The miller effect is used to create a dominant pole in frequency compensated op-amps. Without it, integrated capacitances would be too small to be useful in this manner.
Miller effect can (like it or not) produce a slew-rate-limited
output transition. You can see this in slew-controlled CMOS
logic buffers and in power MOSFET applications - gate
resistor vs Cdg*dV/dt(out) enforce a maximum drain
slew rate.
In simple terms, the Miller effect allows a small capacitance to be "amplified", which is useful for compensation. Using a non-Miller capacitance for compensation means it will have a larger value, hence larger area.