Nobody Understands Quantum Mechanics. Everybody Uses It.

A few days ago, in a boardroom, a director interrupted my briefing on quantum technology with an accusation. “You’re asking us to bet money,” he said, “on something that, by your own admission, nobody in the world understands. Richard Feynman said so himself.”

He wasn’t being hostile. He was doing his job. Directors are supposed to challenge people who show up asking for budget, and he had done his homework: Feynman really did say, in a 1964 lecture at Cornell, “I think I can safely say that nobody understands quantum mechanics.” The line appears in The Character of Physical Law, and it may be the most quoted sentence any physicist has uttered in the past century.

I’ve heard versions of this challenge for years, in boardrooms, at conferences, and lately at dinner tables. This latest one just happened to arrive the same week I decided the objection finally deserves a full written answer.

Here is what I told the director, and what I want to unpack properly in this article: he was conflating two completely different meanings of the word “understand.” In one sense, Feynman was right, and the situation has barely changed in sixty years. In the other sense, quantum mechanics is the best-understood, most rigorously tested, most economically consequential theory in the history of science. Your phone works because of it. So does the MRI machine that found your friend’s tumor, the GPS constellation that routed your taxi, the laser that carried this article to your screen through a fiber under the ocean, and the flash memory it’s stored on.

The confusion between these two meanings does real damage. I meet intelligent, accomplished people who have quietly filed quantum computing in the same mental drawer as astrology, because they’ve absorbed the idea that quantum mechanics is a speculative, half-baked theory that physicists themselves admit they can’t make sense of. That idea is wrong, and it’s worth taking the time to see exactly where it goes wrong.

I’ll admit this article is overdue. For years I wrote for a self-selected audience of security architects, quantum engineers, and researchers who arrived already knowing the basics, so I never covered them. Even my purely educational content from ten years ago was written for 3rd year students with basics already covered. The surge of mainstream interest in quantum technology has changed who reads this site. If you’re one of the newer readers, this piece is the grounding I should have written long ago, and it’s the one I’ll be pointing people to whenever related questions come up.

What Feynman Actually Meant

Feynman was not confessing that quantum mechanics might be mistaken. By 1964 he had already done the work that would win him the Nobel Prize: quantum electrodynamics, the quantum theory of light and electrons, which he helped build precisely because the mathematics works so astonishingly well. A man does not spend his career refining a theory he considers incomprehensible nonsense.

What Feynman meant is that quantum mechanics resists the kind of understanding we get from everyday experience. When you learn classical mechanics, the equations describe things you can picture: balls rolling, planets orbiting, springs compressing. Your intuition, trained by a lifetime of throwing objects and watching them land, eventually clicks into agreement with the math. Quantum mechanics never gives you that click. The math describes electrons that take every available path at once, particles whose properties don’t exist until measured, and pairs of photons that behave as a single object across any distance. You can calculate all of it. You cannot picture it, and the theory gives you no story about what is “really happening” between measurements that all physicists agree on.

So the honest statement of our situation is this: we possess a mathematical framework of extraordinary precision, and we lack consensus on what that framework tells us about the underlying nature of reality. Physicists call the first part the formalism and the second part the interpretation. The formalism is settled science. The interpretation is a live debate.

The director in that boardroom, like most laypeople, heard “nobody understands quantum mechanics” and assumed it referred to the formalism. It refers to the interpretation. That single mix-up is the source of nearly all the witchcraft talk.

The Most Tested Theory in the History of Science

Let me put a number on “settled science.”

An electron behaves like a tiny magnet, and quantum electrodynamics predicts the strength of that magnet from first principles. The prediction requires summing thousands of so-called Feynman diagrams, each representing a way the electron can interact with fleeting virtual particles. It is one of the hardest calculations physicists perform. In 2023, a team led by Gerald Gabrielse at Northwestern measured the electron’s magnetic moment to a precision of roughly one part in ten trillion, using a single electron suspended in a trap for months.

The measurement and the prediction agree. Across roughly twelve significant figures, theory and experiment match. There is no other domain of human knowledge, none, in which we can predict anything with that accuracy. If you measured the distance from New York to Los Angeles with the same relative precision, you would know it to within the width of a bacterium.

