A Note to the Stubborn Physicist

There is a particular kind of objection that surfaces every time the word “quantum” is attached to the word “computing.” It comes, almost always, from people who know the underlying physics better than the people selling the product. They have read the papers. They have done the math. They have sat with decoherence long enough to know what it costs. And their position, when you finally pin it down, is something like this:

“This is not what quantum mechanics is. You are taking a description of nature and bolting it onto an engineering project that does not deserve the name. The states you are talking about are not the states I study. The superposition you are invoking is not the superposition the universe runs on. Stop pretending.”

I want to say something direct to that physicist, because the objection is half right, and the half that is right is worth respecting.

You are correct.

You are also, in the way that matters here, wrong.

The Half That Is Correct

The half that is correct is this: the universe does not owe a computer manufacturer its vocabulary. When a chip designer says “qubit,” they are not making a metaphysical claim about the wavefunction of the cosmos. When a marketing deck says “quantum advantage,” it is not a statement about Bell’s theorem. The physics community has spent a century being precise about what “quantum” means in nature, and watching that precision get stretched thin to sell a product is, legitimately, irritating.

If your objection is that the language is being abused — yes. Often. Frequently. Sometimes embarrassingly.

If your objection is that most of what gets called “quantum” in commercial computing today is engineering scaffolding around a small number of genuinely quantum-mechanical operations, surrounded by an enormous amount of classical control — yes. That is also true.

If your objection is that decoherence is not a marketing problem, it is a thermodynamic one, and pretending otherwise misleads investors — yes again.

So far, the physicist is winning the argument.

The Half That Is Wrong

Here is where the argument turns.

The physicist’s objection assumes that the question on the table is: does the physical universe permit this? And inside that question, they are the expert. They have earned the right to be skeptical, and their skepticism is calibrated.

But that is not the question on the table.

The question on the table is: does a digital universe — a constructed, engineered substrate whose laws we wrote — permit this? And the answer to that question is not governed by the physics of the universe we live in. It is governed by the physics of the universe we built.

A digital universe has its own state space. It has its own conservation laws. It has its own notions of locality, of measurement, of decoherence-by-analogy. Some of those notions were borrowed from physical quantum mechanics because the math was useful. Some of them were invented because the engineering required it. None of them are claims about the cosmos.

When a quantum computing platform talks about a “qubit,” it is not asserting that the physical universe contains a fundamental object called a qubit. It is asserting that within the engineered substrate, there exists a two-level system whose behavior, within tolerances, can be described by the mathematics of a two-level quantum system. The mathematics is the same. The ontology is not.

The physicist is defending the ontology. The engineer is using the mathematics.

These are not the same argument, and they do not contradict each other.

The Argument From What Already Works

There is a harder version of this point that the stubborn physicist has to answer, and I want to put it on the table plainly.

Every computer chip in service today — every classical, binary, “non-quantum” chip — is a quantum-mechanical device. The transistor works because of band theory. The band theory is quantum mechanics. The flash memory in the phone in your pocket stores bits by exploiting quantum tunneling through an oxide barrier. The CMOS sensor in every camera made in the last twenty years is reading the photoelectric effect, one photon at a time. The laser in the fiber connection carrying this article to you is a population-inverted stimulated emission system, which is to say: quantum mechanics, packaged for industrial use.

The “1” and the “0” are not classical objects. They are engineered abstractions on top of quantum behavior. We agreed, decades ago, to draw a threshold across a voltage and call everything above it a 1 and everything below it a 0. Underneath that threshold is electron behavior the classical picture cannot describe. We have been computing on quantum substrate the entire time. We simply chose, for sound engineering reasons, to extract a binary abstraction from it.

So when the stubborn physicist says “this is not really quantum computing, it is engineering on top of quantum effects” — yes. Correct. That is also what every chip on Earth has been doing for sixty years. The only thing changing now is which features of the underlying quantum behavior the abstraction exposes. The binary chip throws away superposition and entanglement and keeps only the threshold. The quantum chip keeps more of them.

That is a difference in how much of the substrate the abstraction surfaces. It is not a difference in whether the substrate is quantum. The substrate has been quantum since the first transistor.

You cannot deny what has already been in working practice for two decades. Longer, actually — closer to seven. The denial would have to retroactively undo the device the denial is being typed on.

The Category Error

What the stubborn physicist is doing — and I say this with respect, because I have done it myself in adjacent domains — is committing a category error. They are taking a claim made inside a constructed system and evaluating it as if it were a claim about the underlying universe.

