even if it is not currently popular. We can update the basic idea of a starter set to incorporate 21st century equipment, and combine it with additional ideas and methods. This can make it a relevant choice again, especially in more difficult and remote locations.
Workshop builders are accustomed to upgrading the shop by building storage, workbenches, jigs, and other extensions using their starter set of tools and machines. Sometimes wood and metal shops will build complete new machines using the ones they already have. But such workshops typically do not produce all of their own parts. Normally they make the easier items, and buy parts like bearings and motors which are harder to make. So far as we know, no one has yet designed and operated a workshop or factory which intentionally uses a starter set, makes most or all of the items for its own expansion, and uses high levels of automation.
- Smart Tools
At first, all work was performed entirely by human muscle. Once work animals were domesticated we added their muscles to ours. Mechanization refers to the replacement of muscle power with machines using other power sources. Mechanization extends back to ancient times in particular cultures and for particular tasks, but only became widespread since the Industrial Revolution, starting around 1760. Where mechanization replaces muscles, Automation, Robotics, Software and Artificial Intelligence (AI), which we can collectively call Smart Tools, further replace the highly flexible arms, senses, and decision-making of people. They make possible mechanical, electrical, electronic, and information systems which can do tasks on their own. This reduces the need for constant human attention for work tasks. By the mid-20th century, electronics had developed to the point they could be programmed and used as sensors, and the field of Control Theory was well enough understood to design self-operating systems that used these electronic devices. Automation, robotics, and software started to be developed on a large scale from that time. AI took until the late 20th century to be well enough developed for practical applications, and is now being widely deployed.
Machines in general involve movement, but in the past the type of movement was fixed by design. Robots are machines which can perform movement under automated or remote control, where the specific movements are not predetermined. You can change the task a robot performs by changing the control inputs. Robotics has developed in parallel with more fixed types of automation. By now (2017) automation and robotics are extensively used in manufacturing, and are starting to be used in other environments. In the past, people have used tools and machines to make more tools and machines. The development of smart tools means the machines are becoming more able to do so on their own, without needing people to do much of the work.
Seed Factory Concept Evolution
We can combine a number of these concepts and processes from nature and human culture into a new type of self-expanding and upgrading production system. Purposely designed starter sets which evolve to larger and more complex systems is the core idea, so we refer to the whole collection as "Seed Factory" technology. The complete collection includes more than just this feature. Other important ones that we include under the Seed Factory name, and their backgrounds, include:
Replication in the general sense has always existed in biology. In technology we often make identical copies of products, but Machine Replication, or Self-Replicating Machines, as a term is reserved for machines or factories which can make complete copies of themselves. The idea of machine replication was evident as soon as control theory and automation became practical. It got serious theoretical study in the same period, the mid-20th century, starting with work on Reproducing Cellular Automata by John Von Neumann. A 2004 book by Freitas and Merkle, Kinematic Self-Replicating Machines thoroughly reviews the literature for replicating systems to that point.
A 1982 NASA report, Advanced Automation for Space Missions (AASM), introduced the concept of a replicating factory for use on the Moon (Figure 2.1). Their approach was to copy the original factory hardware multiple times, using local materials and energy. Once enough copies had been made, the combined production capacity would turn to making an unspecified end product, to support NASA goals. The AASM study introduced the term Seed Factory to mean the first unit of the factory delivered from Earth to the Moon. In this book we use the name for a more general concept of a starter set which can expand by multiple methods besides directly copying its own parts.
Due to low communications bandwidth from Earth at that time, the AASM concept of a Lunar factory assumed it would be fully automated. The resulting computer requirements, estimated at 2 GB memory and 35 GB storage, were far beyond what was available at the time of the study (1980). The seed factory was also assumed to be delivered as a complete functioning unit to the Moon, with an estimated mass of 100 tons. This was beyond the delivery capacity of any lander at the time, and still is today. Lastly, there was no pressing need to build things on the Moon. For these reasons the concept was not developed further.
The study considered availability of raw materials, and processing methods, but did not do a full resource flow accounting of all materials, energy, and data. The first spreadsheet program, Visicalc, had just been introduced in 1980, and Computer Simulation and Numerical Analysis Software were at a relatively early stage of development. The tools available then could not have handled such a complex design. The study also neglected to consider Earth applications of the idea, because NASA's primary goals do not include improving manufacturing on Earth. Despite these shortcomings, the AASM study is probably still the best attempt to date of describing a fully replicating factory.
