Why 80 spins are not automatically impossible classically
A general state of 80 qubits has 280 complex amplitudes. Storing the full statevector is therefore impractical. A classical simulator, however, does not have to use that representation.
A matrix-product state, or MPS, represents the state as a chain of small tensors. Its bond dimension chi limits the correlations the representation can retain. At fixed chi, the number of tensor parameters scales approximately as N times 2 times chi squared. For N=80 and chi=16, this is about 40,960 complex numbers in the raw tensors, excluding working memory and software overhead. All 80 physical spins remain present; it is the permitted state representation that is strongly restricted. Background on MPS.
TEBD evolves an MPS using local two-spin operations and repeatedly truncates the bond dimension. A larger chi can retain more information but requires more computation. TDVP is another tensor-network approach. Pauli propagation takes a different route: it follows the measurement operator in the Heisenberg picture and can be a strong competitor when only a few observables are needed. None of these approaches should be omitted simply because it makes a quantum comparison less favourable.