Key Takeaways
Stop steering by distant goals when their stepping stones are unknowable
The route is hidden. Creative work can be understood as a search through a space of possible ideas, inventions, or careers. A distant objective sounds like a destination, but often gives no reliable clue about the next move. The authors call the intermediate discoveries that open new routes stepping stones. Their examples include vacuum tubes, developed through electrical research long before anyone used them to build computers, and radar technology that unexpectedly helped make microwave ovens possible.
The compass can mislead. A measure that rewards resemblance to the destination may reject the very discoveries that lead there. This is the book’s central paradox: ambitious achievements can become harder to reach when treated as direct targets. That warning is aimed at uncertain, far-off aims, not ordinary tasks such as following a recipe or meeting a clear specification.
The argument is strongest when the route is genuinely unknown. In familiar work, a clear target can coordinate effort and reveal whether a plan is working. In exploratory work, however, a target can narrow attention before anyone understands the terrain. This resembles the difference between navigation on a surveyed road and research in unmapped territory. The useful question is not whether objectives are good or bad, but whether the next step is knowable enough for the objective to guide it.
Let unexpected discoveries become routes others can extend
Pictures became stepping stones. Picbreeder was a website where users selected computer-generated images they liked, then bred variations from them. Each image had an artificial genetic description, so a chosen image could produce related offspring with small changes. Users could also publish an image for someone else to continue, allowing a chain of exploration to outlast any one person’s attention.
The car began as a face. One user continued an alien-like face, not a car. Across generations, its eyes shifted into forms that suggested wheels, and the resulting image became a car. The user had no way to plan this connection. The broader lesson is that discoveries can be valuable as launching points even when their creators cannot explain what they may eventually enable.
Picbreeder makes a powerful case for preserving intermediate work, not just polished outcomes. A failed prototype, odd sketch, or partial result can carry structure that a later explorer recognizes differently. This is also why open repositories and shared tools can matter: they lower the cost of continuing someone else’s line of thought. The approach depends on a space that permits useful variations, though. Without a workable representation, exploration can generate noise rather than fertile options.
When progress scores trap you, explore behaviors the score cannot reward
Novelty search changes the test. In the authors’ experiments, a robot in a maze was rewarded for behaviors unlike those already tried, rather than for getting closer to the exit. It might first crash into a wall, then try different routes. As familiar behaviors were exhausted, avoiding walls and passing through doors became ways to keep discovering something new, although nobody had set those as goals.
The comparison was striking. Across 40 trials, novelty search found a maze-solving behavior 39 times; objective-driven search did so 3 times. A simulated two-legged robot also learned to walk more successfully when searching for novel behaviors than when directly rewarded for walking farther. Yet novelty search could fail in very large mazes. It is a useful alternative in deceptive problems, not a universal solver.
The experiments show how a proxy can distort learning: a robot rewarded for approaching the exit may settle for a nearby dead end, while exploration keeps alternative routes alive. This has practical relevance wherever early performance can be misleading. Still, novelty is not automatically useful. A system can produce difference without insight, and the maze results depend on how behavior and novelty are represented. The method is most persuasive as a way to escape a misleading score, not as a replacement for every form of evaluation.
Build your career by following live curiosity, not a frozen job description
Careers often turn sideways. Johnny Depp joined a band because he wanted to make music; a connection through the band introduced him to acting. John Grisham practiced law for years before testimony in a case stirred the idea of writing fiction. Haruki Murakami ran a jazz bar before writing novels. Their stories do not prove that drifting guarantees success. They show how interests and experiences can create opportunities that a career plan could not have specified in advance.
Serendipity can be cultivated. A study cited by the authors found that nearly two thirds of adults attributed some part of their career choice to an unplanned event. Volunteering, joining clubs, and meeting people can increase the chances of such encounters. Exploration is not passive waiting; it means entering situations that expose a person to unfamiliar possibilities.
These stories challenge the idea that a coherent life must be designed from the beginning. A career can be built through accumulated skills and relationships whose value becomes visible only in retrospect. But famous success stories are vulnerable to survivorship bias: many people change direction without landing in a celebrated profession. The defensible lesson is not to abandon planning, but to avoid treating a plan as evidence that every worthwhile detour is a mistake.
