Six Interpretations — and What Your Quantum Computer Is Doing

Here is the single most important fact in this module, stated plainly: every interpretation of quantum mechanics predicts exactly the same experimental results. No experiment — not a Bell test, not a quantum supremacy run, not anything — can distinguish between them. They disagree about what is real, not about what is measured. That is what makes them philosophy rather than physics. Keep that label on each one as you read.

1. Copenhagen (Bohr, Heisenberg, and the textbooks). The original orthodoxy: the quantum formalism is a recipe for predicting what classical instruments will register. Superpositions describe your knowledge of possible measurement outcomes, not a thing in the world; "collapse" is just you updating the books when you look. Don't ask what the electron is doing between measurements — the question is malformed, because a phenomenon only exists relative to an experimental arrangement. On this view, your quantum computer's intermediate superposed states are calculational devices; only the preparation and the readout are physically describable. The computation "happens" nowhere classical — which is exactly why it looks powerful. (Honest footnote: there was never one Copenhagen view. Bohr and Heisenberg disagreed with each other, and the tidy label was invented later.)

2. Many-worlds / Everett (Everett 1957; DeWitt; Deutsch; Wallace). Delete the collapse postulate entirely. The wavefunction always evolves smoothly, every term of every superposition is equally real, and what we call "measurement" is just the observer becoming entangled with the system — the universe splits, or rather differentiates, and decoherence explains why the branches can't interfere afterward. On this story, Deutsch's account of your quantum computer is literal: Shor's algorithm runs across a vast number of branches at once, and the answer emerges where they interfere. It's the most dramatic story — and it has two famous unpaid bills. First, the preferred-basis problem: a quantum state can be decomposed into "worlds" in infinitely many ways, so which decomposition counts as real seems to depend on what you later choose to measure. Second, the probability problem: if every outcome happens, what does "70% probability" even mean? (Deutsch and others have offered decision-theoretic derivations of the Born rule; they remain contested.) Aaronson's verdict on the whole dispute: supremacy experiments "don't add anything new to this old debate" — they're just more confirmations of the equations everyone already agrees on.

3. Pilot-wave / de Broglie–Bohm (de Broglie 1927; Bohm 1952). The particles are real and always have definite positions — no fuzziness at all. Riding above them is the wavefunction, a real physical pilot wave that choreographs every particle's motion through a guiding equation. The theory is fully deterministic: the apparent randomness of quantum mechanics is just ignorance of exact starting positions, like classical statistical mechanics. The price is nonlocality: the guiding wave lives in configuration space and coordinates distant particles instantaneously. Your quantum computer, on this view, is a swarm of perfectly definite particles being steered by a wave evolving through an exponentially large space — the speedup lives in the wave's choreography.

4. QBism (Fuchs, Schack, Mermin). The most radical reframe: the wavefunction is not a thing in the world at all — it is one agent's personal degrees of belief about their own future experiences. "Collapse" is just a Bayesian update: you learn something, you revise your bets. Nothing in the world jumps. And the famous spookiness of entanglement? "A QBist denies that anything is sent, because there was never an objective state on the distant particle to be disturbed." On this view your quantum computer is a tool an agent uses to place coherent bets — the computation is structured expectation, and the user is part of the story. (Not to be confused with "consciousness collapses the wavefunction" — QBism gives the agent a role without any mind-magic.)

5. Relational quantum mechanics (Rovelli). There are no absolute facts — only facts relative to some physical system. It is meaningless to say "the qubit is in superposition" full stop; you must say "the qubit is in superposition for this observer." Different observers can give different, equally valid accounts, and that is a complete description of the world — there is no God's-eye view to reconcile them. Your quantum computer's intermediate states are real, but only relative to whatever is interacting with them.

6. Consistent histories (Griffiths; Omnès; Gell-Mann & Hartle). Billed as "Copenhagen done right": you may assign probabilities to whole histories — sequences of events through time — provided they satisfy a consistency condition, and you must never mix incompatible descriptions (the single-framework rule). Measurement is just another physical process; on this telling "there is no measurement problem," and apparent paradoxes dissolve. Crucially, it insists that "superluminal influences cannot carry information or anything else, for the simple reason that they do not exist." Your quantum computer has a perfectly ordinary story at successive times — within one consistent framework.

The stress test for all six is Wigner's friend and its modern form, the Frauchiger–Renner thought experiment (2018): nested observers applying quantum mechanics to each other produce a trilemma — universal validity of quantum theory, a single shared world, and consistency between agents' accounts cannot all hold. Every interpretation survives by sacrificing a different one. Which sacrifice bothers you least is, once again, philosophy.

Common myth: "Quantum computers prove the many-worlds interpretation — where else would the computation happen?" No interpretation is confirmed by any computation, because all of them predict identical statistics. Deutsch's multiverse account is his philosophical position, not an experimental result.