And this is one data point among thousands. Atomic spectra, chemical bond energies, semiconductor band gaps, superconducting transition behavior, the anomalous heat capacity of solids that first puzzled Einstein in 1907: quantum mechanics predicts all of it, across a century of increasingly hostile experimental scrutiny. The theory has been tested in tabletop optics labs, in particle accelerators, in the cores of stars (via the spectra we observe), and in the semiconductor fabs that stress-test its predictions billions of times per second on every wafer. It has never failed. Not once, in a hundred years, within its domain of applicability.

Compare that record to the everyday theories we trust without hesitation. Newtonian gravity, the physics we bet lives on every time a plane takes off, is known to be wrong at high precision; GPS satellites would drift by kilometers per day if engineers didn’t correct for the relativistic effects Newton missed. Nobody calls Newtonian mechanics witchcraft. Yet quantum mechanics, which has a strictly better empirical record, gets the reputation problem.

A Century of Quiet Quantum Engineering

The deeper answer to the director’s challenge is not the precision record, impressive as it is. The deeper answer is that his company, like every company, was already running on quantum mechanics. Had been for decades. He just hadn’t been told.

The transistor. In 1947, John Bardeen, Walter Brattain, and William Shockley at Bell Labs built the first transistor, an achievement that earned them the 1956 Nobel Prize in Physics. The transistor is not a device that happens to involve some quantum effects at the margins. It is a quantum device, full stop. Its operation depends on band theory, the quantum-mechanical description of how electrons occupy allowed energy ranges in a crystal, and on the deliberate engineering of those bands through doping. Bardeen and Shockley didn’t stumble onto the transistor by tinkering; they calculated their way to it using the quantum theory of solids developed in the 1930s. Every chip fabricated since is a descendant of that calculation. A modern processor contains tens of billions of transistors, each one a small monument to our ability to compute what electrons do in silicon. The semiconductor industry, now well over $600 billion in annual revenue, is applied quantum mechanics with a supply chain attached.

The laser. Einstein worked out the principle of stimulated emission, the physical mechanism behind every laser, in 1917, decades before anyone built one. It is a purely quantum phenomenon: an excited atom, nudged by a passing photon, releases a second photon that is a perfect copy of the first. Theodore Maiman demonstrated the first working laser in 1960, and today lasers cut steel, correct vision, read barcodes, and carry essentially all intercontinental data traffic through optical fiber. When you stream a video hosted on another continent, quantum mechanics is doing the carrying.

The LED and the solar cell. Both are band-gap engineering, the same quantum theory of solids that gives us the transistor, run in opposite directions. An LED converts electrical energy into photons of a color determined by a quantum energy gap that engineers select by choosing the semiconductor alloy. The blue LED alone, which required taming gallium nitride, earned its inventors the 2014 Nobel Prize and made efficient white lighting possible. A solar cell runs the film backward, using the photoelectric effect, the phenomenon whose quantum explanation won Einstein his own Nobel, to turn photons into current at grid scale.

MRI. Magnetic resonance imaging exploits the quantum spin of hydrogen nuclei in your body’s water. Protons in a strong magnetic field occupy discrete quantum energy states; radio pulses flip them; the signal they emit as they relax back reveals the tissue around them. Spin has no classical analog whatsoever. It is not a tiny ball rotating. It is a purely quantum property, and roughly 100 million MRI scans a year depend on our ability to manipulate it with enough precision to distinguish a tumor from healthy tissue.

Atomic clocks and GPS. Every GPS position fix is a timing measurement, and the timing comes from atomic clocks that count the oscillations of microwave radiation absorbed and emitted by cesium or rubidium atoms as they jump between two specific quantum states. The second itself, the SI unit, has been defined since 1967 as exactly 9,192,631,770 oscillations of the cesium transition. When your phone tells you which side of the street you’re on, it is consulting a quantum standard orbiting 20,000 kilometers overhead.

Flash memory. The storage in your phone holds data using quantum tunneling, the effect by which electrons pass through an insulating barrier that classical physics says is impenetrable. Engineers do not merely tolerate tunneling; they meter it, pushing precise numbers of electrons through an oxide layer onto a floating gate and reading them back years later. Trillions of tunneling events, executed on command, in a device you keep in your pocket next to your keys.

