# Hard and Soft Technology >[!note] >Definition: Technology is the product of the transformation of raw materials, either geological or biological, in a way that is deemed useful in the marketplace or by communities. Naturally, different people define technology in different ways, and this definition is broad. We've chosen it because the data-informed models we're covering here can be used to understand all technologies that fall under this definition, and all of those technologies affect our lives. Also, this broad concept of technology applies throughout history, from Stone Age tools to artificial intelligence, and we want to analyze the progress of technology over these very long time scales to help us develop robust theories. This definition allows us to examine all of the ways that humans are transforming the resources available to us, over time, and changing the way we live in the process. In our current society, the economic market both determines and mediates the usefulness of a technology, as products compete to satisfy a demand. Many definitions of technology only consider technology in the context of the market, but we also want to include technologies used by early pre-market societies. In that context, communities directly determined the usefulness of a technology based on how well it helped them survive and thrive. Our expansive definition of technology encompasses two broad categories: Hardware: - physical tools, such as hammers, stethoscopes, - devices and building blocks such as transistors; - machines, such as cars, printing presses, or computers; - infrastructures, such as the power grid, the transportation system, or networks of medical facilities; - other codified knowledge that takes on a physical form Soft technology: - software, such as operating systems, electronic medical records software, or phone apps; - codified processes, such as user guides, house construction methods, or manufacturing checklists; - codified skills, such as accounting practices, nursing routines, or project management guidelines; - institutional designs, such as organizational charts or governance structures; - other codified knowledge that does not appear in a physical form. Accomplishing many tasks requires both hardware and soft technologies. Consider, for example, technologies to enhance the capacities of our brains to store and process information. Writing systems such as the alphabet, for example, are a soft technology that enables information storage. But writing can't happen without hardware technologies as well: clay tablets, pens, paper, and computer keyboards. Likewise, computers function only through interconnected hardware (hard drives, monitors, keyboards) and soft technology (software), which together increase our ability to store and process information. Stepping to an even higher level, the modern workplace relies on enhanced information processing through a combination of hardware in the form of computers, and soft technologies like management practices or codified routines. > [!important] > Effective, accessible soft technology can decrease the need for hardware. For example, if you have a management system for a shared fleet of personal vehicles (a soft technology), you may be able to reduce the number of automobiles (hardware) needed to serve the community. The overall transportation system might consist of a smartphone application, a predictive model of vehicle demand, urban planning principles, and a range of physical vehicles, from skateboards and scooters to personal cars and trucks. The fleet would be tailored to conveniently meet the needs of community members while reducing the total number of vehicles required to transport people around a city. # Technological Sophistication and Progress What makes one technology more sophisticated than another? Intuitively, it's obvious that, over time, technology has become more sophisticated. A primitive axe made of wood, stone, and a binding is simpler than a chainsaw with a stainless steel blade and an internal combustion engine. Our definition above of technology allows us to pin that idea down more precisely. T==echnology is the transformation of raw materials in a way that is deemed useful. So, a technology is more sophisticated if it requires a greater transformation of raw materials.== So, for example, a simple technology such as an early soccer ball is made of a single material, in this case leather, that requires little processing. An advanced technology such as a laptop, by contrast, requires many different raw materials to go through a series of complex refinements to produce the metals and plastics needed. A second way of understanding the sophistication of technology is to focus on the number of components that must be aggregated to create them. An early soccer ball is an aggregate of just two components, leather and thread. On the other hand, a laptop is an aggregate of the processor, screen, keyboard, battery, mouse pad, camera, and many more, and each of these components is a technology of its own which is, in turn, composed of smaller components. In this sense, technologies are recursive (Arthur, 2009). ==Generally speaking, higher levels of aggregation are observed in technologies that are further from their raw material form.