Every computer you've ever used — laptop, phone, game console — runs on bits. A bit is a switch that's either off or on, written as 0 or 1. Everything your computer does, from rendering video to running a search engine, comes down to flipping billions of these switches according to fixed rules. A classical bit has no in-between state. It's 0, or it's 1, full stop.

A qubit (quantum bit) is the quantum version of that switch, and it behaves fundamentally differently. Instead of being stuck at 0 or 1, a qubit's state is described by a state vector — a mathematical object that assigns a number to each possible outcome, 0 and 1, capturing how much of each outcome is "in" the qubit before you check it. Physicists write this using Dirac notation: the two basic states are written $|0⟩$ and $|1⟩$ (read "ket-zero" and "ket-one"), and a general qubit state is written as

$$|\psi ⟩=\alpha |0⟩+\beta |1⟩$$

Here $\alpha$ and $\beta$ are called amplitudes — numbers that determine how likely you are to get each outcome when you measure the qubit. This is the part that trips people up, so it's worth being precise: amplitudes are not probabilities themselves. To get a probability, you square the amplitude's magnitude. If $\alpha =0.6$, the probability of measuring $0$ is ${0.6}^2=0.36$, or 36%. The two probabilities have to add up to 100%, since the qubit has to come out as something.

Measurement is the act of checking a qubit's value — and it's a one-way door. Before measurement, the qubit's state vector can involve both $|0⟩$ and $|1⟩$ with various amplitudes. After measurement, you get one classical answer, 0 or 1, with odds set by those amplitudes, and the superposition is gone. If you measure again immediately, you'll get the same answer every time — the qubit has "collapsed" into whichever state you found it in.