Key Takeaways
Stop demanding a precise map; build the future through informed participation.
Complexity defeats precision. A complicated system, like an assembly line, follows stable rules and can often be planned. A complex system, like a global economy, shifts as many connected forces interact, and a small event can produce a large effect. A negative online review cut one keyboard maker’s sales in half; a change in public opinion put plastic straws at risk. Global supply chains bring efficiency, but also multiply exposure to political conflict, disasters, and social change.
Uncertainty is not surrender. Forecasts and technology can help people see patterns, but neither can capture every relevant factor. The practical choice is not between perfect knowledge and helplessness. It is whether people participate in shaping what comes next or let others’ predictions and interests narrow their choices.
The distinction between complicated and complex is useful because it changes the question from How do we control this? to What can we influence, and what must we monitor? That shift is especially valuable in organisations that mistake a detailed plan for a reliable one. Yet uncertainty should not become an excuse to dismiss expertise: stable processes still benefit from measurement and planning. The stronger position is selective confidence. Use precision where conditions support it, and keep room to adapt where feedback can change the system itself.
Treat every forecast as a revisable probability, not a promise.
Forecasts have fingerprints. Early economic forecasters Irving Fisher, Roger Babson, and Warren Persons each believed their methods could reveal market laws. Their models left out data, reflected assumptions, and served commercial interests. Fisher wrongly called the market stable just before the 1929 crash. A modern example is the widely repeated claim that 47 percent of US jobs would disappear to automation by 2035. Later estimates varied widely, showing how a precise number can sound more certain than the evidence allows.
Make predictions testable. In Philip Tetlock’s Good Judgment Project, volunteers estimated probabilities, updated their views, and had their results scored. Teams outperformed control groups and professional analysts, but even the strongest forecasters were not reliable far into the future. A forecast is most useful when its assumptions, time horizon, and record are visible, and when it prompts questions rather than obedience.
Scoring forecasts turns confidence into something auditable. It also reduces the advantage of pundits who can make dramatic claims and move on before anyone checks them. The deeper lesson is institutional: organisations should preserve records of what they expected, why, and when they expected it. That practice can expose systematic overconfidence. Still, a forecast can be valuable without being correct if it reveals a vulnerability or triggers preparation. The test is not simply whether the prediction came true, but whether it improved judgment without disguising uncertainty.
Use history to generate questions, not to declare what happens next.
Similarity can mislead. France’s Maginot Line was built to defend against a German attack resembling the First World War. Germany instead invaded through Belgium and the Netherlands, bypassing the fortifications. In the Vietnam War, US leaders compared North Vietnam to Nazi Germany and feared repeating the appeasement of Munich. Familiar analogies made a different conflict look easier to understand than it was.
The past is material, not a script. Historian Roy Foster described events as choices made by people who could not know the ending. That approach makes room for contingency and change. Jacob Dunne, a young man convicted of manslaughter, had seen his life as fixed by failure. Meeting the victim’s parents and returning to education opened possibilities he had not imagined. History can reveal patterns worth investigating, but it cannot erase the differences that shape each new situation.
Historical analogies are powerful partly because they compress complexity into a memorable story. That makes them useful starting points, but dangerous stopping points. A sound comparison asks both what resembles the past and what does not, including the new information and actors absent from the earlier event. Dunne’s experience adds a personal dimension: a past can explain how someone arrived at a moment without dictating what they do next. This does not mean every person has equal freedom to change course. It means that biography should not be mistaken for destiny.
Refuse labels that turn partial data into a person's destiny.
People exceed their profiles. Genetic information can reveal risks, but genes interact with environment and experience. Identical twins do not necessarily share diseases: when one twin develops multiple sclerosis, the other has about a 30 percent chance, not certainty. Personality tests and hiring algorithms make similar mistakes when they treat past patterns as permanent truths. Amazon abandoned a hiring system after it showed a preference for men, partly because it learned from a history of male hires.
Labels can alter the outcome. In the Pygmalion effect, teachers told that randomly selected pupils were especially talented treated them differently, and those pupils later performed better. A label may become a self-fulfilling expectation, shaping opportunities as well as self-image. Profiles can start conversations, but they should not decide who deserves education, employment, or investment.
A profile does not merely describe a person when it is used to allocate opportunity. It changes the conditions under which that person can succeed, making prediction part of the causal system it claims only to observe. This creates a feedback loop: past bias informs a score, the score shapes treatment, and the resulting outcomes appear to confirm the score. The response is not to reject all data, but to ask who is missing from it, who can challenge its use, and whether rejected cases are ever evaluated for errors.
