The Structure of Scientific Revolutions · Thomas S. Kuhn · 1962
Hands down one of the most impactful books I’ve read. Science might just be the best system our species has produced, and this book will help you see it in a new light. The implications of this book are broader than author envisioned, I believe there are lessons in it for those studying technology adoption and change management.
Key Insights
Science presents itself as a constellation of facts, theories, and methods, each building on the last, creating a growing stockpile of scientific technique and knowledge.
Normal science is a slow and methodical exploration of the space. The community focuses on normal problems that ought to be solvable by known rules and procedures. Everything is predicated on the assumption that the scientific community knows what the world is like.
As anomalies accumulate, so begins the extraordinary investigations that lead the profession at last to a new set of commitments. Scientific revolutions are tradition-shattering complements to the tradition-bound activity of normal science.
Copernicus, Newton, Lavoisier, and Einstein all forced the community's rejection of one honored scientific theory in favor of another incompatible with it.
Those scientists who contributed heavily to the old paradigm now have the most to lose.
New discoveries then aren't simply added to a body of work. Instead, they cause a re-evaluation that alters foundational concepts, shifting the perspective and history of an entire scientific community.
Normal science is research firmly based upon past scientific achievements that supply the foundation for further practice.
These foundational insights are called paradigms: they need to attract scientists away from competing modes and be sufficiently open-ended to leave problems for practitioners to resolve.
The study of paradigms (in textbooks) is what prepares students for membership in the scientific community.
The successive transition from one paradigm to another via revolution is the usual developmental pattern of mature science.
In the absence of a paradigm, all facts that could pertain to the development of a given science are likely to seem equally relevant. Early fact-gathering seems far more random and produces a morass.
A paradigm increases the effectiveness and efficiency of research.
A new paradigm implies a new and more rigid definition of the field. Those unwilling or unable to accommodate their work to it must proceed in isolation or attach themselves to some other group.
When scientists can take a paradigm for granted, they don't have to start from scratch. They can pick up a textbook and begin their research where it leaves off. That changes the nature of the work to be more subtle and esoteric. Typically work is addressed only to professional colleagues who share that paradigm, requiring translation for the layman.
A widening gulf separates professional scientists from other fields, but this mechanism is intrinsic to scientific advance.
Achieving a paradigm that guides the whole group's research is almost the definition of 'a field of science.'
Paradigms gain their status because they are more successful than their competitors in solving the important problems recognised by the group.
The paradigm represents work that has been done once and for all.
Confidence in the paradigm, restricts research and focuses attention on a small range of relatively esoteric problems.
The overwhelming majority of the problems undertaken by even the very best scientists usually fall into one of these three: determination of significant fact, matching of facts with theory and articulation of theory.
Normal research problems don’t aim to produce major novelties. Much of the result is known in advance. They retain their significance to scientists because they add to the scope and precision of the paradigm.
The scientific community of a paradigm chooses problems that can be assumed to have solutions.
These conceptual, theoretical, instrumental, and methodological commitments relate normal science to puzzle-solving. The rules allow the practitioner to concentrate on esoteric problems, challenged by how to bring the puzzle to a resolution.
Members of the community learn their trade by studying and practicing the accepted rules of the paradigm.
Debates over legitimate methods, problems, and standards of solution spike before and during scientific revolutions when paradigms are under attack and subject to change.
Discoveries begin with an awareness of an anomaly; when nature violates the paradigms expectations. They continue with extended exploration of the area. They close only when the paradigm theory has been adjusted so that the anomalous has become the expected.
Discovery takes time as it requires the observations to be assimilated into theory. The consequent change of paradigm categories and procedures is often accompanied by resistance. The fact is only seen as scientific after the adjustment of the theory.
The discovery is completed when the theory is adjusted to expect the anomaly.
Despite the fact that normal science is not directed at novelty (it initially tends to suppress it) it nevertheless is effective in causing them to arise.
The first paradigm typically accounts quite successfully for most of the observations and accessible experiments. Further development calls for construction of more elaborate equipment, vocabulary and skills. Professionalisation of this kind leads to a restriction of the scientist’s vision and to resistance to paradigm change.
Professionalisation makes science increasingly rigid, but funnels attention such that observation-theory match achieves great precision.
The more precise a paradigm is, the more sensitive an indicator it provides of anomaly and therefore a paradigm change.
Once a discovery had been assimilated scientists should able to account for a wider range of natural phenomena at greater precision.
Gains are only achieved by discarding previous beliefs and by replacing components of the previous paradigm.
Emergence of new theories sparks a period of professional insecurity with the persistent failure of the puzzles of normal science to come out as they should. Failure of existing rules is the prelude to a search for new ones.