The AASM study assumed full replication of a starter factory. It did not consider starting with a simpler subset and adding new equipment over time. Requiring full replication from the start makes the initial hardware both more complicated to design, and physically larger. Even if you were able to make all the parts needed to replicate, from an engineering standpoint it might not make sense. It is likely that a given location will not have certain rare raw materials. In the case of the Moon, a 1986 study of building solar power satellites found 98% could be sourced from there, but the other 2% had to come from Earth. It is also likely that some parts, like computer chips, are very difficult to make yourself, compared to simply buying them. Therefore a practical design will most likely fall short of 100% replication. These conditions are likely to be true for any location, not just the Moon. Given that some materials and parts would be imported in an optimized design, it is then a small conceptual step to allow the percentage of imported items to vary by time.
A small starter set is able to make some of the parts for new and different equipment, and the balance has to be imported. Once you have added the new equipment, you can now make a wider range of parts, and can therefore contribute to making an even wider range of equipment. This Diversification continues until you you can make the full range of desired products, or reach the full set of equipment which are practical to use. When a factory reaches these full capacities we call it Mature. Compared to a conventional factory, which has the full set of equipment at the start, you begin with a smaller set of equipment, plus a set of design information for the rest. With modern data storage, it costs very little to hold a large amount of information. The seed factory with a subset of equipment is therefore a less expensive way to start production. A fully expanded factory can also end up larger and more capable than what is needed for Practical Replication, i.e. the ability to copy all the parts that are practical to make internally. This happens if certain production processes or materials are not needed for replication, but are used for other end products. So the set of equipment can start well below the ability to output copies of itself, and grow to a level well above it.
One class of new equipment is extensions or attachments for existing equipment. For example, extension rails or a larger bed would allow a CNC mill to make larger parts. Conventional farm tractors are examples of machines which can be used with a multitude of different attachments. A modular vehicle chassis built for factory use could also accept a range of attachments, as could a basic robot arm. Items like assembly jigs, holding fixtures, molds, and custom cutting bits can make a basic machine more flexible. One criterion for choosing a member of the starter set is then how many of these extensions you can use with it, and thus how many different tasks it potentially can do. We call this feature Flexibility. The more flexible the starter machines are, the fewer of them you need for a useful set.
A simplified starter set is one way to reduce the complexity and cost of starting up a new factory. Scaling of machine sizes is another way to reduce cost, and was not considered in the AASM study. It should be evident that machines of a given size can make parts for larger machines. If not for that, our civilization would never have been able to make the larger machines that now exist, like 400 meter long cargo ships. One way to make larger items is by assembling them from smaller pieces, such as by welding or bolting. A common example is that most buildings are made from many smaller parts. Another method besides assembly is to use machines which are open-ended in at least one axis. An example is a rolling mill to make steel shapes. The rolling process does not limit the length of the metal parts you feed through it. It is only limited by how much room you have on either side. An example in two and three dimensions is casting of concrete structures. The forms, mixing, and pouring equipment can be mobile, working on different parts of a structure in series. The resulting structure, such as an airport runway, can be many times larger than the equipment.
Scaling can also be used to make smaller and more accurate devices, but for cost reasons, it will more often be used to produce larger machines. In concept, you can start with whatever size machines are convenient. You can use these machines to make parts for larger machines of the same type, or for larger machines of new types. In turn, the second generation machines are used to make parts for even larger machines, until you reach whatever final size you wanted. In theory you could start with microscopic machines, but for practical reasons there is some lower bound to start from. Whatever operations and maintenance are not yet automated would require human interaction. They then need to be at a scale people can work with. More generations of scaling require more time to grow. They also require some redesign for each generation, since not all parts of a system scale equally. Finally, the design effort for the number of machines in a factory is relatively constant no matter what size they are. So the savings in materials from starting smaller will eventually become negligible relative to the total design cost. These factors in combination will favor a convenient and inexpensive starting size.
- Technology Level
The final assumption made by the AASM study was full automation. This was because with 1980 communications it was not possible to remote control even one factory on the Moon, let alone 1000 copies. We don't need to make that assumption for the current seed factory concept. For one thing, certain tasks are either too hard to automate at present, or are done so rarely it is not worth trying to automate them. For another, trying to automate everything at the start requires more equipment and more design work. Lastly, modern communications has improved enough that remote control is quite possible over long distances. Instead of assuming full automation, we instead assume you start with whatever is a practical level. CNC machine tools and 3D printers are examples of existing equipment that are automated, so they are good candidates to start with. Over time you can add things like robots and automated inventory systems to increase the amount of automation in the factory. Since you are able to partly make your own equipment, adding these later will be less expensive than trying to buy it all at the start. Starting with partial automation allows you to defer the design work of the more complex automation until later. This lowers the cost to get started.
In the extreme limit, you could start with no tools whatsoever, and bootstrap making crude tools by hand. Given modern civilization, this is not necessary. For a particular project, it will make sense to start with some level of hand and power tools, larger stationary and mobile machines, and smart tools. The purchase of such already-designed and in-production items will be worth the time and labor savings over starting from scratch.