== # Technology Innovation We need a precise definition of technology innovation. This is because measurement is the first step in forecasting, and a precise definition is needed to move from qualitative descriptions to quantitative measurements. Our definition of technology innovation is: > [!important] > Technology innovation is a human-driven change in technology that increases its perceived usefulness. Innovations may consist of changes in the function of a technology, the building blocks of a technology, or both. # Design Space and Functional Space ## Design Space Imagine all possible designs of a given technology as existing in an extremely high-dimensional space, where each dimension represents a given aspect of its design. This is called design space. Changes in the building blocks of a technology represent a shift in design space. So, for example, a technology's geometric dimensions (length, width, and height) would be three dimensions in design space. Its materials would give more dimensions, along with various material characteristics, like its weight, texture, or color. Its overall cost would be a dimension, as would the cost of its various components. Its energy consumption or recyclability might be dimensions. Various aspects of performance might be dimensions—its speed, power, or capacity. The idea of design space is more conceptual than precise, because it would be difficult to precisely describe every imaginable aspect of the design of a technology, but that concept is still quite useful. ## Functional Space Similarly, we can imagine all possible functions of a technology as existing in a very high-dimensional functional space. Functional space is again more conceptual than precise — how could we possibly quantify every possible function any technology might have?—but also quite useful. Changes in functional space usually require changes in design space, just as the increased functionality of the internet required the addition of undersea cables. One category of dimensions in functional space is the ability to do physical work. Transportation-related technology is a good example of this, since it amplifies the ability of humans to do work when moving people and objects from one place to another. Speed and distance are two dimensions in the general category, and innovation in transportation technology has progressively enabled travel at faster speeds and over longer distances, improving the position of these technologies in functional space. Improvements to technology have also enhanced the safety and accessibility of transportation services. Other dimensions in functional space allow humans to enhance our ability to process information. The codified system of mathematics is an example of a soft technology that does this. Mathematics dates back to at least the 3rd millennium BCE (Valério & Ferrara, 2022), and has played a central role in underpinning all of science and engineering. The progression of mathematics from arithmetic to geometry, algebra, and number theory, has involved a continuous process of innovation. A third category of dimensions is the preservation of human health. One of the most successful examples of this was the eradication of smallpox by 1980 (World Health Organization, 2022). Smallpox caused suffering, disfigurement, and death over a period that is estimated to have exceeded 3,000 years. The development and distribution of the smallpox vaccine spread immunity worldwide until the virus was driven extinct in human populations. Perhaps the most recent category of technology function to emerge is that of reducing the negative impacts of other technologies. For example, pollution from factories and power plants has caused problems since the late 19th century, and these problems have increased as these facilities increased in size and number. In order to address these concerns, technologies such as flue-gas desulfurization "scrubbers" have been developed and installed. The first such device was installed at the Battersea Station near London in 1931 (Biondo & Marten, 1977). Following investment and legislative measures, flue-gas desulfurization was widely adopted in the 1970s. Modern dry scrubbers can remove SO2 with a 90% efficiency rate (EPA, 2003). > [!cite] > *"So far from being unprepared for in human history, the modern machine age cannot be understood except in terms of a very long and diverse preparation. The notion that a handful of British inventors suddenly made the wheels hum in the eighteenth century is too crude even to dish up as a fairy tale to children.* *Plainly, what is usually called the Industrial Revolution, the series of industrial changes which began in the eighteenth century, was a transformation that took place in the course of a much longer march."