Follow interestingness as a clue, not as a disguised finish line
Novelty points backward. The authors propose novelty as a practical guide when the future is too uncertain to measure. Novelty means difference from what has already been explored. Because the past is visible, it can help identify what is new without pretending to know which direction leads to a distant ideal. In computer experiments, searching for novel robot behavior tended to move from simple actions toward more complex ones, as new behavior required learning about walls, doors, and movement.
People add taste. Human exploration is not merely a hunger for difference. Personal knowledge and experience shape what feels interesting, even when its eventual use is unclear. Newton’s attention to a falling apple mattered because he pursued the observation into deeper questions. The authors argue that curiosity can make serendipity more than blind luck: people notice and investigate possibilities that a fixed score would miss.
Interest can function as a form of informed attention. A person may not be able to justify a hunch in advance, yet years of practice can make certain anomalies, questions, or combinations feel worth pursuing. That is different from treating every impulse as wisdom. Curiosity is more useful when paired with engagement, reflection, and a willingness to revise. The book’s strongest insight is that subjective interest can be evidence of potential, even though it cannot certify future value.
Protect learning from metrics that reward the proxy instead
Measures can change behavior. Campbell’s law describes how a quantitative indicator becomes vulnerable to distortion when it is used to make social decisions. If schools are judged by test scores, teachers have reason to focus on tested material. Scores may rise while broader understanding, creativity, or student independence receive less attention. The authors also point to perverse incentives: a bounty for dead cobras in colonial India reportedly encouraged people to breed cobras for payment, increasing the problem the policy meant to solve.
Accuracy is not enough. A more precise measure of progress toward an ambitious outcome may still point in the wrong direction. In Picbreeder, the images that led to a skull often looked nothing like skulls. Similarly, educational progress may require experiences that do not resemble high test scores in the short term. The proposal is not to reject all assessment, but to stop treating one score as the destination.
The distinction between a measure and the thing measured is central to education, management, and public policy. Once rewards attach to a proxy, people adapt to the proxy, sometimes at the expense of the underlying purpose. Yet the answer cannot be measurement-free governance: without evidence, poor teaching and unequal access can hide behind appeals to intuition. A more robust approach combines quantitative indicators with independent evidence, professional judgment, and room to test alternatives.
Fund science for promising questions, not consensus or predicted impact
Agreement can narrow the frontier. Research funding panels often favor proposals that many reviewers rate highly. The authors argue that consensus can reward safe compromise, while a genuinely unfamiliar idea may divide experts because its value is uncertain. They suggest that some resources should support disagreement and exploratory work, rather than only projects with predicted social impact or a clear route to a large result.
Useful work can look useless. Pure mathematics developed without practical applications in mind later supported fields such as chemistry, physics, and secure online commerce. The US space program did reach the Moon, but depended on earlier discoveries made outside that program, and produced technologies that were not its stated aim. These cases illustrate how the eventual importance of a project can exceed what reviewers could reasonably forecast beforehand.
This is a case for a portfolio of research, not for funding every eccentric proposal. Peer review helps detect weak methods and unsupported claims, but consensus and impact forecasts are poor at recognizing work whose significance depends on future discoveries. Funding systems could preserve both accountability and exploration by supporting varied projects, including some chosen for intellectual promise rather than immediate application. The difficult question is how to distinguish productive uncertainty from mere vagueness without making certainty the admission price.
Separate research exploration from choosing the best tool today
Benchmarks answer a limited question. In AI, researchers often compare a new algorithm with existing methods and favor higher performance or mathematical guarantees. Those tests can help a practitioner choose a tool for a current task. But the authors argue that the same measures can narrow research: an unfamiliar algorithm that performs slightly worse may still open a fruitful new direction, yet reviewers may reject it before others can explore its descendants.
Research needs another lens. The book imagines a journal whose reviewers cannot dismiss an idea solely for weaker benchmark results or missing guarantees. They must assess the idea itself: its novelty, elegance, or ability to inspire further work. This is not a claim that performance and proofs are useless. It is a call to distinguish evidence about present utility from evidence about a research path’s future potential.