Run real experiments where need, resources, and commitment meet.
Test the world you inhabit. An experiment is a practical way to learn in a system whose full workings cannot be seen in advance. Rebecca Hosking’s film about ocean plastic did not change audience behaviour as she had hoped. She then tried a local campaign to remove plastic bags from Modbury, her Devon village. Independent shops joined, then larger retailers, and international attention helped the idea spread. The small intervention revealed possibilities that her original plan had missed.
Start with what is available. Hosking combined a clear need, relevant skills and access to a community, with enough personal commitment to keep going. At Buurtzorg, nurse Jos de Blok tested a simpler homecare model with twelve nurses. Patients recovered in half the time, and the expanded approach cost less than the traditional one. Neither project required a flawless forecast. Each required a measurable attempt, learning, and the willingness to adjust.
Experiments make uncertainty actionable: they exchange speculation for feedback from the real setting. They also limit the cost of being wrong when the test is small and reversible. The important caution is that a local success does not automatically scale. Modbury benefited from an unusual combination of trust, timing, and media attention; a result in one place may not travel unchanged. Replication should therefore mean testing the underlying idea in new conditions, not copying the visible procedure and assuming the same outcome.
Make competing futures vivid before choosing a strategy.
Scenario planning widens the field. It is a method for exploring several plausible futures and asking what each would demand, rather than predicting one outcome. At Shell, teams developed scenarios that challenged assumptions about oil prices. When prices fell in 1973, the company had already considered responses and was better prepared than competitors. Scenarios worked not because Shell controlled events, but because it had rehearsed alternatives.
Include people who see different worlds. Mexico Posibles brought together business leaders, officials, police, activists, Indigenous representatives, and people connected to organised crime to explore possible national futures. Their discussions exposed how differently people experienced the same country. In Ireland, a citizens’ assembly gave ordinary residents structured access to evidence and debate on abortion, before a referendum supported legalisation. Such processes do not guarantee agreement, but can make public choices more informed and legitimate.
A scenario is useful when it changes what people notice and prepare for, not when it masquerades as a forecast. Its value depends on whose knowledge enters the room. Including people with conflicting interests can reveal risks that a technically sophisticated but socially narrow team will miss. However, vivid scenarios can also normalise harmful futures or overwhelm participants. Good facilitation matters: it distinguishes imagining a possibility from endorsing it, and connects each imagined future to decisions that can be made now.
Protect attention and let unresolved questions mature into useful work.
Notice before explaining. Artists gather details without always knowing their purpose. Margaret Atwood collected news clippings that gradually revealed the real-world ingredients of The Handmaid’s Tale. Theatre director Peter Brook carried a memory of a man sitting outside a prison for more than fifty years before it became part of a play. Their work depended on sustained attention, reflection, and time for connections to emerge.
Begin before certainty arrives. Printmaker Norman Ackroyd could not know exactly how an image would look until he made it. Filmmaker Mike Leigh developed scenes through improvisation rather than a fully fixed script. Their process combined discipline with discovery: make something, assess it, revise it, and tolerate periods of doubt. This is not an argument for aimless wandering. It is a case for protecting the attention and freedom that let people notice what a preset plan would filter out.
Creative work shows that not all useful thinking is linear. Incubation can be productive because the mind keeps recombining observations outside a narrow task. But it is easy to romanticise delay: attention alone does not produce a finished work, and artists pair openness with technical mastery, deadlines, and repeated revision. The practical implication is to make space for exploration while retaining a point of contact with reality. A question earns its place not because it feels profound, but because sustained attention eventually helps someone make or see something differently.
Give long projects a purpose broad enough to outlast their founders.
A cathedral project has no neat finish line. The term describes ambitious work built across generations, with outcomes too uncertain to forecast fully. CERN, founded by European nations to pursue particle physics for peaceful purposes, developed the Large Hadron Collider through international collaboration. Its work also produced unexpected practical benefits, including the World Wide Web, which was created to help researchers share information and then released freely.
Purpose needs public trust. Barcelona’s Sagrada Família survived war, political hostility, changing artistic tastes, and the loss of its architect’s plans. Support from local residents, later joined by visitors from around the world, helped sustain construction. These projects need both a clear guiding principle and enough flexibility to adapt. Their legitimacy depends on people outside the organisation believing the work matters and having reason to trust how it is conducted.
