Book summary “New Dark Age”: summary

Introduction

You live in an era of unprecedented technological acceleration where digital tools promise total visibility, yet you face mounting confusion, systemic failure, and environmental crisis. James Bridle argues in New Dark Age that this paradox arises because information abundance does not equal enlightenment, but rather generates profound unknowing and complexity that outstrips human comprehension. You rely daily on opaque systems whose inner workings remain hidden even from their creators.

Our profound dependence on complex and opaque technological systems is plunging us into a new dark age characterized by systemic unknowing, environmental destabilization, and political confusion.

The following chapters examine how we reached this precarious state. You will explore how digital infrastructure rests on physical realities, how historical climate science intersects with modern computation, and how automated governance reshapes your labor and daily choices. The book addresses several critical questions:

• How did computational complexity create a new form of systemic illiteracy?
• What physical limits constrain the cloud and digital networks?
• Why do automated systems encourage dangerous human compliance?
• How can you reclaim agency within opaque technological environments?

1. The Chasm of Ununderstanding

You live enmeshed in technological systems that shape your actions and thoughts. These technologies interact in unexpected and opaque ways that make it impossible to stand outside them.

Technological systems have become so complex and opaque that they outstrip our collective capacity to understand them, creating a dangerous chasm between functional use and deep systemic literacy.

James Bridle experienced a personal crash when his laptop failed during an episode of The West Wing. The screen repeatedly looped a line about technology failing to connect people in an emergency, mirroring the wider failure of technological systems. Solutionism drives the worst trends in modern interactions by reducing societal and practical issues to app-based fixes. It creates an inability to articulate the world in terms that are not computable. True systemic literacy requires fluency in the metalanguage of systems and the capacity to perform and respond to critique.

The Physical Reality of the Cloud

The cloud is not an amorphous, weightless abstraction but a vast physical infrastructure with heavy resource consumption and geopolitical footprints. Engineers originally used a fuzzy circle to reduce complexity, but the metaphor now obscures physical data centres, undersea cables, energy consumption, and corporate power structures.

Google and Facebook strategically build data centres in places like Ireland for low taxes and Scandinavia for cheap energy and cooling. Global empires retain disputed territories where the cloud touches down. Technology relies heavily on the affordances and limitations of nonhuman actors and materials. High-frequency trading infrastructure accommodates the physical properties of silicon, steel, the speed of light through glass, and environmental factors like fog, birds, and squirrels.

Information Abundance and the New Dark Age

The modern abundance of accessible information paradoxically darkens the world instead of enlightening it, fueling fundamentalism and conspiracy theories. The Enlightenment tenet that more information leads to better decisions has broken down. Infinite searchable data fails to produce a coherent consensus reality and instead drives post-factual politics and social division.

In 1926, H.P. Lovecraft wrote about the human mind's inability to piece together dissociated knowledge without inviting terrifying vistas of reality. You must move away from rigid computational solutionism by embracing cloudy thinking and acknowledging a state of productive unknowing. Drawing on historical mysticism and literary references, the author argues that the inability to fully know everything can be a productive space of experience, love, and agency rather than a reason for despair.

When you treat tools strictly as single-purpose instruments like a hammer for nails, you stop thinking about their wider capabilities, histories, and symbolic meanings. Thor's hammer Mjölnir generated weather and served as a protective amulet, while prehistoric stone axes found by later generations were called thunderstones and believed to have fallen from the sky. Re-enchanting your tools means recognizing their broader affordances and nonhuman dependencies.

2. Computation and Automation Bias

Historical attempts to compute and control physical reality have culminated in code/spaces where human operators exhibit dangerous automation bias, trusting machine outputs over their own senses.

Cultural metaphors shape how you perceive large-scale industrial changes. John Ruskin observed a dry, black veil in the skies of England in the late nineteenth century, which he called a plague-cloud or storm-cloud. He believed it was a new phenomenon linked to societal unrest and moral decay.

John Ruskin gave lectures in 1884 at the London Institution detailing observations from his diary of a dry black veil covering the sky that no ray of sunshine could pierce. Alexander Graham Bell transmitted the human voice across 200 metres in Washington, DC, using sunlight bounced off a vibrating mirror. Invented in 1880, the photophone used a beam of light reflected off a surface vibrated by a speaker's voice to achieve wireless transmission of sound. This early integration of light and data prefigured modern fiber-optic cables that govern global computation.

Weather Prediction and Computing Metaphors

Early computational thinking for weather forecasting was conceptualized through manual numerical processes and architectural theater analogies before electronic computers existed. Lewis Fry Richardson performed the first manual calculation of weather forecasting using grid squares and mathematical equations while serving as a pacifist ambulance driver during World War I. He later designed a hypothetical computing theatre where thousands of human computers directed by a conductor would calculate future weather.