The research-practice distinction clarifies why the same evaluation can be sensible in one setting and restrictive in another. A hospital choosing software needs reliable performance today; a laboratory exploring new methods may need permission to investigate ideas before their advantages are measurable. The challenge is institutional: review processes must guard against error without turning familiar success criteria into a filter against novelty. The proposed journal is a thought experiment, but it usefully asks reviewers to explain why an idea matters beyond its score.
Read evolution as branching experimentation, not a ladder toward humans
Survival can be a constraint. The authors offer an alternative interpretation of evolution: surviving and reproducing may define which organisms continue in the process, without being a goal that evolution consciously pursues. On this view, natural selection filters possibilities, while mutation, drift, and the emergence of new ecological niches contribute to diversity. Evolution does not climb toward a predetermined ideal, and humans are not its planned destination.
Old features find new jobs. Feathers first helped dinosaurs regulate temperature before becoming useful for flight. This kind of repurposing is called exaptation: a trait developed in one context later serves another function. The authors compare it with Picbreeder’s alien eyes becoming car wheels and vacuum tubes becoming computer components. They present this as a perspective on evolution’s creativity, not a replacement for established biological theory.
The perspective usefully separates selection’s filtering role from the creative variety of evolutionary outcomes. It also counters the temptation to treat humans as the natural endpoint of life’s history. Still, selection is not merely a brake: it can shape adaptations within ecological niches, and biological evolution has processes that an analogy with algorithms cannot fully capture. The book’s contribution is best read as a complementary lens on how novelty and repurposing accumulate, rather than a claim that fitness and selection do not matter.
Keep practical goals, but release control of distant destinations
Use goals within reach. The authors do not argue that every plan is harmful. Building to specifications, cooking from a recipe, or improving a familiar skill can benefit from clear objectives because the route is relatively known. Their warning applies when the destination lies many uncertain steps away, such as a transformative invention, a fulfilling career, or a major scientific breakthrough.
Choose the next promising step. In those cases, pay attention to what feels novel or interesting, preserve multiple paths, and allow other people to continue discoveries you cannot use yourself. This approach cannot promise a particular prize. The book’s claim is more modest: open exploration can reveal valuable possibilities that a narrow pursuit would overlook, and those discoveries may later make an ambitious objective reachable.
The practical value lies in matching the decision method to the uncertainty of the task. Clear goals support execution when cause and effect are well understood; exploration supports learning when the map itself is missing. This can be applied without romanticizing aimlessness: set constraints, protect time for curiosity, and periodically ask whether the next step is valuable in itself or only because it supposedly advances a distant vision. The destination may remain uncertain, but the quality of exploration is still something people can shape.
Analysis
The book’s central move is not simply to celebrate spontaneity. It argues that ambitious goals can misdirect search because intermediate discoveries often bear little resemblance to the eventual achievement. Picbreeder’s alien face becoming a car gives this abstract claim an unusually tangible form, while robot experiments provide a testable version: a system rewarded for novelty can outperform one rewarded for apparent progress toward a goal. The argument then travels from algorithms into careers, schools, science funding, AI research, and evolutionary biology.
Its most useful distinction is between known routes and unknown terrain. A clear objective is practical when the path is understood; it becomes dangerous when its measurement is mistaken for a reliable map. This distinction helps avoid the book’s most tempting misreading, that planning or evaluation should be abandoned. The authors themselves preserve modest objectives and acknowledge that novelty search cannot guarantee a chosen outcome. Their proposal is better understood as a change in emphasis: maintain constraints and accountability, but make room for exploration whose value cannot yet be scored.
The book is less conclusive when it generalizes from computational demonstrations and selected historical stories to complex institutions. A maze is not a school system, and celebrated career pivots do not show how often wandering ends badly. The analogy is illuminating, but it does not by itself establish how institutions should allocate scarce resources or distinguish fruitful exploration from waste. That gap does not erase the insight. It points toward a practical design problem: systems need both evaluation and protected space for ideas that have not yet learned how to look useful.
Ultimately, the work reframes achievement as a collective process. No single explorer can reliably predict the full chain of prerequisites. Communities make distant advances more plausible by preserving diverse discoveries, allowing others to reinterpret them, and resisting premature convergence on one approved route. The individual lesson is not to abandon ambition, but to loosen the demand that every worthwhile action prove its relevance to a predetermined future.
