Here is a misconception worth killing early: a quantum computer is not fast because it "tries all answers at once." Putting a register into a superposition of every possibility is easy — but a measurement collapses it to just one of them, at random. That is no better than guessing. The thing that makes quantum computing powerful is subtler and stranger: interference.
Amplitudes are waves
A qubit's state is described by amplitudes, and — crucially — amplitudes can be negative (or complex), unlike probabilities, which are always positive. The probability of an outcome is the amplitude squared. This means amplitudes behave like waves: two contributions to the same outcome can be both positive and add up (constructive interference), or one positive and one negative and cancel (destructive interference).
That is the whole game. In a quantum algorithm, each possible answer can be reached by many computational "paths", each contributing an amplitude. If you engineer the paths so that:
- the amplitudes leading to wrong answers cancel out (sum to ~0), and
- the amplitudes leading to the right answer reinforce (sum to something large),
then when you finally measure, the right answer is overwhelmingly likely.
Where phase comes in
Back in Foundations you met phase — the part of a qubit's state that a single measurement can't see. Now it earns its keep: phase is exactly what determines whether two amplitudes add or cancel. Gates like Z and T exist to nudge phases precisely so that, several steps later, the interference lands the way the algorithm designer intended. A quantum program is, in a real sense, a carefully tuned interference pattern.
Why this is hard (and powerful)
Designing that interference is difficult — it is why good quantum algorithms are rare and celebrated. But when it works, the payoff is dramatic: a problem that would take a classical computer an astronomical number of steps can sometimes be solved by letting the wrong paths quietly cancel themselves out. In the next lesson we watch this happen concretely, in the most intuitive quantum algorithm — Grover's search, which finds a needle in an unsorted haystack quadratically faster than any classical method.