Lewis Fry Richardson used complete meteorological records from May 20, 1910, compiled across Europe to manually calculate a numerical forecast from a cold rest billet in France. Vannevar Bush identified that specialized research was overwhelming investigators and proposed the memex, a desk-sized device storing books and records equipped with associative linking capabilities.

Vannevar Bush served as director of the US Office of Scientific Research and Development and helped originate the Manhattan Project. His 1945 article published in the Atlantic directly anticipated hypertext and networked information retrieval. John von Neumann championed the use of high-speed electronic computers to model and manipulate global climate systems, influenced by Vladimir Zworykin's weather proposal and Manhattan Project simulations.

John von Neumann used ENIAC to run the first automated twenty-four-hour weather forecast using 100,000 IBM punch cards. The 1950 Aberdeen meteorological experiment relied on the ENIAC machine, which used 18,000 vacuum tubes. Military-funded weather forecasting models established the foundational philosophy that all stable processes will be predicted and all unstable ones controlled.

Limits of Computation and Code/Spaces

Modern physical environments are built as code/spaces where infrastructure completely ceases to function if underlying computational systems fail. Geographic spaces like airports or transportation networks interweave computation so deeply into the physical structure that software becomes the sole arbiter of operational reality. SABRE evolved from a Cold War air defense network into a multinational reservation system connecting over 57,000 travel agents and hundreds of airlines.

When systems crash, physical venues lose their functional designation entirely. Human beings exhibit automation bias by valuing automated system outputs over their own sensory observations and personal experience. This cognitive shortcut causes pilots, drivers, and users to follow machine instructions blindly into hazardous or fatal situations.

Korean Air Lines flight crews and tourists driving their vehicles into the sea or lakes because their satellite navigation or autopilots instructed them to do so illustrate this danger. In NASA simulations, 75 percent of flight crews shut down the wrong engine when following automated fire warnings, compared with 25 percent when using a traditional paper checklist.

Attempting to compute the world with greater precision reveals an exponentially increasing and unknowable level of complexity. Lewis Fry Richardson discovered differing estimates of national borders while compiling his Statistics of Deadly Quarrels. The coastline paradox shows that more obsessive attempts to compute and measure reality make it appear infinitely more complex.

3. Climate Destabilization and Cognitive Decline

Global warming physically destabilizes digital and ecological infrastructures while elevated carbon dioxide concentrations directly impair human cognitive abilities.

Rising global temperatures are thawing the Siberian permafrost, releasing vast amounts of methane and causing the ground to liquefy, shudder, and occasionally explode. This process creates dangerous positive feedback loops that accelerate global warming and destabilize physical geography.

Researchers in Siberia installed seismic sensors near Sabetta to provide early warning of exploding pingos, and a 2013 explosion created a crater forty metres wide and thirty metres deep on the Taimyr Peninsula. In 2006, Siberian permafrost released 3.8 million tonnes of methane. A 2013 release reached 17 million tonnes. Fully 30 per cent of the world's remaining natural gas reserves are in the Arctic.

Permafrost and Infrastructure Vulnerability

The Svalbard Global Seed Vault serves as an Arctic deep freeze sanctuary to back up global crop biodiversity against the threats of climate change and conflict. It is established 120 metres depth into a sandstone mountain on the Svalbard archipelago, 1,120 kilometres from the North Pole, featuring caverns 150 metres long and 10 wide. The facility relies on minus eighteen degrees Celsius vault refrigeration temperature to preserve genetic diversity. In 2015, the International Center for Agricultural Research in the Dry Areas requested the first withdrawal of 116,000 samples across 130 boxes after its Aleppo headquarters was disrupted by the Syrian civil war.

Melting permafrost in the Arctic is destroying invaluable ancient middens and biological records, acting like a burning library of Alexandria. Ancient sites like the Qajaa settlement in Greenland lose 3,500 years of occupation, erasing the history of the Saqqaq culture established in 2500 BCE.

Climate change directly degrades physical infrastructure and information communication technologies, threatening global networks with physical damage and operational failure. In 2015, global data centres consumed 416.2 terawatt hours of electricity, accounting for 3 per cent of the world's electricity consumption and 2 per cent of total global emissions.

Aviation Disruptions and Cognitive Impairment

Rising atmospheric carbon dioxide concentrations are increasing the frequency and severity of clear-air turbulence, making commercial flights significantly bumpier and harder to predict. United Flight 826 experienced 1.814 G peak normal acceleration and minus 0.824 negative G in 1997. Accidents rose from 0.3 per million departures in 1989 to 1.7 in 2003, with models projecting a 40 to 170 per cent increase in frequency of moderate-or-greater turbulence for transatlantic flights.