Chemistry and materials. All of chemistry is quantum mechanics. The periodic table’s structure, the shapes of molecules, the strength of every bond in every material and every drug: these follow from the quantum behavior of electrons in atoms. Computational chemistry, which pharmaceutical and materials companies use daily, is the business of solving quantum equations approximately and profitably.

Former Lockheed Martin CEO Norm Augustine once remarked, half-jokingly, that a third of GDP is attributable to quantum mechanics. Nobody has produced a rigorous accounting, and I treat the specific fraction as folklore. But walk through the inventory above (semiconductors, lasers, telecommunications, lighting, solar, medical imaging, positioning, data storage, chemistry) and the folklore starts to look conservative. The modern economy is not about to be transformed by quantum mechanics. It was built on it, one calculated device at a time, starting about a hundred years ago.

This is the record the “quantum is witchcraft” crowd has to explain away. Witchcraft does not ship at wafer scale.

What We Genuinely Don’t Understand

Now for the honest part, because the credibility of everything above depends on not hand-waving the mystery.

The unresolved problem in quantum mechanics is usually called the measurement problem. The theory’s central equation, the Schrödinger equation, describes quantum systems evolving smoothly through superpositions, combinations of multiple states at once. An electron can be in a superposition of two locations, and the equation happily tracks both. Yet when we measure the electron’s position, we always find it in one place, with probabilities given by a rule Max Born wrote down in 1926. The formalism tells us with perfect accuracy what probabilities to expect. It does not tell us what happens during the measurement, why we see one outcome rather than the superposition itself, or whether “measurement” is even a fundamental process or just ordinary physics we haven’t finished describing.

Every interpretation of quantum mechanics is an attempt to answer those questions. Four are worth knowing by name.

The Copenhagen interpretation, associated with Niels Bohr and Werner Heisenberg, is the pragmatic elder. It holds that the wavefunction is a tool for computing measurement outcomes and that asking what the electron is “really doing” between measurements is a question physics cannot answer and perhaps shouldn’t ask. Generations of physicists were trained in this tradition, often compressed into the working motto “shut up and calculate.”

The many-worlds interpretation, proposed by Hugh Everett in his 1957 paper in Reviews of Modern Physics, takes the Schrödinger equation with complete seriousness: superpositions never collapse. When you measure the electron, you and your measuring device join the superposition, and every outcome occurs, each in its own branch of an ever-splitting universal wavefunction. The mathematics is arguably the cleanest of any interpretation. The ontological price, an unimaginably vast multiverse, is one many physicists decline to pay.

Pilot-wave theory, developed by Louis de Broglie and revived by David Bohm in his 1952 papers in Physical Review, restores a definite reality: particles always have exact positions, and a physical wave guides their motion. Determinism returns, at the cost of accepting that the guiding wave connects distant particles instantaneously, a feature Einstein would have hated even more than the dice he famously complained about.

QBism, developed by Christopher Fuchs, N. David Mermin, and Rüdiger Schack and introduced formally in 2014, goes the other direction, treating the wavefunction not as a physical object at all but as an expression of an observer’s expectations, something closer to a betting slip than to a description of the world. On this view the measurement problem dissolves, because there was never a physical collapse to explain, only an agent updating beliefs.

There are others, including objective-collapse models that modify the Schrödinger equation itself and, unusually among interpretations, make testable predictions that experimentalists are actively hunting. I’ll confess a personal stake in this corner of the debate: I’ve published my own speculative entry, a thought experiment I call Information-Triggered Collapse (ITC), which conjectures that superpositions reduce when the information content of a quantum system and its environment crosses a critical threshold. (I make no claim that it advances the frontier of physics; I published it as an open notebook entry, in case someone someday shows that collapse really is an informational phase transition and I get to say I wrote it down first.) I mention it here mostly as evidence of how irresistible the measurement problem is. Even people who spend their days on the engineering side of quantum technology end up sketching answers in the margins.

But here is the fact that matters for this article: with narrow exceptions, these interpretations predict identical experimental outcomes. They are not competing theories in the usual scientific sense. They are competing stories about the same flawlessly performing mathematics. A century of debate, including contributions from Einstein, Bohr, Bell, and every generation since, has not produced an experiment that crowns a winner, and it is possible none ever will.