* > Lewis Mumford > _Technics and Civilization_, 1934 # Performance Metrics In order to examine technology innovation, we need to be able to measure technological performance. Such measurements can be made using "performance metrics" that quantify the service provided by technologies. Consider the following examples: The plow is an ancient technology that has been used for thousands of years to increase the power capacity of people growing crops. Prior to plows, farmers tilled their land with hand-held instruments. Plows enabled farmers to augment their own physical strength with that of a working animal. To measure the performance of the plow, one approach would be to measure the person-hours per calorie cultivated. In this case, the service provided is represented by a quantity of food, and the cost is measured as the number of person-hours required. Alternatively, we might consider the total economic costs of pulling the plow through a given area of land. In that case, the plowed field would represent the service. Take now the keyboard. The advent of writing is one of the greatest technological contributions to augmented information processing. Writing augmented humans' communicative capacity greatly by enabling the preservation and transmission of information over time and space. Subsequent developments, including that of typewriters and then computer keyboards, have further increased this capacity. One way of measuring the ability of a writing system to encode information is the amount of information that can be encoded in a given amount of time (e.g., in kilobytes per minute). The keyboard increased performance along this dimension over pen and paper. It must be kept in mind that the metrics we selected to describe the performance of the plow and the keyboard took the form of impact per unit service. In the case of the plow, the impact considered was the cost, measured in person-hours or economic cost. The service was the number of calories cultivated or the area of land plowed. In the case of the keyboard, the impact was the cost in terms of the time required, and the service was the amount of information encoded. In these examples, technology improvement is represented by a declining cost per service. A reduction in the cost of inputs or an increase in service provided would increase technology performance. These metrics could also be inverted to measure the service per cost, in which case technology improvement would be represented by an increasing trend. ## Choosing Performance Metrics In 1994, Bill Gates compared the information storage capacity of a CD-ROM to that of paper by stacking the sheets of paper necessary to hold the same amount of information as one disk, stating that "this CD-ROM can hold more information than all of the paper that's here below me". Implicitly, he was comparing the two technologies using a performance metric with the form of information storage per unit (where the units are the sheets of paper or the number of disks). ![[Pasted image 20250304101155.png]] However, this is only one possible performance metric to use for the comparison. He might also have applied the following metrics: - speed of data retrieval per unit: He might have illustrated this by using the table of contents or the index to sort through the stack of paper, looking for a particular piece of information; - overall cost of production per unit: This would be the cost to manufacture and record information on the CD-ROM compared with the cost to manufacture the paper and print the information on it. The choice of metric matters, because different metrics illustrate different things. For example, if Bill Gates had chosen that last metric, the overall cost of production per unit, the superiority of CD-ROMs probably wouldn't have been as obvious. Other metrics reveal aspects of performance that aren't readily perceived by consumers but that may nonetheless be critical. In the case of paper vs. CD-ROM storage, consider the following metrics: - water usage in production or operation per unit service; - land usage; - carbon emissions; - societal benefits such as job creation. Because these metrics reveal very different things, fully understanding how a particular technology performs requires analyzing a wide range of metrics. An alternate approach is to try to collapse all these metrics into a single number by estimating the economic cost of all impacts, including those impacts internalized in the market price as well as externalities, to arrive at a complete economic cost per unit service for the technology. (This second approach brings some challenges in that the economic costs of some impacts are difficult to estimate.) In either case, it is important to consider the full impact of a technology when allocating resources towards improving it. ## Performance intensity metrics The various types of performance metrics we've discussed so far all measure performance intensity in some way. All performance intensity metrics take the form of impact per unit service. Here are the types of performance intensity metrics. ### Cost intensity Of course, we want technologies to provide a lot of bang for the buck. That's what the metric of cost intensity tells us: how much of some service can be had for a given cost, or equivalently the cost of a given service. "Service" here can mean many different things: saving time, increasing power output, or increasing information processing capabilities, for example. To see how this plays out, suppose we want to evaluate the cost intensity of various forms of personal transport, comparing taxis, personal cars, trains, walking, bicycling, and buses. To do so, we would fix a unit of service, say transporting an