Bad Qubits — Learn Quantum by Doing

Multiple choice

According to the lesson, why are the Clifford gates (generated by H, S, and CNOT) NOT universal for quantum computation on their own?

Clifford gates are not unitary and so cannot represent valid quantum evolution
Clifford gates cannot generate any entanglement between qubits
By the Gottesman-Knill theorem, a Clifford-only circuit on a stabilizer input with computational-basis measurement can be simulated efficiently classically, so it offers no quantum advantage
Clifford gates can only act on a single qubit and never on two qubits