Here's why amps can be more important than speaker

No, the reason amp distortion specs are so low is because it's possible, and because a low number looks good on a spec sheet. I suspect most speakers have just as much higher-order harmonic distortion as typical amps do (along wth a lot more lower-order distortion). Most mainstream amps today have harmonic distortion below audibility thresholds; most speakers do not. Amps are important, but this is not why.

Cdc, Don't be so digital! A real sound from the non-electronic/speaker source will also have it's own harmonic distortions. Bringing it down to approach zero means killing a part of life within that. I stopped looking at specs for THD unless it goes above 0.4%.

Amp is important in the way that it should be properly engineered and its minimal task is not to destroy speakers at its extreams and normal mode and give them always a nice and pleasant drive.
In SS case the heat sinks must be properly matched with transistor parameters, DC offset must be set so that output transistors idle cool(or let's say you can touch them without hurting yourself).
In tube amp case more consideration must be taken in PS reliability and proper noise canceling. Proper output protection should take place especially if tube suddenly blows.
As to the pricing of the speakers v.s. amps, once the minimal task is in effect it can be either more or less costly.

I disagree as well. Ditto Bomarc. I am surprised B&W said something like that - oh well...glad I sold mine!

If you drive your amp hard, distortion will be worse than the spec you see on the paper. Those are measured by a constant resistance load, I think, and the signal level they measured are of course at easy power level (which makes numbers looking good). But for real application, amp drives real speakers at pretty high power/current level. The speakers are much tuffer load, depending on frequency and their own distortions are not small even at easy power level. In engineer's words, both are working at nonlinear regions, and those numbers measured at more linear level (lower current and linear load) do not mean much in real life listening. You can use those numbers as reference but should not trust them more than your ears.
I would trust those spec's on digital equippments more, those are more linear devices, 0 or 1, and you can't lie about how many bits they are and what sampling frequency is.
If they can provide more inf on jittering noise then that would be a good judgement of transport on paper. The only problem is when it gets down to DAC and output state, analog parts come in and the nonlinear effects kick in again.