individual one kilometer. Then we'd calculate the cost of each transportation technology along with the fuel costs. Depending on the scope we want to consider, we could also consider time costs, since walking a mile is likely to take substantially longer than driving a mile. ### Carbon intensity As the world has become more alarmed about the impacts of climate change, we've realized that we need to pay attention to how much carbon is emitted as the products we need are created. That's what carbon intensity measures. Consider food production, for example. We can compare the carbon intensity of two foods by calculating the greenhouse gas emissions per 1000 kilocalories. ### Water intensity Water has become less available in many parts of the world, increasing the importance of water intensity. Consider clothes washing as an example. We can calculate the water intensity of a range of technologies, from hand washing to machine washing to dry cleaning, by calculating the water required to wash a kilogram of clothes. ### Land-use intensity We have many competing demands on land, including agriculture, housing, energy production, and natural habitat. City planners may use this metric when considering zoning and building approvals. Taking a house as the unit of service, planners might weigh the desirability of independent, single-family structures against that of townhouses and apartment buildings by calculating the number of people housed per acre. ### Pollutant intensity This metric quantifies the amount of contaminants emitted to the environment as a result of deploying a technology to provide a service. For example, nitrogen fertilizers can run off of agricultural land, leading to algal blooms in lakes and oceans. This saps the water of its dissolved oxygen content, creating "dead zones" that make the affected area uninhabitable for many organisms. A performance metric such as calories grown/kilogram of nitrogen fertilizer could be used to capture the effects of various agricultural practices and quantify the pollutant intensity. ## Scalability metrics The discussion so far has been focused on performance intensity metrics, but there is another aspect to consider when evaluating technologies: scalability. Can the technology be produced at the scale that would justify the investment in their development? If the natural resources required to produce a technology aren’t available at the level needed, that would limit the technology's growth. Assessing scalability can be key in deciding which technologies to invest time and money in developing. If a technology has excellent performance intensity and no scalability problems, it promises to be a good investment. Similarly, if it has poor performance intensity and cannot scale substantially, it’s almost certainly a bad investment. But the in-between cases, where you have to balance the performance intensity against potential scalability issues, require more careful analyses. For example, consider the development of thin-film photovoltaic technology. Traditional solar cells typically use relatively thick silicon wafers that are crystalline or polycrystalline. Reducing that thickness to the range of microns or even nanometers reduces material usage, manufacturing costs per panel, and some negative environmental impacts while creating lighter, more flexible panels that can be integrated into building materials, including semi-transparent windows. Thin-film solar cells do bring additional challenges, though, particularly maintaining a high enough efficiency to keep the cost of electricity low. Because they can be built in different ways with different materials, choices need to be made about which solar panel technology to bet on. Thin-film cells can be created based on silicon or a combination of cadmium and tellurium, among others. While thicker silicon cells currently outperform the thin-film cells available, there has been some reason to think that with further development cells such as 'cadmium telluride' thin-film cells may provide the best cost per unit service. For instance, (Kavlak et al., 2015) analyzed metal availability and found reasons for concern about the rate at which tellurium production could scale. They estimated that tellurium growth rates could limit cadmium telluride PV cells to a few percent of electricity generation by 2030 if the historical precedents for rates of mining expansion held into the future. By contrast, typical natural resource extraction growth rates would be required for silicon-based cells to supply 100% of electricity generation at that time. That analysis suggests that silicon-based photovoltaics is likely the better bet, even if projections say that cadmium telluride might ultimately outperform silicon on cost per unit service. And this example shows how scalability metrics can be essential in evaluating technology. Another important constraint on scalability in some situations can be the energy resources required to operate a technology. Theoretically, in some cases, even the labor capacity needed to produce or operate a technology could limit its scalability, though that constraint may prove less binding, for example with increasing capabilities in automation. ## Conclusion - Performance metrics help us