Long horizons can make work more meaningful, but duration alone does not justify expense or protect a project from becoming self-perpetuating. The cases here point to two disciplines: a purpose strong enough to coordinate changing contributors, and continuing public scrutiny of whether that purpose still serves a wider good. CERN’s openness and the Sagrada Família’s broad public support are not decorative extras. They are part of the infrastructure that makes costly, uncertain work politically and socially sustainable.
Build trust before crisis; emergency solidarity cannot be summoned on demand.
People are crisis infrastructure. When Nokia’s mobile phone business faltered, chairman Risto Siilasmaa encouraged leaders to develop multiple options rather than cling to a single rescue plan. The company treated departing employees with support for retraining and new work, helping preserve trust among those who stayed. At Barings Bank, a team’s loyalty and daily relationships helped its corporate finance business survive even after the bank collapsed.
Collective action can save lives. During the AIDS crisis, activists learned enough about clinical trials and drugs to challenge medical institutions, while building support networks for people facing stigma and illness. No single leader or tactic solved the emergency. Progress depended on patients, scientists, advocates, and allies working across deep disagreements. The shared lesson is not that conflict disappears in a crisis, but that durable relationships make it possible to keep working through it.
Trust is a reserve of coordination. It lets people share bad news, disagree without immediate defection, and take risks for a collective purpose. That reserve is built through ordinary conduct, not crisis rhetoric. Yet solidarity should not be confused with loyalty to a failing institution or leader. Nokia’s multiple options and AIDS activists’ willingness to confront powerful organisations show that trust can support candour and dissent. Resilient groups need both attachment and the freedom to say when the current approach is not working.
Prepare with backup options, trusted partners, and public legitimacy.
Prepare for what cannot be forecast. The Coalition for Epidemic Preparedness Innovations funds several vaccine candidates for high-risk diseases, since no one can know in advance which candidate will work. It also develops fast manufacturing capacity and builds relationships with affected countries before an outbreak. During an Ebola outbreak in 2018, vaccination of contacts began within eleven days, fast enough to help prevent wider spread.
Protect options, not just plans. Cary Fowler helped create the Svalbard Global Seed Vault to preserve crop diversity against disasters that might destroy national seed collections. Wales’s Wellbeing of Future Generations Act requires public bodies to consider long-term effects across connected areas such as health, equality, and the environment. These efforts differ, but each invests early in resources, relationships, or laws that may matter when circumstances change.
Preparedness is a portfolio, not a prediction. Some capacities should be efficient once the situation is clear, such as scaling vaccine production. Others should be redundant before the danger is known, such as maintaining multiple vaccine candidates or preserving diverse seeds. The hard question is how to allocate limited resources among risks that compete for attention. Public legitimacy helps answer it: people are more likely to sustain preparation when they understand its purpose, can question its priorities, and see who benefits.
Treat death as part of life, and protect choices at its end.
Age can bring new freedom. Researcher Laura Carstensen followed adults from 18 to 94 and found that individuals tended to grow happier over time. As people sense that time is limited, they often focus more on emotionally meaningful relationships and activities. This is not a promise that ageing is painless. It is a reminder that older people remain capable of change, learning, and pleasure. The book’s examples include Glenda Jackson returning to the stage at eighty to play King Lear.
A good death is also a human experience. Hospice workers help patients and families discuss where care should happen, who should be present, and what matters most. Their emphasis is on comfort, dignity, and patient judgment, rather than treating death as a medical failure. By contrast, visions of indefinite life often assume that more years are automatically better, without asking what people value or what losses those years might bring.
The argument is strongest when it challenges the idea that medical intervention is always the highest good. Care should be measured against a person’s own priorities, not only biological duration. Still, the reported pattern of increasing happiness with age should not be universalised: illness, isolation, poverty, and unequal access to care shape later life profoundly. Respecting mortality does not require romanticising it. It requires making room for honest conversations, meaningful support, and choices that remain the patient’s own.
Analysis
Uncharted is both a critique of prediction and a proposal for practical action. Its argument moves from the limits of economic forecasts, historical analogies, personality profiles, and genetic claims to methods for navigating uncertainty: experiments, multiple scenarios, artistic attention, long-term collaborative projects, and preparation. The case studies make the central point tangible. Modbury’s plastic-bag campaign, Mexico Posibles, CERN, the AIDS response, and epidemic preparedness each show people learning and acting without a complete blueprint.
