Hadamard gate (H)
Most quantum algorithms start by taking a qubit that holds a definite answer — say |0> — and 'opening it up' so it is genuinely undecided between 0 and 1. The Hadamard gate is the tool that does this. Hand it a |0> and it produces an equal superposition of |0> and |1>; hand it a |1> and it produces a similar equal mix, but with a different internal sign. Think of it less like flipping a coin and more like setting a coin spinning: the qubit is now poised between both outcomes in a precise, fully describable way, and that exact balance is what later gates rely on.
It helps to be honest about what 'superposition' buys you here. After one H gate, if you measured the qubit immediately you would just get 0 or 1 with 50/50 probability — no magic, no parallel answers. The power comes later: the sign that H tucks into the |1> branch lets the branches interfere with each other as the algorithm runs, so wrong answers can cancel out and right ones reinforce. The Hadamard gate is what creates those branches in the first place; the rest of the circuit choreographs how they interfere.
Two practical notes. H is reversible like every quantum gate, and it is its own inverse — apply it twice and you are exactly back where you started, the two superpositions interfering to undo each other. And it rarely works alone: a layer of Hadamards across many qubits, followed by entangling gates and a final round of interference, is the basic shape of algorithms from Grover's search to Shor's factoring.
H sends each basis state to an equal superposition; note the minus sign on the |1> case — that relative sign is what makes later interference possible.
A Hadamard does not let a qubit 'try all answers at once' — measuring right after H just gives a random 0 or 1; the usefulness comes only from later interference between the branches it opens up.