Review Summary
Why Greatness Cannot Be Planned challenges the notion that setting objectives is the best path to innovation and success. The authors argue that for complex, ambitious goals, rigidly pursuing objectives can be counterproductive. Instead, they propose a "novelty search" approach, emphasizing exploration, curiosity, and collecting "stepping stones" without a fixed end goal. While some readers found the book repetitive, many praised its thought-provoking ideas and their potential applications in various fields, from AI research to personal development. The book's core message resonated with many, encouraging a more flexible and open-minded approach to innovation and problem-solving.
Glossary
Search space
Possibilities available for explorationThe search space is the set of possible outcomes within a domain, such as all potential images, inventions, or robot behaviors. It may contain useful discoveries, ordinary possibilities, and vast regions of little value. The arrangement of possibilities matters because progress often depends on finding intermediate options that connect the present to something not yet imagined.
Stepping stone
Intermediate discovery opening possibilitiesA stepping stone is a discovery or capability that enables further exploration, whether or not it resembles a later achievement. Vacuum tubes, for example, became a component of early computers despite not being developed for computation. A chain of such steps can make ambitious outcomes possible without anyone having planned the whole route.
Deception
A misleading progress signalIn search, deception occurs when a measure of progress points away from a path that could lead to the desired result. A candidate may score poorly because it does not resemble the final target, even though it is an essential intermediate step. The authors use the term for a structural problem in complex search, not intentional trickery.
Novelty search
Search guided by behavioral differenceNovelty search is an algorithmic approach that prioritizes behaviors different from those already encountered instead of rewarding progress toward a specified objective. New behaviors are retained as possible starting points for further exploration. In the authors’ robot experiments, this approach sometimes discovered maze-solving and walking behaviors more reliably than objective-driven search, but it does not solve every problem.
Minimal criteria novelty search
Novelty search with basic viability rulesMinimal criteria novelty search combines exploration of novel possibilities with a limited requirement that candidates meet basic criteria. Instead of ranking all candidates by how well they approach a distant target, it allows diverse candidates to continue if they satisfy minimum conditions. The authors describe it as an algorithmic interpretation of evolution as exploration constrained by the ability to persist.
Exaptation
A trait repurposed for another functionExaptation occurs when a feature that developed or was used in one context later proves useful for a different purpose. The book’s example is feathers, which initially helped dinosaurs regulate temperature and later contributed to flight. Exaptation illustrates how useful structures can emerge without being designed for the eventual function they come to serve.
Campbell's law
Indicators distort when heavily rewardedCampbell's law describes the risk that a quantitative social indicator will be corrupted or will distort the process it is intended to monitor when used extensively for decision-making. In the book, school test scores illustrate the problem: when scores become the target, teaching may shift toward test performance rather than broad learning. The principle applies to metrics used for rewards, penalties, or institutional comparisons.
No Free Lunch Theorem
No search method wins everywhereThe No Free Lunch Theorem, as discussed in the book, states that no search algorithm performs best across all possible optimization problems. An approach that performs well on one class of problems must perform worse on another class under the theorem’s general assumptions. It supports the authors’ caution that no single method can guarantee arrival at every desired outcome.
FAQ
What's Why Greatness Cannot Be Planned about?
- Challenging Objectives: The book critiques the traditional reliance on objectives, suggesting that they can limit creativity and lead to mediocrity.
- Non-Objective Exploration: It introduces the concept of "novelty search," advocating for exploration without predefined goals, akin to a treasure hunt.
- Broader Implications: While initially focused on AI, the insights extend to education, culture, and personal achievement, promoting a mindset of curiosity and discovery.
Why should I read Why Greatness Cannot Be Planned?
- Unique Perspective: It offers a fresh viewpoint on achievement, suggesting exploration over strict goal-setting can lead to better outcomes.
- Practical Insights: The book provides applicable insights for various fields, encouraging a rethink of personal and professional challenges.
- Engaging Examples: Filled with anecdotes and case studies, it makes complex ideas accessible and relatable, challenging conventional wisdom.