Elevated atmospheric carbon dioxide concentrations directly impair human cognitive abilities and clear thinking, compounding the crisis of global warming. Preindustrial atmospheric carbon dioxide measured 275 to 285 parts per million, but atmospheric levels passed 400 ppm in 2015 for the first time in 800,000 years.

Numerous schools in California and Texas measured indoor CO2 levels exceeding 2,000 ppm in 2012. Scientific data demonstrates a 21 per cent drop in human cognitive ability at 1,000 ppm of CO2.

4. Calculation, Crisis, and the Optometrist Algorithm

While Moore's Law drove relentless computing progress, science faces a replication crisis and information overflow that can only be navigated by combining machine learning with human intuition.

William Gibson coined the term steam-engine time to describe how multiple writers or inventors suddenly produce the same idea independently. Magnetism was independently theorized in Greece and India around 600 BCE and in China in the first century CE, while the Haya people of Tanzania have made steel for 2,000 years. Gottfried Wilhelm Leibniz and Isaac Newton independently formulated calculus in the 17th century.

Gordon Moore observed in 1965 that the number of components per integrated circuit doubled every year, which later became a doubling every two years. Semiconductor feature sizes shrank from 10 micrometres in 1971 to 1 micrometre in 1985, and reached 10 nanometres at the beginning of 2017. This rapid miniaturization freed software developers from resource scarcity, leading to bloated code and software speed halving every eighteen months, known as Gates's law.

The Scientific Replication Crisis

In contrast to computing, pharmaceutical research experiences an exponential decline in efficiency despite massive investments, a phenomenon known as Eroom's law. The number of new drugs approved per billion US dollars spent on research and development has halved every nine years since 1950. This decline across a timespan of 60 years was documented by researchers Jack W. Scannell, Alex Blanckley, Helen Boldon, and Brian Warrington in their 2012 study on pharmaceutical R&D efficiency, driven partly by an overreliance on automated high-throughput screening.

Modern scientific research faces a severe crisis of replicability and rising retractions, driven by publication pressures and questionable statistical practices. The Center for Open Science attempted to replicate five landmark cancer studies and found only two were repeatable, two were inconclusive, and one completely failed. Surveys show that 70 per cent of scientists surveyed by Nature failed to replicate other researchers' findings.

Researchers manipulate datasets or statistical methods through data dredging and p-hacking to achieve a p-value below 0.05, turning a statistical threshold into a target for publication rather than a measure of rigor. A hypothetical p-hacking demonstration involves rolling a die where 183 out of 1,000 dice rolls turn up six, yielding a p-value of 0.04. High-profile scientific frauds illustrate how intense professional pressure compromises quality control. Hwang Woo-suk fabricated human embryonic stem cell cloning results in South Korea and was exposed in 2006, while Diederik Stapel fabricated data across numerous social psychology studies before his resignation in 2011.

The exponential growth of data and research output creates overflow that overwhelms traditional quality control mechanisms. Derek de Solla Price studied scientific growth curves in the 1960s, predicting that science would face saturation and senility if it did not change its modes of production. The human genome project was announced in 1990 as the greatest single data-gathering project of its time.

The Optometrist Algorithm

Combining machine learning with human intuition through iterative choices helps navigate complex problem spaces where pure brute force fails. Tri Alpha Energy achieved plasma reactions using hydrogen-boron fuel and deployed the Optometrist Algorithm to guide experimental settings beyond standard automated limitations. The experiment achieved plasma shots during experimental runs with an interval of 8 minutes, building on previous milestones like an 11-millisecond runtime and a 2015 record of 29 hours for a tokamak fusion reactor.

You can operate this feedback loop directly during complex technical runs:

• Collapse thousands of possible experimental settings down to approximately thirty meta-parameters.
• Run the automated experiment or plasma shot to generate new results.
• Present the new results alongside the best preceding shot to a human operator.
• Allow the human operator to select which shot forms the basis for subsequent tests.
• Move the settings a short distance based on human choice and repeat the cycle.

5. Opaque Networks and Algorithmic Labor

High-frequency financial markets, corporate opacity, and algorithmic labor management exploit public infrastructure, concentrate wealth, and trigger flash crashes.

High-Frequency Trading and Dark Pools

Financial markets have evolved from human trading pits to automated algorithms running on private fibre-optic lines and microwave towers. Because the speed of light varies through glass and air, firms spend enormous sums to reduce travel time between exchanges by milliseconds, gaining massive financial advantages over slower participants.

