quantum measurement

Before you look at a qubit, it can sit in a superposition: a blend of both 0 and 1 at once, described by two amplitudes. Measurement is the act of looking, and it forces a decision. The instant you measure, the qubit snaps to a definite 0 or a definite 1, and the blend you had a moment ago is gone for good. Think of a spinning coin: while it spins it is neither heads nor tails, but the moment you slap it flat on the table you get exactly one face, and the spin is over.

Which outcome you get is genuinely random, but the odds are not arbitrary. The Born rule says the probability of reading 0 or 1 equals the square of that outcome's amplitude. So an amplitude is not the chance itself; you square its size to get the chance. Importantly, you only learn one classical bit from the whole event, and the measurement destroys the original superposition rather than revealing it. You cannot peek at the amplitudes directly, you cannot read them off without disturbing them, and you cannot copy the state and try again.

This is the heart of why a quantum computer is not a machine that 'tries all answers at once' and then hands you the best one. The hidden superposition holds rich information, but reading it out collapses everything to a single bit governed by chance. A useful quantum algorithm has to arrange its amplitudes so that, through interference, the right answer becomes the overwhelmingly likely thing to see when you finally measure. The measurement is the narrow doorway every quantum result has to squeeze through.