The book’s most important distinction is between uncertainty and helplessness. Uncertainty means outcomes cannot be fully specified in advance. It does not mean that evidence, expertise, planning, or responsibility are useless. Indeed, the strongest examples combine provisional knowledge with feedback, institutional memory, diverse perspectives, and the capacity to revise decisions. The book also offers a political warning: claims of inevitability can make choices appear closed, while systems that promise frictionless prediction may instead steer behaviour toward the interests of their designers.
One tension deserves attention. The emphasis on agency and participation can understate how unevenly people are positioned to act. A village organiser with local trust, a well-funded research institution, and a person constrained by poverty do not have equivalent resources or influence. Collective methods can widen participation, but they do not automatically correct unequal power. Nor should uncertainty be used to excuse indecision when evidence supports timely action.
The most useful synthesis is not to replace forecasts with intuition, but to make forecasts subordinate to learning. Use data to identify risks, state assumptions, and choose reversible actions where possible. Build reserves where failure would be catastrophic. Invite people with different knowledge into decisions, and make the reasoning visible enough to earn trust. The future remains uncharted because new knowledge and new events alter the terrain. Preparation is therefore less like drawing a final map than building the skills, relationships, and institutions that let people keep navigating together.
Review Summary
Uncharted received mixed reviews, with praise for its engaging writing style and thought-provoking ideas about navigating uncertainty. Readers appreciated Heffernan's insights on embracing complexity, human ingenuity, and the limitations of prediction. However, some found the book's structure disjointed and lacking cohesion. Critics noted that while the book offers interesting stories and perspectives, it fails to provide concrete strategies for navigating the future. Overall, readers found value in the book's exploration of uncertainty and its challenge to conventional thinking about forecasting and planning.
People Also Read
Glossary
complex systems
interacting, changing systemsSystems in which many connected elements affect one another, often nonlinearly. Small changes can produce outsized effects, and the system can change as people respond to it. Unlike a complicated but stable process, a complex system cannot be fully controlled by following a fixed set of rules or extrapolating from past patterns.
Good Judgment Project
forecasting accuracy tournamentA forecasting competition organised for the US intelligence community and studied by Philip Tetlock. Participants estimated the likelihood of world events, revised their judgments as evidence changed, and had their predictions scored. The project found that teams of open-minded volunteers could outperform control groups and professional analysts, while still making conspicuous errors.
automation paradox
automation weakens human skillsThe loss of human capability that can follow when a task is routinely delegated to technology. As people rely on automated systems instead of practising skills such as navigation or judgment, they may become less able to perform those tasks themselves. The paradox is that tools intended to increase capability can also increase dependence.
Pygmalion effect
expectations shape performanceThe tendency for expectations about a person to influence how others treat that person and, in turn, affect performance. In the classroom study described in the book, teachers were told that randomly selected pupils had strong potential. Those pupils later achieved better results, suggesting that expectations had subtly changed the opportunities and attention they received.
scenario planning
exploring plausible futuresA method for developing several evidence-based, plausible accounts of how the future might unfold. Rather than assigning certainty to one prediction, participants examine the forces and uncertainties behind each scenario, then consider what actions would make sense across them. The process aims to expand strategic options and improve readiness.
cathedral projects
ambitions spanning generationsLong-term projects whose purpose and scale exceed a single generation, and whose results cannot be completely predicted at the outset. CERN, the Human Genome Project, and Barcelona’s Sagrada Família illustrate the idea. Such projects depend on a stable guiding purpose, adaptable collaboration, and public legitimacy that sustains support over time.
interpersonal medicine
care built on relationshipsAn approach to healthcare that treats interactions among patients, clinicians, and caregivers as part of treatment, rather than as incidental to it. At the Dell Medical School, experiments included patient-controlled care rooms, longer conversations, and questions designed to understand patients’ lives. The aim is to improve decisions and outcomes through more human relationships.
Late Style
late-life creative reinventionA quality Margaret Heffernan identifies in some artists’ later work: rather than simply repeating earlier achievements, they distil experience and break toward new forms or questions. Examples include Beethoven’s late quartets and Matisse’s cut-paper works. The concept challenges the assumption that old age is only decline, while applying to a minority rather than everyone.
Wellbeing of Future Generations Act
law requiring long-term thinkingA Welsh law requiring public bodies to consider the long-term consequences of their decisions and work toward broad goals including health, equality, resilience, and environmental responsibility. The Act created a Future Generations Commissioner with powers to review decisions and make recommendations, giving future impacts a formal place in public policy.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.