David Deutsch discusses the four strands of his “fabric of reality” — the universality of computation, Popperian epistemology, evolution by natural selection, and quantum theory (specifically the many-worlds interpretation) — and how they interconnect to form a worldview in which knowledge creation is the central explanatory principle of the universe. He argues that most people, including experts, misunderstand the deepest implications of each strand, and that the connections between them reveal something profound about free will, wealth, optimism, and how to live.
The universality of computation
Turing’s key discovery is that computation is physically universal: a universal computer can simulate any physical process, not just other computers. This is a claim about physics, not mathematics, and it means there is an absolute limit to what can be computed by anything in the universe — aliens from Alpha Centauri cannot compute anything beyond what a Turing machine can, even with more speed or memory.
Mathematicians often miss this because they treat computation as a branch of pure math governed by unchangeable logical laws, when in fact it is a physical process governed by physical laws that could in principle be different.
The common objection “how do we know the brain is a computer?” reflects this misunderstanding. Turing’s thesis says any physical process can be computed; the tree is not a general-purpose computer, but a Turing machine can capture anything the tree does computationally. The brain’s status as a computer follows from the universality of computation, not from an analogy like “the brain is a steam engine.”
People also fail to grasp that universality implies a hard ceiling: nothing in the universe can exceed the computational power of a Turing machine. Quantum computers, if built, will also be universal — they won’t transcend Turing’s limit, just reach it more efficiently for certain problems.
The growth of knowledge begins with problems
The most important concept in Popper’s epistemology, Deutsch argues, is the concept of a problem. All previous epistemologies assumed knowledge must be justified by tracing it back to something certain — the senses, divine revelation, pure reason — and this has led to millennia of philosophical error, religious violence, and the torture of people over interpretations of ancient texts.
Popper’s insight is that knowledge begins with problems, not with foundations. A problem is an unsolved clash — once you recognize that two ideas cannot both be true, you have a problem, and the drive to solve it is what generates knowledge. There is no need to justify knowledge by going back to first principles.
This reframing eliminates the urge to wage wars over whose ancient authority is correct. The truth of Christianity, for example, does not depend on constructing a genealogy back to King David — that entire way of thinking is an epistemological error.
Problems are clashes between ideas
A problem, in Popper’s sense, is a clash between ideas, interpretations, or theories — not just about empirical phenomena but about anything: morality, mathematics, science. The key realization is that the clashing ideas cannot both be true (though they may both be false).
Even when a clash appears to be between an observation and a theory, it is really between theories: the theory of what the observation means versus the theory being tested. When Mercury’s orbit conflicted with Newtonian gravity, there were competing theories about the observations themselves (some claimed Eddington’s eclipse data was wrong), theories about auxiliary hypotheses (an unseen planet), and the two main theories.
Clashes between ideas are beneficial even if never resolved, because the confrontation changes the ideas themselves. Even if you leave an argument thinking you’ve won, you’ve developed a new angle on your own view. This is a beneficial side effect of Popper’s concept.
The clash between Newtonian gravity and general relativity was resolved because Einstein’s theory was the only good explanation left — an infinite number of bad explanations always remain, but only one good one survived the experimental and theoretical scrutiny.
Evolution is not the survival of the fittest
Most people who say they accept Darwinism immediately misstate it as “survival of the fittest,” which is tautological and not what the theory says. The theory is about the differential replication of gene variants — genes that encode knowledge about how to survive are the ones that propagate.
Deutsch argues the Lamarckism-vs.-Darwinism debate is more important than creationism-vs.-evolution, because the former is about what counts as a scientific explanation, while the latter is about whether one wants a scientific explanation at all.
Dawkins’s formulation: “Fitness is that quantity which appears to be maximized if what is actually maximized is the survival of genes.” This is a deep point disguised as a simple one — Darwin could have written the theory on one page, but it took a book to explain, and even now it is not fully understood.
Despite enormous computing power, no one has successfully simulated an open-ended artificial ecosystem. Simulated evolution always improves and then stops; real evolution is open-ended, accelerating, and constantly branching into new species. Robots using evolutionary algorithms only solve the specific problem they were programmed to solve — they don’t generate their own problems, which is the hallmark of genuine evolution.
Bad philosophy in quantum theory
Quantum theory is the strand people understand least, and the misunderstanding originated inside physics itself — a small founding subculture adopted positivism and mysticism (e.g., the observer’s consciousness changes reality), which was then transmitted to the public.