It is the same shape of error as objecting to the rules of chess because knights do not move in L-shapes in physical reality. The chess knight is not making a claim about horses. The quantum bit is not making a claim about electrons.

The constructed universe — the digital one, the one we are actually arguing about — has its own physics. We wrote it. We can change it. The rules of that universe are not constrained by the rules of the universe outside it, except at the seam where they meet: the physical hardware that runs the abstraction.

And that seam is exactly where the physicist’s expertise becomes valuable again, and exactly where most of the legitimate criticism of quantum computing lives. How well does the physical substrate maintain the abstraction? That is a physics question. That is the question the physicist should be asking. That is the question worth being stubborn about.

But “the abstraction is illegitimate because the universe does not work that way” is not that question. It is a different question, and it has a different answer, and the answer is: the abstraction does not have to correspond to the universe to be a coherent system. It only has to correspond to itself.

Why This Matters Beyond Quantum

The reason I am writing this down is not because I think one more essay will move the stubborn physicist. They are stubborn for a reason, and the reason is often that they have seen too many bad pitches and developed an immune response. Fine.

I am writing it down because the same shape of error appears everywhere computing meets a discipline that owns a vocabulary.

• Neuroscientists who object to “neural networks” because the units do not behave like biological neurons.
• Linguists who object to “language models” because the systems do not have a theory of grammar.
• Cognitive scientists who object to “memory” in software because there is no consolidation, no forgetting curve, no hippocampus.

In each case, the objection is partially correct — the language is borrowed, sometimes abused, often overclaimed. And in each case, the objection becomes a category error the moment it asserts that the engineered system must conform to the natural system to deserve the borrowed word.

It does not. It only has to be internally coherent. It only has to do, in its own substrate, what the mathematics says it does.

The universe we built has its own rules. Defend the universe you study. But do not assume the rules of that universe are the rules of every universe we are allowed to construct.

A Short Truce

To the physicist who has read this far and is still unconvinced: I am not asking you to stop being stubborn. Stubbornness in physics is a feature, not a bug. The discipline runs on the refusal to accept claims that are not earned.

I am asking only that the stubbornness be aimed at the right target.

Aim it at the hardware. Aim it at the decoherence budgets. Aim it at the gate fidelities. Aim it at the marketing decks that overclaim physical quantum advantage when the underlying system is doing something far more modest. There is real work for your skepticism there, and the field is better when you do it.

But do not aim it at the abstraction itself. The abstraction is not a claim about your universe. It is a description of a different one — one we wrote, one we maintain, and one whose physics we are allowed to define.

You can be right about the cosmos and wrong about the construct at the same time. Most of the stubborn objections to quantum in computing are exactly that: a correct argument, addressed to the wrong universe.

A Longer Timeline Than the Argument Assumes

One last thing, and then I will stop.

The objection that constructed systems are not “really” the thing they describe is older than computing. It is older than physics as a discipline. It is, at minimum, a two-thousand-year-old argument, and the side that says “the construct has its own legitimate reality, governed by its own internal rules” has been winning that argument, slowly, for most of that time.

Euclid’s geometry was not a claim about the surface of the Earth. It was a constructed system, internally consistent, useful precisely because it did not have to correspond to any physical surface to be true within itself. The objection “but real triangles are not made of pure lines” was available in 300 BCE. It was, even then, a category error.

Formal logic in the nineteenth century, set theory in the twentieth, computability in the 1930s, information theory in the 1940s — each one of them is a constructed universe with its own physics, and each one of them was met by some version of the same objection. This is not real. The universe does not work this way. And each one of them, in the end, turned out to be doing real work — running the world, in some cases — precisely because the question of “real” was the wrong question to ask of a construct.

Quantum-in-computing is one more entry on that list. The argument we are having now is the same argument Euclid was having, the same argument Frege was having, the same argument Turing was having. The stubborn physicist is the most recent voice on a very long side of a very long discussion.

The reason work like this succeeds — the reason a constructed-universe framing of computation, governance, cognition, or anything else can land — is not that the construct is finally being recognized as physical. It is that the two-thousand-year-old distinction between the universe we live in and the universes we build is, at last, becoming load-bearing infrastructure rather than a philosopher’s footnote.

We are not arguing about technology. We are arguing about an old idea that technology finally made operational.

This is part of an occasional series on the foundations of computational vocabulary — what we borrow, what we owe, and where the borrowing breaks.