What Is a Qubit?
A classical bit has exactly two possible states: 0 or 1. A qubit's state is described by two numbers (technically, complex amplitudes): one weighting how "0-like" it is, one weighting how "1-like" it is, and both can be nonzero at the same time. That's superposition. It isn't that the qubit is secretly 0 or 1 and we just don't know which yet (that would be ordinary uncertainty, like a coin under your hand). The mathematics and the experiments both show it's genuinely in a combined state until measured.
α and β are the amplitudes from Figure 1's vector. Squaring their sizes gives the probability of measuring 0 or 1, and those two probabilities always add to 1.
Physically, a qubit can be built from very different things: a superconducting circuit cooled near absolute zero (IBM, Google), a single trapped ion held in an electromagnetic field (IonQ, Quantinuum), a photon of light, or the spin of a single electron. They all behave differently as engineering problems, but they all implement the same qubit mathematics underneath, which is why "how many qubits" alone is a weak way to compare quantum computers; how good those qubits are (how long they stay coherent, how low their error rates are) matters just as much.
The moment a qubit is measured, its superposition collapses to a definite 0 or 1, and that answer comes out probabilistically, weighted by those amplitudes. This is why quantum algorithms are usually run many times and the results interpreted statistically, rather than run once like a normal program.
Try it livePut a real qubit into superposition and watch the vector move on an interactive Bloch sphere in the Bloch Sphere simulation.
Work with real qubit states and watch superposition and measurement happen in F1 — Quantum States.