Even Physicists Scratch Their Heads: A Candid Confession About Quantum Physics

A scholar named Doug from the University of Southampton wrote a short essay titled "I Don't Understand Quantum Physics," which garnered 206 upvotes and 146 comments on the tech community Hacker News. That number is interesting in itself—an article admitting "I don't understand" sparked the liveliest discussion.

When Experts Also Admit Confusion

The mathematical framework of quantum mechanics is unsettlingly precise. From semiconductors to nuclear magnetic resonance, from lasers to quantum computing, this theory underpins half of modern technology. But the core argument of Doug's essay is quite plain: being mathematically useful does not equal truly understanding.

He describes this state as a kind of "split"—we can skillfully calculate, predict, and verify, yet when asked "what is a particle, really?" we are at a loss for words. This is not a personal failing but a long-standing public embarrassment across the entire physics community. Feynman once said "no one truly understands quantum mechanics," and this repeatedly quoted line precisely shows that confusion is not the exclusive domain of beginners.

Where Does the Confusion Come From

The essay highlights several key tensions. The first is wave-particle duality: light sometimes behaves like a wave, sometimes like a particle, and the phrase "sometimes" itself dodges the question—what is it, actually? The second is the measurement problem: before observation, a system is in a superposition, and measurement causes collapse. Who is measuring? What counts as a measurement? Is collapse a physical process or an epistemological description? The third is non-locality: correlations between entangled particles seem to span space instantaneously, yet do not allow information transfer, challenging intuitive notions of causality.

These are not philosophical embellishments but real contradictions written into textbook formulas. The Copenhagen interpretation chooses to "shut up and calculate," the many-worlds interpretation introduces branching universes, and decoherence theory attempts to explain the emergence of the classical world—each approach avoids or redefines some problem.

The Copenhagen interpretation is the most mainstream orthodox interpretation of quantum mechanics, established by Bohr and Heisenberg in the 1920s. Its core claim is that the wave function of a quantum system does not describe objective reality but only provides probabilistic predictions for measurement outcomes; discussing "where a particle actually is before measurement" has no physical meaning. The so-called "collapse" is not a real physical process but an update of our state of knowledge after obtaining observational information. This pragmatic stance—"shut up and calculate"—allowed physics to bypass ontological controversies and move forward, but it is also the institutionalized source of what Doug calls the "split": it excludes confusion from within the theory rather than resolving it.

The many-worlds interpretation was proposed by Hugh Everett in 1957. It eliminates the collapse postulate: the wave function always evolves unitarily according to the Schrödinger equation, never collapses, and all possible measurement outcomes are realized in continuously branching parallel universes. This approach is mathematically extremely simple, requiring no additional collapse mechanism, at the cost of extreme ontological inflation—the number of universes grows exponentially with each quantum event. It has long been a minority position in physics but has gained new attention in the contexts of quantum computing and quantum gravity.

Decoherence theory emerged in the 1980s. It does not provide a new interpretation but explains why the classical world "emerges." The core mechanism is: interaction between a quantum system and a vast environment dilutes interference terms to unobservability within an extremely short time (imperceptibly short for macroscopic objects), the system's density matrix tends toward diagonalization, and superpositions practically degrade into classical probability mixtures. Decoherence solves the question of "why we don't see macroscopic superpositions" but does not solve the single-measurement problem of "why measurement yields only one result"—the latter remains contested among various interpretations.

Why Admitting Confusion Is Worth Writing About

In the tech community, this essay resonated beyond pure popular-science traffic. The reason may be that acknowledging cognitive boundaries is an underrated professional virtue.

What makes quantum physics special is that it is already the foundation of engineering practice. Engineers use quantum mechanics to design devices, chemists use it to explain bonding, yet controversy over the underlying picture has never subsided. This state of "using without understanding" parallels what many technical fields face—we make heavy use of deep learning tools yet lack a complete explanation of the model's internal representation mechanisms; we rely on compiler optimizations yet may not be able to explain the logic of each step.

Doug's essay offers no answers, not even new questions, but it offers an attitude: drawing an honest line between precise calculation and true understanding is more valuable than piling up terminology to create an illusion of "known."

The "lack of complete explanation of representation mechanisms" mentioned in the essay is called the interpretability problem in deep learning research. Current mainstream models (such as large-scale neural networks) are black-box: we can precisely compute their outputs, optimize their behavior through training objectives, yet it is difficult to describe in human-understandable concepts what internal neurons and attention heads actually encode. Mechanistic interpretability attempts to reverse-engineer the computational circuits inside networks, similar to physicists trying to find intuitive pictures for the quantum formalism. Both face structurally similar predicaments—formal systems operate extremely successfully, but explanations of "what is actually happening inside" always lag behind predictive power. This state of "using without understanding" is not uncommon in engineering, and precisely for this reason, Doug's candor sparked cross-disciplinary resonance.

Where the Discussion Diverges

The 146 comments on Hacker News show a typical divide. One camp holds that "understanding" is itself a dispensable philosophical requirement and that the goal of physics is only prediction and control—a position that can be called instrumentalism. The other camp insists that the endpoint of science is understanding the world, and that a theory which calculates accurately but cannot explain clearly merely hides confusion inside mathematics.

Some readers also noted that the way quantum mechanics is taught creates confusion—giving formulas first and discussing interpretation later, or simply skipping interpretation and teaching only calculation. Doug's original text precisely confirms this comment: confusion often does not come from mathematical difficulty but from the gap between "calculating correctly" and "missing a picture."

This is not a debate that can be won. The confrontation between instrumentalism and realism in quantum mechanics has lasted nearly a century, and Doug's contribution is to bring this confrontation from the footnotes of academic papers to the table, saying in the first person something many privately acknowledge but rarely write publicly: I don't understand.

For any technical practitioner learning complex systems, this may be the most easily overlooked lesson—being able to calculate an answer and being able to explain the reasoning are two different things.

Instrumentalism is a position in the philosophy of science holding that the value of a theory lies in its predictive and operational capability, not in whether it depicts the true structure of the world. In the context of quantum mechanics, instrumentalists hold that the wave function is merely a tool for calculating probabilities, and that asking "what a particle actually is" is a pseudo-problem. The opposing scientific realism holds that successful scientific theories should approximately describe an objective reality independent of observers, otherwise science's explanatory power would be groundless. This debate is not unique to quantum mechanics, but quantum mechanics pushes it to its sharpest point: a theory with prediction accuracy to more than ten decimal places, yet no consensus on "what it describes." The divide in the HN discussion is essentially a spontaneous reappearance of these two philosophical positions in the tech community.

One-line Summary

Quantum mechanics calculates accurately but cannot explain clearly—admitting confusion is an underrated professional virtue.

Paragraph Summary

University of Southampton scholar Doug's essay "I Don't Understand Quantum Physics" sparked heated discussion on Hacker News. The essay points out that quantum mechanics' mathematical framework is precise and effective, yet the underlying picture—wave-particle duality, the measurement problem, and non-local entanglement—remains controversial; "calculating correctly" does not equal "truly understanding." The Copenhagen interpretation, many-worlds interpretation, and decoherence theory each avoid or redefine certain problems. The HN discussion reveals a divide between instrumentalism and realism: one side holds that prediction and control suffice, the other insists science must provide understanding. The author argues that acknowledging cognitive boundaries is an underrated professional attitude, with general relevance for practitioners of deep learning and other technologies that are heavily used yet whose internal mechanisms are difficult to explain.