I won’t pretend the debate is idle philosophy. John Bell’s theoretical work in the 1960s on quantum entanglement, motivated squarely by these foundational questions, led to the experiments by Alain Aspect, John Clauser, and Anton Zeilinger that earned the 2022 Nobel Prize in Physics and closed the door on the “hidden local reality” Einstein had hoped for. Those experiments confirmed that entanglement is exactly as strange as the formalism says. Foundational research has repeatedly sharpened our grip on the theory and even seeded new technology. The mystery is real, it is profound, and chasing it has been productive.

But notice what kind of mystery it is. Nobody disputes what the equations predict. Nobody has ever caught the equations in an error. The argument is entirely about what the equations mean.

The Mystery Never Touches the Machines

This is the crux, so let me state it as plainly as I can.

An engineer designing a transistor, a laser, an MRI sequence, or a quantum processor uses the formalism. She writes down Hamiltonians, solves the Schrödinger equation (or lets software solve it), applies the Born rule, and gets numbers: energy levels, transition rates, gate fidelities, error probabilities. At no point does the calculation ask her to declare whether the wavefunction physically collapsed, branched into parallel worlds, rode a pilot wave, or updated somebody’s beliefs. The interpretations are, by construction, invisible to the engineering. A Copenhagen loyalist and an Everettian sitting at adjacent lab benches will design the same device, predict the same behavior, and measure the same results, then argue about what it all meant over beer.

Humanity has been in this position before, and it has never stopped us. Roman engineers built aqueducts that still stand without a theory of stress tensors. Sadi Carnot worked out the efficiency limits of heat engines in 1824 while believing heat was a fluid called caloric, a substance that does not exist; his results were correct anyway and became the second law of thermodynamics. Nineteenth-century chemists synthesized dyes and drugs for decades while eminent colleagues insisted atoms were a useful fiction. In every case, a reliable calculational framework outran the settled account of what the framework described, and the technology worked fine while the philosophers caught up.

Quantum mechanics is the most extreme instance of this pattern, in both directions: the calculational framework is the most accurate ever constructed, and the gap between calculation and settled meaning is the widest. The strangeness of the second fact says nothing against the first.

So when someone tells you quantum computers can’t be trusted because “even physicists admit nobody understands quantum mechanics,” the response has two parts. First, the mechanism a quantum computer relies on, superposition and entanglement evolving under the Schrödinger equation, is precisely the part of the theory that has been verified to twelve decimal places and has powered a trillion-dollar electronics industry for seventy-five years. Second, whatever quantum computing’s genuine challenges are (and I’ve written extensively about the engineering hurdles, which are serious), “the underlying physics might be witchcraft” is not among them. The hard questions about quantum computers are engineering questions: error rates, scaling, control systems, manufacturing. They are the same kind of question that faced the transistor in 1948. The physics beneath them is the surest ground in all of science.

Mysterious and unreliable are different properties. Quantum mechanics is maximally the first and, on all evidence, minimally the second. Holding both facts at once is the entire trick.

What I Told the Director

I told him a version of the Carnot story, and then I put it this way: “You already trust quantum mechanics with your life. From pacemakers and MRI suite, your cell phone, when your plane navigated by GPS, and when all of your tech ran on silicon. The only question on the table is whether we can control quantum systems in a new way, not whether the physics behind them is sound. That question was settled before you were born, by every device you’ve ever owned.”

The conversation turned, as these conversations eventually do, to better questions: error correction, vendor claims, timelines. Those are the questions that actually deserve a director’s skepticism, and I was glad to spend the rest of the meeting on them.

Feynman’s line will keep circulating, and it should; it’s true, in the sense he meant it. We still do not know what quantum mechanics is telling us about reality, and I find that genuinely wonderful, one of the great open questions our species gets to work on. There is even a pleasing possibility that the machines will return the favor: large-scale quantum computers, which are in effect the most controlled quantum systems ever built, may eventually probe the foundations, testing whether superposition survives at scales where some collapse models predict it shouldn’t. The tool built from the mathematics may end up interrogating the mystery.

Until then, the working position is the one physics has occupied, profitably, for a century. We don’t fully understand quantum mechanics. We understand how to use it better than we understand how to use almost anything else. Those two sentences are both true, and the entire modern world is the proof.