Remember analog computers? You would set up a circuit so that the voltage in one place was the answer to your computation, and then instead of calculating the answer you would *measure* the answer. We stopped thinking about them because it was tricky to set up the circuit for each calculation, but once you had it set up the computation would happen at the speed of electrons.
The D-Wave computer is similar to this. Given a polynomial in many variables (with positive real coefficients, and the variables only take the values 0 and 1), you might like to find the assignment to the variables that minimizes the polynomial. So D-Wave sets up a thermodynamic system whose steady state can be *measured* and gives an assignment to the variables that makes your polynomial small. Systems will naturally try to minimize their energy, and so the assignment is likely to be your perfect minimum (repeat 100 times, and the best assignment is likely to have appeared).
The question is whether the system minimizes its energy by classical thermodynamic flow (super fast), or by quantum effects (super-duper fast). It is, for that particular sort of problem, *much* faster than anything else ever. It had seemed to be so much faster and accurate that it had to be using quantum effects. But now somebody has found a faster way to do it classically, so that it isn't *that* much faster. For those of you in the know, the question isn't speed but the rate of growth of the speed: is the ratio of speed-up growing polynomially in the input, or exponentially?