You only need a linear increase in number of error-corrected qubits to get exponential gains on the limited subset of problems they can improve, whereas Moore's

You only need a linear increase in number of error-corrected qubits to get exponential gains on the limited subset of problems they can improve, whereas Moore's law has provided an exponential increase in the number of elements for classical computers. So, Moore's law isn't needed for something similar to Moore's law gains in quantum computing, a linear Moore's law will do if quantum error correction doesn't face scaling laws for the whole ensemble. That might mean they only keep pace with classical computing if Moore's law continues to hold there, but eventually that hits the Landauer limit without new physics (or possibly reversible computing).