Interference as the Source of Speedups

Quantum interference lets amplitude paths add or cancel, concentrating probability on correct answers; it is what makes quantum algorithms fast.

The real engine

Superposition creates many possibilities and entanglement correlates them, but interference is what turns a quantum computer into a useful one. Because amplitudes are complex and can be positive, negative, or imaginary, contributions to the same outcome can add (constructive) or cancel (destructive).

Amplitudes, not probabilities

This is the decisive break from classical randomness. Classical probabilities are non-negative and only ever add. Quantum amplitudes can cancel, so a path leading to a wrong answer can be arranged to vanish. Probability is the squared magnitude of the summed amplitude, so cancellation before squaring produces effects impossible for any coin-flipping classical machine.

A minimal example

Apply a Hadamard to |0>: you get (|0>+|1>)/sqrt(2). Apply it again and the two amplitude paths to |1> have opposite signs and cancel, while the paths to |0> reinforce, returning |0> with certainty. Interference, not measurement luck, produces the deterministic result.

How algorithms use it

Every quantum speedup is an interference pattern engineered to make correct answers likely and wrong answers unlikely:

• Grover's search rotates amplitude toward the marked item each iteration
• Shor's algorithm uses the Fourier transform to make period information interfere constructively
• Quantum simulation lets physical amplitudes interfere as they naturally would

Why coherence is required

Interference needs well-defined relative phases. Decoherence randomises those phases, which is exactly why it destroys computational power: without stable phase relationships the amplitudes no longer cancel and add cleanly, and the algorithm degrades to a classical random guess. Protecting phase coherence is protecting the very mechanism of speedup.