The novel theory seems a direct response to crisis.
In the absence of crisis, better solutions have been ignored. You need to make contact with a troubled spot of an existing paradigm.
Retooling is an extravagance to be reserved for the occasion that demands it.
A scientific theory is only declared invalid if an alternate candidate is available to take its place. The decision to reject one paradigm is to simultaneously accept another.
Defenders of a theory will devise many articulations and ad hoc modifications in order to eliminate any apparent conflict caused by anomaly.
Normal science continually strives to bring theory and fact closer, it’s the search for confirmation or falsification.
Science students accept theories on the authority of teacher and text, not because of evidence.
If an anomaly is to evoke crisis, it must usually be more than just an anomaly.
An anomaly can cause a transition from normal science to crisis to extraordinary science. More and more attention is devoted to the anomaly, creating a focal points for scientific scrutiny. Soon scientists view its resolution as the subject matter of their discipline.
Einstein on paradigm change
Crisis begin with the blurring of a paradigm and the consequent loosening of the rules for normal research. Research during crisis resembles research during the pre-paradigm period.
A crisis in science can be resolved either by the existing paradigm proving capable of handling the problem, by the problem being set aside for future generations with more advanced tools, or by the emergence of a new paradigm that sparks a battle over its acceptance.
The transition from an old paradigm to a new one involves a fundamental reconstruction of the field, changing its theoretical foundations, methods, and applications, rather than a cumulative process of extending the old paradigm. It is usually achieved by young scientists or those new to the field, as they are less committed to the traditional rules and more likely to conceive of a new set of rules to replace them.
Scientific revolutions, like political revolutions, occur when existing institutions or paradigms fail to adequately address new problems, leading to a crisis. Society divides into competing camps supporting either the old or new paradigm.
Proponents of each paradigm use their own framework to argue its defence, and while persuasive, these arguments alone cannot conclusively settle the debate. The choice between paradigms is also a choice between incompatible modes of community life.
Textbooks obscure the fact that during revolutions, the entire network of fact and theory shifts. They can do so because they are rewritten each time.
Theories do not evolve piecemeal to fit pre-existing facts; instead, they emerge together with the facts from a revolutionary reformulation of the preceding scientific tradition.
New paradigms emerge in the minds of a few individuals, usually young or new to the field, who are less committed to the old paradigm and have focused intensely on the crisis-provoking problems.
Paradigm testing involves competition between rival paradigms for the allegiance of the scientific community. New paradigms incorporate much of the old paradigm's vocabulary and apparatus but often use them differently. Proponents of different paradigms practice their trades in different worlds, seeing different things and relationships.
The transition between paradigms cannot be made gradually or forced by logic and neutral experience; it must occur all at once, like a gestalt switch. Resistance is inevitable, and paradigm change cannot be justified by proof, but arguments can still persuade scientists to change their minds.
Welcome to the Era of Experience
David Silver, Rich Sutton. 2025. (View Paper → )
We stand on the threshold of a new era in artificial intelligence that promises to achieve an unprecedented level of ability. A new generation of agents will acquire superhuman capabilities by learning predominantly from experience. This note explores the key characteristics that will define this upcoming era.
AI is transitioning from an era dominated by human data to one driven by experiential learning. While current AI systems like large language models have achieved impressive capabilities by training on human-generated data, they're approaching the limits of what can be learned from existing human knowledge. The authors propose that the next breakthrough will come from agents that learn continuously through direct interaction with their environments, generating their own data that improves as they become stronger. This new paradigm will feature agents that operate in continuous streams rather than discrete episodes, with actions and observations grounded in real-world environments, rewards derived from measurable outcomes rather than human judgments, and reasoning processes that aren't limited to human modes of thinking. The authors cite AlphaProof's medal-winning performance in the International Mathematical Olympiad as evidence this transition has already begun, and suggest experiential learning will ultimately enable superhuman capabilities across many domains by allowing AI to discover knowledge beyond current human understanding.
/w=3840,quality=90,fit=scale-down)
Book Highlights
After code is written, it is very difficult to throw it out. Like writers in love with their prose, programmers tend to have emotional attachments to their algorithms. Alan Cooper · The Inmates Are Running the Asylum
No needless parts. Every design element, from the largest to the smallest, must have a purpose, and contribute to the purpose of a larger element it is part of. If you can’t explain what an element is for, most likely it shouldn’t be there. Alla Kholmatova · Design Systems
Most people do not listen with the intent to understand; they listen with the intent to reply. Roman Pichler · How to Lead in Product Management