Spread Networks spent $300 million to build a private fibre link between Chicago and New Jersey in 2009-10. This project reduced message travel time from seventeen to thirteen milliseconds and saved $75 million per millisecond. Firms can generate an estimated $100 million in additional annual revenue from a single millisecond advantage.

Dark pools are private trading venues that conceal large transactions from public exchanges, frequently exposing ordinary investors to predatory algorithms. A 2015 SEC report noted that dark pool trading accounted for one-fifth of all stock trades on public exchanges. Barclays and Credit Suisse were fined $154 million in 2016 for allowing high-frequency traders access to their private dark pools.

Invisible financial and digital networks physically parasitize public urban infrastructure, starkly contrasting high-frequency capital with struggling public services. McKay Brothers and other firms placed microwave dishes on top of Hillingdon Hospital in London. During the same period, the UK National Health Service faced a funding shortfall of 700 million pounds for 2017.

Modern technological progress and financialization act as key drivers of extreme wealth inequality rather than broad societal uplift. Thomas Piketty analyzed that in 2014 the US richest 0.01 percent controlled 11.2 percent of total wealth, while the top 0.1 bracket held 22 percent. The top 1 percent captured 95 percent of income growth from 2009 to 2012.

Accelerated and opaque financial markets are inherently prone to sudden, violent flash crashes triggered by feedback loops among automated trading algorithms. On May 10, 2010, the Dow Jones Industrial Average dropped 600 points in five minutes and reached a maximum deficit of 1,000 points below the previous day's average. During this flash crash, 2 billion shares changed hands for a total value of $56 billion, and the market recovered most of its losses within twenty-five minutes. A hacked AP tweet in 2013 caused the Dow Jones to drop 150 points in under two minutes.

Algorithmic Labor and Corporate Deception

Digital platforms and automated warehouses subject human workers to algorithmic management that treats them like replaceable machine parts. Amazon workers walk up to fifteen miles a day following pings on hand-held devices in chaotic storage warehouses, where a fully loaded truck is dispatched from a UK facility every three minutes. Dunfermline workers have resorted to sleeping in freezing tents near the facility to afford travel costs.

Corporate opacity and hidden technological systems allow companies to systematically deceive regulators and endanger human health. Volkswagen installed defeat devices in diesel cars to pass US emissions tests while emitting nitrogen oxide at forty times the legal limit on the road. This practice leads to an estimated 1,200 early deaths per decade in Europe.

Automated generation of digital products and content creates product spam and artificial stupidity that degrades culture and search environments. Zazzle claims to sell 300 million unique products generated by vendors like LifeSphere. Amazon was forced to remove 30,000 auto-generated phone cases and offensive austerity nostalgia T-shirts.

The integration of poorly understood computation into the Internet of Things inserts technological vulnerability into the most basic human needs. Researchers found vulnerabilities in Samsung smart fridges, Philips Wi-Fi Hue lightbulbs that could trigger epilepsy, and smart locks that allowed strangers into homes.

The mythology of Epimetheus and Prometheus illustrates how technological progress driven purely by efficiency and foresight ignores human understanding and social justice. Technology prioritizes the tyranny of techne over social justice, creating a dark age where you face systemic opacity.

Living responsibly in this environment requires concrete commitments:

• Take full responsibility for the toxic products of our technological culture.
• Commit to doing the least harm in the present.
• Acknowledge our responsibility to future generations without presuming we can control them.
• Think clearly and act with justice in the immediate present.

10 Key Ideas

1. Technological systems have become so complex and opaque that they outstrip our collective capacity to understand them, creating a dangerous chasm between functional use and deep systemic literacy.
2. The cloud is not an amorphous, weightless abstraction but a vast physical infrastructure with heavy resource consumption and geopolitical footprints.
3. Historical attempts to compute and control physical reality have culminated in code/spaces where human operators exhibit dangerous automation bias, trusting machine outputs over their own senses.
4. Rising global temperatures physically destabilize digital and ecological infrastructures while elevated carbon dioxide concentrations directly impair human cognitive abilities.
5. While Moore's Law drove relentless computing progress, science faces a replication crisis and information overflow that can only be navigated by combining machine learning with human intuition.
6. Combining machine learning with human intuition through iterative choices helps navigate complex problem spaces where pure brute force fails.
7. High-frequency financial markets, corporate opacity, and algorithmic labor management exploit public infrastructure, concentrate wealth, and trigger flash crashes.
8. The modern abundance of accessible information paradoxically darkens the world instead of enlightening it, fueling fundamentalism and conspiracy theories.
9. When you treat tools strictly as single-purpose instruments like a hammer for nails, you stop thinking about their wider capabilities, histories, and symbolic meanings.
10. Living responsibly in this environment requires taking full responsibility for the toxic products of our technological culture while committing to doing the least harm in the present.