shift from describing technology with adjectives (qualitatively) to describing technology with numbers (quantitatively). - Performance intensity metrics cover a wide range of impacts, including those that are reflected in the price of a technology and those that are not. - Scalability metrics allow us to evaluate which emerging technologies are likely to be able to be produced at large quantities, which would tend to make them more valuable investments. - By broadening the dimensions along which we measure the performance of a technology, we are often better equipped to influence innovation in directions that are useful and beneficial to society. - If we adopt a broad set of metrics to evaluate a technology's performance, we can understand impacts that are not currently reflected in price. # References - Abraham, D. S. (2015). _The elements of power: Gadgets, guns, and the struggle for a sustainable future in the rare metal age_. Yale University Press. [URL](https://www.amazon.com/Elements-Power-Gadgets-Struggle-Sustainable/dp/030022690X). - Arthur, W. B. (2009). _The nature of technology: What it is and how it evolves_. Free Press. [URL](https://sites.santafe.edu/~wbarthur/technologybook.html). - Azevedo, M., Campagnol, N., Hagenbruch, T., Hoffman, K., Lala, A., & Ramsbottom, O. (2018). _Lithium and cobalt—a tale of two commodities_. McKinsey & Company. [URL](https://www.mckinsey.com/industries/metals-and-mining/our-insights/lithium-and-cobalt-a-tale-of-two-commodities). - Bettmann. (n.d.). _Portrait of Lewis Mumford_ [Photograph]. Corbis Images. [URL](https://www.wnyc.org/story/206665-lewis-mumford/). - Biondo, S. J., & Marten, J. C. (1977). A history of flue-gas desulfurization systems since 1850. _Journal of the Air Pollution Control Association, 27_(10). [URL](https://www.tandfonline.com/doi/epdf/10.1080/00022470.1977.10470518?needAccess=true&role=button). - Denning, P. (1989). The Science of computing: The ARPANET after twenty years. _American Scientist_, _77_(6), 532–534. JSTOR. [URL](https://www.jstor.org/stable/27856002). - Donnellan, A. (2020, December 3). Breakingviews - Vaccine bottlenecks are main obstacle to recovery. _Reuters_. [URL](https://www.reuters.com/article/us-health-coronavirus-vaccine-breakingvi-idUKKBN28D1JA). - EPA. (2003). _Air pollution control technology fact sheet_. [PDF](https://www3.epa.gov/ttncatc1/dir1/ffdg.pdf). - European Commission. (n.d.). _About the energy label and ecodesign_. [URL](https://commission.europa.eu/energy-climate-change-environment/standards-tools-and-labels/products-labelling-rules-and-requirements/energy-label-and-ecodesign/about_en). - The Editors of Encyclopaedia Britannica. (2019). Stove. In _Encyclopædia Britannica_. [URL](https://www.britannica.com/technology/stove). - Gapminder & Dollar Street. (n.d.). _Life on the four income levels_ [Image]. Gapminder. [URL](https://www.gapminder.org/fw/income-levels/). - Imgur. (n.d.). _Bill Gates, 1994_ [Photograph]. [URL](https://i.imgur.com/kSq9xDX.jpg). - Mumford, L. (2018, December 31). In _New World Encyclopedia_. [URL](https://www.newworldencyclopedia.org/entry/Lewis_Mumford#Technics). - Mumford, L. (1934, December 1). _Technics and civilization_. Harcourt. [URL](https://www.amazon.com/Technics-Civilization-Lewis-Mumford/dp/0226550273). - Leyden, J. (2021, June 16). _The evolution of ARPANET, 1969–1982_ [Image]. The Daily Swig. [URL](https://portswigger.net/daily-swig/arpanet-anniversary-the-internets-first-transmission-was-sent-50-years-ago-today). - National Oceanic and Atmospheric Administration. (2017, August 2). _Gulf of Mexico 'dead zone' is the largest ever measured_. [URL](https://www.noaa.gov/media-release/gulf-of-mexico-dead-zone-is-largest-ever-measured). - Poore, J., & Nemecek, T. (2018). Reducing food’s environmental impacts through producers and consumers. _Science, 360_(6392), 987–992. [URL](https://www.science.org/doi/10.1126/science.aaq0216). - Psihoyos, L. (1994) _Bill Gates illustrating storage of CD-ROM_ [Photograph]. Psihoyos.com. [URL](https://www.psihoyos.com/image/I0000jtF1ui2j79Q). - Rabalais, N. (2017, August 2). _Gulf of Mexico dead zone in July 2017_ [Image]. NOAA. [URL](https://www.noaa.gov/media-release/gulf-of-mexico-dead-zone-is-largest-ever-measured). - Ritchie, H., & Roser, M. (2020, January). _Environmental impacts of food production_. Our World in Data. [URL](https://ourworldindata.org/grapher/ghg-kcal-poore). - Rosling, H., Rosling R., A., & Rosling O. (2018, April 3). _Factfulness: Ten reasons we're wrong about the world—and why things are better than you think__._ Flatiron Books. [URL](https://www.amazon.com/Factfulness-Reasons-World-Things-Better/dp/1250107814). - Valério, M., & Ferrara, S. (2022). Numeracy at the dawn of writing: Mesopotamia and beyond. _Historia Mathematica, 59_, 35–53. [URL](https://www.sciencedirect.com/science/article/pii/S0315086020300665?via%3Dihub). - Winkelman, S. (2018, July 10). Appy birthday: A brief history of the App Store’s first 10 years. Digital Trends. [URL](https://www.digitaltrends.com/news/apple-app-store-turns-10/). - World Health Organization. (2022). _History of smallpox vaccination_. www.who.int. [URL](https://www.who.int/news-room/spotlight/history-of-vaccination/history-of-smallpox-vaccination).