Physicists have used instrumentalism, positivism, bad empiricism (“shut up and calculate”), and sheer intimidation to resist the many-worlds interpretation. Students who question the orthodoxy face mistreatment and reduced standing.
The existence of many universes is actually one of the least surprising conclusions of quantum theory — it follows directly from any interference experiment like the double slit. Deutsch finds it far less counterintuitive than the curved spacetime of general relativity, which is genuinely hard to visualize.
Entanglement is more counterintuitive than parallel universes. The resistance to many-worlds seems driven by a lingering anthropocentrism — the desire to be at the center of everything, reemerging as “the observer” in quantum theory.
Free will is intimately connected with knowledge-creation
Free will is not solved by the multiverse, but arguments against free will based on physics are wrong because they misconstrue physics. The claim that “you’re just atoms, so free will is an illusion” is a category error.
Some theories only exist at certain levels of emergence. Thermodynamic irreversibility is real and law-like even though every individual molecule moves reversibly. You cannot detect irreversibility by looking at a single atom. Similarly, chemistry contains real explanations (acids, reactions) that cannot be usefully reduced to physics even in principle, because the computation required would exceed the resources of the universe.
The domino example: you could in principle trace every domino and predict which one never falls, but you would still not understand that it’s about prime numbers. Some explanations are not accessible from lower-level descriptions.
Free will is intimately connected with the creation of new explanatory knowledge. Einstein creating general relativity was not implicit in the Big Bang or in Einstein’s brain beforehand — it was genuinely created. This is the quintessential act of free will: creating something new that is not random (it’s the solution to a problem) but is also not predictable from prior conditions.
The “think of a random number” example is the opposite of free will, not an illustration of it. Free will is about creating knowledge that solves a problem, not about generating randomness.
Wealth is not a number, it is a set of transformations
Wealth, in Constructor-theoretic terms, is the set of all possible transformations an entity could bring about. It cannot be quantified as a single number — Mozart’s knowledge of musical beauty and Rothschild’s knowledge of banking are incommensurable.
Wealth grows when an entity has the right problems and the right solutions, enabling physical transformations. Creating knowledge grows wealth because knowledge is what makes transformations possible.
Knowledge is a form of information that is necessary for a physical transformation — it encompasses both genetic knowledge and explanatory knowledge in human minds. So far, the only known creators of knowledge are biological evolution and human thought.
The principle of optimism
The principle of optimism states that all evils are due to insufficient knowledge. If a transformation is physically possible, it can be brought about with the right knowledge. There is no law of nature that says problems are unsolvable.
This emerges from Constructor Theory’s insight that in the set of all possible physical transformations, the overwhelming majority can only be brought about by people creating knowledge. Without humans (or similar knowledge-creators), the universe would contain almost nothing beyond stars and black holes.
Deutsch is critical of techno-optimism that treats progress as inevitable — this is dangerous because it causes people to ignore real dangers. We can destroy ourselves; the fall of Athens and Florence show that civilizations do decline. Progress depends on our choices, not on laws of physics.
Constructor Theory
Constructor Theory is an attempt to reformulate all of science in terms of what transformations are possible and impossible, rather than the traditional framework of initial conditions and laws of motion. The traditional framework is time-asymmetric and depends on knowing initial conditions, which makes it ill-suited to explaining a universe whose future contains unbounded knowledge.
A law in Constructor Theory says: no matter what resources you bring, you cannot make a Tesla go faster than light. This is a statement about all possible transformations, not about what a system will do in isolation.
Constructor Theory can express the laws of thermodynamics exactly, including the first law in terms of information (not just the second law). Chiara Marletto has shown how to define the second law at a microscopic level using an adiabatic/non-adiabatic process distinction.
The goal is a timeless, explanatory framework rather than a predictive one. Good explanations should apply at any time, not depend on a specific initial moment.
How to make a better world
For an individual who wants to make a better world, the answer is the same as how to live a better life: follow your own problems. Norman Borlaug solved the problem of agricultural productivity because it figured large in his mind, not because he set out to save the world. Faraday discovered electromagnetic induction because he was fascinated by the problem, not because he intended to create generators that would become matters of life and death.
It is dangerous to adopt someone else’s problem or a problem you think is important in an abstract sense rather than one that genuinely engages your creativity. Newton described himself as a boy picking up particularly nice pebbles — he followed what interested him, and it turned out to be among the most important scientific problems of the age.
Nothing is guaranteed — doing the right thing can still lead to disaster, and doing the wrong thing can sometimes succeed. But the process of creative thought, driven by genuine problems, is what leads to knowledge growth, and knowledge growth is what makes a better world possible.