What are the key takeaways of Why Greatness Cannot Be Planned?
- Objectives Can Mislead: The book argues that objectives can act as a "broken compass," leading to dead ends rather than fruitful exploration.
- Value of Novelty: Emphasizing novelty and creativity over rigid objectives can lead to unexpected and valuable discoveries.
- Diversity in Exploration: Encouraging diverse approaches fosters a richer environment for discovery, avoiding the pitfalls of uniformity.
What is "novelty search" as defined in Why Greatness Cannot Be Planned?
- Definition: Novelty search prioritizes discovering new and interesting behaviors or ideas over achieving specific objectives.
- Mechanism: It rewards behaviors different from previous attempts, leading to a natural progression from simple to complex.
- Applications: Demonstrated in AI experiments, it shows effective problem-solving without explicit objectives, promoting creativity.
How does Why Greatness Cannot Be Planned relate to education?
- Critique of Testing: The book critiques standardized testing, arguing it stifles creativity and leads to superficial learning.
- Encouraging Exploration: It suggests education should prioritize curiosity over rigid objectives, allowing organic knowledge discovery.
- Long-Term Implications: A non-objective approach can foster deeper understanding and creativity, leading to meaningful achievements.
What are some examples of deception in objective-driven pursuits mentioned in Why Greatness Cannot Be Planned?
- Chinese Finger Trap: Used as a metaphor for how rigid objectives can lead to counterproductive outcomes.
- Perverse Incentives: Historical examples, like the British snake bounty, show how objectives can create unintended consequences.
- Misleading Metrics: Focusing on metrics like test scores or GDP can lead to misleading conclusions about progress.
What are the best quotes from Why Greatness Cannot Be Planned and what do they mean?
- "Life is what happens...": This John Lennon quote highlights the theme that focusing too much on objectives can blind us to unexpected opportunities.
- "We don’t need objectives...": Emphasizes that exploration driven by curiosity can lead to greater achievements than rigid goals.
- "To achieve our highest goals...": Encourages letting go of rigid goals to unlock potential, suggesting true innovation arises from unexpected paths.
How does Why Greatness Cannot Be Planned suggest we can achieve greatness without objectives?
- Focus on Exploration: Prioritizing exploration and curiosity over rigid objectives can lead to discovering new paths and opportunities.
- Embrace Serendipity: Some of the best ideas come from unexpected moments rather than strict goal pursuit.
- Collect Stepping Stones: View experiences as stepping stones, building upon each discovery without being constrained by specific objectives.
What is the "treasure hunter" concept in Why Greatness Cannot Be Planned?
- Exploration Without Objectives: Represents a mindset focused on discovery rather than achieving specific goals.
- Collecting Stepping Stones: Each new idea leads to further exploration, contrasting with the linear path of objective-driven pursuits.
- Encouraging Creativity: Fosters creativity and innovation, allowing for organic development of ideas and solutions.
How does Why Greatness Cannot Be Planned relate to the field of artificial intelligence?
- Critique of AI Research: Argues that focusing too heavily on performance metrics can stifle innovation in AI.
- Importance of Novelty: Encourages exploring novel algorithms rather than just improving existing ones for breakthroughs.
- Meta-Search in AI: Discusses the complexity of searching for algorithms that themselves search, highlighting the depth of exploration needed.
What are the implications of Why Greatness Cannot Be Planned for innovation in business?
- Rethinking Objectives: Encourages businesses to abandon rigid objectives in favor of exploratory approaches for unexpected breakthroughs.
- Fostering Creativity: Advocates for environments that encourage diverse ideas and experimentation, helping businesses adapt and thrive.
- Learning from Failure: Emphasizes the value of learning from failures, fostering a resilient and innovative organizational culture.
How does Why Greatness Cannot Be Planned address the concept of serendipity?
- Serendipity in Discovery: Highlights the role of serendipity in natural evolution and human innovation, with significant discoveries often occurring unexpectedly.
- Encouraging Open-Mindedness: Promotes a non-objective approach, suggesting openness to unexpected findings can lead to valuable insights.
- Exaptation as a Mechanism: Discusses exaptation, where features evolve for one purpose but are used for another, illustrating flexible thinking leading to innovation.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.